EDUCATIO NA L IHC Guidebook Sixth Edition Immunohistochemical Staining Methods Education Guide Immunohistochemical Staining Methods Sixth Edition Editors „„Clive R. Taylor, MD, D.Phil, Past Chair, Department of Pathology, University of Southern California; Past Council Member and President, Association of Pathology Chairs „„Lars Rudbeck, PhD, MSc, Scientiic Editor, Dako Denmark Graphic Design „„Anja H. Sjørup, MSc, Graphic Designer, Dako Denmark © Copyright 2013 | Dako Denmark A/S, An Agilent Technologies Company. All rights reserved. No part of this book may be reproduced, copied or transmitted without written permission. Immunohistochemical Staining Methods | Table of Contents Table of Contents Immunohistochemical Staining Methods Part I: The Staining Process Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Chapter 1 | Introduction to Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 by C.R. Taylor 1.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1.2 | History of Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.3 | Standardization in Clinical Immunochemistry vs. Immunohistochemistry . . . . . . . . . . . . . . . . . . 15 1.4 | Growing Consensus for Standardization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.5 | Standardization Starts in the Surgery Room . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 1.6 | Future Aspects for Standardization of Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Chapter 2 | Fixation and Other Pre-Analytical Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 by E.C. Colley & R.H. Stead 2.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.2 | Tissue Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 2.3 | Fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 2.4 | Tissue and Slide Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 2.5 | Special Tissue Preparations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2.6 | Control Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 2.7 | Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.8 | Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Chapter 3 | Antigen Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 by S-R Shi & C.R. Taylor 3.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.2 | Major Factors that Inluence the Effect of Antigen Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3.3 | Standardization of AR – The “Test Battery” Approach. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.4 | Application of AR Techniques – The Basic Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.5 | AR-IHC-based Research and Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.6 | Reagents and Protocols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.7 | Water Bath Methods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3.8 | Pressure Cooker Heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 3.9 | Autoclave Heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.10 | Microwave Oven Heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.11 | Proteolytic Pre-treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 3.12 | Combined Proteolytic Pre-treatment and Heat-Induced Antigen Retrieval. . . . . . . . . . . . . . 40 3.13 | Combined Deparafinization and Antigen Retrieval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 3.14 | Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Table of Contents | Immunohistochemical Staining Methods Chapter 4 | Selection of the Primary Antibody . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 by S. Nielsen 4.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.2 | Seletion of the Proper Antibody . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 4.3 | A New (Replacement) Antibody for an Old Marker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.4 | A New Antibody for New Marker in the Pathology Laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 4.5 | A New Experiment in the Pathology Laboratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 4.6 | Examples of Good and Poor Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.7 | Current and Future Challenges. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Chapter 5 | Staining Protocol Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 by L. Jacobsen, M. Nielsen, S. Månsson & L. Rudbeck 5.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 5.2 | The Basis for an Optimal Staining Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 5.3 | The protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 5.4 | Reproducibility of IHC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5.5 | Requirements for Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 5.6 | Veriication and Validation of a Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 5.7 | Guide to the Development of an Optimal Staining Protocol. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 5.8 | Protocol Performance for Abnormal Tissue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 5.9 | Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Chapter 6 | Immunohistochemistry Staining Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 by K. Petersen & H.C. Pedersen 6.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.2 | Avidin-Biotin Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.3 | Polymer-Based Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 6.4 | Catalysed Signal Ampliication (CSA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 6.5 | Fluorescyl-tyramide Ampliication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 6.6 | Improved Catalysed Signal Ampliication (iCSA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 6.7 | Multi-Staining Immunohistochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83 6.8 | Selection of Dyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 6.9 | Automated Image Acquisition and Analysis in Multiple Staining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 6.10 | Immunoluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87 6.11 | Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Chapter 7 | Digital Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 by J. Schmidt & C.R. Taylor 7.1 | Microscopy – From Analog to Digital . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 7.2 | Automation of Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 7.3 | Slide Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 7.4 | Digital Slide Review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 7.5 | Applications in Digital Pathology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Immunohistochemical Staining Methods | Table of Contents 7.6 | Image Analysis and Quantiication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 7.7 | Analysis of IHC stains – ’Eye vs Algorithm’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 7.8 | Quality Assurance in IHC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 7.9 | Future Trends in Digital Pathology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Part II: The Potentials and Pitfalls Chapter 8 | Optimization of Immunohistochemical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 by G.E. Pace 8.1 | Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 8.2 | Tissue Digestion using Antigen or Heat-Induced Epitope Retrieval. . . . . . . . . . . . . . . . . . . . . . . . . . 103 8.3 | Tissue Digestion using Proteolytic Enzymes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 8.4 | Endogenous Enzyme Blockers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 8.5 | Protein Blocking Reagent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 8.6 | Endogenous Biotin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 8.7 | Antibody Diluents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 8.8 | Antibody Concentrations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 8.9 | Incubation times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 8.10 | Wash Buffers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 8.11 | Chromogen Enhancers for DAB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 8.12 | Type of Glass Slides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Chapter 9 | Automation in IHC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 by O.F. Rasmussen 9.1 | History of IHC Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 9.2 | Key Advantages of IHC Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 9.3 | Staining Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 9.4 | Requirements for Effective Automated Staining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 9.5 | Automation vs. Worklow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 9.6 | Key Features for ‘Complete Staining Solutions’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 9.7 | Next Steps in Automation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Chapter 10 | Optimizing Laboratory Worklow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .122 by D. MacDonald 10.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 10.2 | Specimen Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 10.3 | Specimen Collection and Identiication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 10.4 | Specimen Accessioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 10.5 | Grossing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 10.6 | Tissue Processing and Microtomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 10.7 | Routine Staining and Coverslipping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 10.8 | Slide/Block Reconciliation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Table of Contents | Immunohistochemical Staining Methods 10.9 | Laboratory Asset Tracking and Worklow Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 10.10 | One Worklow Does Not Fit All. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 10.11 | General Sample Labeling and Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Chapter 11 | Companion Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .132 H.C. Pedersen & J.T. Jørgensen 11.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 11.2 | History of Companion Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 11.3 | Companion Diagnostics and ‘Personalized Medicine’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 11.4 | Co-Development of Drug and Companion Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 11.5 | Clinical Validation of the Companion Diagnostics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 11.6 | Companion Diagnostics and Regulatory Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 11.7 | Learnings from EGFR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 11.8 | Conclusion and Future Perspectives. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Chapter 12 | Tissue Microarray - Construction and QA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .144 by R. Saxena & S. Badve 12.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 12.2 | Advantages and Disadvantages of TMAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 12.3 | Types of TMAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 12.4 | Team Required for TMA Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 12.5 | TMA Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Chapter 13 | IHC Visualization of Molecular Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 by J. Mollerup & J.T. Jørgensen 13.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 13.2 | FISH versus CISH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 13.3 | Principle of the CISH Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 13.4 | HER2 CISH in Clinical Decisions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 Chapter 14 | Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .160 by O.F. Rasmussen & R. Jørgensen 14.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 14.2 | Purpose of Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 14.3 | Categories of Controls and Control Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 14.4 | Use of Controls in Daily Routine Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 14.5 | Tissue Process Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 14.6 | Cell Line Controls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 14.7 | Monitoring the Staining Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 14.8 | External Quality Assurance Programs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 14.9 | Future aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169 Immunohistochemical Staining Methods | Table of Contents Chapter 15 | Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .170 by H.G. Wendelboe 15.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 15.1 | Detection Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 15.3 | Antigen Retrieval (Heat-Induced Epitope Retrieval) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 15.4 | General Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174 15.5 | Miscellaneous Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 15.6 | General Aspects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 Chapter 16 | Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .180 by H.G. Wendelboe, A. Lykke & G.E. Pace 16.1 | Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Section 1 – Common Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181 Section 2 – Systematical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Section 3 – Troubleshooting Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191 Section 4 – Speciication Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Section 5 – Automated Platform Performance Checks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 Appendix A | Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .198 Revised by S.S. Jensen App A.1 | Immunoglobulins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 App A.2 | Antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 App A.3 | Polyclonal Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 App A.4 | Monoclonal Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 App A.5 | Polyclonal Antibodies versus Monoclonal Antibodies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 App A.6 | Antibody Afinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 App A.7 | Antibody Cross-Reactivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204 App A.8 | Antibody Reaction Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 App A.9 | Antibody Stability and Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205 Appendix B | Basic Immunochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208 Revised by S.S. Jensen App B.1 | Antibody Titer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 App B.2 | Antibody Dilution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 App B.3 | Incubation Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 App B.4 | Incubation Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 App B.5 | Basic Enzymology in IHC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .214 Biography | Dako’s Guidebook to Immunohistochemical Staining Methods Biography Dako’s Guidebook to Immunohistochemical Staining Methods Editor: Clive Taylor MD, D.Phil. Dr. Clive Taylor started his medical education at the University of Cambridge and completed his doctoral studies in Immunology at University of Oxford. After his education, he accepted a position at University of Southern California where he for 25 years functioned as Chair of the Department of Pathology and Laboratory Medicine, and for 10 years as Dean for Educational Affairs. During his many years of devoted work for improving standardization and quantiication of immunohistochemstry for cancer diagnostics, he has published 400 papers and 20 books. Currently he is Editor in Chief of Applied Immunohistochemistry and Molecular Morphology. “ IHC is a precise immunoassay that must be performed only with a high degree of technical rigor and control where ready-to-use reagents, coupled with proven detection systems, ixed and validated protocols, recommended controls and automation, represent a pathway that could, if widely adopted, lead to improved levels of reliability and performance for IHC. ” Clive Taylor, MD, D.Phil Contributors Many people were involved in the creation of the Sixth Edition of Dako’s Guidebook to Immunohistochemical Staining Methods, and Dako would like to thank everyone who contributed. Special thanks go to: Clive Taylor, Elizabeth Colley, Ronald Stead, Shan-Rong Shi, Søren Nielsen, Lars Jacobsen *, Majken Nielsen *, Sofie Månsson*, Kenneth Pedersen*, Hans Christian Pedersen*, Joachim Schmidt*, Gale E. Pace*, Ole F. Rasmussen*, David McDonald*, Jan T. Jørgensen*, Jens Mollerup*, Rashmil Saxena, Sunil Badve, Rikke Jørgensen*, Helle G. Wendelboe*, Anette Lykke*, Susie S Jensen* and Lars Rudbeck*. Sections, in whole or parts thereof, from the previous editions of this Guidebook are used in the 6th edition. We sincerely thank and acknowledge the contribution of the authors. Special acknowledgements to: Thomas Boenish, James F. Happel, Mark Key, Jim Farmilo*, J. Paul Robinson, Jennifer Sturgis, George L. Kumar*, Ulla Henriksen*, Sven Müller*, Andreas Schönau*, Nanna K. Kristensen*, Lars Winther*, Kirsten Bisgaard*, Kenneth Bloom and Karen Atwood*. We truly value everyone’s contribution, and appreciate such dedicated participation. * Current or former Dako employee 8 Innovation and Quality | Preface Preface Innovation and Quality Welcome to the Sixth Edition of Dako’s Guidebook to Immunohistochemical Staining Methods. Dako earned its reputation for innovation and quality by introducing antibodies with standardized titer almost 50 years ago, and it is still Dako’s goal to continue to expand knowledge within the ield and continue on our path of scientiic advancement. The focus of this book is therefore to provide a comprehensive immunohistochemistry (IHC) resource for lab managers, lab technicians, learning pathologists, and students from around the world. For readers familiar with the previous editions of this guidebook, it should be noted that the structure of the new edition has changed slightly, so that the irst part covers the entire staining process from biopsy to inal analysis. The second half is comprised of the many supporting aspects within the ield of immunohistochemistry. Since the focus is solely on IHC, the in situ hybridization (ISH) method is not covered in this edition. Part I covers the immunohistochemical staining process, and includes a general introduction as well as chapters covering pre-analytical factors, antigen retrieval, selection of the primary antibody, staining protocol optimization, IHC staining methods and analysis of IHC stains. Part II examines the potentials and pitfalls in immunohistochemistry, with chapters on optimization of immunohistochemical reactions, automation in IHC, optimizing the laboratory worklow, companion diagnostics, tissue microarray, IHC visualization of molecular tests, controls, background and troubleshooting. We sincerely hope that the publication of this book will further enhance the advancement of the ield of immunohistochemistry, and will help new and practitioners within the ield continue to progress and drive the standardization process within IHC to improve diagnostic certainty. Treatment decisions are heavily inluenced by the immunohistochemistry results, thus making IHC important for the ultimate goal of better care of the patient. Lars Rudbeck PhD, Scientiic Editor, Dako 9 Part I: The Staining Process Chapter 1 Introduction to Immunohistochemistry Clive R. Taylor, MD, D.Phil Im • mu • no • his • to • chem • is • try (n.) Microscopic localization of speciic antigens in tissues by staining with antibodies labeled with luorescent or pigmented material. The American Heritage® Medical Dictionary Introduction to Immunohistochemistry | Chapter 1 Chapter 1.1 Introduction tochemical staining result start in the surgery operating room and end at the interpretation of the stain by the pathologist, Immunohistochemistry (IHC) is a method used to determine which ultimately leads to treatment decision by the oncologist. the expression of biomarkers in tissue. This educational guidebook will describe immunohistochemistry as it is used in the For those new to the world of immunohistochemistry here pathology laboratory as an aid in the differential diagnosis is a brief outline of the steps needed to localize antigens in and classiication of cancer, and for certain other diseases, tissues using antibodies for cancer diagnosis: including infections. The factors that inluence the immunohis- Grossing Accessioning Tissue processing & embedding Biopsy Sectioning Reporting Screening, interpretation & archive Staining Enzyme Secondary antibody Dextran Backbone Primary antibody Tissue antigen DAB STEP C STEP D STEP E Figure 1.1 From biopsy to reporting. 11 Chapter 1 | Introduction to Immunohistochemistry Pre-Analytical Steps Post-Analytical Steps 1. A Biopsy (surgically removed tissue specimen or needle 7. In the post-analytical process, the pathologist interprets the biopsy) from the surgery room arrives in ixative at the stains in context with positive and negative tissue controls, pathology laboratory. 2. In the Accessioning room the sample details are entered into the laboratory information system (LIS). A barcoded using bright ield microscopy. 8. The results are reported to the oncologist for treatment decision. label can ensure track and trace capabilities. 3. During Grossing, the specimen is visually examined for Processing Decalcification suspicious areas that require further examination. Samples from the specimen that require further microscopic testing are excised as tissue blocks and placed in barcoded cassettes. Embedding Vax, type, storage 4. Tissue processing and embedding are the steps where the tissue block is processed into a form and condition suitable Pre-analytic Tissue Type, dimensions, biological variation cauterization for making ultrathin microscopic sections. Typically, the tissue is ixed in formalin then dehydrated before it is embedded in parafin. 5. Sectioning is the ine art of cutting the parafin-embedded Sections Thickness, drying, storage Fixation Delay, time, type, volume tissue blocks into ultrathin (~4 µm) sections and placing them onto glass slides. A barcode on the slide can ensure trace- Platform Manual, automated ability and may also contain protocol information for the requested test for that particular section. Analytical Steps 6. Staining is the analytical part of the IHC process. It encompasses antigen retrieval, application of the primary antibody and visualization system, ending with counterstaining: Counterstain Time, color Pre-treatment Proteolysis, HIER, time, temp, pH Analytic Visualization system Sensitivity, specificity enhancement Primary antibody Clone, dilution buffer, time, temp a. Antigen retrieval is performed to recover the antigens that Chromogen Sensitivity, localization may have been altered by ixation; b. Endogenous enzymes are blocked (this step can also be performed after primary antibody incubation); c. A primary antibody is applied that speciically binds to the antigen of interest; d. The secondary antibody carries the label (enzyme); upon application it binds to the primary antibody; Interpretation Quantification, localization, pos./neg. def. cut-off level, panels, algorithms Control Internal/external, critical stain quality indicators Post-analytic e. Chromogen is applied to visualize the antibody/antigen complex; f. Counterstaining is performed to visualize nuclei and overall tissue architecture; Reporting Diagnostic context g. Sections are dehydrated, mounted and coverslipped. Figure 1.2 Many factors may inluence the IHC staining result. With just 3 choices at each of 14 steps there are 4.8 million different procedures! 12 Introduction to Immunohistochemistry | Chapter 1 Chapter 1.2 History of immunohistochemistry cein and visualized by ultraviolet light (luorescence microscopy) (1). During the next 25 years, the Coons method was used with This IHC Educational Guidebook will describe the potentials different modiications, including labeling with heavy metals, but and pitfalls in the immunohistochemical staining process from it was not until the introduction of enzyme-labeled antibodies (2) biopsy to interpretation, with special attention to the analytical that the method overcame many of the inherent issues with luo- processes and how to improve certainty in the staining result by rescein and heavy metal labeling of antibodies. In the early 1970s, employing standardization to the processes. application of the ‘immunoperoxidase’ method to formalin parafin embedded tissues by Taylor, Mason and colleagues in Oxford, Before immunohistochemistry reached its now widespread use was a critical step in extending use of the method into ‘routine’ as an important method in routine cancer diagnosis, the tech- diagnosis in anatomic pathology. The direct labeling method had nology had a long history of technological developments out- the drawback that each individual primary antibody, or the secon- lined in the table below. dary antibody, had to be labeled with enzyme. That problem was circumvented by the development of an unlabeled antibody en- Table 1.1 The major milestones in the history of immunohistochemistry. zyme method, the peroxidase anti-peroxidase (PAP) method, which had the further advantage of increased sensitivity, facili- Year Method References 1941 Fluorescence-labeled primary antibodies Coons et al (1) 1967 Enzyme-labeled primary antibodies Nakane & Pierce (2) 1970 Secondary un-labeled antibodies (PAP) Sternberger et al (3) and improved detection systems for visualization of antigens in 1970 Detection of antigens on ultrathin sections Kawarai & Nakane (4) tissue, IHC suffered from lack of reproducibility, due in part to poor 1974 Application to routine formalin paraffin sections Taylor et al (5-7) 1975 Invention of monoclonal antibodies Köhler & Milstein (8) Increased demand led to better quality reagents from the com- 1978 Double staining using un-labeled antibodies (APAAP) Mason & Sammons (9) mercial sector, with improved quality control of production 1979 Monoclonal antibodies to human antigens McMichael et al (10) 1988 Capillary gap semi-automated staining Brigati et al (11) 1991 Heat-induced antigen retrieval Shi et al (12) Dr. Niels Harboe, founder of Dako, realized in the early 1970s 1993 Standardization efforts as ‘Total Tests’ Taylor (13) the need for standardized antibody preparations for safe and 1995 Dextran-polymer-based detection system Dako reproducible diagnoses and began producing puriied poly- 1998 Immunohistochemistry as companion diagnostics Dako (HER2) 2007 Recommendations for improved standardization of IHC Goldstein et al (14) Wolff et al (15) Even with the puriied and highly speciic polyclonal antibodies Molecular HER2 CISH Tests in the IHC lab Invitrogen there was a need for improved speciicity of antibodies and a tating use in routine tissues. A related parallel development was the introduction of the alkaline phosphatase anti-alkaline phos- 2008 phatase (APAAP) in 1978 (9). Even with the development of new quality antibody reagents, and in part to the inconsistent and adverse effects of ixation. methods. Polyclonal antibody preparations differ between serum samples in afinity and speciicity, as the immuneresponse changes with time and immunization preparations, and as one animal is replaced by another as the source. clonal antibodies that had the same strength (as measured by titer) from batch to batch. greater variety in terms of target proteins. The invention, in 1975, of hybridomas that could produce monoclonal antibodies (8) resulted in the production of the irst monoclonal antibody that was Professor Albert H. Coons and co-workers demonstrated in 1941 highly speciic for human thymocytes using hybridoma technolo- that it was possible to localize antigens in tissue slices using an- gy (10). Monoclonal antibodies paved the way for a rapid growth tibodies against Streptococcus pneumoniae labeled with luores- in the use of IHC in research and diagnosis of cancer. 13 Chapter 1 | Introduction to Immunohistochemistry One other consequence of the lack of reproducibility was the Label development of automated instruments (11). Automation was invented with the fundamental thought that a properly functioning and maintained instrument will consistently perform its Primary antibody pre-programmed instructions in the same way – slide after slide- Tissue antigen Direct method (one step) The primary antibody (green) is labeled with an enzyme or luorescence. which is the principal reason why an instrument potentially can give superior reproducibility, compared with manual methods. However, progress was slow until 1991, when Shi et al (12) in- Enzyme troduced ‘antigen retrieval’ (or heat-induced epitope retrieval), thereby facilitating extension of IHC to a much broader range Secondary antibody of applications in formalin parafin sections, but at the cost of adding yet another variable to the process. This important publication on antigen retrieval thus gave new insights and impetus to efforts in standardization of IHC, leading to the intro- Indirect method (two steps) An enzyme-labeled secondary antibody reacts with unconjugated primary antibody bound to tissue antigen. Primary antibody Tissue antigen duction of the ‘Total Test’ concept (13) as a result of a series of meetings sponsored by the Biological Stain Commission and the FDA in the early 1990s. Peroxidase anti-peroxidase complex The standardization efforts, coupled with attempts to use IHC in a semi-quantitative setting raised demands to a new level, exempliied by the introduction, in 1998, of the HercepTest™ (Dako), which was the irst cancer companion diagnostic, in this instance designed for selection of breast cancer patients Secondary antibody for treatment with the new drug Herceptin® (Genentech/Roche). Clinical trials had shown that patients whose tumors over- Primary antibody expressed HER2 would beneit the most from Herceptin® treat- Tissue antigen ment. The HercepTest™ assay uses IHC on patient samples, Unlabeled method Pre-formed enzyme immune complex reacts with secondary antibody. in combination with control cell lines having known HER2 expression to determine if a breast cancer overexpresses HER2. Some 15 years later, this assay together with similar HER2 assays from other vendors, still serves as a rare example of a Dextran backbone HRP enzyme semi-quantitative IHC assay used in routine clinical pathology. The polymer-based visualization system, introduced shortly before HercepTest™, is the most widely used detection method in IHC today, with advantages of stability and high sensitivity. Secondary antibody, mouse/rabbit The technical advances in IHC in the last decade have been Primary antibody incremental, with little impact on the basics of the method. Tissue antigen Labeled polymer A long dextran polymer is labeled with both the secondary antibody and multiple enzyme molecules. Automation has become more advanced, including laboratory information system integration, with track and trace of samples, while whole slide digital imaging is slowly being integrated into the analysis of stain result. These advances can best be regarded as improvements in standardization; a process that started back 14 Figure 1.3 The development of detection systems used for IHC. Please see Chapter 6 for a full description of the many different detection methods. Introduction to Immunohistochemistry | Chapter 1 in the early 1990s and was re-emphasized in the 2007 publications by Goldstein et al (14) and Wolff et al (15), but also by the work being done e.g. estrogen receptor assessment (16, 17). The Chapter 1.3 Standardization in Clinical Immunochemistry vs. Immunohistochemistry critical importance of IHC standardization became evident with the revelation of disturbingly high numbers of false negative or For more than 30 years, clinical immunochemistry has em- false positive results in IHC determinations of ER (estrogen recep- ployed blood or urine samples to determine the concentra- tor) and PR (progesterone receptor) expression, and also HER2. tion of certain biomarkers, e.g. creatinine and cystatin C for In one example, a re-testing in 2007 of 1,023 breast cancer sam- evaluation of kidney function, and C-reactive protein as a ples from Newfoundland revealed that approximately 1 out of 3 marker of inlammation. Although clinical immunochemis- samples was scored falsely ER negative (17). As a consequence try covers a multitude of assay types, most of these tests of the false negative ER test results, these women were not ac- are based on the ELISA (enzyme-linked immunosorbent as- corded the potential beneit of anti-hormonal therapy. say) method, a method that closely parallels IHC in principle. One major difference is that International Reference Materials and Calibrators are used in clinical immunochemistry Immunohistochemistry publications (ELISA) to achieve quantitative results from these assays. 140.000 122198 115006 120.000 100.000 Immunohistochemistry is based on principles similar to the ELISA method, yet it is at best a semi-quantitative method for 94391 determination of the expression of biomarkers in tissue sam- 80.000 75352 60.000 ples. However, IHC should not be regarded as simply anoth- 61315 40.000 er ‘special stain’, like a PAS stain or a silver stain. IHC is essentially an ELISA method applied to a tissue section. In this 32626 respect, when correctly performed, IHC has the potential 14319 20.000 9044 3009 7423 to perform as a reproducible and quantitative tissue-based 2 01 7 20 08 -2 00 2 -2 03 20 19 98 -2 00 7 2 99 19 93 -1 99 7 19 88 -1 98 2 19 83 -1 98 7 19 78 -1 97 2 -1 97 73 19 -1 68 19 19 63 -1 96 7 0 Interval (Years) ELISA assay; much more than a simple stain. That the IHC method does not perform to this level, relects deiciencies in the application of the method, speciically inconsistent sample preparation, lack of reference or calibration standards, Figure 1.4 The number of IHC publications in the last 50 years. The data are from Pubmed using the search term “immunohistochemistry”. and inadequate validation of reagents (18, 19). If ELISA can use a standard curve to convert the measured immunoreactivity into a quantitative amount of tested protein, then IHC – in theory – can also convert the IHC intensity observed in The latest development in cancer diagnosis is the inclusion FFPE tissue sections into the amount of tested protein by an of molecular tests (FISH/CISH) in anatomic pathology labs, equivalent standard ruler. Comparative studies of IHC inten- driven by HER2 assessment requirements. Other techno- sity on frozen tissue vs. FFPE tissue have shown identical logies also are entering into the pathology lab and into rou- intensity by using an optimized AR protocol (20, 21), and tine diagnosis, and technologies such as array comparative similar protein quality is evident when examined by mass genomic hybridization or next generation sequencing will likely spectrometry (22), leaving no theoretical reason for lack of be a fundamental part of cancer diagnosis in the future. One true quantitative IHC assays. Nonetheless, today IHC as- ongoing goal is to interface these newer methods of molecu- says are at best no more than semi-quantitative, for reasons lar analysis with existing and improved morphologic criteria, a that are more of a practical nature. ield termed ‘Molecular Morphology’. 15 Chapter 1 | Introduction to Immunohistochemistry Chapter 1.4 Growing Consensus for Standardization Table 1.2 The Total Test: An IHC stain should be managed in the same rigorous manner as a clinical laboratory analysis. Modiied from Taylor (14, 24). Pre-analytic From the beginning there has been concern relating to the reproducibility of immunohistochemical methods as applied Test selection to formalin-ixed, parafin-embedded (FFPE) tissue sections. Specimen type A consequence of not controlling all parameters (in ixation, Acquisition, pre-fixation/transport time processing and staining) is poor day to day reproducibility within a single laboratory, and poor reproducibility among Fixation, type and total time Processing, temperature different laboratories. In recent years these concerns have increased and lack of standardization, well shown in inter- Test selection laboratory quality assurance surveys performed by NordiQC Antigen retrieval procedure and UK NEQAS, is now recognized as a major impediment to basic research, clinical trials, and direct patient care. Over the past three decades a number of conferences have been Selection of primary antibodies Protocol; labeling reagents held to address this topic and to seek constructive resolutions. Reagent validation Among the most productive were a series of meetings spon- Control selection sored by the Biological Stain Commission and the FDA in the Technician training/certification early 1990s, that led to recommendations for manufacturers concerning the precise description and validation of IHC rea- Laboratory certification / QA programs gents (23), and also highlighted the necessity to pay attention Post-analytic to all aspects of the IHC test procedure. The latter recommen- Assessment of control performance dation, borrowed from the much more rigorous protocols ap- Description of results plied to immunologic assays in clinical laboratories, became Interpretation/reporting known as the ‘Total Test’ approach (Table 1) (23, 24). A decade later a meeting of the FDA and NIST (National Institute of Pathologist, experience and CME specific to IHC Standards and technology) focused upon standardization of HER2 IHC assays, and the need for universal control materials In the analytical steps, antigen retrieval is the irst challenge. (reference standards) (25). Different antigens require different antigen retrieval for optimal staining results, and the different variations of the AR process add another variable that must be controlled. Antigen retrieval Chapter 1.5 Standardization Starts in the Surgery Room is described in detail in Chapter 3. Selecting the right antibody for the right marker is one of the While Table 1.2 only mentions a few of the major steps in a To- key steps in the analytical process. Some monoclonal anti- tal Test, the pre-analytical process alone contains at least 62 body clones are more speciic than others against the same identiiable steps of which 27 have been examined in published biomarker. In other cases a polyclonal antibody may be the research. Out of these 27 steps, 15 pre-analytical variables are best choice. Selection of the primary antibody is described in capable of impacting the immunohistochemistry staining result detail in Chapter 4. including ixation delay, ixative type, time in ixative, reagents 16 and conditions of dehydration, clearing, parafin impregnation Using a protocol that is optimized for the detection of the bio- and conditions of slide drying and storage (26). Pre-analytical marker is vital. The optimal protocol must be able to identify variables are described in detail in Chapter 2. the antigen of interest in cells and structures with both low and Introduction to Immunohistochemistry | Chapter 1 high expression. Optimization of the staining protocol is de- Post-analytical standardization is essential for prognostic or pre- scribed in detail in Chapter 5. dictive biomarkers, e.g. HER2 and ER/PR, adhering to speciied stain interpretation guidelines to give the sample a scaled score The inal step of the analytical process is the visualization of the (e.g. from 0-3+). However, most biomarkers are used for cell line- antigen/antibody reaction. Here the selection of the detection age and tissue identiication, where expression levels are usually system must consider the complexity of the visualization and the not as critical and interpretation is not linked to a semi-quantita- required ampliication needed to visualize the biomarker. The var- tive scoring system, but is reported as a binary ‘Yes’ or ‘No’ sys- ious detection systems are described in detail in Chapter 6. tem (positive or negative) for the tested biomarker. Digital analysis of IHC stains is described in Chapter 7. Table 1.3 Major steps affecting the immunohistochemistry staining result. Step Effect on IHC Biopsy Depending on the suspected cancer type, tissue samples can be obtained in different ways such as punch/core biopsy, excisional/incisional biopsy, etc. Tissue degradation begins at the time of sample removal. Fixation Embedding Sectioning and Mounting Antigen Retrieval The sample should be fixed as soon as possible after surgery, ideally within less than an hour. The chemical fixation crosslink proteins in the sample thereby stopping the degradation process. Too short or too long fixation can affect the staining result. After fixation, the sample is embedded in paraffin for long-term storage and to enable sectioning for subsequent staining. Once embedded in paraffin, samples can be stored (almost) indefinitely. Formalin-fixed, paraffin-embedded tissues are sectioned into thin slices (4-5 μm) with a microtome. The sections are then mounted onto adhesive-coated glass slides. Due to the fixation process, an antigen retrieval treatment is applied to unmask the epitopes, either by heat (heat-induced epitope retrieval; HIER) or enzymatic degradation (proteolytic-induced epitope retrieval; PIER). Incorrect antigen retrieval for the biomarker of interest will adversely affect the staining result. Chapter 1.6 Future Aspects for Standardization of Immunohistochemistry The consensus arising from the standardization efforts is that the reliability and reproducibility of IHC methods in routine surgical pathology have been greatly hindered by two key factors. 1. While reagents available for IHC have increased in quality, there has been an even greater increase in number of sources and variety of staining methods. This plentitude of reagents contributes to lack of standardization in signiicant ways, that in theory are manageable by good technique and use of proper controls, but in practice have led to requirements for such high standards of excellence in the technical process, that many laboratories cannot ind suficient, or suficiently skilled, staff to comply. 2. The usual method of sample preparation for tissue remains as formalin ixation and parafin embedment (FFPE). This venerable approach may be satisfactory for the preservation of morphologic detail, but does adversely affect the antigenicity of many target molecules in the tissue, to degrees that are unknown. The enormous variation in protocols (including Primary Antibody An antibody with specificity for the biomarker of interest is applied. The specificity and sensitivity of the antibody affect the staining result. ixation times) employed for FFPE among different laborato- Visualization The antigen/antibody complex signal is amplified and visualized using a detection system. The strength of amplification of the reaction affects the staining result (intensity). current poor reproducibility. Interpretation The staining pattern is assessed by a pathologist in context with other biomarkers, controls and other tests (e.g. H&E, special stains. Inter- and intra-observer variability is common, especially for semi-quantitative assays. This variability highlights the importance of training and inter-calibration. ries, or within the same laboratory from specimen to specimen, compounds the problem and contributes to the While several decades have passed, these issues have not been satisfactorily addressed. Legions of investigators, and many manufacturers, have addressed different aspects of the problem, focusing upon better sample preparation (ixation), more effective methods of antigen retrieval, improved 17 Chapter 1 | Introduction to Immunohistochemistry reagents, more sophisticated automated platforms, more sen- „„High-quality reagents are available, with highly sensitive sitive detection methods, and the development of reference detection methods, but they must be employed pro- standards or controls (13, 23-25). perly in controlled fashion, and currently often are not. Participation in EQA schemes can help laboratories In order to improve the quality and reproducibility from sample to improve the reproducibility; sample, and lab to lab, the accreditation process for many pa- „„There is a pressing need for tissue-based IHC controls (or thology laboratories now includes participation in external quality ‘reference standards’) (19, 25) that can be made available assurance (EQA) schemes. EQA organizations, like NordiQC, UK to all laboratories performing IHC assays, somewhat ana- NEQAS and CAP, are independent organizations not associated logous to the international reference standards and cali- with commercial suppliers. Their role is to promote the quality of brators that are available to clinical laboratories performing immunohistochemistry (and in situ hybridization) by arranging ELISA testing. external QA schemes for pathology laboratories. Similar EQA schemes are now available in many countries and regions around From this brief discussion it follows that to improve standardi- the world. The purpose of EQA schemes is to improve the quality zation to the point that all laboratories would carry out the IHC of staining results in the participating laboratories; thus it is the in identical fashion for every phase of the ‘Total Test’; it would individual labs that are being assessed. It is their choice of an- require them to use the same ixative and ixation time (adjusted tibody, visualization system, instrumentation and protocol that is to tissue type), the same antigen retrieval process, the same pri- the basis for the EQA organization's evaluation and feedback. A mary antibodies and detection systems, with the same automat- lab volunteers to participate in the assessment runs. Laboratories ed stainer and common controls. Clearly this perfect option will typically enroll for a year, during which they receive approximate- never happen, and we therefore must do what we can to reduce ly 16 unstained tissue slides (NordiQC), or 7-8 different modules, the consequences of the variables in the process. where each module usually has two tissue slides (UK NEQAS), to stain using their own internal standard protocols for those markers Ultimately the overriding factor in effecting signiicant change designated by the QA organization. The labs return the stained must be to transform the mindset of pathologists, at least of slides to the QA organization for assessment, which is conducted the next generation, to the view that the end result of an IHC by experts engaged by the organization. The labs receive either protocol is not just a ‘stain’, with intensity to be adjusted at the a “Passed” rating or “Not Passed” rating. Both NordiQC and UK whim of the pathologist. Rather IHC is a precise immunoassay NEQAS inform all participants of their individual scores and pro- that is strictly quantiiable, and must be performed only with a vide suggestions for protocol optimization when required. Both or- degree of technical rigor and control that matches any other ganizations present the anonymous results on their web sites, with immunologically-based assay of like principle (namely ELISA). statistics and best method for the particular marker. ELISA is a ‘gold standard’ method for quantitative assays in the clinical laboratory. ELISA reagents are purchased in prepared CAP (College of American Pathologists) in the US, has a simi- form, with all of the necessary reagents, deined protocols, lar QA process, but requires only the return of stain results and and reference or calibration standards, for use with speciied interpretation, not the stained slides. instrumentation. Ready-to-use reagents, coupled with proven detection systems, ixed and validated protocols, recommend- Some broad conclusions are possible: „„Resolution of the problem of pre-analytical sample prep- aration is not imminent; the practical aspects of developing tissue handling and ixation procedures that it the daily routine of every hospital are challenging. Importantly the logistical obstacles to implementation of standardized sample preparation procedures worldwide are formidable; 18 ed controls and automation, represent an analogous pathway that could, if widely adopted with appropriate controls, lead to improved levels of reliability and performance for IHC. Introduction to Immunohistochemistry | Chapter 1 References 1. Coons AH, Creech HJ, Jones RN. Immunological properties of an antibody containing a luorescent group. Exp Biol Med 1941; 47:200-2. 2. Nakane PK, Pierce GB. Enzyme-labeled antibodies for the light and electron microscopic localization of tissue antigens. J Cell Biol 1967; 33:307-18. 3. Sternberger LA, Hardy PH, Cuculis JJ, Meyer HG. The unlabeled antibody enzyme method of immunohistochemistry preparation and properties of soluble antigen-antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use in identiication of spirochetes. J Histochem Cytochem 1970; 18:315-33. 4. Kawarai Y, Nakane PK. Localization of tissue antigens on the ultrathin sections with peroxidase-labeled antibody method. J Histochem Cytochem 1970; 18:161-6. 15. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American society of clinical oncology/college of american pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med. 2007;131:18-43. 16. Yaziji H, Taylor CR, Goldstein NS, Dabbs DJ, Hammond EH, Hewlett B, et al. Consensus recommendations on estrogen receptor testing in breast cancer by immunohistochemistry. Appl Immunohistochem Mol Morphol 2008; 16:513-20. 17. Hammond ME, Hayes DF, Wolff AC, Mangu PB, Temin S. American society of clinical oncology/college of american pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Oncol Pract 2010; 6:195-7. 5. Taylor CR. The nature of reed-sternberg cells and other malignant "reticulum" cells. Lancet 1974;2:802-7. 18. Taylor CR. Quantiiable internal reference standards for immunohistochemistry: The measurement of quantity by weight. Appl Immunohistochem Mol Morphol 2006; 14:253-9. 6. Taylor CR, Burns J. The demonstration of plasma cells and other immunoglobulin-containing cells in formalin-ixed, parafin-embedded tissues using peroxidase-labelled antibody. J Clin Pathol 1974; 27:14-20. 19. Taylor CR, Levenson RM. Quantiication of immunohistochemistry — issues concerning methods, utility and semiquantitative assessment ii. Histopathol 2006; 49:411-24. 7. Taylor CR, Mason DY. The immunohistological detection of intracellular immunoglobulin in formalin-parafin sections from multiple myeloma and related conditions using the immunoperoxidase technique. Clin Exp Immunol 1974;18:417-29. 8. Kohler G, Milstein C. Continuous cultures of fused cells secreting antibody of predeined speciicity. Nature. 1975;256:495-7. 9. Mason DY, Sammons R. Alkaline phosphatase and peroxidase for double immunoenzymatic labelling of cellular constituents. J Clin Pathol 1978;31:454-60. 10. McMichael AJ, Pilch JR, Galfre G, Mason DY, Fabre JW, Milstein C. A human thymocyte antigen deined by a hybrid myeloma monoclonal antibody. Eur J Immunol. 1979;9:205-10. 11. Brigati DJ, Budgeon LR, Unger ER, Koebler D, Cuomo C, Kennedy T, et al. Immunocytochemistry is automated: Development of a robotic workstation based upon the capillary action principle. J Histotech 1988; 11:165-83. 12. Shi SR, Key ME, Kalra KL. Antigen retrieval in formalin-ixed, parafin-embedded tissues: An enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 1991; 39:741-8. 13. Taylor CR. An exaltation of experts: Concerted efforts in the standardization of immunohistochemistry. Appl Immunohistochem 1993; 1:232-43. 20. Shi SR, Liu C, Pootrakul L, Tang L, Young A, Chen R, et al. Evaluation of the value of frozen tissue section used as "gold standard" for immunohistochemistry. Am J Clin Pathol 2008; 129:358-66. 21. Shi SR, Liu C, Taylor CR. Standardization of immunohistochemistry for formalin-ixed, parafin-embedded tissue sections based on the antigen-retrieval technique: From experiments to hypothesis. J Histochem Cytochem 2007; 55:105-9. 22. Shi SR, Liu C, Balgley BM, Lee C, Taylor CR. Protein extraction from formalin-ixed, parafin-embedded tissue sections: Quality evaluation by mass spectrometry. J Histochem Cytochem 2006; 54:739-43. 23. Taylor CR. Quality assurance and standardization in immunohistochemistry. A proposal for the annual meeting of the biological stain commission, june, 1991. Biotech Histochem 1992;67:110-7. 24. Taylor CR. Report from the biological stain commission: Fda issues inal rule for classiication/reclassiication of immunochemistry (ihc) reagents and kits. Biotech Histochem 1998; 73:175-7. 25. Hammond ME, Barker P, Taube S, Gutman S. Standard reference material for her2 testing: Report of a national institute of standards and technology-sponsored consensus workshop. Appl Immunohistochem Mol Morphol 2003;11:103-6. 26. Engel KB, Moore HM. Effects of preanalytical variables on the detection of proteins by immunohistochemistry in formalin-ixed, parafin-embedded tissue. Arch Pathol Lab Med 2011; 135:537-43. 14. Goldstein NS, Hewitt SM, Taylor CR, Yaziji H, Hicks DG. Recommendations for improved standardization of immunohistochemistry. Appl Immunohistochem Mol Morphol 2007; 15:124-33. 19 Part I: The Staining Process Chapter 2 Fixation and Other Pre-Analytical Factors Elizabeth C. Colley, MLT, ART Ronald H. Stead, PhD, FRCPath Pre-an • a • lyt • i • cal (n.) Describing any variable whose value can affect the outcome of a subsequent analysis. The American Heritage® Science Dictionary Fixation and Other Pre-Analytical Factors | Chapter 2 Chapter 2.1 Introduction Chapter 2.2 Tissue Handling Immunohistochemistry (IHC) and in situ hybridization (ISH) tech- The pre-analytical stage begins as soon as a piece of tissue is re- niques have advanced considerably since the irst edition of this moved from its nutritional source (blood supply) and the time to ix- handbook was published in 1983. The need for standardization ation is critical (1). Degeneration is caused primarily by autolysis, was emphasized in that publication and is still very relevant today, which is a process of self-digestion by enzymes contained within to ensure meaningful interpretation and allow quantitative micros- cells; and this begins immediately. This process is accelerated by copy. The histological process, which begins with the acquisition increased temperatures. Fixatives are used to stop degeneration, of tissue samples and continues through to the interpretation of while preserving the structure and integrity of the tissue elements IHC or ISH results, can be broadly broken down into three main as much as possible. However, ixation itself introduces artifacts stages: pre-analytical, analytical and post-analytical. This chap- and the ideal ixative would also maintain the structure of all of the ter will address some of the pre-analytical issues that should be epitopes in the tissues. This is not achievable, as the alteration in considered when preparing sections for IHC and/or ISH. Pre-an- chemical structure caused by ixation necessarily modiies at least alytical variables can signiicantly and adversely affect the accu- some epitopes. For IHC and ISH procedures it is critical that the rate detection of targets. Many of the comments made below are tissue does not dry out during any stage of the tissue handling and based on the large number of publications about prognostic/pre- slide preparation (1). Drying may cause morphological changes, dictive (Class II) markers, such as ERBB2 (HER2), however, most such as poorly deined chromatin; and subsequently alter the of the points are broadly applicable regardless of the target under structure of the target particularly along the edge of the tissue. investigation or the method of choice. This could inhibit ligand binding and is particularly applicable to small specimens such as endoscopic biopsies. Additionally, dry tissue is more adsorbent, which increases the risk of non-speciic or unwanted adsorption of reagents during staining procedures, thereby interfering with interpretation of results. Surgery (cold ischemia) De-waxing and hydration Fixation Cold Ischemic Time Recently, there has been more of a focus on “cold ischemic Accessioning (documentation) time” and the impact this may have on IHC and ISH results. The duration of cold ischemia is calculated from when the tissue is removed from the body to when the tissue is placed into ixative. Collecting and drying Grossing (fixation) This time should be as short as possible, with published guidelines of one hour or less (2, 3). The deleterious effects of delayed ixation are illustrated in Figures 2.2 and 2.3; and may include in- Further fixation Processing creased, decreased or de-localized immunoreactivity. It should be noted that deterioration of an epitope due to ischemia cannot Embedding (paraffin) Floating Sectioning be recovered using antigen retrieval techniques. Relatively little has been published on the ischemic effects for speciic antigens or molecular targets which are Class I. Nonetheless, Figure 2.2 shows that the pattern of staining change with the three Class I targets illustrated. Perhaps, a broader under- Figure 2.1 An overview of the processing steps included in the pre-analytical phase. standing of the interrelationship between ischemic time and different targets will be easier, once the recording of ischemic times becomes a part of required documentation for all specimens. 21 Chapter 2 | Fixation and Other Pre-Analytical Factors 0 hour 1 hour 2 hour 4 hour Accessioning and Documentation When the specimen is received in the laboratory it is ‘acces- HER2 sioned’ and given a unique, traceable number. The documentation (requisition) which accompanies each surgical specimen should include: patient and physician information, date of procurement, clinical information, specimen site and type, collection Figure 2.2 Cold ischemia alters the staining intensity of HER2 in MDAMB-453 cells (2+ cell line). Weak to moderate membrane immunoreactivity on approximately half of the cells is observed in a cell pellet ixed immediately in 10% NBF (0 hour). With as little as one hour cold ischemia (the cell pellet was kept moist under saline-damped gauze), the morphology is already deteriorating and there appears to be increased numbers of cells with membrane staining. After two hours, the staining is even stronger. Following four hours cold ischemia time, much of the membrane staining is lost and the preservation of the cells is poor. This illustrates the need for prompt ixation and that different cold ischemic times can give rise to over staining or under staining of the cell membranes. Cells were stained using IHC and HER2 antibody (Code A0485, Dako) and the Autostainer Link 48, Dako platform. time, cold ischemic time, type of ixative and duration of ixation (3). If it is necessary to decalcify a specimen, then that information must also be recorded, including: time in ixative before decalciication, the type of decalciication used, the length of decalciication and any post-decalciication treatment (4). Part of the sample veriication process during accessioning is to conirm that the information on the requisition matches that on the specimen container. The specimen container should have a minimum of two identiiers such as patient name and date of birth. It is well published that the length of time in 10% neutral buffered 0 hour 1 hour 2 hour 4 hour formalin (NBF) has a variable impact on IHC and ISH results. Yet Ki-67 getting the required information for all specimens can be a challenge. For example a survey conducted in 2010, to determine the compliance of 757 laboratories with the ASCO/CAP (American Society of Clinical Oncology / College of American Pathologists) guidelines for HER2 testing, showed approximately 28% p53 of respondents did not include ixation information in case reports (5). Reporting the ixation time is valuable for interpretation and troubleshooting aberrant or unexpected results. It may also Cyclin D1 inluence protocol choice, such as the type or timing of antigen Figure 2.3 MDA-MB-453 (Ki-67 & Cyclin D1) and MDA-MB-231 (p53) cell pellets were ixed immediately in 10% NBF (0 hour) or held for one hour, two hours or four hours before transferring to NBF (the cell pellets were kept moist under saline damped gauze prior to ixation). For Ki-67 and p53 progressively more 'connective tissue' staining is observed with increased cold ischemia time, due to de-localization of the antigens from the nuclei. For Cyclin D1, progressive loss of staining is observed, with some de-localization. Sections were stained using the Autostainer Link 48, Dako platform, with FLEX detection and the following Dako RTU primary antibodies: MIB-1 (Ki-67), DO-7 (p53) and EP12 (Cyclin D1). retrieval or enzyme digestion required and the choice of control material (6). It is the responsibility of the collection team (often the clinical or surgical team) to provide the sample information, including the duration of cold ischemia and the time the specimen was placed in ixative. It is the responsibility of the pathology team to set guidelines which clearly outline criteria that determine if a specimen is acceptable for IHC or ISH staining. Grossing Once a specimen is deemed acceptable, it is examined macroscopically. This is referred to as grossing and it is a critical pre-analytical step which requires proper training. Larger specimens should be ‘bread loafed’ (sliced) into approximately 5 mm sections and placed in 10% NBF. Gauze or paper towel may be placed between the slices to facilitate exposure to the ixative. Care must be taken to handle each type of tissue in a standard- 22 Fixation and Other Pre-Analytical Factors | Chapter 2 ized manner and not to physically damage the tissue. Usually, it during ixation. The important point is that the rate at which each is necessary to select areas of interest from a larger specimen. of these reactions takes place is different; and the reaction rates These pieces of tissue, or blocks, should be trimmed such that are all slower than the penetration rates. It is these differences the size does not exceed 20 mm in length and width, or 4-5 which have led to confusion about what is an acceptable length mm in depth. Trimmed tissue is then placed into a processing of time for 10% NBF properly to ix a variety of specimens. Tissue cassette and submerged immediately into the desired ixative measuring 4 mm thick should be ixed for at least 24 hours at (usually 10% NBF). The volume of ixative should be approxi- room temperature (10). The ASCO/CAP recommendation guide- mately 10 to 20 times that of the specimen (3, 7). Formalin enters lines for ixing tissue state that 10% NBF ixation, prior to IHC or the tissue relatively quickly, but the chemical processes which ISH labeling of ER, PgR and HER2 cannot be less than 6 hours actually ix the tissue are more time consuming, taking at least and should be no more than 72 hours (2b, 3). The reader should 24 hours (8). When calculating total time in ixative, the time the be aware of the recently published update to the 2007 ASCO/ specimen sits in 10% NBF in the grossing area and on the auto- CAP HER2 breast cancer guideline (2a) which has changed the mated tissue processor must be included. ixation recommendations to 6-72 hours for HER2 (2b), aligning this with the recommendations for hormone receptors. The recommendation for HER2 ixation was originally made for breast Chapter 2.3 Fixation cancers but has been extended to include gastric cancer. These recommendations represent a compromise for the sake of speed, Part of the challenge when choosing a ixative is the fact that ignoring data, noted above, that ixation should be for at least 24 the amounts of antigens or molecular targets within tissue hours (also see below). specimens are inite. Although ixatives are meant to preserve these elements from elution, migration or degradation The ASCO/CAP guidelines also recommended some exclusion by changing the protein structure, they may destroy or mask criteria for HER2, ER and PR ixation, speciic to different types these targets. There are many ixative recipes, but most of of samples, which are unchanged from 2007 to 2013 (2a, 2b). these can be grouped into three main categories: those These exclusion criteria are: “tissues ixed in other than neutral containing formalin; those containing alcohol; and those buffered formalin; needle biopsies ixed for less than one hour containing a combination of both. Regardless of the ixative in neutral buffered formalin; and excisional biopsies ixed in for- of choice, the preparation and use of that solution must be malin for less than six hours or longer than 48 hours” (2a). It is consistent. Of the many pre-analytical variables which affect assumed, however, that the last exclusion criterion should have IHC and ISH results, ixation is probably the most signiicant, been changed in the 2013 Guideline Update (2b) to 6-72 hours impacting many other variables such as antigen retrieval and to be consistent with the new ixation recommendations for HER2. epitope binding. Unfortunately, to date, no single ixative has It is further emphasized that ixation outside of the recommended proven to be ideal for all targets and detection methods. How- parameters (especially with other ixatives) must be fully validated ever, it is generally more deleterious for tissue to be ‘under- in the user’s laboratory; and that if testing is performed on tis- ixed’, rather than ‘overixed’. sues ixed outside of the recommended ranges, this fact must be included in the report (2a, 2b). In the opinion of the authors, 10% NBF these recommendations and exclusion criteria are inadequate. A The most frequently used ixative is 10% NBF with pH 7.0 to 7.4 minimum of 24 hours ixation should be applied to all samples. It (9). This ixative is the ‘gold standard’ and has traditionally been is very important that tissue be “properly” ixed and that suficient used by pathologists, perhaps because the ingredients are rela- time is given to ensure completion of this process. tively inexpensive and the solution is simple to prepare and stable when stored. Formalin ixes by penetrating the tissue and form- Due to its cross-linking characteristic, NBF is an especially good ing cross linkages between reactive amino groups in proteins. Of ixative for small molecules such as hormones (1). It is the pro- course, this is an oversimpliication of what is actually happening gressive cross-linking nature of formalin ixation that often leads 23 Chapter 2 | Fixation and Other Pre-Analytical Factors to the masking of potential IHC or ISH targets, meaning that an Antigen retrieval techniques have been developed to help end point for ixation is almost as important as a start time. There ‘de-mask’ many targets after NBF ixation; however, without are many examples of situations that have led to incorrect inter- complete information, it is dificult to choose a suitable an- pretation of staining patterns caused by the time in ixative. One tigen retrieval protocol that will produce accurate results. example is elution of Ki-67 and cyclin D1 protein from the nuclei Modern heat-retrieval methods seem to work over relatively of ‘improperly ixed’ samples (Figure 2.4). broad ixation times (Figure 2.5); however, the timing of heat retrieval may need to be adjusted in tissues ixed for short 10% NBF 70% Alcohol or for excessive times, which is somewhat of a moot point in Ki-67 laboratories using the recommended 24 hour minimum NBF time. With proteolytic retrieval methods, it has been known for years that the necessary duration of trypsinization or pep- Cyclin D1 sinization is directly proportional to the duration of ixation, as illustrated in Figure 2.5. It must be emphasized that the need for retrieval and the optimum protocol may differ from epitope to epitope – not by antigen, so different monoclonal Figure 2.4 SK-BR-3 cell pellets were ixed in 10% NBF overnight before transferring to 70% alcohol (left), as the irst step in tissue processing, or directly into 70% alcohol before processing (right). The lack of ixation in the latter resulted in loss of immunoreactivity for Ki-67 and cyclin D1. Sections were stained using the Autostainer Link 48, Dako platform, with Flex detection and the following Dako RTU primary antibodies: MIB-1 (Ki-67) and EP12 (Cyclin D1). antibodies may require different methods. Since antigen retrieval methods are discussed in Chapter 3, these will not be given additional attention here. Formalin is considered a Class II carcinogen. Even formalin fumes have the ability to ix. This makes it a prime target for safety concerns. Good laboratory practices should always be 32 hour 56 hour 104 hour High pH 8 hour employed when handling this product. Many alterative ixatives to 10% NBF have been proposed, and perhaps there will be a trend away from formalin in the future, towards more ‘molecular friendly’ ixatives. The logistical and practical issues of switch- Trypsin ing to another ixative are, however, enormous. Alcohol Fixation When a tissue is ixed in 10% NBF for less than 6 hours prior Figure 2.5 Length of ixation affects immunoreactivity for IgM in tonsil mantle zone B cells. This igure illustrates reduced immunoreactivity for IgM in B lymphocytes in the mantle zone of secondary lymphoid follicles in tonsils ixed for several days. It further illustrates the fact that different methods of antigen retrieval produce different results, when attempting to recover the immunoreactivity. In the four images in the top row, the tissues were ixed for 8, 32, 56 and 104 hours and the mantle zone staining is approximately the same in the sections which were treated with Dako target retrieval solution (high pH) for 20 minutes at 97 °C. However, in the bottom row, the target retrieval was performed using trypsin (0.1% for 30 minutes). In this case, the antigen is unmasked when tissue is ixed for 8 and 32 hours but signiicantly depleted in tissues ixed for 56 and 104 hours. Extending the trypsin digestion time would damage the shorter ixed samples but result in stronger staining for IgM in the tissues ixed for two or four days (not shown). This type of comparison needs to be made for each new antigen being validated in the laboratory. 24 to being processed through to parafin, the tissue is probably ixed in alcohol, or has a variable combination of formalin and alcohol ixation. This non-standardized type of ixation may cause false negative or positive results (see Figure 3). Alcohol ixes by coagulating and precipitating proteins and tends to extract tissue elements such as low molecular weight carbohydrates. It also tends to dehydrate the tissue which causes shrinking and hardening. Alcohol ixation has an advantage over formalin ixation, in that it generally eliminates the need for antigen retrieval. It initially penetrates and ixes tissues more readily than formalin (although penetration slows down subsequently), and is often recommended for nucleic acid work. Fixation and Other Pre-Analytical Factors | Chapter 2 Enhancing Fixation modern rapid tissue processors which employ alternative ixa- There are some ixation methods which incorporate microwave tives and microwave enhanced processing (as well as small or ultrasound technology. Heat is generated by the excitation specimen size). This combination allows an H&E diagnosis on of molecules in both methods and accelerates the rate of re- parafin sections the same day. Nonetheless, the morphology actions (1). This effect also speeds up the penetration of the will differ from routine FFPE processed samples; and IHC and solutions by relaxing the cell structure. Microwave ixation may, ISH methods will require complete re-validation, as some of however, induce uneven tissue ixation that can vary with the these will not need pre-staining antigen retrieval, whether this is size and composition of the specimens and type of microwave of the heating or proteolytic type (personal observation). used (10). Directly, microwaving tissue causes protein coagulation and can lead to hard or ‘over-cooked’ tissue. Section Preparation Generally, unless otherwise speciied by a protocol of choice, sections for IHC or ISH are cut at 3 µm, 4 µm or 5 µm. Thicker Chapter 2.4 Tissue and Slide Processing sections may cause dificulty during staining, and also problems in interpretation due to the multi-layering of cells. After Not all tissue processors are the same and these differences sections are cut they are usually loated on water and picked are most apparent when tissues are not ixed adequately. How- up onto glass slides that are coated with some adherent ma- ever, most instruments will produce satisfactory results, most terial. Sections must lay lat against the glass to prevent lifting of the time. Some of the basic principles of processing and during staining or bubble formation, which may trap staining slide preparation are discussed below. reagents. The more points of adhesion the more likely the tissue will remain ixed to the slide, supporting the need for thin- Tissue Processing ner sections. Some commercially available slides come with a During tissue processing, ixation reagents containing water positive charge that attracts the negative charges of tissue pro- are replaced by wax (polymer, non-polymer and microcrystal- teins. These charged slides are especially effective following line formulas exist) which is done through a series of passages formalin ixation of tissues, since formalin blocks amino acids through increasing concentrations of alcohol, up to 100% (ab- in tissues, rendering the tissue more acidic and therefore more solute) alcohol. This process is followed by clearing the alcohol negatively charged. Different manufactures of staining plat- from the tissue (for example by using xylene) and replacing it forms may recommend the use of particular slides to achieve with molten wax. Low melting temperature (45 °C) as opposed optimal results. As with every other pre-analytical step, cutting to higher melting temperature (65 °C) waxes have been re- and mounting sections onto glass slides, and all steps prior to ported to produce better staining results for IHC, particularly in staining must be standardized. For example, if the slides are to T-lymphocyte staining (7). Next, the parafin iniltrated pieces left at room temperature for 15 minutes, in an upright position of tissue are embedded to form blocks, which are easily han- to allow draining of excess water and then heated in staining dled, cut and subsequently stored. rack at 60 °C for 30 minutes prior to staining, this step must be repeated every time sections for IHC or ISH are prepared. Fi- Rapid Tissue Processing nally, the changes resulting from block and section storage pri- There is increasing pressure to shorten turnaround times (TATs) or to IHC and ISH staining may also affect staining results (11). in tissue pathology, so that patients do not have to wait days to For example, it is recommended that sections cut for HER2 receive their pathological diagnoses. However, the laboratory testing should not be used if they are more than 6 weeks old. staff still need to ensure that samples are properly ixed (>24 hours in NBF, even for needle biopsies), to make sure that va- Dewaxing and Hydration lidated IHC and ISH methods are used. As noted above, tests Wax must be removed completely from the tissue sections, so used after alternative ixation and processing must be fully that aqueous antibodies, molecular probes and detection re- re-validated. This requirement is particularly applicable to the agents can penetrate and adhere to the tissue. Traditionally, 25 Chapter 2 | Fixation and Other Pre-Analytical Factors dewaxing was done by immersing the sections into a dewax- a large number of sponge-containing specimens are pro- ing solution (such as xylene), with or without prior brief heating. cessed simultaneously. Consideration must be given to this This step was followed by absolute and graded hydrating contingency when choosing appropriate processing sched- solutions (generally alcohols), until the inal solution: water. If ules and reagent replenishing practices. To eliminate this xylene is used to dewax sections, approximately 50 slides per variable, the use of nylon biopsy bags, which do not retain 50 mL of xylene is the limit, before it is no longer effective and signiicant amounts of reagent, is recommended. residual wax begins to cause artifacts in the stained tissue. Biopsy specimens can often be very dificult to see, espeToday, there are many commercially available staining plat- cially after parafin processing. Adding a small amount of forms which include onboard removal of wax and rehydration alcoholic eosin to the processing alcohols can make the tis- of the tissue sections. The accumulation of residual wax may sues more visible, by tinting them slightly. This simple pro- be a problem with these instruments, if rinsing is insuficient or cedure does not appear to have any detrimental effect on if solutions are not replenished regularly. subsequent IHC or ISH staining, however, the use of eosin or other biopsy coloring luids should be tested prior to implementation. An example of another stain used for this Chapter 2.5 Special Tissue Preparations purpose is Mercurochrome. Notwithstanding the fact that this mercurochrome should not be used for safety reasons, Biopsy Specimens its application diminishes or abolishes the signal in FISH There is a trend away from invasive surgical procedures to- assays and causes excessive background luorescence wards less invasive biopsy techniques for pathological diag- (personal observation). nosis (12). The cost effectiveness of these procedures and 26 improved imaging capabilities during biopsy procurement, Frozen Sections support the notion that these types of specimens will become If tissue targets cannot be demonstrated following ixation and even more prominent in pathology laboratories. Handling of parafin processing, then alternative methods such as rapid these biopsy specimens can present challenges for the la- freezing of the tissues, may be considered, keeping in mind that boratory. For example, these tiny fragments or cores of tissue frozen tissue is not routinely available, and is dificult to obtain require greater dexterity during grossing, embedding, cutting for reasons of logistics and expense. As with ixed sections, the and sometimes staining, making them more time consuming. protocols for obtaining suitable rapidly frozen tissues sample and Also, as noted above, they are particularly susceptible to dry- preparing frozen sections must be standardized and validated for ing artifacts and often exhibit structural damage from the bi- all antibodies, antigens, or molecular targets; as well as detection opsy procedure per se. They limit the amount of tissue avail- methodologies. An example of a situation when a frozen section able for microscopic evaluation and limit the tissue elements may be preferred to a formalin ixed specimen is during the in- available for IHC or ISH targeting. These types of specimens vestigating of autoimmune or inlammatory diseases or disorders are often processed using accessories such as biopsy bags of the skin. Using the Direct Immunoluorescence (DIF) technique or sponges, in an attempt to reduce the risk of sample loss and frozen sections, immunoglobulins and complement in skin during processing and to help maintain the architecture of biopsies can easily be demonstrated. An advantage of frozen the specimen by mitigating folding or wrinkling, which can sections is that they can be prepared in less time than traditional interfere with reagent low, staining and interpretation. If, parafin processed sections. The most challenging part of prepar- for example, sponges are used to secure biopsies in tissue ing these types of sections is freezing the tissue rapidly (quench- cassettes, the sponges themselves will absorb and retain ing or snap-freezing using liquid nitrogen). Freezing, thawing reagents more readily than tissue alone. Hence, excess and then re-freezing specimens causes freezing artifacts, which amounts of absorbed processing luids may be transferred destroy morphology and subsequently may affect the integrity of from one processing container to another, particularly when IHC or ISH. This problem is particularly applicable to control ma- Fixation and Other Pre-Analytical Factors | Chapter 2 terial, which is often re-cut multiple times. Frozen sections cut for uous supply of some these types of tissues, especially tumors; IHC and ISH should be between 4 µm and 6 µm thick, although and an alternative approach is to use deined cell lines. thicker sections may be required. Fatty specimens, particularly breast tissue, are dificult to cut because fat does not freeze as Control Cell Lines well as the rest of the tissue (except at temperature low enough to The use of cell lines would theoretically, provide a long-term cause shattering of the tissue). Fat continues to be a nemesis of supply of material that contains speciied levels of expression frozen section preparation. for various antigens (11). Using cells as controls is not novel and is incorporated in the HercepTest™ kit (Dako, Denmark) Post-ixation of frozen sections must be standardized for each as well as assays from other manufacturers. According to the target under investigation. Many of the colorimetric IHC and ISH HercepTest™ kit insert, if the cell controls provided are used protocols used on FFPE sections include the use of an enzyme, in association with the recommended platform and in-house such as peroxidase, to produce the inal color. Hence, quenching controls, they provide valuable information on assay validity for of endogenous enzyme (i.e. peroxidase) activity is often incorpo- the semi-quantitative assessment of HER2 overexpression. Cell rated into IHC and ISH protocols; however, these steps are not lines are commercially available and can be cultured, harvested usually included when frozen sections are stained. This omission and pelletized, before being ixed and processed in the same potentially can interfere with interpretation of the results. manner as test specimens (11, 13; see Figure 2.6). The selection of appropriate cells and the method of processing these Unixed, frozen tissue may contain viable human pathogens or for each antibody, molecular marker, detection protocol, and in- toxins. Universal precautions must to be employed when hand- terpretation method must be validated prior to implementation. ling these types of specimens. Formalin is known to inactivate 0 tissue samples (except prions) and archived tissue should be formalin ixed, prior to discard. Chapter 2.6 Control Material Positive Tissue Controls The interpretation of IHC and ISH results relies on the use of method controls and a general acceptance of what is consid- 1+ 2+ 3+ HER2 by far the majority of pathogens that might contaminate human Figure 2.6 Different levels of expression of HER2 on four different breast cancer cell lines. Four cell lines with different levels of HER2 expression, ranging from no expression (0; MDA-MB-231) through 1+ (MDAMB-175), 2+ (MDA-MB-453) and high expression (3+; SK-BR-3) can be used as controls/standards for HER2 IHC. Other cell lines may be used for different antigens. Cells were stained using IHC and HER2 antibody (Code A0485, Dako) and the Autostainer Link 48, Dako platform. ered to be appropriate staining. Control material must be ixed and processed in the same manner as the test material, to as- Unfortunately, while this approach should produce control sure the accuracy of the results. In the authors’ laboratory, this material that is equivalent to the test samples processed requirement includes replicate tissue control blocks ixed in NBF in the same manner, this is not always the case. The au- for various times (e.g. 8, 32, 56 and 104 hours; see Figure 2.5); thors and Dako Research and Development staff have all and the control block is selected to match the length of ixation observed that control cell lines seem to be inferior to solid of the test specimen. Traditionally, tissues with known expres- tissues as controls for FISH assays (personal communica- sion of the target are used as controls. For example, breast tu- tion); and the cell lines may be more easily damaged dur- mor control tissues with different levels of HER2 expression (as ing staining. This problem serves to emphasize the need for well as normal breast tissue) are employed as controls for HER2 thorough validation of the control materials. IHC and ISH. A similar approach is taken for all other antigens, and low level expressing normal tissues are particularly useful Tissue Microarrays as method controls. However, it is dificult to maintain a contin- The use of tissue microarrays (TMAs) has become popular for 27 Chapter 2 | Fixation and Other Pre-Analytical Factors IHC/ISH method development and research purposes, where organizations such as NordiQC or CAP, and product inserts multiple different tumors and normal tissues can be combined from manufacturers. This information can narrow the scope of into a single block, vastly reducing the number of slides needed unknown variables, thereby facilitating more eficient, precise for staining. TMAs are also used for control material for routine testing, and potentially reducing effort and costs. However, such work. Thus small samples of a range of controls (e.g. as noted information is not a substitute for in-house validation, which must above for HER2) can be put together into a single small block. still occur, albeit with reduced numbers of reference samples. Sections from such blocks can be placed alongside every test When validating a primary antibody for IHC or an ISH method section, to give the optimum “on-slide” control, assuring that for Class II targets (prognostic or predictive markers), a mini- each slide received all reagents during the staining run. The mum of 25 to 100 cases must be tested, according to ASCO/ small size of the samples in TMA blocks conserves the con- CAP recommendations. Both positive and negative cases are trol tissue. However, care must be taken to ensure that hetero- to be included and some of the positive cases should have low geneity within tumors does not result in inappropriate controls; expression of the target (4). As part of the ongoing assessment and that ixation and processing conditions are matched. and monitoring for any “drift” in test results, participation in external quality assessment programs is critical and the correlation between the internal and external testing should be very Chapter 2.7 Validation high (90% for positives and 95% for negatives) (4). Validation is an essential step when establishing protocols and Instruments such as tissue processors and automated IHC when choosing reagents. Validation ensures that a certain stand- and ISH staining platforms must also be validated. Methods of ardized procedure will give consistent and diagnostically useful validating instruments vary, but the purpose is to prove that the results on tissues processed in a standardized manner. It also instrument is working as expected, repeatedly and reliably. It is tests the limits of changes in the procedure that will continue to pro- helpful and cost effective when manufacturers of these instru- vide optimum results. For example, users can validate the effect ments provide validation documentation to customers upon of 10% NBF ixation on a particular antigen by preparing multiple, installation. Customers can then verify that the information pro- uniform pieces of tissue from the same tissue specimen. These vided is reproducible at their own facilities. Different countries blocks are then ixed in 10% NBF for different lengths of time, and regulatory bodies have different standards regarding vali- spanning multiple days (see Figure 2.5). They are then processed dation and these local rules should be consulted. using identical protocols, cut and then stained simultaneously. The results help to establish the range of ixation times that produces acceptable staining for a particular target (epitope), while employing a particular detection method. The work involved in validation is often dificult, time consuming and expensive. Alternatively, users can choose a system with an existing standardized and validated protocol, the socalled RTU (Ready-to-Use) approach. Commercially available RTU kits, when utilized exactly as described in the kit inserts, are guaranteed (within limits) to provide diagnostically useful results. Examples of such kits are: ER/PR pharmDx™ Kit; and being made „„ When an alternative ixative is being introduced „„ When a new decalcifying protocol is being introduced „„ When a new staining platform is being introduced (auto- mated or manual) „„ When a new release of software for a staining platform is being introduced „„ When a new IHC or ISH staining detection protocol is be- ing introduced HercepTest™ (c-erb-B2 oncoprotein – HER2 protein) from Dako, „„ When a new antibody or molecular marker is being intro- Denmark, while similar RTU reagents are widely available from duced, including when an alternative clone or nucleic acid other manufacturers. Another option is to use published information provided by high quality peer-review studies, external 28 Examples of when validation would be required include: „„ When changes in the ixation or processing protocols are probe for an existing test is being introduced „„ When new control tissues or cells are being introduced. Fixation and Other Pre-Analytical Factors | Chapter 2 Once initial validations are complete, it is not necessary to re-validate unless there has been a signiicant change in the test system, such as a new water supply. However, veriication of staining performance is an on-going process. Examples of when veriication may be required include: „„ When new lot numbers of antibodies or detection reagents are put into use „„ When unexpected or aberrant staining results occur; and „„ When several different staining platforms are all being used to perform the same IHC or ISH tests (veriication should be done at pre-set intervals to determine if each platform is producing comparable staining results, regardless of instrument used) Chapter 2.8 Conclusions Patient safety based on accurate interpretation of results depends heavily on this standardization of all pre-analytical variables. Prognostic tests using IHC and ISH are being developed and they will independently forecast clinical outcomes for patients. HER2, ER and PgR are considered predictive markers that inluence the selection of patients who will respond more favorably to therapies, emphasizing further the need for standardization. Even if it is not possible to perfectly optimize every pre-analytical step, it is possible to perform each step in the same manner each time it is done. Rigorous adherence to this approach will yield more meaningful results and will, if necessary, facilitate problem solving. 2b. Wolff AC, Hammond EH, Hicks DG, Dowsett M, McShane LM, Allison KH, et al. Recommendations for Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Update. J Clin Oncol 2013. 10.1200/ JCO.2013.50.9984 (Ahead of print). 3. Hammond ME, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendation for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Clin Oncol 2010; 28:2784-95. 4. Torlakovic EE, Riddell R, Banerjee D, El-Zimaity H, Pilavdzic D, Dawe P, et al. Canadian Association of Pathologists-Association Canadienne des Pathologistes National Standards Committee/ Immunohistochemistry: best practice recommendations for standardization of immunohistochemistry Tests. Am J Clin Pathol 2010; 133:354-65. 5. Nakhleh RE, Grimm EE, Idowu MO, Souers RJ, Fitzgibbons PL. Laboratory compliance with the American Society of Clinical Oncology/College of American Pathologists guidelines for human epidermal growth factor 2 testing: a College of American Pathologists survey of 757 laboratories. Arch Pathol Lab Med 2010; 134: 728-34. 6. Shi SR, Shi Y, Taylor CR. Antigen retrieval immunohistochemistry; review and future prospects in research and diagnosis over two decades. J Histochem & Cytochem 2011; 59: 13-32. 7. Engel KB and Moore HM. Effects of preanalytical variables on the detection of proteins by immunohistochemistry in formalin-ixed, parafin-embedded tissue. Arch Pathol Lab Med 2011; 135:537-43. 8. Goldstein NS, Ferkowicz M, Odish E, Mani A, Hastah F. Minimal formalin ixation time for consistent estrogen receptor immuno histochemical staining in invasive breast carcinoma. Am J Clin Path 2003; 120:86-92. 9. Grizzle WE. Models of ixation and tissue processing. Biotech Histochem 2009; 84:185-93. Acknowledgements The authors would like to thank Jill Vandenberg for assisting with the preparation of the igures included in this chapter and 10. Goldstein NS, Hewitt SM, Taylor CR, Yaziji H, Hicks DG. Recommendation for improved standardization of immunohistochemistry. Appl Immunohistochemistry Mol Morphol 2007; 15:124-33. References 11) Rhodes A, Jasani B, Couturier J, McKinley MJ, Morgan JM, Dodson AR, et al. A formalin-ixed, parafin-processed cell line standard for quality control of immunohistochemical assay of Her-2/neu expression in breast cancer. Am J Clin Pathol 2002; 117:81-89. 1. Immunohistochemical Staining Methods, 5th Edition, editors: Kumar, GL. Rudbeck, L. Dako North America, Inc. Carpenteria, U.S.A. and Dako Denmark A/S, Glostrup, Denmark. 2009. 12. Kwan SW. Bhargavan M, Kerlan RK Jr, Sunshine JH. Effect of advanced imaging technology on how biopsies are done and who does them. Radiology 2010; 256:751-58. 2a. Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, et. al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med 2007; 131:18-43. 13) Xiao Y, et al. Cell lines as candidate reference materials for quality control of ERBB2 ampliication and expression assays in breast cancer. Clinical Chemistry 2009; 55:1307-15. Aaron Hess for preparing cell lines. 29 Part I: The Staining Process Chapter 3 Antigen Retrieval Shan-Rong Shi, MD Clive R. Taylor, MD, D.Phil Re • triev • al (n.) The act or process of getting and bringing back something. Merriam-Webster Online Dictionary Antigen Retrieval | Chapter 3 Chapter 3.1 Introduction cient degree for recovery of antigenicity? With these key questions in mind, Shi spent many days and nights in 1988, prior In the majority of cases, tissue specimens for immunohisto- to online data access, searching the chemical literature the chemical (IHC) staining are routinely ixed in formalin and sub- old fashioned way! The answer was inally found in a series of sequently embedded in parafin. Because of the long history studies of the chemical reactions between protein and forma- of the use of formalin-ixed, parafin-embedded (FFPE) tissue lin, published in the 1940s (2-4). These studies indicated that sections in histopathology, most of the criteria for pathological cross-linkages between formalin and protein could be disrupt- diagnosis have been established by the observation of FFPE ed by heating above 100 °C, or by strong alkaline treatment. tissue sections stained by hematoxylin and eosin. Additionally, With this knowledge of high temperature heating as a potential a great number of FFPE tissue blocks, accompanied by known retrieval approach, the heat-induced AR technique was devel- follow-up data, have been accumulated worldwide, providing oped in 1991 (5). an extremely valuable resource for translational clinical research and basic research that cannot easily be reproduced. Subsequently, this AR technique has been applied to in situ The major drawback of FFPE tissue is that formalin-induced hybridization, TUNEL, immunoelectron microscopy, blocking molecular modiication of proteins (antigens) may result in loss cross-reactions for multiple immunolabeling, aldehyde-ixed of the ability of the antibody to react with the antigen, a loss frozen tissue sections, mass spectometry on FFPE tissue sec- that can only be corrected by the restoration (retrieval) of the tions, and the development of a series of novel techniques for ‘formalin-modiied’ antigen molecular structure. Since the early successful extraction of nucleic acids and proteins from FFPE 1970s, many active pioneers, mostly practicing pathologists tissues (6). Arguably this contribution to protein extraction has who were acutely aware of the need to enhance the capabili- proved critical to the development of modern tissue proteom- ties of IHC on FFPE tissue sections while retaining morphologic ics on FFPE tissues (7, 8). features, have been searching for an effective retrieval technique (1). Some retrieval methods, such as enzyme digestion, As a result, FFPE archival tissue collections are now seen improved IHC staining only for limited antigens. One of the au- as a literal treasure of materials for clinical and translational thors (Shi) began a different approach, based upon the prac- research, to an extent unimaginable prior to the introduction tical and theoretical issues to be addressed. A key scientiic of heat-induced antigen retrieval two decades ago. The ad- question was whether ixation in formalin modiied the struc- vantages of FFPE tissues in terms of preservation of both ture of antigens in a reversible or an irreversible manner. To morphology and molecules in cell/tissue samples are broad- be more speciic, was there any prior scientiic evidence that ly recognized. For example, there is a growing body of liter- the effects of formalin ixation on proteins could be reversed? ature demonstrating successful application of FFPE tissue And if reversed, was the structure of protein restored to a sufi- samples for molecular analysis, using AR based methods Table 3.1 Comparison of frequency concerning application of different terms of heat-induced AR according to OVID Medline data of the 1st week of July & August 2013. Different terms used Total articles Percentage (%) 1st week of July 1st week of August 1st week of July 1st week of August Antigen retrieval 138 140 63.9 63.9 Epitope retrieval 22 22 10.2 10.1 Heat-induced epitope retrieval 15 15 6.9 6.9 Microwave treatment 41 42 19 19.1 Total 216 219 100 100 31 Chapter 3 | Antigen Retrieval for extraction of DNA/RNA, and proteins from FFPE tissues. sections produces hydrolysis that contributes to break down Today, twenty years on, the AR technique is widely, almost cross-links (14, 15). In the very irst article on AR, Shi and universally, used in surgical pathology, including veterinary colleagues (5) showed a strong keratin-positive staining pathology, in all morphology based sciences, and in pharma- result simply by boiling sections in distilled water in a cology drug related research, with thousands of original ar- microwave oven. While the composition of the AR solution ticles published worldwide (6). The enormous impact is re- plays a part, it is the presence of heat and water that is lected in the need to divide all publications with respect to critical: immersing FFPE tissue sections in pure 100% IHC on FFPE tissue into two phases: the pre-AR and post-AR glycerine followed by the IHC staining procedure gives a eras, with the dividing line in the early 1990s (9). The term negative result, adding water to the glycerine and boiling ”antigen retrieval” (AR) was originally adopted by Shi and again, gives satisfactory IHC staining (16). That high colleagues in 1991. Other terms exist, such as heat-induced temperature heating is the most important factor for AR epitope retrieval (HIER) or antigen unmasking/demasking, technique has been conirmed by numerous subsequent but have no particular merit to cause replacement of the publications (17, 18). There are several critical technical original term (10). Table 3.1 is a comparison of frequency points with respect to the combination of heating tempe- with respect to usage of different terms for this technique. rature and heating time (heating condition = heating tempe- Clearly the original term, antigen retrieval, has greatest acceptance and will be employed in this chapter. rature x heating time): „„For many antigens, almost any kind of heating treatment, including microwave oven, water bath, pressure cooker, or The earlier introduction of enzymatic pre-treatment of tissue sections (11) remains in use for certain selected applications, but these methods are much more dificult to control and have been largely replaced by heat-induced AR. autoclave may generate similar results, if adjusted appropriately for time „„There is generally an inverse correlation between heating temperature (T) and heating time (t), as expressed by the formula: AR = T x t (19) „„For most antigens, higher temperature heating, such as Chapter 3.2 Major Factors that Inluence the Effect of Antigen Retrieval boiling FFPE tissue sections for 10-20 minutes, may be an optimal heating condition. However, a few antigens require lower temperature heating conditions, over a longer period Following wide application of the heat-induced AR, numerous of time (20). modiications of the AR technique and various protocols have „„It has been recommended that to preserve tissue mor- been documented in literature. As a result, there is a grow- phology, a lower temperature (90 °C) with an elongated ing need for standardization of the AR technique itself. The time may be preferable (21) critical importance of standardization of AR-IHC has been „„ Within the above generalizations, for some antigens the emphasized by the American Society of Clinical Oncology most extreme conditions of temperature and time (e.g. and the College of American Pathologists in their Guideline pressure cooker for hours) gives the greatest staining, but at Recommendations for HER2 testing in breast cancer, as well the cost of morphology. Such methods should be considered as numerous subsequent documents (12a, 12b, 13). In order as a last resort. to understand the key issues with respect to standardization of AR, it is critical irst to study the major factors that inluence pH Value of the AR Solution the effectiveness of AR-IHC. The following conclusions are The pH value of the AR solution is another factor that sig- based on our more than twenty year experience of research, niicantly inluences the result of AR-IHC. In 1995, we (22) and upon literature review. tested the hypothesis that pH of the AR solution may inlu- „„Heating is the most important factor: high temperature heating of formaldehyde-ixed proteins in FFPE tissue 32 ence the quality of immunostaining of a panel of antibodies, by comparing seven different AR buffer solutions at different Antigen Retrieval | Chapter 3 pH values ranging from 1 to 10. The conclusions of this study 5. A higher pH AR solution, such as Tris-HCl or sodium acetate are relevant when choosing the optimal AR method for any buffer at pH 8.0-9.0, may be suitable for most antigens (see particular antigen/antibody pairing: circle in Figure 3.1). 6. Low pH AR solutions, while useful for nuclear antigens may give a focal weak false positive nuclear staining; the use of Staining Intensity negative control slides is important to exclude this possibility. A Numerous investigators have independently reached similar conclusions (23-26). B Chemical Composition of the AR Solution Other potential factors have been examined for their effect on AR. In considering citrate buffer, it is generally accepted that effectiveness is not dependent so much on the chem- C ical, “citrate”, as upon the high temperature heating. Stud1 10 pH value ies have tested various additives to AR solutions, including metal salts, urea and citraconic anhydride; the last of these Figure 3.1 Schematic diagram of the three patterns of pH-inluenced AR immunostaining. Line A (pattern of Type A) shows a stable pattern of staining with only a slight decrease in staining intensity between pH 3 and pH 6. Line B (pattern of Type B) shows a dramatic decrease in staining intensity between pH 3 and pH 6. Line C (pattern of Type C) exhibits an ascending intensity of AR immunostaining that correlated with increasing pH value of the AR solution. Circle (right) indicates the advantage of using an AR solution of higher pH value. With permission, reproduced from Shi S-R, et al. J Histochem Cytochem 1995;43:193-201. showed promise in achieving stronger intensity by testing 62 commonly used antibodies, indings conirmed by others (28, 29). In our comparative study between citrate buffer and citraconic anhydride, using 30 antibodies, more than half (53%) showed a stronger intensity of IHC when using citraconic anhydride for AR, whereas 12 antibodies (43%) gave equivalent results; only one antibody gave a stronger intensity using citric buffer alone for AR (28). 1. Three types of patterns, relecting the inluence of pH, are indicated in Figure 3.1. Today many commercial retrieval solutions are available, often as part of an RTU approach to an automated platform (see Chapter 2. A, several antigens/clones showed no signiicant variation 5), and some products contain secret ingredients. Under pre- utilizing AR solutions with pH values ranging from 1.0 to 10.0 scribed conditions many of these reagents give good results, but (L26, PCNA, AE1, EMA and NSE); B, other antigens/clones care should be exercised in applying commercial AR solutions, of (MIB1, ER) showed a dramatic decrease in the intensity of unknown composition, to targets other than those described by the AR-IHC at middle range pH values (pH 3.0-6.0), but the vendor, or in protocols other than those recommended; both strong AR-IHC results above and below these critical zones; false positive and false negative results can occur. and C, still other antigens/clones (MT1, HMB45) showed negative or very weak focally positive immunostaining with With the growing use of automated staining platforms, the choice a low pH (1.0-2.0), but excellent results in the higher pH range. of ‘autostainer’ to a large degree dictates not only the selection 3. Among the seven buffer solutions at any given pH value, of the primary antibody (see Chapter 4), and its concentration, the intensity of AR-IHC staining was very similar, except that but also the detection system, and the protocol (see Chapter 5 Tris-HCl buffer tended to produce better results at higher and Chapter 6), including the method of antigen retrieval. The pH, compared with other buffers. vendors of automated stainers generally offer recommended 4. Optimization of the AR system should include optimization of the pH of the AR solution. AR protocols for (almost) all of the primary antibodies in their portfolio, usually a high pH method (pH 9), a mid/low pH meth- 33 Chapter 3 | Antigen Retrieval od (pH 6), and an enzyme-based method for a small number of antibodies. The recommendation usually includes the use In practice, the method may be further simpliied in the following ways; of proprietary AR solutions, and deined heating conditions, as „„Test three pH values by using one temperature (boiling), part of the protocol. As noted above, departure from these rec- select the best pH value and then test various temperatures; or, ommendations requires a full revalidation process. „„ Test several commonly used AR solutions (or those recom- For new antibodies (see Chapter 4), and for antibodies pro- mended for the autostainer in use in the laboratory), such as duced by other vendors (other than the manufacturer of the par- citrate buffer pH 6.0, Tris-HCl + EDTA of pH 9.0 ticular automated stainer in use) the laboratory must undertake a study to establish the optimal retrieval method. For this pur- Although this later method is not a complete test, it is more pose it is recommended that the laboratory use some variation convenient for most laboratories. If satisfactory results are not of the Test Battery approach introduced by Shi and colleagues. obtained other variations may be tested, including citraconic anhydride, or enzyme-based digestion methods. Numerous recent articles have emphasized that the application of test Chapter 3.3 Standardization of AR – The “Test Battery” Approach battery for establishing an optimal AR protocol is also dependent on the primary antibody and the subsequent detection system. In other words, if an optimal AR protocol is good for In 1996, a “test battery” approach was recommended as a antibody clone ‘1’ to protein ‘A’ employing detection system ‘B’, method for quick examination of the two major factors that it is not necessarily good for antibody clone ‘2’ to protein ‘A’, affect the outcome of AR, namely the heating condition and using the same or different detection systems; but a different pH value, in order to reach the strongest signal of AR-IHC AR protocol might give acceptable results. (maximal level of AR) (30). This test battery serves as a rapid screening approach to optimize the AR protocol and in so Specially prepared tissue microarrays (TMAs), incorpora- doing achieve some degree of standardization (31). In the ting a range of tissues and tissue cores ixed for differing initial recommendation the test battery included three levels times, are also of value in helping establish the optimal AR of heating conditions (below-boiling, boiling and above-boil- method for a particular antibody, by staining of only a ing), and three pH values (low, moderate, and high), such few TMA slides. The advantages are further enhanced by that a total of nine FFPE tissue sections were used (Table 3.2). application of recently developed image analysis software (AQUA) that is designed for quantitative IHC in TMA using Table 3.2. Test battery suggested for screening an optimal antigen retrieval protocol. Temperature Tris-HCl buffer pH 1.0-2.0 (Slide #)a pH 7.0-8.0 (Slide #)a pH 10.0-11.0 (Slide #)a Super-high (120 °C)b #1 #4 #7 High (100° C), 10 min #2 #5 #8 Mid-high (90° C), 10 minc #3 #6 #9 (a) One more slide may be used for control without AR treatment. Citrate buffer of pH 6.0 may be used to replace Tris-HCl buffer, pH 7.0 to 8.0, as the results are similar, and citrate is most widely used. (b) The temperature of super-high at 120°C may be reached by either auto claving or pressure cooker, or microwave heating at a longer time. (c) The temperature of mid-high at 90°C may be obtained by either a water bath or a microwave oven, monitored with a thermometer. Modiied from Shi SR, et al. J. Histochem. Cytochem. 45: 327-343. 1997. 34 an automatic scan (32). Antigen Retrieval | Chapter 3 Table 3.3 Major applications of antigen retrieval technique and principle. Areas of application of AR Application of AR technique and/or principle Reference Immunoelectron microscopy (IEM) AR pre-treatment of Epon-embedded ultra-thin sections after etching the grids by solutions(a) to achieve satisfactory positive results; or, directly heating the grid and followed by washing procedures including 50 mM NH4Cl and 1% Tween 20 39, 40 In situ hybridization (ISH) High temperature heating FFPE tissue sections prior to ISH to achieve satisfactory results 41-43 TUNEL Optimal heating time, as short as 1 min to improve the signal 44, 45 Multiple IHC staining procedures Adding a microwave heating AR procedure (10 min) between each run of the multiple IHC staining procedure effectively blocks cross-reactions, by denaturing bound antibody molecules from the previous run 33 Human temporal bone collections Combining sodium hydroxide-methanol and heating AR treatment provides an effective approach for IHC used in celloidin-embedded temporal bone sections. This method is also used for plasticembedded tissue sections, including IEM 46, 47 Immunoluorescence To enhance intensity and reduce autofluorescence 48 Cytopathology AR pre-treatment of archival PAP smear slides promotes satisfactory IHC staining 49 Flow cytometry (FCM) Enzyme digestion followed by heating AR treatment was adopted to achieve enhancement of FCM of FFPE tissue 50 Floating vibratome section Microwave boiling of vibratome sections improves IHC staining results; further extended for use with whole mount tissue specimens 51 En Bloc tissue AR heating of 4% paraformaldehyde-fixed animal brain or testis tissue blocks enhances immunoreactivity for most antibodies tested 52 Frozen tissue section Aldehyde-fixed frozen tissue section with use of AR treatment gives both excellent morphology and IHC staining 34, 35 DNA extraction from FFPE tissue sections Boiling AR pre-treatment prior to DNA extraction gives improved results compared to enzyme treatment 53-56 RNA extraction from FFPE tissue sections Heating AR treatment prior to RNA extraction gives improved results compared to enzyme treatment 57, 58 Protein extraction from FFPE tissue sections AR pre-treatment with AR solution including 2% SDS and/or other chemicals improves efficiency of protein extraction from FFPE tissue compared to enzyme digestion. Combining with elevated hydrostatic pressure may increase extraction of up to 80-95% of proteins from FFPE tissue sections 59-62 Imaging mass spectrometry (IMS) AR pre-treatment gives satisfactory results of IMS . Based on comparison among different AR solutions, Gustafsson et al summarized that citrate acid AR method is an important step in being able to fully analyze the proteome for FFPE tissue 36-38 AR = antigen retrieval; FFPE = formalin-ixed parafin-embedded; IEM = immunoelectron microscopy; ISH = in situ hybridization; TUNEL = terminal deoxynucleotidyl transferase dUTP nick end labeling; FCM = low cytometry; IMS = imaging mass spectrometry. (a) 10% fresh saturated solution of sodium ethoxide diluted with anhydrous ethanol for 2 min or with a saturated aqueous solution of sodium metaperiodate for 1 hour. Reprinted with permission from Shi SR, et al. J. Histochem. Cytochem. 59:13-32, 2011. 35 Chapter 3 | Antigen Retrieval Chapter 3.4 Application of AR Techniques – The Basic Principles Chapter 3.6 Reagents and Protocols Sections 3.6-3.12 will describe the following retrieval protocols: In addition to its use in IHC, AR has increasingly been adopted in the following related applications: „„In situ hybridization (ISH) and in situ end-labeling (TUNEL) of apoptotic cells in FFPE tissue sections; as well as in low cytometry to achieve stronger positive signals while reducing „„Water Bath Methods – Dako PT Link – Water Bath (conventional) Heating „„Pressure Cooker Heating „„Autoclave Heating non-speciic background noise „„Microwave Oven Heating reaction from the previous run (33) „„Combined Proteolytic Pre-treatment and Antigen Retrieval „„In IHC multi-stains, AR has been used to block the cross„„In addition to FFPE tissue sections, AR has been adopted „„Proteolytic Pre-treatment „„Combined Deparafinization and Antigen Retrieval for aldehyde-ixed fresh tissue sections, plastic-embedded tissue sections, cell smear samples for cytopathology, and The composition and the pH of retrieval buffers are crucial for loating vibratome sections (33) optimal retrieval. Although citrate buffers of pH 6 are widely „„Modiied AR methods have been used successfully for used retrieval solutions, high pH buffers have been shown to extraction of DNA and RNA from FFPE tissue sections for be widely applicable for many antibodies, as discussed previ- PCR-based methods and sequencing ously (22, 64). It is the responsibility of the individual laboratory „„Imaging mass spectrometry (IMS) has been applied to to determine which of the available buffers perform optimally proteins extracted from FFPE tissue sections by AR ap- for each antigen/antibody and then to use them consistently. proaches, providing an avenue to fully analyze the proteome Although 0.01 M citrate buffers of pH 6 have historically been of archival FFPE tissue (36-38) the most widely used retrieval solutions, high pH buffers have started being implemented when showing improved end results for some antigens. The following protocol descriptions Chapter 3.5 AR-IHC-based Research and Diagnostics should serve as guidelines only. It is the responsibility of the individual laboratory to compare methods and select the optimal protocol for consistent use. It is recommended for the Over the past two decades AR has found extensive application, AR methods to control temperature settings and to measure not only for IHC, but also for molecular methods applied to FFPE the actual temperature at regular intervals. The following proto- tissues, so called tissue proteomics, as well as standardization cols focus mostly on Dako reagents and systems, with detailed and quantiication of IHC. For further details the reader is referred input from Dako; other manufacturers supply similar reagents to the multi-author text edited by Shi and colleagues (6), which and protocols, which should be followed scrupulously. includes discussion of a proposal for quantitative IHC, based upon the use of AR. This hypothesis proposes to minimize the variation in IHC that is observed in clinical FFPE tissue sections, Chapter 3.7 Water Bath Methods by using optimal antigen retrieval (AR) in a test battery approach. 36 The intent is to use AR to reduce the loss of antigenicity observed A. Dako PT Link for many proteins, following variable ixation, to a level compara- Dako PT Link instrument simpliies the water bath antigen ble to frozen tissue sections, at which point a standard calibration retrieval process by performing automated retrieval using curve could be developed using internal proteins. This approach speciied protocols, which incorporate preheat temperature, is similar to that of enzyme-linked immunosorbent assays (ELISA) antigen retrieval temperature, and time as well as cool down where a standard curve is used to convert the immunoreaction settings. Typically, antigen retrieval is performed for 20 min- signal into a quantitative amount of protein (63). utes at 97 °C. Antigen Retrieval | Chapter 3 6. Press [RUN] button for each tank and the CYCLE will show PREHEAT. Allow solution to reach the selected preheat temperature. 7. Open the PT Link and immerse the Autostainer Slide Rack with deparafinized tissue sections into the preheated target retrieval solution.* 8. Place tank lids on tanks. Close and lock main lid with external latch. 9. Press [RUN] button for each tank to start run. CYCLE will show WARM-UP and the lid lock will engage. 10. PT Link will warm up to preset temperature and then start the countdown clock for target retrieval cycle. 11. When target retrieval cycle is inished, CYCLE will show COOL. The COOL cycle is inished when temperature reaches Preheat SET temperature, even if Preheat is disabled. 12. When COOL cycle is inished, CYCLE will show DONE and lid will unlock automatically. 13. Open the PT Link and remove each slide rack with the slides Figure 3.2 Dako PT Link is a water bath method for antigen retrieval from the PT Link Tank and immediately immerse slides into the PT Link Rinse Station containing diluted, room temperature Dako Wash Buffer (10x). Materials Required „„ Dako PT Link* „„ Dako Autostainer Slide Rack „„ Retrieval solution „„ FLEX IHC Microscope Slides or slides coated 14. Leave slides in the diluted, room temperature Dako Wash Buffer for 1-5 minutes. 15. Proceed with IHC staining. *As an alternative, a 3-in-1 solution can be used for both deparafinization and target retrieval. See Section 3.13 | Combined Deparafinization and Antigen Retrieval. with other suitable adhesives „„ Personal protective equipment *Dako Omnis has onboard pre-treatment module. See User Manual for protocol. Protocol Wear chemical-protective gloves when handling parts immersed in any reagent used in PT Link. 1. Deparafinize and rehydrate tissue sections. 2. Prepare a working solution of the selected target retrieval solution according to speciications. 3. Fill tanks with 1.5 L of desired target retrieval solution. 4. Place tank lids on tanks. Close and lock main lid with external latch. 5. See Operator’s Manual for instrument set-up details: a. Recommended time is 20-40 minutes. b. Set antigen retrieval temperature to 97 °C. c. Set preheat temperature to 65 °C (allows up to 85 °C). Figure 3.3 Antigen retrieval in conventional water bath. 37 Chapter 3 | Antigen Retrieval B. Water Bath (conventional) Heating solutions. As an alternative, individual plastic container(s) can One of several advantages of the water bath heating method be illed with retrieval solution and placed in reagent grade wa- is the ready availability of water baths in most clinical laborato- ter in the pressure cooker pan. ries. Temperature settings just below the boiling point of water (95-99 °C) are most commonly used. Materials Required „„ Temperature-controlled water bath „„ Slide rack „„ Incubation container and cover „„ Retrieval solution „„ Tris-Buffered Saline „„ Silanized Slides or slides coated with other suitable adhesives „„ Thermometer „„ Personal protective equipment Protocol It is recommended to wear insulated gloves when handling parts immersed in any reagent used in a water bath. 1. Deparafinize and rehydrate tissue sections. 2. Fill container with enough retrieval solution to cover slides and equilibrate to 95-99 °C in water bath. 3. Immerse racked slides in preheated retrieval solution, cover container with lid, and incubate for speciied time Figure 3.4 Pressure cooker for antigen retrieval. within the 20-40 minutes range after the set temperature has been reached. 4. Remove the container from the water bath and cool the contents with the lid in place for 20 minutes at room temperature. 5. Rinse with Tris-Buffered Saline (TBS) or distilled water at room temperature. 6. When removing the slides from the container it is very important that the slides do not dry out. 7. Transfer slides to TBS. 8. Proceed with IHC staining. Materials Required „„ Stainless steel pressure cooker, preferably electrically programmable „„ Metal or plastic slide racks „„ Retrieval solution „„ Silanized Slides or slides coated with other suitable adhesives „„ Tris-Buffered Saline „„ Incubation container (optional) „„ Personal protective equipment Protocol Chapter 3.8 Pressure Cooker Heating It is recommended to wear a safety face shield and insulated gloves. 38 Pressure cookers set to approximately 103 kPa/15 psi will 1. Deparafinize and rehydrate tissue sections. achieve a temperature of approximately 120 °C at full pressure. 2. Fill the pressure cooker with enough retrieval solution to Alternatively, setting at 125 °C can be used for antigen retrieval. cover slides. Alternatively, ill individual plastic container(s) Stainless steel pressure cookers are recommended as the alu- with retrieval solution and add at least 500 mL of reagent minum models are susceptible to corrosion by some retrieval grade water to pressure cooker chamber. Antigen Retrieval | Chapter 3 3. Bring contents to near boiling point, place racked slides into 7. Rinse slides in Tris-Buffered Saline (TBS) or reagent grade retrieval solution, seal the pressure cooker, and again bring water. When removing the slides from the container it is very the solution to a boil. For programmable pressure cookers, important that the slides do not dry out. set target temperature and heating time, place racked 8. Transfer slides to TBS. slides in retrieval solution, seal the pressure cooker, and be- 9. Proceed with IHC staining procedure. gin antigen retrieval procedure from room temperature. 4. Boil for 30 seconds to 5 minutes and allow the pressure cooker to cool for 20 minutes prior to opening. (Note: Vent any Chapter 3.10 Microwave Oven Heating residual pressure before opening). Open programmable pressure cooker when antigen retrieval procedure is completed. Microwave ovens are very eficient for the heating of aqueous 5. Transfer slides to room temperature Tris-Buffered Saline. solutions, however, the standardization of procedures is impor- When removing the slides from the container it is very im- tant when used for antigen retrieval (and for the retrieval of DNA portant that the slides do not dry out. for in situ hybridization, i.e. target retrieval). In an effort to main- 6. Proceed with IHC staining procedure tain consistency of AR protocols and to ensure reproducibility of staining results, the following elements should be standardized: „„ Wattage of the microwave oven Chapter 3.9 Autoclave Heating „„ Presence of a turntable When set to 15 psi, an autoclave, like a pressure cooker, will „„ Number of slides per container „„ Volume of retrieval buffers per container achieve a temperature of about 120 °C at full pressure (65, 66). „„ Number of containers Materials Required Materials Required „„ Bench top autoclave „„ Plastic or metal slide rack 750-800 W microwave oven with turntable (please note that the effective power may decrease over time) „„ Incubation container „„ Incubation container for microwave oven „„ Silanized Slides or slides coated with other suitable adhesives „„ Retrieval solution „„ Retrieval solution „„ Tris-Buffered Saline „„ Personal protective equipment Protocol „„ Plastic slide holder for microwave oven „„ Silanized Slides or slides coated with other suitable „„ adhesives „„ Tris-Buffered Saline „„ Personal protective equipment It is recommended to wear safety face shield and insulated gloves. Protocol 1. Deparafinize and rehydrate tissue sections. Never use the microwave oven with metallic material present. It 2. Place slides in plastic or metal slide rack. is recommended to wear insulated gloves when handling parts 3. Fill the incubation container with enough retrieval buffer (typi- immersed in any reagent. cally 250 mL) to cover slides. Insert the slide rack and cover. 4. Place the container in the autoclave and follow Autoclave Manufacturer’s Operating Instructions. 5. Set the temperature to 120 °C/15 psi and the time to 10-20 minutes. Start operation. 6. After venting pressure, open the lid and remove the slide container from the autoclave. 1. Deparafinize and rehydrate sections. 2. Place slides in slide holder. Fill empty positions with blank slides. 3. Fill incubation container with enough retrieval solution to cover slides and insert slide holder. 4. Cover the container to minimize evaporation. Use a lid with minimal opening to avoid build-up of pressure and reduce evaporation. 39 Chapter 3 | Antigen Retrieval 5. Place container in the middle of the turntable and heat to near boiling point. 6. Incubate for ixed amount of time, typically 10 minutes. 7. Pepsin: Digestion for 10 minutes at room temperature is generally suficient, but may be prolonged to 15 minutes. Remove the container from the microwave oven, remove the lid, and allow to cool at room temperature for 15-20 minutes. Proteolytic Enzyme, Ready-to-Use: 8. Rinse with distilled water. Digestion for 5-10 minutes at room temperature is suficient. 9. Place in Tris-Buffered Saline. For details, please refer to the product speciication sheets. 10. Proceed with staining protocol. Chapter 3.11 Proteolytic Pre-treatment Chapter 3.12 Combined Proteolytic Pre-treatment and Antigen Retrieval As with other pre-treatment methods, procedures for pro- Some antigens are more eficiently retrieved by a combination teolytic pre-treatment vary due to laboratory-speciic differ- of heating and enzyme digestion, e.g. some cytokeratins and ences in formalin ixation. Proteolytic pre-treatment must be immunoglobulin light chains. The protocol below describes a optimized (dilution and time – speciic elevated temperature method of irst treating with Proteinase K and then AR by either may also be selected) according to the particular ixation water bath or microwave method. process used in each laboratory. Examples of antigens most often treated with proteolytic enzymes include cytokeratins and immunoglobulins. Materials Required „„ Silanized Slides or slides coated with other suitable adhesives „„ Target Retrieval Solution, pH 6, Dako* Materials Required „„ Tris-Buffered Saline „„ Silanized Slides or slides coated with other suitable adhesives *Other target retrieval solutions will work with a similar protocol optimized according „„ Humidity chamber „„ Proteolytic Enzyme, Ready-to-Use „„ Tris-Buffered Saline „„ Tris-buffered NaCl Solution with Tween 20 (TBST), pH 7.6 to individual laboratory requirements. Protocol Protocol 1. Deparafinize and rehydrate tissue sections. 2. Place slides horizontally and apply enough enzyme working solution to cover tissue section(s), typically 200-300 µL. 3. Incubate for deined time, typically 5-15 minutes. 4. Stop enzymatic reaction by rinsing with distilled water or Tris-Buffered Saline. 5. It is recommended that enzyme digestion is included in the It is recommended to wear insulated gloves when handling parts immersed in any reagent. 1. Deparafinize and rehydrate tissue sections. 2. Cover tissue sections with Proteinase K and incubate for 5-10 minutes. 3. Rinse with distilled water and place in Tris-Buffered Saline. 4. Proceed to antigen retrieval using either PT Link, another water bath or microwave method below. relevant Autostainer protocols. For the RTU series antibodies, enzyme digestion is included. AR – Water Bath 5. Fill container with enough retrieval solution (200 mL) to For Dako Proteolytic Enzymes, the following guidelines apply: cover slides and equilibrate to 95-99 °C in water bath. Proteinase K, Concentrated and Ready-to-Use: Place the incubation container into the water bath and in- Digestion for 6 minutes at room temperature is generally sufi- cubate for 20-40 minutes. cient, but may be prolonged to 15 minutes. 6. Remove the container from the water bath and cool the contents with the lid removed for 20 minutes at room temperature. 40 Antigen Retrieval | Chapter 3 7. Rinse with Tris-Buffered Saline (TBS) or distilled water at room temperature. 8. Transfer slides to Tris-Buffered NaCl Solution with Tween 20 (TBST), pH 7.6 Wash Buffer. Table 3.4 Dako Products for Antigen Retrieval** Product Name Dako Code Target Retrieval Solutions FLEX Target Retrieval Solution, High pH K8004 FLEX Target Retrieval Solution, Low pH K8005 Target Retrieval Solution, pH 6.1, 10x Concentrated S1699 mL) to cover slides and insert slide holder. Insert slides in Target Retrieval Solution, pH 6.1, Ready-to-Use S1700 holder and cover. Target Retrieval Solution, pH 9, 10x Concentrated S2367 Target Retrieval Solution, pH 9, Ready-to-Use S2368 Target Retrieval Solution, pH 9, 10x Concentrated, (3-in-1) S2375 9. Proceed with IHC staining. AR – Microwave 5. Fill incubation container with enough retrieval solution (200 6. Place the incubation container into microwave oven and incubate for 2 x 5 minutes. 7. In between steps 4 and 5, ill up the container with enough distilled water (50 mL) to cover slides. 8. After the second treatment, leave the sections in the retrieval Proteolytic Enzymes Proteinase K, Concentrated S3004 Proteinase K, Ready-to-Use S3020 Pepsin S3002 Proteolytic Enzyme, Ready-to-Use S3007 solution at room temperature to cool for 15-20 minutes. 9. Rinse with distilled water. 10. Proceed with IHC staining. Chapter 3.13 Combined Deparafinization and Antigen Retrieval Buffers Dako Wash Buffer (10x) S3006 Tris-Buffered Saline S3001 Tris-buffered NaCl Solution with Tween 20 (TBST), pH 7.6, 10x Concentrated S3306 signiicantly and provides added convenience without sacriicing staining quality. Using Dako PT Link instrument simpliies the Instruments and Other Products Combining deparafinization and AR reduces slide handling time combined deparafinization and target retrieval process by performing automated deparafinization and retrieval in a single step. Materials Required „„ PT Link „„ PT Link Rinse Station „„ Silanized Slides or slides coated with other suitable adhesives „„ Target Retrieval Solution, pH 9, 10x Concentrated, (3-in-1)* „„ Dako Wash Buffer (10x) *When used in PT Link for 3-in-1 specimen preparation procedure, the diluted deparafinization / target retrieval solution can be used three times within a ive day period, if stored at room temperature. Dako Omnis GI100 Dako PT Link PT100/PT101 PT Link Rinse Station PT109 PT Link Tank PT102 Dako Autostainer Slide Rack S3704 FLEX IHC Microscope Slides K8020 Silanized Slides S3003 **Note that other manufacturers provide similar products; the user should bear in mind that commercial products generally are designed and tested to be used in the speciied format, within a deined protocol, and speciied instrumentation. Products are not freely interchangeable across detections systems, and any change from the recommended protocol requires complete revalidation. 41 Chapter 3 | Antigen Retrieval Protocol three choices of antigen retrieval are programmed for the in- Wear chemical-protective gloves when handling parts im- strument, with the appropriate AR recommended reagents. If mersed in any reagent used in PT Link. Recommended 3-in- satisfactory results are not obtained, it is advised then to revert 1 specimen preparation procedure using PT Link and above to a test battery approach. target retrieval solution: 1. Prepare a working solution of the selected target retrieval solution according to the speciications. 2. Fill PT Link Tanks with suficient quantity (1.5 L) of working solution to cover the tissue sections. 3. Set PT Link to preheat the solution to 65 °C. 4. Immerse the mounted, formalin-ixed, parafin-embedded tissue sections into the preheated target retrieval solution (working solution) in PT Link Tanks and incubate for 20-40 minutes at 97 °C. The optimal incubation time should be de termined by the user. 5. Leave the sections to cool in PT Link to 65 °C. 6. Remove each Autostainer Slide Rack with the slides from the PT Link Tank and immediately dip slides into a jar/tank (PT Link Rinse Station) containing diluted, room temperature Dako Wash Buffer (10x). 7. Leave slides in the diluted, room temperature Wash Buffer for 1-5 minutes. 8. Place slides on an automated instrument and proceed with staining. The sections should not dry out during the treatment or during the immunohistochemical staining procedure. 9. After staining, it is recommended to perform dehydration, clearing and permanent mounting. Chapter 3.14 Conclusion As discussed above, an effective AR protocol is based on the major factors that inluence the effect of AR-IHC. Thus, for new antibodies, a test battery approach is recommended for establishing the optimal AR protocol for each antigen/antibody pair in FFPE tissue sections. Although citrate buffer of pH 6 is a widely used retrieval solution, high pH buffers have been shown to be widely applicable for many antibodies. It is the responsibility of the individual laboratory to determine which of the listed AR solutions perform optimally for each antigen/antibody pair. In an automated system a new antibody can be ‘plugged’ into an existing automated protocol, and run with whatever two or 42 Antigen Retrieval | Chapter 3 References 1. Taylor CR, Cote RJ. Immunomicroscopy. A Diagnostic Tool for the Surgical Pathologist. 3rd ed. Philadelphia: Elsevier Saunders, 2005. 2. Fraenkel-Conrat H, Brandon BA, Olcott HS. The reaction of formaldehyde with proteins. IV. Participation of indole groups. J Biol Chem 1947;168:99-118. 3. 18. Igarashi H, Sugimura H, Maruyama K, Kitayama Y, Ohta I, Suzuki M, et al. Alteration of immunoreactivity by hydrated autoclaving, microwave treatment, and simple heating of parafin-embedded tissue sections. APMIS 1994; 102:295-307. Fraenkel-Conrat H, Olcott HS. Reaction of formaldehyde with proteins. VI. Cross-linking of amino groups with phenol, imidazole, or indole groups. J Biol Chem 1948;174:827-843. 19. Shi S-R, Cote RJ, Chaiwun B, Young LL, Shi Y, Hawes D, et al. Standardization of immunohistochemistry based on antigen retrieval technique for routine formalin-ixed tissue sections. Appl. Immunohistochem. 1998;6:89-96. 4. Fraenkel-Conrat H, Olcott HS. The reaction of formaldehyde with proteins. V. Cross-linking between amino and primary amide or guanidyl groups. J Am Chem Soc 1948;70:2673-2684. 20. 5. Shi SR, Key ME, Kalra KL. Antigen retrieval in formalin-ixed, parafin-embedded tissues: an enhancement method for immunohistochemical staining based on microwave oven heating of tissue sections. J Histochem Cytochem 1991;39(6):741-748. Shi S-R, Cote RJ, Liu C, Yu MC, Castelao JE, Ross RK, et al. A modiied reduced temperature antigen retrieval protocol effective for use with a polyclonal antibody to cyclooxygenase-2 (PG 27). Appl. Immunohistochem. Mol. Morphol. 2002;10(4):368-373. 21. Biddolph SC, Jones M. Low-temperature, heat-mediated antigen retrieval (LTHMAR) on archival lymphoid sections. Appl. Immunohistochem. Mol. Morphol. 1999;7(4):289-293. 22. Shi SR, Imam SA, Young L, Cote RJ, Taylor CR. Antigen retrieval immunohis tochemistry under the inluence of pH using monoclonal antibodies. J Histochem Cytochem 1995;43(2):193-201. 23. Evers P, Uylings HB. Microwave-stimulated antigen retrieval is pH and temperature dependent. J Histochem Cytochem 1994;42(12):1555-63. 24. Pileri SA, Roncador G, Ceccarelli C, Piccioli M, Briskomatis A, Sabattini E, et al. Antigen retrieval techniques in immunohistochemistry: comparison of different methods. J Pathol 1997;183(1):116-123. 25. Kim SH. Evaluatin of antigen retrieval buffer systems. J. Mol. Histol. 2004;35:409-416. 26. Yamashita S, Okada Y. Mechanisms of heat-induced antigen retrieval: analyses in vitro employing SDS-PAGE and immunohistochemistry. J. Histochem Cytochem 2005;53(1):13-21. 27. Namimatsu S, Ghazizadeh M, Sugisaki Y. Reversing the effects of formalin ixation with citraconic anhydride and heat: A universal antigen retrieval method. J Histochem Cytochem 2005;53(1):3-11. 28. Shi S-R, Liu C, Young L, Taylor CR. Development of an optimal antigen retrieval protocol for immunohistochemistry of retinoblastoma protein (pRB) in formalin ixed, parafin sections based on comparison of different methods. Biotech Histochem 2007 82(6):301-309. 29. Leong AS-Y, Haffajee Z. Citraconic anhydride: a new antigen retrieval solution. Pathology 2010;42(1):77-81. 30. Shi SR, Cote RJ, Yang C, Chen C, Xu HJ, Benedict WF, et al. Development of an optimal protocol for antigen retrieval: a 'test battery' approach exempliied with reference to the staining of retinoblastoma protein (pRB) in formalin-ixed parafin sections. J Pathol 1996;179(3):347-352. 31. O'Leary TJ. Standardization in immunohistochemistry. Appl Immunohistochem Mol Morphol 2001;9(1):3-8. 32. Cregger M, Berger AJ, Rimm DL. Immunohistochemistry and quantitative analysis of protein expression. Arch Pathol Lab Med 2006;130:1026-1030. 33. Lan HY, Mu W, Nikolic-Paterson DJ, Atkins RC. A novel, simple, reliable, and sensitive method for multiple immunoenzyme staining: use of microwave oven heating to block antibody crossreactivity and retrieve antigens. J Histochem Cytochem 1995;43(1):97-102. 34. Yamashita S, Okada Y. Application of heat-induced antigen retrieval to aldehyde-ixed fresh frozen sections. J Histochem Cytochem 2005;53(11): 1421-1432. 35. Shi S-R, Liu C, Pootrakul L, Tang L, Young A, Chen R, et al. Evaluation of the value of frozen tissue section used as "gold standard" for immunohistochemistry. Am J Clin Pathol 2008;129(3):358-366. 6. Shi S-R, Taylor CR. Antigen Retrieval Immunohistochemistry Based Research and Diagnostics. Hoboken, New Jersey: John Wiley & Sons, 2010. 7. Tanca A, Pagnozzi D, Addis MF. Setting proteins free: Progresses and achievements in proteomics of formalin-ixed, parafin-embedded tissues. Proteomics Clin. Appl. 2011;6:1-15. 8. Shi S-R, Taylor CR, Fowler CB, Mason JT. Complete Solubilization of Formalin-Fixed, Parafin-Embedded Tissue May Improve Proteomic Studies. Proteomics Clin Appl. 2013 . PROTEOMICS - Clinical Applications 2013; doi: 10.1002/prca.201200031. 9. Gown AM. Unmasking the mysteries of antigen or epitope retrieval and formalin ixation. Am. J. Clin. Pathol. 2004;121:172-174. 10. Taylor CR, Shi S-R. Antigen retrieval: call for a return to irst principles (Editorial). Appl. Immunohistochem. Mol. Morphol.2000;8(3):173-174. 11. Huang S-N. Immunohistochemical demonstration of hepatitis B core and surface antigens in parafin sections. Lab. Invest. 1975;33:88-95. 12a. Wolff AC, Hammond MEH, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American Society of Clinical Oncology/College of American Pathologists Guideline Recommendations for Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer. Arch. Pathol. Lab. Med. 2007;131(1):18–43. 12b. Wolff AC, Hammond EH, Hicks DG, Dowsett M, McShane LM, Allison KH, et al. Recommendations for Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Update. J Clin Oncol 2013. 10.1200/JCO.2013.50.9984 (Ahead of print). 13. 14. Taylor CR, Shi S-R, Barr NJ. Techniques of Immunohistochemistry : Principles, Pitfalls, and Standardization. In: Dabbs DJ, editor. Diagnostic Immunohistochemistry: Theranostic and Genomic Applications, 3rd Edition. Philadelphia, PA, USA: Saunders-Elsevier Inc., 2010:1-41. Mason JT, Fowler CB, O'Leary TJ. Study of Formalin-Fixation and Heat-Induced Antigen Retrieval. In: Shi S-R, Taylor CR, editors. Antigen Retrieval Immunohistochemistry Based Research and Diagnostics: John Wiley & Sons, 2010:253-285. 15. Bogen SA, Sompuram SR. A Linear Epitopes Model of Antigen Retrieval. In: Shi S-R, Taylor CR, editors. Antigen Retrieval Immunohistochemistry Based Research and Diagnostics: John Wiley & Sons, 2010:287-302. 16. Beebe K. Glycerin antigen retrieval. Microscopy Today 1999(9):30. 17. Kawai K, Serizawa A, Hamana T, Tsutsumi Y. Heat-induced antigen retrieval of proliferating cell nuclear antigen and p53 protein in formalin-ixed, paraf in-embedded sections. Pathol Int 1994;44(10-11):759-764. 43 Chapter 3 | Antigen Retrieval 36. Ronci M, Bonanno E, Colantoni A, Pieroni L, Di Ilio C, Spagnoli LG, et al. Protein unlocking procedures of formalin-ixed parafin-embedded tissues: Application to MALDI-TOF Imaging MS investigations. Proteomics 2008;8(18):3702-3714. 51. Evers P, Uylings HBM. Microwave-stimulated antigen retrieval in neuroscience. In: Shi S-R, Gu J, Taylor CR, editors. Antigen Retrieval Techniques: Immunohistochemistry and Molecular Morphology. Natick, Massachusetts: Eaton Publishing, 2000:139-150. 37. Groseclose MR, Massion PP, Chaurand P, Caprioli RM. High-throughput proteomic analysis of formalin-ixed parafin-embedded tissue microarrays using MALDI imaging mass spectrometry. Proteomics 2008;8(18):3715-3724. 52. Ino H. Antigen Retrieval by Heating En Bloc for Pre-ixed Frozen Material. J Histochem Cytochem 2003;51(8):995-1003. 53. 38. Gustafsson JOR, Oehler MK, McColl SR, Hoffmann P. Citric acid antigen retrieval (CAAR) for tryptic peptide imaging directly on archived formalin-ixed parafin-embedded tissue. J Proteome Res 2010;9(9):4315-4328. Coombs NJ, Gough AC, Primrose JN. Optimisation of DNA and RNA extraction from archival formalin-ixed tissue. Nucleic Acids Res 1999;27(16):e12-e17. 54. Stirling JW, Graff PS. Antigen unmasking for immunoelectron microscopy: labeling is improved by treating with sodium ethoxide or sodium metaperiodate, then heating on retrieval medium. J Histochem Cytochem 1995;43(2):115-23. Frank TS, Svoboda-Newman SM, Hsi ED. Comparison of methods for extracting DNA from formalin-ixed parafin sections for nonisotopic PCR. Diag Mol Pathol 1996;5(3):220-224. 55. Shi S-R, Cote RJ, Wu L, Liu C, Datar R, Shi Y, et al. DNA extraction from archival formalin-ixed, parafin-embedded tissue sections based on the antigen retrieval principle: heating under the inluence of pH. J Histochem Cytochem 2002;50(8):1005-1011. 56. Shi S-R, Datar R, Liu C, Wu L, Zhang Z, Cote RJ, et al. DNA extraction from archival formalin-ixed, parafin-embedded tissues: heat-induced retrieval in alkaline solution. Histochem Cell Biol 2004;122:211-218. 57. Masuda N, Ohnishi T, Kawamoto S, Monden M, Okubo K. Analysis of chemical modiicaiton of RNA from formalin-ixed samples and optimization of molecular biologyy applications for such samples. Nucleic Acids Res1999;27(22): 4436-4443. 58. Shi S-R, Taylor CR. Extraction of DNA/RNA from Formalin-Fixed, Parafinembedded Tissue Based on the Antigen Retrieval Principle. In: Shi S-R, Taylor CR, editors. Antigen Retrieval Immunohistochemistry Based Research and Diagnostics. Hoboken, New Jersey: John Wiley & Sons, 2010:47-71. 59. Shi S-R, Liu C, Balgley BM, Lee C, Taylor CR. Protein extraction from formalin-ixed, parafin-embedded tissue sections: quality evaluation by mass spectrometry. J Histochem Cytochem 2006;54(6):739-743. 60. Ikeda K, Monden T, Kanoh T, Tsujie M, Izawa H, Haba A, et al. Extraction and analysis of diagnostically useful proteins from formalin-ixed, parafin-em bedded tissue secitons. J Histochem Cytochem 1998;46:397-404. 61. Fowler CB, O'Leary TJ, Mason JT. Techniques of Protein Extraction from FFPE Tissue/Cells for Mass Spectrometry. In: Shi S-R, Taylor CR, editors. Antigen Retrieval Immunohistochemistry Based Research and Diagnostics: John Wiley & Sons, 2010:335-346. 62. Fowler CB, Chesnick IE, Moore CD, O'Leary TJ, Mason JT. Elevated pressure improves the extraction and identiication of proteins recovered from formalin-ixed, parafin-embedded tissue surrogates. PLoS ONE 2010; 5(12):e14253. doi: 10.1371/journal.pone.0014253. 63. D'Ambra-Cabry K, Deng DH, Flynn KL, Magee KL, Deng JS. Antigen retrieval in immunoluorescent testing of bullous pemphigoid [see comments]. Am J Dermatopathol 1995;17(6):560-563. Shi S-R, Shi Y, Taylor CR. Antigen Retrieval Immunohistochemistry: Review of Research and Diagnostic Applications over Two Decades following Introduction and Future Prospects of the Method. J Histochem Cytochem. 2011. 64. Boon ME, Kok LP, Suurmeijer AJH. The MIB-1 method for ine-tuning diagnoses in cervical cytology. In: Shi S-R, Gu J, Taylor CR, editors. Antigen Retrieval Techniques: Immunohistochemistry and Molecular Morphology. Natick, Massachusetts: Eaton Publishing, 2000:57-70. Bankfalvi A, Navabi H, Bier B, Bocker W, Jasani B, Schmid KW. Wet autoclave pretreatment for antigen retrieval in diagnostic immunohistochemistry. J Pathol 1994;174(3):223-8. 65. Shin R-W, Iwaki T, Kitamoto T, Tateishi J. Hydrated autoclave pretreatment enhances TAU immunoreactivity in formalin-ixed normal and Alzheimer's disease brain tissues. Lab Invest 1991;64:693-702. 66. Shin R-W, Iwaki T, Kitamoto T, Sato Y, Tateishi J. Massive accumulation of modiied tau and severe depletion of normal tau characterize the cerebral cortex and white matter of Alzheimer's disease. Am J Pathol 1992;140:937-945. 39. 40. Wilson DF, Jiang D-J, Pierce AM, Wiebkin OW. Antigen retrieval for electron microscopy using a microwave technique for epithelial and basal lamina antigens. Appl Immunohistochem 1996;4:66-71. 41. Lan HY, Mu W, Ng YY, Nilolic-Paterson DJ, Atkins RC. A simple, reliable, and sensitive method for nonradioactive in situ hybridization: use of microwave heating to improve hybridization sficiency and preserve tissue morphology. J Histochem Cytochem 1996;44:281-287. 42. McMahon J, McQuaid S. The use of microwave irradiation as a pretreatment to in situ hybridization for the detection of measles visus and chicken anaemia virus in formalin-ixed parafin-embedded tissue. Histochem J 1996;28:157-164. 43. Sibony M, Commo F, Callard P, Gasc J-M. Enhancement of mRNA in situ hybridization signal by microwave heating. Lab Invest 1995;73:586-591. 44. Strater J, Gunthert AR, Bruderlein S, Moller P. Microwave irradiation of parafin-embedded tissue sensitizes the TUNEL method for in situ detection of apoptotic cells. Histochemistry 1995;103:157-160. 45. 46. Shi SR, Cote C, Kalra KL, Taylor CR, Tandon AK. A technique for retrieving antigens in formalin-ixed, routinely acid- decalciied, celloidin-embedded human temporal bone sections for immunohistochemistry. J Histochem Cytochem 1992;40(6):787-792. 47. Shi S-R, Cote RJ, Taylor CR. Antigen retrieval immunohistochemistry used for routinely processed celloidin-embedded human temporal bone sections. In: Shi S-R, Gu J, Taylor CR, editors. Antigen Retrieval Techinques: Immunohistochemistry and Molecular Morphology. Natick, Massachusetts: Eaton Publishing, 2000:287-307. 48. 49. 50. 44 Lucassen PJ, Labat-Moleur F, Negoescu A, van Lookeren Campagne M. Microwave-enhanced in situ end-labeling of apoptotic cells in tissue sections; pitfalls and possibilities. In: Shi S-R, Gu J, Taylor CR, editors. Antigen Retrieval Techniques: Immunohistochemistry and Molecular Morphology. 1st ed. Natick, Massachusetts, USA: Eaton Publishing Co., 2000:71-91. Redkar AA, Krishan A. Flow cytometric analysis of estrogen, progesterone receptor expression and DNA content in formalin-ixed, parafin-embedded human breast tumors. Cytometry 1999;38(2):61-9. 45 Part I: The Staining Process Chapter 4 Selection of the Primary Antibody Søren Nielsen, Scheme Manager, NordiQC An•ti•bod•y (n.) Any of numerous proteins produced by B lymphocytes in response to the presence of speciic foreign antigens, including microorganisms and toxins. The American Heritage® Science Dictionary Selection of the Primary Antibody | Chapter 4 Chapter 4.1 Introduction This chapter will focus on the considerations to which users must pay special attention when selecting the right antibody Immunohistochemistry (IHC) has now become an indispensable for a given test. assay and is consistently performed in anatomic pathology in order to give a speciic diagnosis and subclassiication of neoplasms. IHC serves at present as a diagnostic, prognostic and Chapter 4.2 Selection of the Proper Antibody predictive assay and the results contribute to the choice of The selection of the primary antibody and the analytical con- treatment of patients in a clinical setting. ditions applied for the analyte have a signiicant impact on the IHC is a complex assay, where the end result is inluenced IHC result, regarding the diagnostic sensitivity and speciicity by multiple parameters in the pre-analytic phase, the analytic of the test. phase and the post-analytic phase (1). At least 4 million different protocols can be generated for one IHC analysis of one Laboratories are consistently introduced to new antibodies speciic antigen (Figure 4.1, see also Chapter 1). from scientiic publications and manufacturers, and in each situation the beneits of the new antibody must be carefully Depending on the selection and performance of these para- considered before implementation in a diagnostic setting. meters, the inal IHC result using the same primary antibody can show a range from negative to positive for the target antigen. In First of all the diagnostic potential and application areas order to provide a robust and diagnostically useful IHC assay, it is must be evaluated. For example, is the antibody labeling important that the assay is based on a solid foundation, anchored a new target, a supplementary antibody for a well-known by the most important factors inluencing the assay. Special atten- target, a replacement for an existing antibody within a cer- tion should be given to the following ive parameters: tain area, or a new antibody clone replacing an old clone, etc.? From a questionnaire performed by the College of „„ The tissue ixation process American Pathologists and submitted to American labora- „„ The antigen retrieval method tories, it was seen that typically 4-12 new markers were im- „„ The primary antibody selection plemented annually in the 727 laboratories responding to „„ The detection system applied the survey (2). „„ The choice of the tissue control material Pre-analytical Fixation Delay Time Type Volume Processing Decalcification Analytical Tissue Type +/- Heat-mediated Dimension Sections Pre-treatment Thickness Drying HIER Proteolysis Platform Post-analytical Antibody Visualization Report Interpretation Internal Context Positive/Negative Manual Clone Sensitivity Automated Buffer Enhancement External Qualitative Cut-off Quantitative Localization Time Time Specificity Temp Temp Localization Storage Controls Quality indicators pH Figure 4.1 The multiple parameters inluencing the IHC result. 47 Chapter 4 | Selection of the Primary Antibody Each and every new antibody introduced and implemented for single speciicity and single afinity. High speciicity may occa- diagnostic use in a laboratory must be technically optimized to sionally be accompanied by a low afinity, consequently reducing identify the ‘best practice’ protocol that gives the highest sig- the sensitivity of the antibody. Polyclonal antibodies, typically nal-to-noise ratio. The antibody must subsequently be validat- produced in rabbits by traditional immunization techniques, with ed, providing data of the diagnostic sensitivity and speciicity, booster immunizations to maximize the reactivity against the using the protocol identiied and selected (3). For new antibod- target antigen, frequently give a higher sensitivity (avidity) com- ies replacing old clones or existing markers, the laboratories pared to monoclonal antibodies, as the many antibody ‘species’ may exploit data from established protocols and publications, present react with more antigen sites. Polyclonal antibodies may facilitating the optimization and validation process, and sup- thus minimize the deleterious impact of tissue ixation and pro- porting the inal evaluation. For antibodies against new targets, cessing, thus providing a more robust assay. However, because both the technical optimization and the validation process can more antigen sites are recognized by polyclonal antibodies, the be much more challenging, especially if no benchmark data risk of cross-reaction to other proteins is increased. are available, and there is only limited information about positive and negative controls. Recently, rabbit monoclonal antibodies have been introduced to IHC (6). Some scientists hold that these antibodies combine The ongoing development of new and improved antibodies, the high speciicity of monoclonal antibodies, being generated introduced to the laboratories by the manufacturers, is very by the hybridoma technique, with greater sensitivity that results valuable and contributes to the continuous enhancement and from improved recognition of human antigens by the immune expansion of IHC. At the same time, the many choices for the systems of rabbits. Certainly, some rabbit monoclonal antibod- selection of the primary antibody also contribute to a lack of ies have signiicantly improved the quality for the immunohis- standardization of IHC and require increased resources from tochemical demonstration of challenging antigens, such as laboratories to comply with the antibody optimization and val- cyclin D1 (7), estrogen receptor (8) and CDX2 (9), providing idation process. The test process gives rise to an increased increased sensitivity and robustness, but also an analogous workload in the laboratory, and also requires a high level of speciicity comparable to the corresponding mouse mono- technical expertise to interpret the tests performed. clonal antibodies. The inal choice whether to use a mouse monoclonal antibody, a rabbit monoclonal antibody, or a pol- At present more than 180 companies offer in excess of 350,000 yclonal antibody must be determined by the individual labo- primary antibodies for clinical and research use (4), a huge ratory, as the inal performance is highly dependent on many number which underlines the complex issue. For one target pre-analytical and analytical parameters in the total test. antigen more than 100 primary antibodies may be candidates for implementation. In the IHC external quality program, NordiQC (www.nordiqc. org), it has been shown that for certain epitopes, many differ- 48 For primary antibodies the sensitivity and speciicity are the ent antibodies, including both mouse monoclonals and rabbit core elements. Ideally, the primary antibody must provide both polyclonals, may be used to obtain similar good staining re- a high sensitivity and a high speciicity to produce an accurate sults. For other epitopes the choice of the antibody is much and robust IHC assay. more critical. Monoclonal antibodies have become widely used because of their For prostate speciic antigen (NordiQC, run 27, 2009), the high speciicity, consistency, purity and commercial availability. selected IHC protocol could be based on many monoclonal Monoclonal antibodies, produced in mice according to the in and polyclonal antibodies to provide an optimal staining result, vitro hybridoma method developed by Köhler (5), or by molecu- whereas for cyclin D1 (run 33, 2011) only the rabbit monoclo- lar engineering, contain a single ‘species’ of antibody molecule, nal antibodies, clones SP4 and EP12, could be used to gener- where every antibody molecule is identical by idiotype, with a ate an optimal result. Selection of the Primary Antibody | Chapter 4 The choice of the primary antibody is highly dependent on Examples of technical issues and problems of an existing anti- the various steps of the general analytical protocol and the body and reasons to replace this: IHC stainer platform used by the laboratory. The chemical „„ An antibody from a mouse ascites harvest may give a composition of the reagents, sequence of reagent applica- positive staining reaction in human tissue with blood group tion, and mechanical handling of the sections by the stainer A (10, 11) complicating the interpretation platform, can have a deleterious effect on some antibodies, in „„ An antibody against a nuclear antigen may give an aberrant particular for some mouse monoclonal antibodies. For exam- staining reaction in the membranous or cytoplasmic com- ple, the mouse monoclonal antibody clone 1F6 for CD4 may be adversely affected by the blocking step for endogenous peroxidase. If this step is performed by using a standard 3% partment, a reaction not anticipated from available data „„ An antibody may be unstable in the diluted format, showing a signiicant loss of afinity after storage for a few days hydrogen peroxidase solution after heat-induced epitope „„ An antibody may require enzymatic proteolysis as the retrieval (HIER) and immediately before incubation with the method of epitope retrieval, compromising the robustness primary antibody, the detection of antigen is signiicantly im- of the protocol, because performance often is greatly in- paired. This protocol setting has typically been the backbone luenced by ixation time in formalin (in contrast to protocols for many fully automated IHC stainer platforms and conse- based on HIER, which is less inluenced by the formalin quently this clone will not provide the desired staining result ixation time) (12) under these circumstances. With the use of proper external and internal tissue controls the Due to the complexity of identifying the most appropriate above mentioned issues and similar technical issues should primary antibody and the high demands of optimizing and not have a diagnostic impact. However, they do present a daily validating the IHC protocols, many laboratories choose to challenge and may warrant replacement with another antibody apply ready-to-use (RTU) antibodies and RTU IHC systems having superior performance characteristics. as an alternative. It is also possible to purchase a total IHC system where the primary RTU antibody is calibrated, along Concerning practical utility in diagnosis, a new antibody may with the detection system and a protocol which is optimized provide an improved sensitivity and/or speciicity compared and validated by the manufacturer. The laboratories must with the old antibody. Many antibodies have been commer- still verify and validate the end results on their own tissue cially available for years or decades and still provide excellent samples, but the optimization process to identify the best staining results, whereas other ‘troublesome’ antibodies should epitope retrieval method, antibody titre, choice of detection be replaced when a new and improved antibody becomes system etc., has been carried out by the manufacturer, an available. The mouse monoclonal antibody cocktail AE1/AE3 approach which can lead to a better standardization of IHC. against ‘Pan-cytokeratin’ and a rabbit polyclonal antibody against S100 are both examples of excellent antibodies, which have been commercially available for more than 20 years and Chapter 4.3 A New (Replacement) Antibody for an Old Marker yet still provide optimal IHC results (www.nordiqc.org). They have thus maintained their utility through the general transition and development of the IHC reagents and methods, including The continuous focus to develop and to optimize IHC fre- introduction of antigen retrieval/HIER and IHC automation. quently encourages laboratories to test and evaluate new antibodies, as replacements for existing and already imple- Other antibodies with low speciicity and/or sensitivity have re- mented antibodies in their daily diagnostic practice. The in- mained for diagnostic use for a long time because no better al- citements to test and compare data concerning a new anti- ternatives were available. With new immunogens and improved body can be manifold and relate to technical, diagnostic or antibody production methods many improved antibodies have economic issues. become available, a fact that may be overlooked by labora- 49 Chapter 4 | Selection of the Primary Antibody Figure 4.2 A) Intestinal neuroendocrine tumor. Staining for synaptophysin using the mouse monoclonal antibody DAK-SYNAP. Virtually all the neoplastic cells are distinctively demonstrated. The staining reactivity is as expected and conirms the neuroendocrine differentiation of the neoplasm. Figure 4.2 B) Intestinal neuroendocrine tumor. Staining for synaptophysin using the mouse monoclonal antibody clone SY38. An inferior staining pattern is seen compared to the result expected and obtained by new antibodies against Synaptophysin (Fig 4.2A). From a technical aspect, the antibody was optimized to give the best possible result, but still is inferior, underlining the low afinity of the clone SY38. tories. The mAb clone SY38 against synaptophysin and the This means that if the laboratory uses ixatives other than for- mAb clone 35βH11 against cytokeratin 8 were also introduced malin (e.g. B5) then tissues ixed with both reagents must be to the laboratories more than 20 years ago. However, when included to monitor the performance of both. It is essential that comparing the performance of these two antibodies with new tissues ixed for different times are included in the technical antibodies, the data (www.nordiqc.org) clearly indicate that the optimization process, in order to evaluate the impact of the for- new antibodies for these two target antigens provide superior malin ixation time. In a routine situation, the material sent to staining performance and should replace the old ones (Figure pathology departments may undergo an enormous range of 4.2 and Figure 4.4). ixation times, as much as 2-168 hours, with clear effects on the performance of some antibodies (see Chapter 2). Due to the comprehensive workload and demands for techni- 50 cal expertise associated with antibody selection and protocol For the technical optimization process it is very valuable to per- optimization, some laboratories hesitate to perform evaluations form the tests on tissue microarrays (TMA) composed of cores and comparisons of new antibodies (typically new clones) with of different normal tissues. These should be processed and existing antibodies, while other laboratories perform antibody ixed for different times according to the routine and standard evaluations regularly. conditions used by the laboratory. When a new antibody is tested and compared to an existing Using serial sections of a TMA as shown in Figure 4.3, identi- reagent, the laboratory must carefully design the frames and ication of the protocol giving the best technical result is facil- conditions of the test. The technical optimization process must itated. The antibody can be applied in different concentrations be performed on material/tissue processed in the same way as e.g. 1/50, 1/200 and 1/800 as a starting point, and for all the ti- the diagnostic material. Processing factors to include are ix- ters selected the different epitope retrieval methods used by the ative(s) used, time range in ixation, decalciication methods, laboratory can be tested (this method corresponds to the ‘test if relevant, and dehydration conditions used in the laboratory. battery’ approach advocated by Shi and colleagues to deter- Selection of the Primary Antibody | Chapter 4 mine optimal retrieval conditions – see Chapter 3). In this testing ent samples are recommended for inclusion in the validation set phase, the laboratory should also include the protocol settings (3), relecting the range from no expression to high expression of recommended by the manufacturer of the primary antibody. the antigen of interest. In both the optimization and validation process it is highly recommended that normal tissues are included as these tissues in general will express a more consistent level of antigen when compared with neoplasms. During this process, it is critical that the staining patterns, the distribution and sub-cellular localization of staining, the number of cells demonstrated and the staining intensity are evaluated for the antibodies being compared. In this evaluation phase, focus should be on the reasons for initiating the test; whether related to technical issues or diagnostic issues. It is also important that other observations are registered and taken into consideration. If the new antibody meets predetermined expectations, and functions well within the basic laboratory protocols, the antibody can be implemented. However, the replacement of an existing antibody that has a well-described and long history of performance will at least for a period compromise intra-laboratory standardization until more experience is gained with the new reagent. When a new antibody has been implemented it is recommended to monitor the staining results and register any aberrant results on a regular basis. Communication with vendors and with other laboratories, in addition to studies of the literature describing Figure 4.3 TMA for the technical optimization process of primary antibodies. All specimens are ixed in 10% neutral buffered formalin and processed accordingly to the SOP in the laboratory. the antibody in question, is crucial to gain reliable information of When the laboratory has identiied the protocol giving the best Chapter 4.4 A New Antibody for a New Marker in the Pathology Laboratory technical result for the new antibody, this protocol and the current- performance and interpretation of a new primary antibody. ly used antibody should be tested simultaneously and validated on the same material. In this context, it is of high importance that IHC is an indispensable assay in the armamentarium of diag- different tissues/diseases/neoplasms with a wide range of ex- nostic tools for the pathologist, primarily to immunophenotype pression levels of the target antigen are tested. Tissues with high and classify the neoplasms concerning lineage or origin (such level, low level, and no expression must be represented in order as carcinoma, melanoma, lymphoma etc.). Secondary IHC is to evaluate both sensitivity and speciicity. A key question is; how used as a predictive and prognostic asset (e.g., in breast car- many samples are needed to secure a solid validation? No spe- cinoma to demonstrate and quantify hormone receptors). IHC ciic number can be pre-determined, as this will depend on the is also rapidly inding application in theranostics/companion marker in question and its usage. If the marker is only used within diagnostics where the therapy selected for the individual pa- a restricted diagnostic area, or the target antigen is rarely encoun- tient in part is based on the outcome of the IHC tests. A related tered, then the number might be less compared to a marker used use is for 'genomic IHC', where primary antibodies are used extensively in a number of different areas. A minimum of 25 differ- to identify speciic proteins that are indicative for a genomic 51 Chapter 4 | Selection of the Primary Antibody abnormality or mutation. As a consequence of these novel de- These central topics must be thoroughly studied before the mands and possibilities, many new markers are being intro- marker is validated against control tissues and inally imple- duced to IHC. Laboratories must be prepared to establish and mented for clinical use. validate these central and vital assays. When new diagnostic markers are introduced, the focus areWhen setting up and evaluating an IHC marker as replacement as and utility of these markers are frequently directed towards for an existing marker, the laboratory has a set of benchmark very speciic ields, and they highlight the diagnostic beneits data that facilitates interpretation of the comparative study (Sec- and challenges within these ields. Additional studies must be tion 4.2). For a brand new marker with no history and no bench- performed to elucidate the utility of the marker both within in mark data in the laboratory, the optimization and validation the initial areas described and in a broader perspective. It has process is much more challenging and a number of additional often been shown that new emerging markers initially have been considerations have to be taken into account. The clinical and reported to be diagnostically very valuable, providing high diag- diagnostic application of the new marker must be deined and nostic sensitivity and/or speciicity for a certain target. However, relevant literature and publications must be carefully reviewed subsequent testing by other groups may generate radically dif- concerning both diagnostic potential and the IHC methodologi- ferent results due to IHC methodological issues, including the cal parameters applied in the published studies. choice of a different primary antibody or clone. In the review paper by Ordonez (13), it is shown that the irst IHC studies pub- If available, detailed data must be collected concerning the IHC lished for PAX8 (paired box protein 8) reported B-cell lympho- staining results reported in the literature, with focus on the di- mas and neuroendocrine carcinomas as being positive, while agnostic sensitivity and speciicity. The number, selection, and later studies reported these neoplasms to be negative. The dif- diversity of samples included in the reported studies all have an ferent conclusions and staining patterns were mainly related to impact on the results obtained. For a laboratory considering im- the reactivity of the antibodies used in the studies. The reason plementation of an antibody to a new marker, which is described appears to be that antibodies reacting with the N-terminal of the of value in a speciic differential diagnosis, it is crucial to have PAX8 antigen also show a cross reaction with PAX5 expressed information about the reactivity of the antibody in a broad range by normal and neoplastic B cells, as the antigen site on the of clinically relevant samples. Descriptions of staining patterns, N-terminal is shared by all PAX family members. If antibodies both for the sub-cellular (nuclear, cytoplasmic, membranous) directed towards the C-terminal of the PAX8 antigen are utilized, and tissue distribution are essential. These data are used for B cells and neuroendocrine carcinomas test negative. the internal optimization and validation process, including the identiication of positive and negative tissue controls. It is also central to search for guidelines or discussions concerning the ‘cut-off’ value or staining threshold, for interpretation and how to Chapter 4.5 A New Experiment in the Pathology Laboratory integrate and report the result in the inal description to the clinicians. Additionally, focus must be centered on details of the IHC Having focused on the issues and considerations related to the methods given in reported studies. First of all, the studies must requirements for the test set-up of an antibody as replacement be performed on samples processed identically, or in a manner of an existing antibody, and how to implement a new marker, similar to the conditions used by the laboratory. The IHC ana- this section will describe the considerations of how to start the lytical parameters will have a signiicant impact on the staining process of implementing a marker, for which there is only lim- results. For this reason, focus should be on three central issues: ited or no data in the literature. The focus will be on the pitfalls „„Choice of antigen retrieval method „„The selection of primary antibody „„IHC stainer platform adopted (or manual protocol if that is the choice) 52 associated with this process. As mentioned previously, benchmark data and antibody performance history are of high value when laboratories evalu- Selection of the Primary Antibody | Chapter 4 ate new clones and markers to serve as replacements, or new functionality in different applications such as IHC, WB, im- tests in a well-described area. When a new undescribed mark- munoprecipitation (IP) and enzyme-linked immunosorbent er is evaluated for possible incorporation as a laboratory test assays (ELISA). Occasionally, they may also provide informa- many other obstacles and methodological issues arise. Some tion on documented cross-reactivity in different species. If no of the challenges and questions are as follows: information is listed about the functionality of the antibody for „„ Which samples (tissues/cells) can be used as negative and positive controls? „„ Does the antibody work on formalin-ixed, parafin-embedded (FFPE) material? „„ What is the expected sensitivity and speciicity of the antibody? „„ How should the staining results be interpreted? „„ What is the reproducibility of the antibody and the test protocol? IHC on FFPE material, this does not mean that the antibody will not work on this type of material, but the laboratory must perform studies to conirm (or disprove) effectiveness. For these purposes antibodies showing narrow bands for the molecular weight of the expected target antigen by WB should be preferred; in most instances such antibodies will provide a higher signal-to-noise ratio in the IHC assay. In this context, it has been shown that only about 50% of antibodies that are effective in other assays, will function in IHC (4). In addition to the challenges listed, it is dificult to determine in Cell lines can also be used as control material in the initial test which order to approach these questions. It is not possible to phase for antibodies that purport to demonstrate proteins gen- evaluate the overall performance of the marker on FFPE mate- erated by gene alterations, such as translocation proteins (ana- rial or the impact of the pre-analytical conditions, if the proper plastic lymphoma kinase (ALK) in lung carcinoma and anaplas- negative and positive controls are not identiied. In practice tic large cell lymphoma), or fusion proteins (ASPL-TFE3 in renal this usually means that the laboratory, or the researcher, will papillary carcinoma), or other proteins not expressed in normal start the evaluation process on one set of materials and dur- adult tissues, such as fetal transcription factors (myogenin). The ing the optimization and validation process they may have to cell lines can be used as checkpoints to evaluate the afinity and go back and repeat some studies on other materials, as more functionality of the antibody for the target antigen, as expressed in knowledge about these selections is generated. the cell lines processed under selected conditions. Subsequently, the antibody must be evaluated on TMAs with a wide range of The identiication of appropriate tissue samples suited to be in- normal and neoplastic samples, potentially showing the antigen cluded in the test set-up may be dificult for antibodies with no at different expression levels ranging from negative to high. The solid reference data concerning tissue expression. As an alter- complexity of the implementation and optimization process of an native, cell lines with known expression levels of the antigen can antibody against a protein not expressed by normal cells is re- be the best, or only, type of material that is useful in the initial lected in the set-up for the ALK translocation protein. No normal phase of the process. In order to evaluate how pre-analytical cells express this protein and in the neoplasms a high-level of conditions will affect the afinity of the primary antibody, the cell expression is typically seen in anaplastic large T-cell lymphomas, lines can be processed by different methods e.g. adjusting the whereas lung adenocarcinomas only express low-levels of ALK type and length of ixation etc., and FFPE cell line pellets with the protein (14). To ensure the development of a protocol demonstrat- characteristic of tissue processing standards can be generated. ing ALK in a wide range of tissues and different tumor types, the laboratory must have access to samples with this range of anti- It may be beneicial to test, side by side, more than one an- gen expression to perform a test of the inal protocol. tibody/clone for the same epitope. For example, two or three antibodies from different vendors may be selected as candi- When testing a new antibody with no or only limited history, it is dates for optimization and validation tests. Vendors typically also important to establish data for the reproducibility of the anti- give information about the speciicity of their antibodies by body test. For the individual laboratory this includes inter-run re- showing western blot (WB) panels and information about the producibility, using same reagents, protocol, tissue, and inter-lot 53 Chapter 4 | Selection of the Primary Antibody reproducibility to see if the antibody provides identical results can inluence the inal diagnosis and management of patients when different lots and aliquots of the antibody are used. in a clinical setting, which underlines the importance of using high quality antibodies in clinical laboratories. As described, It is also necessary to evaluate the speciicity of a new antibody. the overall goal is to use antibodies that are speciic, sensi- This aspect is described in depth in section 4.2, and only sup- tive, robust (not inluenced by pre-analytical parameters) and plemental considerations are listed here. A standard method is reproducible (inter- and intra-laboratory), as these prerequi- to employ absorption by blocking peptides. Identical peptide se- sites are the foundation for standardization in the ield of IHC. quences, to those used to generate the antibody, are incubated The standardization of IHC is, in fact, compromised by the with the antibody in great excess. The antibody with and without large number of reagents, including primary antibodies, that the blocking peptide is then applied to two serial sections of tissue are available from the many manufacturers. It can be dificult, samples expressing the target of interest. If the antibody is specif- or virtually impossible, for the individual laboratory to consis- ic, the addition of the blocking peptide will result in loss or at least tently select and use the optimal antibodies in their total IHC ar- a major reduction of staining on the tissue sample, compared to mamentarium. As mentioned in the previous sections, new an- the positive result obtained by the antibody without blocking re- tibodies are continuously being introduced to the market and a agent. This test documents the speciicity of the antibody to the balance must be found for replacement of existing antibodies in immunogen, but it does not test for any staining in ‘off-target’ (cross current use. From a practical point or view, the skills and exper- reactive) sites, and aberrant (unexpected) staining may still occur. tise necessary to conduct optimization and validation studies are Negative cell lines and negative tissue controls are essential and considerable. In consequence, laboratories might refrain from must be included in the evaluation of antibody speciicity. performing complete validation due to lack of resources, therefore they continue to use already implemented antibodies, even New initiatives from research groups have generated publicly though they may produce inadequate staining results. available tools that facilitate the validation of new markers. Antibodypedia (www.antibodypedia.com) is a searchable database When some laboratories rapidly integrate new antibodies, of antibodies against human proteins. It aims to provide the re- and especially new clones for the same target antigen, while search community and scientists with information on the effec- other laboratories continue to use the old well-established anti- tiveness of speciic antibodies in speciic applications, including bodies, huge differences in the IHC results can be obtained by IHC. The database is generated by academic groups submitting the two scenarios. The differences can be related to all the core their validation data, and by information from the commercial pro- issues of speciicity, sensitivity, robustness and reproducibili- viders of the antibodies. ty. Some antibodies were initially designed and brought to the market for one target area but later studies may have revealed Another resource is the Human Protein Atlas project other possible areas of use. This aspect, in combination with (www.proteinatlas.org) where information on IHC-based pro- the continuous optimization of IHC technical methods, with more tein expression proiles are available for a large number of nor- eficient antigen retrieval methods, and detection systems hav- mal human tissues, cancers and cell lines. For the majority of ing improved sensitivity, continues to expose previously unde- proteins the sub-cellular distribution and transcription expres- scribed reaction patterns for many antibodies. sion levels in cell lines are also available. One of the irst antibodies against cytokeratin, high molecular weight (CK-HMW), was the mouse monoclonal antibody Chapter 4.6 Examples of Good and Poor Antibodies clone 34βE12. It was introduced for IHC in FFPE material in 1982 (15) and primarily used to demonstrate the CK-HMW subtypes 1, 5, 10 & 14. After 30 years, this antibody is still 54 The selection and choice of the primary antibody will have a the most widely used marker for CK-HMW, and diagnostic signiicant impact on the IHC result. Consequently, this choice utility has been documented in a remarkably high number of Selection of the Primary Antibody | Chapter 4 publications. In particular, it has been shown to be valuable different use is critical. It has been shown that the mAb clone for the demonstration of CK-HMW in basal cells of prostate 34βE12 is less speciic for this diagnostic context, compared glands for the differential diagnosis of non-invasiveness and with newly introduced antibodies to CK-HMW (run 16 and 38, invasiveness. The demonstration of CK-HMW is now also per- www.nordiqc.org); a cross reaction with a cytokeratin low mo- formed in breast pathology to identify the basal cell like sub- lecular weight is seen. Consequently, the mAb clone 34βE12 type of breast carcinoma and to differentiate this entity from cannot be recommended to be used as a general antibody other subtypes. The speciicity of the primary antibody for this for CK-HMW in breast (see Figure 4.4A and 4.4B). Figure 4.4 A) Breast ductal carcinoma. Staining for CK-HMW using the mouse monoclonal antibody clone D5/16B. The myoepithelial cells decorating the carcinoma in situ component show a cytoplasmic staining reaction, while the iniltrating tumor cells are negative. Figure 4.4 B) Breast ductal carcinoma. Staining for CK-HMW using the mouse monoclonal antibody clone 34βE12. The iniltrating tumor cells show a positive staining reaction due to a cross reaction of the primary antibody to an unidentiied target antigen. Figure 4.5 A) Mantle cell lymphoma. Staining for CD5 using the rabbit monoclonal antibody clone SP19. Virtually all the neoplastic B cells show a moderate staining reaction. In the centre a normal T cells show a strong staining reaction. Figure 4.5 B) Mantle cell lymphoma. Staining for CD5 using the mouse monoclonal antibody clone CD5/54/F6. Only the normal T cells are demonstrated, while all the neoplastic B cells are false negative. 55 Chapter 4 | Selection of the Primary Antibody For cells with low expression of a target antigen, the central A issue in a diagnostic setting is to provide high sensitivity. In the immunophenotyping of small B-cell lymphomas the demonstration of CD5 is important. The neoplastic B cells will typically only express limited amounts of CD5, compared with the high-level CD5 expression in any admixed normal Tcells. It is therefore mandatory to visualize the entire range of expression of the target antigen in the tissue sample tested. Low afinity antibodies may show an acceptable IHC staining result in cells with high levels of the target antigen (T cells), but inferior performance in cells with low levels (neoplastic B cells). This observation further emphasizes the need to optimize and validate the total IHC protocol on tissue with both high- and low-level antigen expressions. Liver. Small cell lung carcinoma. Many CD5 antibodies are commercially available and a huge difference regarding the effective sensitivity is seen for the B most commonly used antibodies. Using optimized protocols, the mouse monoclonal antibody clone CD5/54/F6 detects CD5 in high-level expression sites but not in low-level expression (see Figure 4.5A and 4.5B). The mouse and rabbit monoclonal antibodies clone 4C7 and SP19, respectively, detect CD5 in both low- and high-level expression sites. The selected antibody must also provide high robustness and give a consistent and reliable result only minimally inluenced by the variations in the tissue handling process, and other luctuations inevitably seen in an IHC analysis. Antibodies that require heat-induced antigen retrieval (AR or HIER – Chapter 3) should be preferred to antibodies that require enzymatic pre-treatment, as HIER reduces the impact of variations in formalin ixation time compared with enzymatic pre-treatment. Enzymatic pre-treatment must be adjusted to the duration of formalin ixation to provide a consistent IHC result, which can be dificult to accomplish in a routine setting, where the tissue samples inevitably show a wide variation in the ixation time. When comparing the data sheets of the two mouse monoclonal antibodies against broad spectrum cytokeratin, clones AE1/AE3 and MNF116, both react with the most relevant subtypes of cytokeratin and can be used on FFPE material. However, the IHC protocol for the clone MNF116 must be based on enzymatic pre-treatment, reducing the robustness of the IHC assay compared with the IHC assay using clone AE1/AE3 (see Figure 4.6A and 4.6B). 56 Liver. Small cell lung carcinoma. Figure 4.6 A) Staining for broad spectrum cytokeratin using the mouse monoclonal antibody clones AE1/AE3. The IHC protocol is based on HIER. In the liver sample hepatocytes and the bile epithelial cells are demonstrated and in the small cell lung carcinoma, all the neoplastic cells are distinctively demonstrated. A consistent and optimal staining reaction is seen in both samples tested using the same IHC protocol. B) Staining for broad spectrum cytokeratin using the mouse monoclonal antibody clone MNF116. The IHC protocol is based on enzymatic pre-treatment. The staining reaction in the liver sample is as expected and comparable to the staining obtained by the clone AE1/AE3. As the hepatocytes are distinctively demonstrated, the staining result is optimal according to e.g. NordiQC criteria. The staining of the small cell lung carcinoma is inadequate, as only few cells are demonstrated. The IHC protocol based on enzymatic pre-treatment is in this scenario less robust as the sensitivity is signiicantly reduced in the lung sample and the protocol must be adjusted for the individual samples tested to provide an optimal result. Selection of the Primary Antibody | Chapter 4 Chapter 4.7 Current and Future Challenges in Pathology Laboratories IHC result is often unique to each individual IHC automated system. Typically, the different IHC systems provide different sensitivity levels depending on, e.g. the detection systems and It has been recognized in several publications that stand- HIER settings that are used. For the RTU antibodies in gener- ardization of IHC is fundamental for reproducible and re- al, the vendors can adjust the concentration and/or incubation liable results. In order to fully exploit IHC in a diagnostic time of the primary antibody to regulate the results of IHC tests, setting, consistency must be achieved both in the individual such that the end results on their different stainer systems will laboratory and in between different laboratories. Because be closely comparable. “IHC is technically complex, and no aspect of this complexity can be ignored, from the moment of collecting the Transition from manual IHC and LDT to fully automated IHC, specimen to issuance of the inal report” (1), attention must using RTU antibodies in closed IHC stainers (see chapter 9), be given to all variables inluencing the result. This chapter with vendor-determined preset analytical parameters, is es- has mainly focused on the selection of the primary antibody sential to standardize the ield of IHC. In general, overall IHC and the challenges associated. Due to the many pitfalls and performance will be improved as a result of higher reproduci- extensive requirements for laboratories to accomplish the bility, and the use of high-quality and state-of-the art reagents optimization and validation process, implementation of RTU being offered by the vendors. However, it is recognized that antibodies has expanded in the ield of diagnostic IHC. RTU for some antibodies performance might be impaired as preset antibodies are typically accompanied by speciic protocols analytical conditions and reagents selected for the IHC stainer and guidelines concerning the choice of reagents and the system may fail to optimize performance of certain antibodies tissue controls to be used in conjunction with the RTU prod- under certain conditions of processing and ixation. The mAb uct, thereby facilitating implementation in the laboratory for clones 1F6 for CD4 and PG-B6 for BCL-6, are examples of clinical use. For optimal performance, a RTU antibody must antibodies that are adversely impacted by some aspects of be used within a total IHC system, where the product is used reagent composition and by the protocol used to quench en- strictly according to the guidelines given by the vendor. It dogenous peroxidase. If the quenching step is performed af- is important to stress that if a RTU antibody is used with a ter HIER and before the incubation with the primary antibody, modiied protocol, then a full optimization process must be and by the use of 3% hydrogen peroxide (a standard setting performed by the laboratory. in many fully automated IHC systems) then the afinity of these two antibodies is signiicantly compromised and the IHC re- The class II and III IVD tests, such as HercepTest™ and PATH® sult is unreliable. For other markers, different staining patterns WAY for HER2 demonstration in breast cancer, have for years can be seen when applied on different IHC stainer platforms, been used by the laboratories as RTU systems, and have prov- even though comparable protocols regarding sensitivity lev- en to be superior to laboratory developed tests (LDT). A gen- els, antigen retrieval conditions and detection reagents are eral transition to convert class I antibodies from LTD to RTU being used (see Figure 4.7A and 4.7B). Such differences may is now an ongoing process. Related to the standardization be the result of several small differences, not having a signif- of diagnostic IHC, RTU systems from the vendors should be icant impact in isolation, but in combination antibody binding aligned to give comparable results for a speciic target anti- to target antigen is impaired. gen, irrespective of different reagents, methods and IHC automated platforms being used by the individual vendors of the In these circumstances, alternative antibodies to the same RTU systems. In principle, the exact methodology applied to target might resolve the issue. Laboratories changing from generate the inal IHC result is of minor importance, providing manual to automated IHC, or making a change from one IHC that the result is concordant to the expected, regarding the stainer platform to another must take such parameters into overall sensitivity and speciicity of the test. The combination of consideration and should monitor and manage any differenc- reagents and procedural parameters that generates the best es observed. 57 Chapter 4 | Selection of the Primary Antibody Figure 4.7 A) Liver. Staining for low molecular weight cytokeratin using the mouse monoclonal antibody clone 5D3. The IHC protocol is performed on a semi-automated IHC stainer from vendor A using the reagents and protocol settings linked to the system. All hepatocytes and bile duct epithelial cells are demonstrated as expected. Figure 4.7 B) Liver. Staining for low molecular weight cytokeratin using the same mouse monoclonal antibody clone 5D3. The IHC protocol is performed on a fully-automated IHC stainer from vendor B using the reagents and protocol settings linked to the system. The bile duct epithelial cells are demonstrated, whereas the hepatocytes are negative. From a practical point of view, and in relation to issues as primary antibody concentration and protocol must be opti- seen for the endogenous peroxidase blocking step, it is im- mized separately for each of the three areas. Using a pro- portant that the 'backbone' of the closed RTU stainer system tocol optimized for surface IgK on lymphocytes produces a still has some lexibility and is modular. Adjusting the pro- much too intense and un-interpretable result for demonstra- tocol should be the rare exception for a ‘dificult’ antibody. tion of plasma cells in bone marrow samples. Other antibodies and markers should irst be explored. If no other marker is available, the laboratory must then be able No vendor can offer all primary antibodies as RTU, and RTU to modify the manufacturer’s preset protocol settings to es- systems from different vendors frequently give different IHC tablish a modiied antibody-directed protocol to give the de- staining results. This underlines the need for some degree of sired staining result. If this approach fails, the particular test lexibility and modularity retained in the IHC stainer systems to may be performed manually or by another stainer system. be used judiciously by the laboratories. At present, there are no uniform guidelines concerning the staining requirements Typically, RTU systems will support standardization of IHC, and no descriptions for the expected levels and staining pat- but occasionally it can be impossible to develop a single terns of the vast majority of immunoreactions. Dako Atlas of test protocol that is applicable to all tissue samples pre- Stains gives examples, but the lack of a complete library with sented to the laboratory, for all diagnostic purposes. Some recommended tissue controls complicates the development of antigens show such a wide range of expression levels that properly calibrated RTU systems, as well as the inal validation one protocol will not it all diagnostic applications. As an and implementation of the markers cross all laboratories. example, the demonstration of immunoglobulin kappa (IgƘ) light chains can be based on use of the same primary antibody, for quite different purposes, e.g. visualization of membranous IgƘ located on lymphocytes, intra-cytoplasmic IgƘ in plasma cells, and extracellular IgƘ deposits in autoimmune diseases. However, for these different purposes the 58 Selection of the Primary Antibody | Chapter 4 References 1. Taylor CR. The total test approach to standardization of immunohistochemistry. Arch Pathol Lab Med 2000;124(7):945-51. 2. Hardy LB, et al. Immunohistochemistry validation procedures and practices: a College of American Pathologists survey of 727 laboratories. Arch Pathol Lab Med 2013;137(1):19-25. 16. Lacroix-Triki M, et al. Value of cytokeratin 5/6 immunostaining using D5/16 B4 antibody in the spectrum of proliferative intraepithelial lesions of the breast. A comparative study with 34betaE12 antibody. Virchows Arch 2003;442(6):548-54. 3. Goldstein NS, et al. Recommendations for improved standardization of immunohistochemistry. Appl Immunohistochem Mol Morphol 2007;15:124–133. 4. Björling E, Uhlén M. Antibodypedia, a portal for sharing antibody and antigen validation data. Mol Cell Proteomics 2008;7(10):2028-37. 5. Kohler G, Howe SC, Milstein C. Fusion between immunoglobulin -secreting and nonsecreting myeloma cell lines. Eur J Immunol. 1976;6:292-295. 6. Rossi S, et al. Rabbit Monoclonal Antibodies. A comparative study between a novel category of immunoreagents and the corre sponding mouse monoclonal antibodies. Am J Clin Pathol 2005; 124(2):295-302. 7. Pruneri G, et al. SP4, a novel anti-cyclin D1 rabbit monoclonal antibody, is a highly sensitive probe for identifying mantle cell lymphomas bearing the t(11;14)(q13;q32) translocation. Appl Immunohistochem Mol Morphol 2005;13(4):318-22. 8. Cheang MC, et al. Immunohistochemical detection using the new rabbit monoclonal antibody SP1 of estrogen receptor in breast cancer is superior to mouse monoclonal antibody 1D5 in predicting survival. J Clin Oncol 2006;24(36):5637-44. 9. Borrisholt M, Nielsen S, Vyberg M. Demonstration of CDX2 is highly antibody dependant. Appl Immunohistochem Mol Morphol 2013; 21(1):64-72. 10. Kliman et al. A mucin-like glycoprotein identiied by MAG (mouse ascites Golgi) antibodies. Menstrual cycle-dependent localization in human endometrium. Am J Pathol 1995;146(1):166-81. 11. Spicer et al. Some ascites monoclonal antibody preparations contain contaminants that bind to selected Golgi zones or mast cells. J Histochem Cytochem 1994;42(2):213-21. 12. Webster JD, et al. Effects of prolonged formalin ixation on diagnostic immunohistochemistry in domestic animals. J Histochem Cytochem 2009; 57(8): 753–761. 13. Ordonez GN. Value of PAX 8 Immunostaining in tumor diagnosis: a review and update. Adv Anat Pathol 2012;19:140–151. 14. Kenudson MM. A novel, highly sensitive antibody allows for the routine detection of ALK-rearranged lung adenocarcinomas by standard immunohistochemistry. Clin Cancer Res 2010;16(5):1561–71. 15. Gown AM, Vogel AM. Monoclonal antibodies to intermediate ilament proteins of human cells: unique and cross-reacting antibodies. J Cell biol 1982;(95):414-424 59 Part I: The Staining Process Chapter 5 Staining Protocol Optimization Lars Jacobsen, PhD, MSc Majken Nielsen, PhD, MSc Soie Månsson, PhD, MSc Lars Rudbeck, PhD, MSc Pro •to •col (n.) The plan for a course of medical treatment or for a scientiic experiment. The American Heritage® Medical Dictionary Staining Protocol Optimization | Chapter 5 Chapter 5.1 Introduction thology laboratories, it was reported that one out of three laboratories does not have written validation procedures Ever since the irst use of immunohistochemistry (IHC) for for introducing new, non-predictive antibodies. The survey the diagnosis of cancer, reproducibility has been a key con- also showed that 14% of the laboratories did not validate cern. Lack of reproducibility, both inter- and intra-laborato- the performance of the most recently introduced IHC test. ry, shows the need for improved standardization, yet many Of those laboratories that did validate, 75% used fewer than attempts to improve the quality of IHC staining worldwide 21 cases, and 42% did not include weak or focally positive are limited by two main factors; pre-analytical and analytical cases (1). factors in the form of the quality of IHC reagents and the staining protocols used. The survey points to the unclear deinitions of the terms validation and veriication as one reason for the relatively high Diagnostic IHC tests can be divided in two main types; those percentage of labs that do not validate new IHC assays. The that are qualitative and those that are semi-quantitative. The FDA deinition of validation is “conirmation by examination majority of IHC tests are qualitative, where the resulting stain is and provision of objective evidence that the particular re- interpreted only as positive or negative. These types of stains quirements for a speciic intended use can be consistently can to some extent involve quantitation in the form of a cut- fulilled” whereas veriication is “a study used to determine off point or threshold for positivity (e.g. >10% stained cells is whether a test system meets speciications”. CAP guidelines indicative of a positive result). Interpretation of the qualitative (2) recommend the use of 10 positive and 10 negative neo- IHC tests is focused on the correct cellular localization of the plasm samples for previously well-described antibodies, but staining reaction and staining of the correct tissue structures. do not specify the need to include weak or focally positive Appropriate positive and negative controls are pivotal for opti- samples. Either way, there seems to be a lack of testing (ver- mization of these tests. iication and/or validation) of new assays to ensure optimal staining; not only when introducing new antibodies, but also Semi-quantitative IHC tests are interpreted according to an when changing ixative, antigen retrieval, detection system, arbitrary scoring range (e.g. from 0 to 3+) that relects anti- and even instrumentation. gen expression by means of the staining intensity and distribution, as well as the percentage of positive cells. Ideally, the Data from NordiQC EQA schemes also suggest that the semi-quantitative tests are optimized and calibrated against concept of optimal staining protocols is not implemented reference control materials, having known levels of expression throughout pathology laboratories. Accumulated assess- of the target antigen. This way the staining result of a patient ment run data show that ~30% of all stains assessed by sample can be itted into the scoring range. NordiQC are scored as insuficient, with some of the reasons being new antibodies, new techniques and new platforms. This chapter focuses on the optimization of qualitative IHC A weak or false negative stain accounts for 85-90% of the tests, where the purpose of the IHC protocol is to ensure that insuficient stains (3). the IHC staining unambiguously determines if the antigen is, or is not, expressed in the tissue. Optimal staining protocols This chapter will describe the requirements for an optimal are fundamental to reduce the risk of false negative results, staining protocol with focus on signal transfer and signal gen- while at the same time not introducing a false positive result. eration in each step from antigen retrieval to visual chromoge- An optimized protocol for a qualitative IHC test should thus nic deposition. The chapter is based on the research done for not aim to accurately relect the expression level (quanti- Dako’s FLEX Ready-to-Use solution and automation, on both tative) of the antigen in the tissue, but should ensure that Dako Omnis and Autostainer Link instruments. Other vendors the staining accurately relects whether or not the antigen may have different approaches to protocol optimization. is expressed at all. In a recent survey of 727 American pa- 61 Chapter 5 | Staining Protocol Optimization Chapter 5.2 The Basis for an Optimal Staining Protocol Correctness Correctness relates to the protocol’s ability to correctly visualize the antigen in the tissue, at both low and high expression An optimal staining protocol is characterized by two main levels. A correct representation of antigen expression in a qual- properties: itative IHC test is not a linear representation of staining intensity „„ Robustness as a function of antigen expression levels. Correctness in a „„ Correctness qualitative IHC test depends on whether the antigen is truly expressed or not, thus balancing on the ine line of speciic Robustness signal versus noise. Robustness is the protocol’s ability to minimize variation in the staining result by ‘absorbing’ variations related to the Optimal Signal Transfer pre-analytical factors, including the degradation and recov- It is essential that the visualization of the antigen is performed ery of the antigen (Chapter 2 and Chapter 3), as well as the with little or no signal loss, meaning that all antigen/antibody analytical factors, by ensuring optimal signal transfer in the complexes are converted into a signal. A signal becomes vis- staining reactions. As illustrated in Figure 5.1, the pre-analyt- ible when enough antigen/antibody complexes are converted ical variations in tissue handling (red-colored bar) are carried into a chromogenic deposit at a single site in the tissue. By se- through the analytical steps (grey-colored bars) adding to lecting incubations that result in visually saturated signals, the the total variation. When non-optimized analytical protocols varying factors such as time and eficiency are not transferred are used, the total variation becomes even greater. Optimal into the visual signal. signal transfer reactions in the analytical steps ensure that all available targets in the tissue are detected and thereby all Each step in an IHC protocol is necessary for the conversion contribute to the visual signal. This optimized staining reduc- of signal from the antibody/antigen/enzyme complex into a es the contribution from analytical variation, thus increasing chromogenic deposit that the human eye can see. Each active the robustness of the IHC test. step (not including wash/blocking steps) in this conversion to a Non-optimal incubations Optimal signal transfer incubations IHC protocols with unsaturated incubations add to the total variation IHC protocols with saturated incubations minimize the total variation + Pre-analytical variation + + Analytical variation = + Total variation Pre-analytical variation + + Analytical variation = Total variation Figure 5.1 Optimal signal transfer protocols ensure that analytical variation has as little as possible inluence on the total variation. “Unsaturated/saturated” refers to the protocols’ ability to convert available targets into signal. 62 Staining Protocol Optimization | Chapter 5 Number of accessible epitopes Optimal protocol Surgery-Warm Transport-Cold ischemia ischemia Lab-Fixation and Processing Lab-Paraffin block/cutting Sub-optimal protocol Analytical phase Epitope loss A Epitope loss B Biological epitop variation A B Positive Detection level variation False negative Figure 5.2 An optimal staining protocol must ensure that as many available epitopes as possible are converted into visible signals to avoid false negative staining results while not introducing false positive results. In scenario A, the minor loss of epitopes in the pre-analytical phase does not affect the result which is above the detection level for both the optimal (red line) and sub-optimal protocol (blue line). However, in scenario B the loss of epitopes is not recovered by the sub-optimal protocol, which could result in a false negative result. The same could occur when having a sample with low expression due to biological epitope variation. The igure illustrates the challenges when creating an optimal staining protocol that is affected by pre-analytical factors, which may differ from laboratory to laboratory, together with the biological nature of the protein being examined. visible signal consists of incubations which basically follow the Staining intensity saturation curve shown in Figure 5.3. Visual ‘Saturation’ Obtaining intense staining without background is a delicate Variation Minor impact balance between chemical saturation parameters in the antigen/antibody reaction, plus the subsequent visualization reactions, versus the visual saturation of the brown end-product Variation Severe impact A robust protocol requires incubations at the saturated level for the DAB reaction. As shown in Figure 5.3, the FLEX RTU protocol has been developed with the goal of always reaching the best signal to noise ratio possible. Some pathologists may consider the resulting stain to be “overstained”, but the proper philosophy is to stain as intensely as possible with- Saturation curve of visualization Incubation time/ concentration out introducing background staining. The intense staining is a result of optimal signal transfer, where all available targets contribute to the signal. This is the aim of the protocol in a qualitative IHC test. Correctness in a qualitative IHC test relates to whether the antigen is truly expressed or not. In this Figure 5.3 Each step in the analytical phase basically follows an incubation/concentration curve. A robust protocol ensures optimal signal transfer in all steps of the analytical phase. The green box represents the visual saturation area of a FLEX RTU protocol which is the goal in order to minimize the impact of variation. context, correctness does not imply any direct proportional measurement of antigen expression. 63 Chapter 5 | Staining Protocol Optimization With the saturation curve in mind, it seems obvious that the vis- to a valid diagnosis. The analytical part of the IHC process is ualization (detection system) part of the staining protocol should controllable from a protocol point of view. It consists of antigen not be used to compensate for sub-optimal antigen retrieval, retrieval (Chapter 3), antibody selection and incubation (Chap- or insuficient antibody concentration and incubation time, i.e. ter 4), and the visualization (Chapter 6). sub-optimal chemical signal transfer. Thus, the addition of extra ampliication steps should only be used if the previous steps in In-House Protocols vs. RTUs From Vendors the protocol have been optimally conducted and the speciic Creating an optimal protocol is – in theory – quite simple: Op- protocol has not reached the optimal signal to noise ratio. timize the protocol parameters so that the antigen of interest is stained in normal tissue elements with high expression of the Chemical ‘Saturation’ antigen and at the same time – using the same protocol set- An optimal protocol should ensure that all available targets in tings – will stain the antigen in normal tissue elements with low the tissue contribute to the signal. Chemical ‘saturation’ in this expression of the antigen. Abnormal tissues of interest should context relates to ensuring that all available targets in the tis- preferably express the antigen within the upper and lower limits sue have been found and are subsequently converted equal- of the control tissue elements. Background should not cause ly into a signal. Visual ‘saturation’ is the conversion of these risk of misinterpretation of any positive signal (false positive). chemical targets into chromogenic deposits. As noted in Chapters 4 and 6, this result may be achieved either by use of RTU reagents, or by in-house optimization of antibody concentrates and detection systems, using either au- Chapter 5.3 The protocol tomated platforms or manual methods. In order for a stain to be optimal a number of requirements FLEX RTU Core Protocol must be fulilled. The variation in the staining intensity is a sum As mentioned in the Introduction, this chapter is based on the of the total variation of all possible inluencing factors. research and development of the FLEX RTU series of antibod- „„ Biological variances. The protocol must be able to identify the protocol adjustments for each of the individual antibodies expression, and more importantly unknown levels of in the series. The core protocol has been developed by testing expression in abnormal tissue. Tumors are known to exhibit a panel of antibodies and thereby inding the common denom- very heterogeneous antigen expression. inator that ensures labeling of all available targets in the tis- „„ Ischemic time. The protocol should ideally ‘iron out’ different sue. From here, each ready-to-use antibody is tested for low or degrees of antigen degradation and retrieval (to a certain high pH target retrieval requirement and inally the incubation point) (Chapter 3). time of the primary antibody. If needed for visual saturation, an „„ Time of ixation. The protocol must be able to identify the ampliication step (Linker) may be added. Concentrated an- same level of antigen expression, independent of the time tibodies have one other parameter to consider and to adjust, of ixation when in the overall validated timeframe. namely the concentration (see also Chapter 4). ixatives. The widespread use of formalin ixation has nar- The aligned core protocol will ensure an eficient worklow with rowed this requirement. minimal risk of false positive or negative results. „„ Fixative. The protocol must be usable with commonly used „„ Tissue thickness. The protocol should encompass the ef- fects of differences in section thickness on staining intensity. The aim when developing optimal IHC protocols is to achieve a robust and correct visualization of the target antigen in clinical samples with unknown levels expression, thereby contributing 64 ies. In this development, a Dako core protocol is the basis for the antigen in normal tissue, with both high and low Staining Protocol Optimization | Chapter 5 Antigen Retrieval Staining intensity with different AR times As discussed in Chapter 3, antigen retrieval is a process that is inluenced by many factors; most importantly heat and time. Anti-S100 Staining intensity However, factors such as pH, molarity, content and concen- Very high tration of the retrieval buffer, may all have major impact on the High eficiency of the antigen retrieval process, which further affects the choice of clone and dilution of the primary antibody (4). Medium Low Time and Temperature Negative Minutes 15 20 25 30 respond better to either low or high pH target retrieval solution Anti-Ki-67 Staining intensity The selected antigen retrieval heating time (and temperature) secures optimal target retrieval for groups of antibodies that 35 (Chapter 3). The antibodies are grouped and tested over vary- Very high ing incubation times. In the FLEX RTU method, one incubation time is inally selected for a group of antibodies, in order to High maximize throughput in the laboratory, while maintaining optiMedium mal or near optimal retrieval for each antibody. Low Negative Minutes 15 20 25 30 35 High expression structure Low expression structure Background Staining intensity with different antibody dilutions Staining intensity Anti-Chromogranin A Very high Figure 5.5 No variation in target retrieval eficiency from 15 min to 35 min for Anti-S100 (top). However, due to the low expression (LE) structures labeled by Anti-Ki-67 (bottom) and other antibodies (not shown) an incubation of 35 min at 97 °C is necessary for effective target retrieval of a broad range of antibodies. High Medium Low Negative Dilution 2:1 1:1 1:2 1:4 1:8 Neuroendocine (high expression structure) Axons (low expression structure) Background Figure 5.6 Anti-BCL2 staining of tonsil (FFPE). Target retrieval performed for 15 min (left) and 30 min (right). Figure 5.7 Control tissue, normal colon stained with Anti-Chromogranin A. Staining intensity of the high expression (HE) structures is strong and stable across the dilutions. The low expression (LE) structures are moderate in staining intensity at the 2:1 and 1:1 dilution with a steep fall off at higher dilutions. The optimal antibody dilution (1:1) is selected as the dilution that delivers crisp staining with the strongest possible intensity, while maintaining acceptable background. The 2:1 dilution provides a more intense staining but background staining is unacceptably high. The 1:2 dilution is not chosen as the LE structure is too weakly stained. See also Figure 5.8. 65 Chapter 5 | Staining Protocol Optimization Figure 5.5 exempliies some of the issues associated with ob- Antibody Dilution taining the optimal target retrieval setting and applying it to the Deinition of the optimal antibody dilution and protocol must daily worklow in a busy pathology laboratory. While the target take both speciic staining intensity as well as background retrieval time for Anti-S100 is optimal at just 15 minutes, the staining into consideration (see also Chapter 4). The overall optimal time for Anti-Ki-67 is 35 minutes. When many different principle is to deine an optimal antibody dilution and protocol, antibodies are used in the lab, an alignment of incubation time resulting in a crisp staining with the strongest possible intensi- is needed in order to keep the most eficient worklow in the ty, but without generating background staining. During devel- daily routine. Antibodies should be tested with a wide range opment each FLEX RTU antibody is tested in different dilutions of incubation times so that optimal worklow is obtained, with- along with its recommended protocol, and then tested for suf- out unduly sacriicing the optimal staining result. This can be a icient and robust staining performance on a broad range of cumbersome task for a laboratory when introducing new con- clinical samples with variable antigen expression. centrated antibodies. An RTU version of the antibody will have should have aligned the incubation times, while ensuring opti- Antibody Incubation Times – ‘Adjust the antibody incubation time, not the core protocol’ mal staining quality. The rationale behind optimal signal transfer reactions in the been tested for these factors before release, and the vendor analytical step of an IHC stain is to minimize the impact of sample variation (Figure 5.1). The visualization (detection) part of 1:1 1:2 1:4 Medullary thyroid carcinoma LE HE 2:1 Figure 5.8 The impact on staining intensity when altering the concentration of the primary Chromogranin A antibody. The optimal dilution was selected as 1:1. HE: The neuroendocrine cells in colon show a strong cytoplasmic staining reaction independent of antibody dilution. The epithelial cells are negative. Note that there is some background around the crypts when staining with 2:1 concentration. LE: The axons and ganglion cells in colon show an increased staining intensity when using 2:1 concentration and a profound decrease in staining when using a 1:2 or 1:4 dilution. Medullary thyroid carcinoma: The majority of the neoplastic cells show an increased staining intensity when staining with 2:1 concentration and a profound decrease in staining intensity when using a 1:2 or 1:4 dilution. 66 Staining Protocol Optimization | Chapter 5 the protocol should not be used to compensate for suboptimal 17.5 minutes before visual saturation is obtained. When test- antigen retrieval, insuficient antibody concentration, or insufi- ing a wide panel of sensitive antibodies even the most chal- cient incubation time. Once the optimal conditions for antigen lenging structure obtains a staining intensity plateau at 20 retrieval, visualization and counterstaining (the core protocol) minutes of HRP incubation. have been determined, only the incubation time of the RTU primary antibody is adjusted to ensure proper staining of both low When testing the impact on intensity of IHC staining of chro- and high antigen-expressing cells and structures in the tissue mogranin A using an alternative protocol, the most profound (Figure 5.9). reduction in staining intensity was seen in the LE structures, as well as in the clinical tissue. Visualization System Incubation Times The overall aim for selecting the right incubation time is to Using the FLEX RTU protocol, the high expression structures ensure that all primary antibodies associated with the spe- may seem “over-stained” (Figure 5.11, top left), but this visual ciic target are transferred into signal as effectively and with saturation is needed to minimize the risk of false negative re- as little variation as possible. As shown in Figure 5.10, some sults (Figure 5.11, bottom right). Again, these results highlight LE structures are visually saturated after just 7.5 minutes the importance of evaluating not only the HE structures, but horseradish peroxidase (HRP) polymer incubation (CD20, also the LE structure since they are generally more sensitive ST-M3-02), while the LE structures for S100 staining requires to changes in the protocol than the HE structure. LE Follicular lymphoma Adjusted 10-15-10-2x5 Optimal 20-15-20-2x5 HE Figure 5.9 The impact of reducing the antibody incubation time (from 20 min to 10 min (red text)) together with a shorter ampliication step (from 20 min to 10 min text) for EnVision™ FLEX+ with Linker ampliication. Note that changing two steps at the same time, in isolation, is usually not a good idea, because any observed effect may be due to either one, or both, of the changes. In some instances where the primary antibody does not saturate, then the detection system may also be adjusted as part of a larger experiment, often termed a chess board or chequer board titration, to achieve mutual optimization and saturation (Chapters 4 and 6). 67 Chapter 5 | Staining Protocol Optimization Staining intensity with different HRP times Anti-CD20 Staining intensity Very high Very high High High Medium Medium Low Low Negative Negative Minutes 0 5 7.5 10 Anti-S100 Staining intensity 15 Minutes 0 20 ST-M3-02 (high expression structure) ST-M3-02 (low expression structure) 5 10 15 17.5 20 ST-M3-05 (high expression structure) ST-M3-05 (low expression structure) Figure 5.10 Some speciic structures are visually saturated faster than others (7.5 min HRP incubation for CD20 versus 17.5 min for S100. However to obtain optimal worklow the incubation times for FLEX RTU protocols are aligned while still securing optimal signal transfer and thereby robust protocols. LE Medullary thyroid carcinoma Alternative protocol EnVision™ FLEX RTU protocol HE Figure 5.11 The impact on the staining intensity of the quality indicators and on clinical tissue when altering the protocol of the FLEX RTU chromogranin A antibody. HE: The neuroendocrine cells in colon show a negligible reduction in staining intensity when using the alternative protocol. The alternative protocol is optimized using the HE structures, whereas the FLEX RTU protocol is optimized using the LE structures. LE: The ganglion cells and axons in colon, as well as the medullary thyroid carcinoma show a profound reduction in staining intensity when using the alternative protocol compared to using the FLEX RTU protocol. 68 Staining Protocol Optimization | Chapter 5 Section Thickness Counterstaining Incubation Times The ine art of tissue sectioning requires skilled technicians, Hematoxylin is commonly used as a counterstain for IHC due and consistent tissue thickness may sometimes be an over- to a number of advantages it offers in the interpretation phase looked factor in the IHC standardization efforts. The standard of the IHC stain. The blue nuclear coloring by hematoxylin section thickness of 4-5 µm used in IHC may only partially produces a high level of contrast for the human eye, between include the nucleus, which is ~6 µm on average. This varia- the brown DAB chromogen, or red AEC chromogen. This tion in depth of included nucleus can potentially lead to re- facilitates improved visualization of tissue on the slide, as- duced staining intensity of a nuclear IHC stain, but is rarely sessment of tissue morphology, and determination of stained important in the diagnostic result since the typical scoring structures and cells. The counterstaining step of the protocol criterion is qualitative (yes/no). However, it may be important must also be optimized to reduce variability in intensity, which in semi-quantitative scoring, where a tangential cut of the nu- may otherwise lead to dificulty in recognizing nuclear locali- cleus may not register as positive, potentially leading to a zation of the chromogen. lower score. The tissue section thickness is important when the protocol settings are being optimized. A thicker section Too intense counterstaining can lead to an optical distortion, will tend to give stronger staining intensity than a thin sec- especially for nuclear antigens, and too weak counterstaining tion, simply because there is more tissue, more antigen, and may impair the tissue morphology assessment. For some as- more ‘space’ into which the DAB end-product may precipi- says the counterstaining can directly affect the diagnostic out- tate. Also, uneven thickness over the section may be encoun- come, such as Ki-67 assessment, where the number of neg- tered, but is usually only detectable with quantitative digital ative nuclei (denominator) directly affects calculation of the imaging (5). As illustrated in Figure 5.12, the same protocol percentage of Ki-67-positive cells, i.e. a weak counterstain can can give remarkably different staining intensity depending on lead to overestimation of the Ki-67 index (6). the thickness of the tissue. Frequent microtome calibration and skilled technicians are necessary for obtaining consist- Worklow Alignment ent tissue section thickness. Laboratories have to meet the increasing demand for more tests and faster results, with fewer resources. The use of pre-deined protocols for FLEX RTUs is a balance between worklow and turnaround time without compromising optimal staining result. As shown in Figure 5.5, it is possible to use an antigen retrieval time of just 15 minutes for Anti-S100, while the optimal time for Anti-Ki-67 is 35 minutes. For the alignment of worklow, a 35 minute incubation is used without negatively 8 µm 5 µm impacting the staining of S100. Depending on the instrumentation in the laboratory, the use of different target retrieval solutions for antibodies that are used often, or used in the same panel, can be a challenge for the optimal worklow. For example, a frequently used antibody 3 µm 1 µm Figure 5.12 The impact of section thickness. A thicker tissue section can give stronger staining. All stains are done with the same protocol on serial sections. Courtesy of Søren Nielsen, NordiQC. may give optimal staining using a low pH target retrieval buffer, but acceptable staining can also be achieved using a high pH target retrieval buffer. Many labs are tempted to align the target retrieval protocol to the high pH protocol, used for 8590% of all antibodies. The result is a better aligned worklow in the laboratory, at the expense of a less optimal, but still ac- 69 Chapter 5 | Staining Protocol Optimization ceptable staining result for the frequently used antibody. Al- reliability that is suficiently robust to allow comparison of IHC ternatively, laboratories would have many different protocols results, both intra- and inter-laboratory. each suitable for only a few antibodies, which would thereby negatively affect the worklow. As discussed in Chapter 2, the inluence of pre-analytical factors now has a higher attention in the standardization efforts than previously, e.g. the issues with ixation are being Chapter 5.4 Reproducibility of IHC addressed with practical worklow efforts in most laboratories, to ensure immediate and proper ixation time of tissue spec- The issues of reproducibility and standardization of IHC have imens. This step is important for correct diagnosis because been addressed for years by researchers, pathologists and epitopes where the primary amino acid structure is degraded manufacturers using different approaches to solve one or more cannot be restored. Not even an optimal protocol with the most of the factors affecting the staining result (Chapter 1). Similar speciic and sensitive reagents can relect the in vivo antigen pre-analytical preparation methods, uniform antigen retrieval expression when epitopes are lost prior to ixation. In other methods, high quality reagents, sensitive visualization systems words, garbage in = garbage out. and use of positive and negative tissue controls are central elements for improving staining quality. Despite these efforts, high level of reproducibility and reliability has not yet been Chapter 5.5 Requirements for Controls achieved, as evidenced by data from EQA schemes, e.g. UK NEQAS and NordiQC). For many antibodies, a variety of tissue structures are positive with high and low expression of the antigen in the normal control Finding the cause of inconsistent IHC results can be ad- tissues. It is therefore important to choose structures that are both dressed in different ways. Is variation caused by biological, stable and relect changes in protocol performance. A proper technical, or observational differences? Very few studies have been designed to ind the actual root cause of inconsistency in IHC staining results. One study comparing ive antibodies at two institutions - each with two observers - found that the largest source of variation comes from differential expression of control tissue should preferably fulill the following parameters: „„ Show stable antigen expression between samples of the same tissue type „„ Relect changes in protocol performance by changes in staining intensity the molecule being tested (biological), rather than technical or observational differences (7). This conclusion differs from other In normal tissue, antigen expression is generally relatively sta- studies, e.g. the ER reproducibility issues where both technical ble between different samples of the selected control tissues, and observational issues contributed to unreliable results (8). whereas clinically relevant abnormal tissues often show a large variation in antigen expression, between samples as well as These two papers illustrate some of the challenges emerg- within the same tissue. As a consequence of these require- ing from both the pre-analytical and post-analytical phases, ments, abnormal tissue samples are usually not suitable for together with the biological nature of the protein under exami- protocol optimization studies, as they by nature have unknown nation. When a pathologist evaluates a stained tissue sample, and unstable antigen expression. Instead optimization must be the staining protocol must be suficiently robust to minimize conducted on normal samples that ‘resemble’ the diagnostic variations in the pre-analytical phase and give a visual rep- material as closely as possible, including weak or focally posi- resentation of the expression of the antigen in the tissue that tive cases. Interestingly, only 40% of laboratories include such the human eye will be able to interpret comparably across weak/focally positive cases in validation procedures (1). various patient samples. To date, standardization approach- 70 es are still needed for an overall IHC system that assures uni- For the technical optimization process, it is very valuable to form, high-quality staining, with a level of reproducibility and perform the tests on TMAs (see Chapter 12) with different nor- Staining Protocol Optimization | Chapter 5 mal tissues processed according to the routine and standard the cause (other than personal preference of staining intensity) conditions used by the laboratory. The use of RTU systems should protocol modiications be attempted. If the laboratory where both reagents and matching protocols are validated by dilutes the RTU, changes reagents or adjusts the protocol out- the vendor, can make the introduction of new assays much side the manufacturer’s guidelines, the laboratory must do a simpler. The laboratory can simply select positive control tis- full revalidation of the assay. sue with pre-analytical factors matching the anticipated diagnostic samples, and stain these strictly according to the man- Using too few or inappropriate controls, which do not relect ufacturer’s validated protocol. The protocol has already been variations in the pre-analytical and biological factors, jeopard- tested with different pre-analytical and analytical factors on an izes the diagnostic staining result. As illustrated in igure 5.13, automated staining system, but variations in sample prepara- an optimized protocol should have a large ‘window’, where tion in individual laboratories may inluence the staining result samples with variable expression levels and variable pre-an- compared with the vendors’ tested settings. When an optimal alytic factors will be positively stained. If an altered protocol is staining result has been successfully achieved, then the ro- tested with only a few controls or with cases that only represent bustness and correctness of the assay can be further validat- few variables, the risk of having samples that fall outside the ed on additional cases in the user laboratory. If the vendor’s acceptable window increases. validated protocol does not give the preferred staining result, the irst option should be to identify pre-analytical factors in In order to serve as a reliable control tissue, the antigen the laboratory that could potentially have a negative effect on expression must be predictable and stable between inde- the staining result. Only when the laboratory cannot identify pendent samples. Normal tissue that fulills these criteria Window of Visual Staining Intensity Window of Visual Staining Intensity Variation of Antigen Expression nce fere Pre l a son Per A Variation in Pre-Analytic Factors B Few protocol controls with limited variation Figure 5.13 A) An optimized protocol ensures that samples with variation in antigen expression and pre-analytical factors produce a visible signal, due to a rigorous validation procedure. Many samples with differing expression levels from different sources in optimal signal transfer incubation reactions are tested. B) When only a few samples with little variation and poor signal transfer reactions are used to optimize a staining protocol, the window in which the protocol generates visual signals is reduced and some samples may fall outside the window, increasing the risk of false negatives and false positives results. 71 Chapter 5 | Staining Protocol Optimization should be chosen as control tissue. When optimizing the However, evaluation of various normal tissues may identify protocol it is important to identify normal tissues that express cases where antigen expression varies too much to be deined the antigen in high expression (HE) and low expression (LE) as stable. Using the same Anti-Actin antibody as an example, structures. Below is an example of Anti-Actin, clone HHF35, normal liver tissue shows staining of perisinusoidal smooth where the deined HE and LE structures are present in differ- muscle cells – but only in some samples (Figure 5.14C). This ent tissue types. In colon, all smooth muscle cells in vessel inding serves to underline that all normal tissues may not, by wall, muscle layers and lamina muscularis mucosa were de- default, be suitable as control tissue. Detailed analysis should ined as HE (Figure 5.14A), whereas LE was observed in the be carried out when selecting a normal tissue that is optimal myoepithelial cells of the mucous/salivary glands in tongue as control tissue, by using a validated protocol that is able to (Figure 5.14B). identify variations in antigen expression. A Colon, HE Tongue, LE Liver Case 3 B Case 2 C Case 1 Figure 5.14 A) Normal colon from three different patients stained with Anti-Actin, clone HHF35. Smooth muscle cells in vessel wall, muscle layers and lamina muscularis mucosa, deined as high expression (HE) structure show comparable staining intensity in all three samples. B) Normal tongue from three different samples Anti-Actin, clone HHF35. The myoepithelial cells of the mucous/salivary glands, deined as low expression (LE) structure, show comparable staining intensity in all three samples. C) Normal liver from three different patients stained with Anti-Actin, clone HHF35. The staining intensity of perisinusoidal smooth muscle cells varies from weak to negative, and is consequently a poor control tissue due to the variable antigen expression between tissue samples. 72 Staining Protocol Optimization | Chapter 5 Having a ‘catalog’ with normal external control tissues with cellular structures/elements, as well as clinically relevant (ab- high and low expression of the antigen is useful for a number normal) control tissue are included in the tests. of critical requirements for the optimal staining: „„ High expressers are required to evaluate the process (right antibody) „„ Low expressers are required to evaluate sensitivity (right protocol) „„ Non expressers are required to evaluate speciicity (right signal to noise) „„ Variation is required to evaluate robustness (right staining, repeatedly) Validation Validation is performed on production lots and validates the intended use of the product. The validation testing is performed according to the instructions for use of the antibody on an extensive set of positive and negative clinical specimens related to the intended use of the product. The test is performed as a method comparison study, and the positive and negative agreement is compared to a second antibody of similar speciicity and documented performance, if obtainable. Insuficient staining results in NordiQC assessment runs are often related to the use of inappropriate positive tissue controls (3). Increased focus on external tissue controls is needed to standardize and optimize IHC. See Chapter 4 and Chapter 14 Chapter 5.7 Guide to the Development of an Optimal Staining Protocol and for more information on the use of controls in IHC. This chapter has focused on minimizing the risk of false negative and false positive results by ensuring that the protocol transfers Chapter 5.6 Veriication and Validation of a Protocol all available targets into a visual signal without introducing background staining. The cornerstone of FLEX RTUs is the core protocol. To reach the core protocol the antibody has been tested Veriication using a protocol decision tree. If the staining is not optimal the Once the protocol has been optimized and ‘locked’ using the protocol is adjusted in the next step and so forth (Figure 5.15). optimal signal transfer incubation principles explained above, the antibody enters the veriication phase where both the anspeciications related to analytical speciicity and sensitivity. Chapter 5.8 Protocol Performance for Abnormal Tissue The FLEX RTU antibody is tested on a minimum of 30 different When evaluating the deined HE/LE structures in variable pro- normal tissues from three independent sources, according to tocol settings, variation in protocol performance should be requirements from the FDA (9). This test is evaluated by an ex- relected by changes in staining intensity. Typically, staining ternal, qualiied pathologist and the results are included in the intensity of the HE structure is fairly stable, whereas the LE instructions for use for the antibody (package insert). Relevant structure varies in intensity. When ine-tuning a protocol, inten- stains are cataloged in the Dako Atlas of Stains book. sity changes in HE structure are usually minor, whereas the tibody and the protocol are tested to ensure that they meet intensity changes in the LE structure are generally larger and The antibody undergoes extensive precision testing to ensure easier to detect. reproducibility and repeatability of antibody performance. The test is performed as a Gauge R&R study and includes intra-run The changes in IHC performance monitored by the HE/LE reproducibility, inter-run reproducibility and inter-instrument re- structures are also relected in the staining intensity in the producibility, to conirm that the antibody performance is the clinical tissue. In Figure 5.16, two different T-cell lymphoma same on the same day, between days, and when used on dif- cases show variation between samples of clinical tissues with ferent (Dako) platforms. Normal control tissue with well-deined respect to antigen expression. Compared with the optimal pro- 73 Chapter 5 | Staining Protocol Optimization Primary Antibody, Recommended Concentration or RTU Antigen retrieval Low pH High pH Visualization system Visualization system With ampliication Without ampliication With ampliication Without ampliication Ab incubation, x min Ab incubation, x min Ab incubation, x min Ab incubation, x min Wash Wash Wash Wash Peroxidase block, 2 min Peroxidase block, 2 min Peroxidase block, 2 min Peroxidase block, 2 min Wash Wash Wash Wash Linker, 10 min Linker, 10 min Wash Wash HRP, 20 min HRP, 20 min HRP, 20 min HRP, 20 min Wash Wash Wash Wash Wash Wash Wash Wash DAB, 5 min DAB, 5 min DAB, 5 min DAB, 5 min Wash Wash Wash Wash HTX, 3 min HTX, 3 min HTX, 3 min HTX, 3 min Wash Wash Wash Wash DI, 2 min DI, 2 min DI, 2 min DI, 2 min Wash Wash Wash Wash Core protocol Figure 5.15 The core protocol is reached by testing the parameters using a protocol decision tree. Step 1: Assess the optimal AR. Test both high pH and low pH AR using the vendor’s recommended primary antibody concentration (incubate for 20 min) and a Linker visualization system. Select the antigen retrieval solution which retrieves most targets. Step 2: Assess staining protocol. Titrate primary antibody concentration (never for RTUs) or adjust incubation time. Test with and without ampliication in the visualization protocol. Select the protocol that stains both HE and LE expression structures with optimal signal to noise ratio. 74 Staining Protocol Optimization | Chapter 5 tocol, the staining intensity of the HE structure is only slightly Recommended protocol Sub-optimal protocol protocol. Thus, in the sub-optimal protocol, Case 2 might be interpreted negative – a false negative result that would only Case 1 the LE structure is dramatically decreased in the sub-optimal T-cell lymphoma weaker in the sub-optimal protocol, while staining intensity of be recognized by evaluating the HE and LE structures in the LE structures, it would be impossible to identify whether the staining quality of the clinical tissue was due to the biology of Case 2 Without including the normal control tissue with associated HE/ T-cell lymphoma normal control tissue. the cancer or inluenced by the quality of the protocol. These results emphasize the importance of monitoring IHC protocol performance by carefully evaluating the deined HE and LE structures within the deined normal control tissue in order to ascertain optimal staining in clinical tissues. This process in turn helps to ensure the correct diagnosis. Some laboratories Figure 5.16 Staining intensity of the LE structure is dramatically decreased in the sub-optimal protocol. Thus, in the sub-optimal protocol, Case 2 is false negative. Stained with Anti-CD4, Clone 4B12 using the recommended Linker ampliication (EnVision™ FLEX+) versus a protocol that is sub-optimal for this antibody (EnVision™ FLEX). Pancreas (control) average, both the assay and the protocol are scrutinized to A identify possible issues with the test. Using an ampliication step (Figure 5.17) is sometimes necessary to increase the sensitivity of the protocol to minimize the risk of false negative results. Ampliication steps should only be considered when adjustment of parameters outside nal to noise reaction. Therefore, ampliication should not to be used to create a more intense staining when the problem of weak signal may be due to a non-optimal protocol. Ampliication may produce a very intense staining of structures in the control tissues, which some pathologists ind undesirable (“over-stained”). However, as exempliied in igure 5.17, a protocol that produces intense staining of control tis- B the core protocol (Figure 5.15) does not result in optimal sig- Papillary thyroid carcinoma Clead positive result es. In the case that the percentages differ too much from the Risk of false negative result also monitor the monthly distribution of HER2 score percentag- Figure 5.17 Using an ampliication (Linker) step for the protocol will produce strong, intense staining of the control tissue (bottom, left), and can reduce the risk of false negative results (bottom, right). Protocol A: Primary Ab: 20 min, HRP: 20 min, DAB: 2x5 min. Protocol B: Primary Ab: 20 min, Linker: 10 min, HRP: 20 min, DAB: 2x5 min. sue structures is sometimes needed to minimize the risk of false negative results. 75 Chapter 5 | Staining Protocol Optimization Chapter 5.9 Concluding Remarks A more widespread use of RTUs can contribute to some improvement in standardization, but cannot solve the pre-ana- This chapter has described the requirements for creating an lytical standardization issues. RTU reagents are extensively optimal protocol with focus on robustness and correctness, tested internally by the manufacturers, and vendors should using proper tissue controls with high and low expression provide protocols that establish optimal performance of the of the antigen. The FLEX RTU solution, with a deined vis- system, including guidance to correct control tissue. In-house ualization system and validated protocol has been used to testing, within the user laboratory, must include a wide range of exemplify the many contributing factors to which laborato- FFPE tissues that have been subject to different pre-analytical ries must pay attention, when trying to optimize protocols factors, including time and conditions of ixation. A correct and for concentrated antibodies. The use of an RTU solution re- robust RTU protocol, from antigen retrieval to counterstaining, duces the resources necessary to verify performance when can and should minimize the effects of lack of standardiza- introducing new antibodies in the laboratory. It also forces tion by ‘absorbing’ the variations. To ensure visual saturation the standardization of reagents, dilutions, detection systems of both the LE and HE structures in different tissue types, the and staining protocols among those different laboratories IHC stain may sometimes appear “over-stained” for HE struc- using the same system. Optimal protocols minimize the im- tures. This high intensity is an intended ‘drawback’ (the ‘price pact of pre-analytical factors, due to optimal signal transfer to pay’) when improving diagnostic certainty in tumor samples incubations, and thus increase the diagnostic conidence, with heterogeneous antigen expression. by detection of cells and structures with both low and high expression of the antigen. Skilled and experienced staff in many laboratories are no doubt capable of matching a concentrated antibody to a de- While an optimal protocol, with optimal signal transfer incuba- tection system. They can perform the necessary titrations, and tions, can to some extent ‘absorb’ or ‘iron out’ pre-analytical test incubation times to establish the optimal protocol for the vari-ations, these variations are still one of the most important FFPE tissues available to each antibody. However, the resourc- hurdles to overcome for achieving standardization among pa- es and experience necessary to conduct such optimization thology laboratories (10, 11). Neutral buffered formalin has studies are substantial, and may not be available in smaller become the standard choice of ixative in most laboratories laboratories. Larger institutions, on the other hand, may not worldwide but many laboratories still do not have standard easily ind the required time to conduct such extensive opti- procedures for the preparation of formalin, the source (ven- mization, validation and veriication tests. RTUs are an option dor/home-made), or even the ixation time. Practical daily representing a path to improved diagnostic certainty for both procedures in the hospital can cause luctuations in ixation experienced, high volume laboratories as well as for smaller time, which can inluence eficacy of antigen retrieval, and laboratories without skilled and experienced staff. ultimately the staining result. Standardization starts in the operating room, where prolonged duration of pre-ixation time When up to 90% of insuficient staining results are due to weak (ischemic time) may permit degradation of tissue resulting in staining intensity (3), it seems obvious that a greater focus on loss of epitopes. Not even the best protocol can restore the appropriate tissue controls and optimal staining protocols is a primary sequences of an epitope lost in this manner. Similarly, step in the right direction for standardization of IHC. the effect of other steps in the pre-analytical phase, including processing, dehydration, parafin embedding, storage before staining, de-parafinization and rehydration are not fully understood (12). Something as simple as the water, or type of parafin, used in the laboratory may have an impact on the parameters in the optimal staining protocol. 76 Staining Protocol Optimization | Chapter 5 References 1. Hardy LB, Fitzgibbons PL, Goldsmith JD, Eisen RN, Beasley MB, Souers RJ, et al. Immunohistochemistry validation procedures and practices: A college of american pathologists survey of 727 laboratories. Arch Pathol Lab Med 2012;137:19-25. 2. CAP, Anatomic Pathology Checklist. 2013, College of American Pathologists: www.cap.org. p. 34-38. 3. Nielsen S. Validating the analytical power and parameters of an immunohistochemical test, in Molecular Diagnostics: The Key Driver in Personalized Cancer Medicine, J.T.J.H. Winter, Editor. 2010, Pan Stanford Publishing: Singapore. p. 185-206. 4. D'Amico F, Skarmoutsou E, Stivala F. State of the art in antigen retrieval for immunohistochemistry. J Immun Methods 2009; 341:1-18. 5. Grube D. Constants and variables in immunohistochemistry. Arch Histol Cytol 2004;67:115-34. 6. Dowsett M, Nielsen TO, A’Hern R, Bartlett J, Coombes, CR, Cuzick J, et al. Assessment of Ki67 in Breast Cancer: Recommendations from the International Ki67 in Breast Cancer Working Group. J Nat Cancer Institute. 2011; 103:1656-64. 7. Lawrie CH, Ballabio E, Soilleux E, Sington J, Hatton CSR, Dirnhofer S, et al. Inter- and intra-observational variability in immunohistochemistry: A multicentre analysis of diffuse large b-cell lymphoma staining. Histopathol 2012;61:18-25. 8. Hammond ME, Hayes DF, Wolff AC, Mangu PB, Temin S. American society of clinical oncology/college of american pathologists guideline recommendations for immunohistochemical testing of estrogen and progesterone receptors in breast cancer. J Oncol Pract 2010;6:195-7. 9. FDA, Guidance for Industry and FDA Staff, Statistical Guidance on Reporting Results from Studies Evaluating Diagnostic Tests. www.fda.gov, 2007. Document issued on: March 13, 2007: p. 1-39. 10. Taylor CR, Levenson RM. Quantiication of immunohistochemistry – issues concerning methods, utility and semiquantitative assessment ii. Histopathol 2006;49:411-24. 11. Goldstein NS, Hewitt SM, Taylor CR, Yaziji H, Hicks DG. Recommendations for improved standardization of immunohistochemistry. Appl Immunohistochem Mol Morphol 2007;15:124-33. 12. Engel KB, Moore HM. Effects of preanalytical variables on the detection of proteins by immunohistochemistry in formalin-ixed, parafin-embedded tissue. Arch Pathol Lab Med 2011;135:537-43. 77 Part I: The Staining Process Chapter 6 Detection Methods Kenneth Petersen, PhD, MSc Hans Christian Pedersen, MSc De•tect (n.) To discover or determine the existence or presence of <something>. Merriam-Webster Online Dictionary Detection Methods | Chapter 6 Chapter 6.1 Introduction more intense staining. The secondary antibodies used in the indirect methods are typically raised in goat against either Immunohistochemistry (IHC) has emerged as a powerful in- mouse (GaM) or rabbit (GaR) antibodies. With the successful vestigative tool that can provide supplemental information to application of IHC methods to formalin-ixed specimens, new the routine morphological assessment of tissues. The use of staining methods were rapidly developed including the immu- IHC to study cellular markers that deine speciic phenotypes noperoxidase bridge method (6) and the peroxidase anti-per- has provided important diagnostic, prognostic, and predic- oxidase (PAP) complex method (7). tive information relative to disease status and biology. The application of antibodies to the molecular study of tissue Table 6.1 Complexity vs. Sensitivity of Detection Systems. pathology has required adaptation and reinement of IHC iCSA 5 techniques, particularly for use in ixed tissues. In contrast to native proteins, in ixed tissues the preservation of antigen is variable and unpredictable. Thus, the history of IHC has evolved so that we today are able to detect proteins in tissue with great sensitivity, and also provide a semi-quantitative assessment, with the ultimate goal of integrating tissue-based 4 Number of Steps solution-based immunoassays that detect relatively abundant Polymer (Doublestain) 3 LSAB/ABC 2 Indirect 1 The irst staining with an antibody to ind an antigen in tissue Polymer + Linker Polymer Direct 1 analysis with proteomic information. Immunohistochemistry: In the Beginning CSA II 2 3 4 5 6 Sensitivity LSAB=Labeled streptavidin-biotin; ABC=Avidin-biotin complex; CSA II=Catalyzed signal ampliication II, iCSA = Improved catalyzed signal ampliication. was reported in 1941, using a luorescence-labeled antibody (1). Twentyive years later, the enzyme horseradish peroxidase (HRP) together with 3,3’-diaminobenzidine (DAB) was used to study mouse kidneys (2). The following year, an antibody linked Chapter 6.2 Avidin-Biotin Immunohistochemistry to HRP was used to visualize antigens in tissue using the indirect method, where a second antibody is used to recognize the The next generation of IHC methods emerged in 1981 with the irst or primary antibody which is attached to the antigen (Figure avidin-biotin-based methods (Figure 6.2) (8). These methods 6.1). The secondary antibody recognize the constant part (Fc) are still used to a limited degree in some pathology laborato- of the primary antibody, which makes it possible to recognize all ries and rely on the strong afinity of avidin or streptavidin for primary antibodies as long as they are from the same species. the vitamin biotin. These pioneering studies using enzyme labels instead of luorescent dyes and the application to formalin-ixed, paraf- Streptavidin (from the bacteria Streptomyces avidinii) and avi- in-embedded tissue (FFPE) (3) opened the door to the use of din (from chicken egg) both have four binding sites for bio- immunoperoxidase methods for routine diagnosis in anatomic tin. The biotin molecule is easily conjugated to antibodies and pathology (4, 5), and led to the development of modern meth- enzymes. In the avidin-biotin complex (ABC) method second- ods of IHC (see Chapter 1). ary antibodies are conjugated to biotin and function as links between tissue-bound primary antibodies and an avidin-bio- The good preservation of features and improved morphology tin-peroxidase complex. The four binding sites for biotin make of FFPE of tissues, makes this method the preferred choice in ‘lattice’ complexes possible, where the avidins are linked to- almost every clinical pathology laboratory. The indirect staining gether via the enzyme (8). The only requirement is that the en- methods are likewise the preferred staining methods because zyme has at least two biotin molecules attached so that it can labeling of the primary antibody is avoided, and they give a function as a link between the avidins. A colorless substrate, 79 Chapter 6 | Detection Methods e.g. DAB, is subsequently added, and is converted to a brown Enzyme end-product by the multiple peroxidase enzyme molecules now attached at the site of the target antigen. Label Secondary antibody In a similar method the labeled streptavidin-biotin (LSAB) method also utilizes a biotinylated secondary antibody that links pri- Primary antibody mary antibodies to a streptavidin-peroxidase conjugate (Figure 6.3). This approach has the advantage that preassembly of the Tissue antigen Direct staining Indirect staining ABC complex is not needed. In both methods a single primary antibody is subsequently associated with multiple peroxidase molecules, and because of the large enzyme-to-antibody ratio, Figure 6.1 Direct vs. indirect staining. a considerable increase in sensitivity is achieved compared to direct peroxidase-conjugate methods. When using these methods it is important to be aware of their limitations. Avidin has a tendency to bind non-speciically to Avidin-biotin enzyme complex Must be prepared 30 minutes prior to use lectin-like and negatively charged tissue components at physiological pH. For streptavidin less non-speciic tissue binding is observed. Another challenge is the presence of endogenous biotin in tissues. Formalin ixation and parafin embedding has been Biotinylated secondary antibody shown to signiicantly reduce the level of endogenous biotin, but residual activity can still be observed in tissues such as liver and kidney. Methods to block endogenous biotin are partially effec- Primary antibody tive, but add another layer of complexity to an already complex Tissue antigen procedure. In frozen tissue sections, the level of endogenous biotin is usually even higher than that encountered in FFPE specimens, giving troublesome non-speciic binding of the avidins. Figure 6.2 Avidin-Biotin Complex (ABC) method. Chapter 6.3 Polymer-Based Immunohistochemistry Streptavidin enzyme complex Biotinylated secondary antibody, mouse/rabbit The limitations associated with the avidin-biotin system, led to the development of detection systems with higher sensitivity and speciicity, employing polymer-based IHC techniques (9). These methods utilize a polymer backbone to which multiple antibodies and enzyme molecules are conjugated. As many Primary antibody as 70 enzyme molecules and about 10 primary antibodies Tissue antigen can be conjugated to a single dextran backbone. This construct allowed the entire IHC staining procedure, from primary antibody to enzyme, to be accomplished in a single step Figure 6.3 Labeled Streptavidin-Biotin (LSAB) Method (10). On the other hand, one limitation of this method was its restriction to a select group of primary antibodies provided 80 Detection Methods | Chapter 6 by the manufacturer, and lack of utility for many user-supplied cule. The amine can be used to add biotin, or other molecules of primary antibodies. interest, to tyramine through an amide bond, hence the tyramide ampliication name also used for this method. When tyramide To overcome this limitation a new type of visualisation system, is oxidized, it will react rapidly with electron-rich aromatic com- EnVision™, was introduced (Figure 6.4). This indirect visualiza- pounds, such as the amino acid tyrosine found in protein mole- tion system also contains a dextran backbone to which multi- cules (13). This reaction can be used in IHC to bind biotinyl-tyr- ple enzyme molecules are attached. However, the EnVision™ amide to protein molecules in the immediate vicinity of peroxidase system contains secondary antibodies with anti-mouse Ig and enzymes. This reaction results in the deposition of numerous bio- anti-rabbit Ig speciicity. This ‘universal’ reagent could be used tin signals around the primary antibody. to detect any tissue-bound primary antibody of mouse or rabbit origin. The broad applicability of this method opened the door In a typical CSA-based IHC procedure, peroxidase en- to a new family of polymer-based IHC methods. The sensitivity zymes are irst associated with primary antibodies by any of these methods when compared to LSAB and ABC methods of the standard IHC methods (Figure 6.5). Biotinyl tyramide is comparable or even slightly greater in most cases (11). By and hydrogen peroxide are applied as a substrate to gen- adding an additional linker step, the sensitivity can be improved erate numerous biotin signals. These biotin molecules can further. However, because of the large molecular size of the then be used to capture subsequent streptavidin-peroxi- polymer conjugates, accessibility to certain epitopes can be a dase enzymes to produce the desired staining by addition challenge, presumably due to steric hindrance. of the appropriate substrate (14). Another possibility is repetition of the biotinyl-tyramide reaction, which will increase the numerous biotin signals even further. This cycling of the Chapter 6.4 Catalysed Signal Ampliication (CSA) reaction is practically limited to two or three cycles before background staining becomes too high. CSA is a highly sen- This ampliication technique is based on the ability of peroxidase sitive ampliication technique, but has several disadvantag- enzyme to oxidize phenolic compounds to highly reactive and es that prevent its general use. The method is time consum- unstable intermediates called radicals (12). The commonly used ing, results can be hard to reproduce, and as in previous substrate in this technique is tyramine. It has a phenol in one end, biotin-based methods endogenous biotin can give a high used by peroxidases, and an amine in the other end of the mole- background staining. Dextran backbone Avidin-biotin enzyme complex HRP enzyme Secondary antibody, mouse/rabbit Secondary antibody Primary antibody Primary antibody Tissue antigen HRP – conjugated streptavidin Tissue antigen Biotinyl-tyramide deposition Figure 6.4 Two-step polymer method (EnVision™). Figure 6.5 The CSA system. 81 Chapter 6 | Detection Methods Chapter 6.5 Fluorescyl-tyramide Ampliication Chapter 6.6 Improved Catalysed Signal Ampliication (iCSA) Fluorescyl-tyramide can replace biotinyl-tyramide to avoid endogenous biotin background. In this procedure peroxidase is The latest improvement of the CSA method to increase sensi- associated with a tissue-bound primary antibody by applica- tivity and improve signal to noise ratio introduces a new more tion of a secondary anti-mouse antibody to which peroxidase soluble substrate. It entails a background-reducing effect, has been conjugated. The peroxidase catalyzes the conver- combined with a crosslinker that enhances the precipitation sion and deposition of luorescyl-tyramide onto the tissue sec- of the substrate in step 3 (Figure 6.6) The luorescein is con- tion. At this point the reaction can be terminated and viewed served in the substrate while the tyramine is substituted with by luorescence microscopy, or the signal can be converted to ferulic acid, which is a much better peroxidase substrate. To- a colorimetric reaction by the sequential application of an an- gether these changes improve CSA method by maintaining ti-luorsecein antibody conjugated to peroxidase followed by a high sensitivity and reducing background, giving high sig- diaminobenzidine-hydrogen peroxide substrate. nal-to-noise ratio. Furthermore, the incubation time in each step can be reduced signiicantly making it possible to stain In comparison to standard IHC methods, tyramide ampliica- a tissue in less than one hour. tion methods have typically increased sensitivity by at least 50fold or greater (15). As with any ampliication method, background tends to increase along with signal. Dextran backbone Substrate and cross-linker Enzyme + Primary antibody Secondary antibody Tissue antigen STEP 1 STEP 2 STEP 3 Red ed ch chromogen AP conjugated F(Ab’) antibody STEP 4 STEP P5 Figure 6.6 Improved CSA system (iCSA). A proprietary methodology developed by Dako. 82 Detection Methods | Chapter 6 Chapter 6.7 Multi-Staining Immunohistochemistry expressed in benign prostate tissue. Thus it is used as a positive cancer marker, often in a multiplex stain with keratin and In some cases there is a need for knowledge about the relative p63 (see example in Figure 6.8). If single stains are done on localization of targets, which context can only be obtained by vis- serial sections, interpretation is much more dificult and ambig- ualizing multiple targets in one slide. In other cases, the material uous lesions may be absent in adjacent cuts, especially when available for staining is scarce and there is a need for multiplexing dealing with small foci, with the result that some malignancies to retrieve all possible information out of material available. may remain undiagnosed. In this context, multiple staining protocols signiicantly improve the ability to distinguish between Deinition of Multi-Staining IHC benign and malignant lesions. This approach, which reduces Multiple staining can be deined as the detection of two or the percentage of ambiguous lesions and the need for addi- more targets on one slide, thus increasing the information ob- tional biopsies, is being extended to facilitate recognition of tained from each slide and reducing turnaround time, com- other invasive cancers, as in breast. pared to single staining or sequential staining (see deinition below). This technique also makes it possible to assess the Technical Challenges topographic relationship of two or more targets, for exam- Before embarking on a multi-staining project, some important ple, to determine whether targets are present in different cell issues should be considered: populations, in different cells, in the same cell, or even in the „„ Most primary antibodies used today originate from either same cellular compartment. In addition, multiple staining al- mouse or rabbit and are visualized using systems based lows the combination of in situ hybridization (ISH) and IHC, on anti-mouse and anti-rabbit secondary antibodies. The giving information about a particular target both at protein challenge of distinguishing between two primary anti- level and DNA/mRNA level. Information can also be obtained bodies of the same species (mouse-mouse, or rabbit- on possible cell-to-cell spatial contacts of different cell types. rabbit) must be addressed, because separate mouse and Furthermore, with an increasing demand for less invasive rabbit primary antibodies to the chosen targets often are sampling techniques and smaller and fewer specimens avail- not available. Utilizing two primary antibodies of the same able, multiple staining has the advantage not only of conserv- species can require quite elaborate protocols. ing tissue, but also saving time and reagents. „„ Spectral differentiation of stain colors may be dificult, especially if the targets are co-localized leading to a mix- Examples of Multiple Staining ture of colors (18). The ‘mixed’ color should contrast well The diagnosis of prostatic intra-epithelial neoplasia (PIN) is with the two basic colors. In the case where a rare target just one example of the clinical importance of multiple staining. is co-localized, the color reaction of the more abundant Prostate needle biopsy is the preferred method for diagnos- target will tend to dominate the other. ing early prostate cancer, but in some cases the diagnosis is „„ Even if targets are not co-localized it is dificult to balance uncertain because the biopsy includes only a few malignant signals so as to enable visualization of a rare target in glands, or a few hyperplastic or dysplastic glands that are the same slide as highly expressed targets. An adjustment dificult to distinguish from cancer (16, 17). Since basal cells in concentration of the primary antibodies may solve this typically are present in hyperplastic, and dysplastic glands, as problem. well as around ‘in situ’ (PIN) lesions, but absent in malignant in- „„ If different targets are viewed under different magni- vasive glands, the demonstration of basal cells can be used to ications, it may be dificult to obtain the desired assist recognition, or exclusion, of invasive cancer. Basal cells topographic information. are labeled using high molecular weight cytokeratin, cytokera- „„ Image analysis approaches, such as ‘spectral separation’, tin (e.g. CK5/6 - cytoplasmic) or p63 (nuclear) immunostaining, are generally superior to the human eye in segregating the or both. In addition, AMACR/P504S, is expressed in a high per- different color reactions in a multiplex stained slide. centage of prostate carcinomas, but is negative or only weakly 83 Chapter 6 | Detection Methods Pre-treatment double staining due to incomplete elution of unreacted rea- Multiple staining, like single staining, can be performed on gents from the irst staining sequence, before application of any of FFPE tissue sections, frozen sections, cell smears and the next reagents. cytospin preparations. Multiple staining may be constrained by the fact that it may not be possible to ind a single tissue pre-treatment (retrieval) protocol that is optimal for all targets. In this case, it may be necessary to determine a method that allows all targets to be stained, although the method may be sub-optimal for some targets. Multi-Staining Method Selection To ensure success, IHC staining using multiple antibodies must be carefully planned. If primary antibodies of the desired speciicity for the two (or more) targets are commercially available, and made in different species, then there are several different staining methods that one can choose. However, very often the choice may be limited by the reagents available (19). Care must be taken to avoid cross-reactivity between reagents; in the event that avoidance is not possible, then measures must be taken to minimize the risk, including additional controls to Figure 6.7 Sequential double staining method performed with the EnVisionTM G|2 Doublestain Kit using polyclonal anti-kappa light chains (red) and polyclonal anti-lambda light chains (brown) as primary antibodies. FFPE tissue sections from tonsils. detect signiicant cross reactivity if present. In general, staining methods can be divided into the following Elution may become an issue with some high-afinity primary classes: antibodies, as these may remain at their binding-site, leading to spurious double stained structures. Elution also risks de- Sequential staining naturing epitopes of antigens to be visualized subsequently. By this method, one staining procedure succeeds another. For Furthermore, for some chromogens there is a risk that the irst example, the irst antibody is applied to the tissue section fol- chromogen (DAB in particular) may shield other targets. This lowed by a labeled detection system such as streptavidin-bi- technique is, therefore, not recommended for evaluation of otin horseradish peroxidase (HRP), with a chromogen such as mixed colors at sites of co-localization, because not all reac- DAB. The second primary antibody is applied only after the tion products are capable of surviving the rigorous washing excess DAB is rinsed off, followed by labeling with a streptavi- required to remove the antibodies. To avoid such problems din-biotin alkaline phosphatase (AP) detection system and and blurry staining results, it is recommended to use the most a colored chromogen. The biggest advantage of sequential ‘robust’ dyes such as DAB, Fast Red, AEC and Blue chromog- staining is that by this procedure problems related to cross-re- en irst, followed by other less ‘robust’ dyes. activity are minimized, possibly due to steric interference. Simultaneous staining 84 A sequential staining is shown in Figure 6.7. Here, the primary In a simultaneous double stain, the primary antibodies can and secondary antibodies from the irst staining were eluted be applied simultaneously. The advantage of this method is before the staining of the next target was performed. The dis- that it is less time-consuming because the reagents can be advantages of sequential staining are: the method cannot be mixed together. However, the technique can only be used if used for co-localized targets, the technique often leads to a the primary antibodies are from different species, or are di- long staining protocol and carries an inherent risk of incorrect rectly labeled with different enzymes (20). Detection Methods | Chapter 6 A simple example of the direct method is when the primary antibodies are luorescence-labeled with luorochromes emitting different colors to allow direct visualization of two or more targets. This avoids cross-reactivity, but is rarely practical since some form of ampliication is necessary to get suficient luorescent signal. Alternatively, the primary antibodies may be conjugated directly with enzymes, biotin or haptens, subsequently employing the corresponding secondary antibody or streptavidin reagent. This approach is less time-consuming than the sequential method, because primary and secondary antibodies can be mixed together in two incubation steps. However, it requires avoiding all cross-reactivity. With the indirect method it is also possible to apply time-saving antibody ‘cocktails’ because the primary antibodies are recognized by different secondary antibodies. Generally, it Figure 6.8 Simultaneous double staining performed with EnVisionTM DuoFLEX using an antibody cocktail containing monoclonal rabbit anti-AMACR (red), monoclonal mouse anti-HMWCK and monoclonal mouse anti-CK 5/6 (brown/black). FFPE tissue sections from prostate. is advantageous to use secondary antibodies raised in the same host in order to prevent any unexpected interspecies cross-reactivity at the level of the secondary antibody. One TM Multi-step staining can be used when the selection of primary DuoFLEX from antibodies is limited. However, when using this method it is not Dako. This system applies a mixture of primary antibodies possible to mix reagents. Users will often ind that the choice of mouse and rabbit origin, followed by a mixture of the of staining method is limited by the availability of the primary secondary goat-anti-mouse and goat-anti-rabbit antibodies antibodies with respect to species origin or label. example of such a system is the EnVision labeled with HRP and AP, respectively. Finally, the chromogens are applied sequentially. The result is a double stain Dificulties arise when targets are known or suspected to be co-lo- where the primary mouse antibodies are stained brown with calized, and the only available primary antibodies are unlabeled DAB and the primary rabbit antibodies are stained red with monoclonal mouse antibodies of the same IgG subclass. In that Permanent Red (for an example, see Figure 6.8). The sys- case, none of the techniques described above are applicable. tem has been developed for Dako’s line of RTU cocktails of primary antibodies, but may also be used with other anti- One solution for such circumstance is the Dako Animal Research body cocktails or individual antibodies that are sequentially Kit (ARK™), which contains reagents for labeling mouse prima- incubated on a single slide. ry antibodies with a biotinylated anti-mouse Fab fragment, followed by blocking of the remaining reagent with normal mouse Multi-step technique serum. This approach can be applied to the tissue as part of the This is an indirect/direct method combining unlabeled primary multi-step technique (21). The kit uses a non-covalently labeled antibodies with directly-conjugated antibodies (3). The meth- antibody, thus avoiding the risk of reducing afinity. In addition, od starts with staining of the unlabeled antibody/antibodies only small amounts of primary antibody are needed and the kit with the appropriate detection system, but without performing does not require time-consuming puriication steps. the inal enzymatic staining reaction. The tissue is blocked with normal serum from the host of the irst primary antibody Another solution is Zenon Technology (Invitrogen) developed for before the second, directly-labeled primary antibody is add- low cytometry. It essentially uses the same technique and offers la- ed. The staining ends with the two enzymatic reactions being beling kits for mouse primary antibodies, available as enzyme con- performed sequentially. jugates or conjugated to one of a wide variety of luorescent dyes. 85 Chapter 6 | Detection Methods Finally, it is important to be aware of the fact that visualization color, from each individual color. Double staining using chromo- systems with dual recognition such as the EnVision™+ Dual genic dyes is well-established, but if the targets are co-localized Link system do not discriminate between species, and thus are then a percentage of the single colors cannot be easily identiied. only suitable for multiple staining when using the sequential For triple staining, it is naturally more dificult to choose colors that method. Visualization kits with ampliication layers that are not can be unambiguously differentiated, and even more so if targets clearly speciied should be avoided, since possible cross-re- are co-localized. In such cases, a technique known as spectral activity cannot be predicted. imaging may be applied (18,19). Spectral imaging allows images of the single stains to be scanned and by using specialized software algorithms the colors are unmixed, thereby displaying the Chapter 6.8 Selection of Dyes distribution and abundance of the individual chromogens. The primary choice to make when deciding how to make the Visualizing Low Expressed Targets targets visible is whether to use immunoenzyme staining or lu- A narrow dynamic range is a disadvantage for immunoenzymatic orescence. Both have advantages and disadvantages and in staining. The precipitation process, which is crucial for this meth- the end, decisions should be made based on conditions of the od, is only triggered at a certain threshold concentration of sub- individual experiment. strate and product. On the other hand, at high concentrations the precipitated product may inhibit further reaction. Therefore, it is Chromogenic Dyes dificult to visualize rare targets and highly abundant targets in the Table 6.2 Examples of enzyme/chromogen pairs suitable for triple staining. same slide. To ease this problem, catalyzed signal ampliication Enzyme Chromogen Color - an extremely sensitive IHC staining procedure can be used (Fig- Gal X-Gal Turquoise ures 6.5 and 6.6). The method can bring low expressed targets AP Fast Blue BB Blue HRP AEC Red HRP DAB Brown Gal X-Gal Turquoise AP Liquid Fast Red Red It is equally necessary to select dyes with distinguishable HRP DAB Brown spectral properties. However, there are more colors availa- AP New Fucsin Red HRP TMB Green within the same dynamic range as high expressed targets. Gal (beta-galactosidase); X-Gal (5-bromo-4-chloro-3-indolyl β-galactoside); AP (alkaline phosphatase); HRP (horseradish peroxidase); AEC (3-amino-9-ethylcarbazole); DAB (3,3’-diaminobenzidine); TMB (3,3′,5,5′-tetramethylbenzidine) Fluorescent Dyes Double immunoluorescence labeling is quite well established (22). Some of the same considerations as for chromogenic dyes apply when working with immunoluorescence. ble and the emissions spectra of the luorescent molecules are narrower than the spectra of the chromogenic dyes. It is possible to have more stains on one slide with luorescent dyes than it is with chromogenic dyes, which is one of the main advantages of luorescent dyes in multistaining. The use of multiple luorescent colors is also well established in FISH and low cytometry. When using luorescence dichroic excitation/emission, ilters are employed to separate the 86 When selecting color combinations for multiple staining with chro- different luorescent signals. The spectral separation can mogenic dyes, it is advisable to choose opposing colors in the be aided by digital compensation for overlapping emission color spectrum such as red and green to facilitate spectral differ- spectra. In addition, new luorescence microscope systems entiation. If using a counterstain, this color must also be included can separate the spectral signatures of up to eight luoro- in the considerations. When working with co-localized targets, chromes without any problems, using multi-spectral imaging dyes must be chosen so that it is possible to distinguish the mixed techniques such as emission ingerprinting (23). Detection Methods | Chapter 6 When staining co-localized targets, luorescent dyes may al- within the desired range. Visualizing a combination of several low separate identiication of targets. This makes it possible to gates, with the color selected independently of the dyes used discern targets even in very different concentrations, whereas for staining, may clarify pictures and facilitate interpretation. subtly mixed colors from chromogenic dyes may easily pass This capability also makes it possible to set a threshold on sig- unnoticed with immunoenzyme staining. nal intensity, to exclude non-speciic staining or background staining from inal images. A more thorough discussion of im- Thus immunoluorescence has some advantages, but there age acquisition and analysis can be found in Chapter 7. are also inherent problems; mainly loss of morphologic detail, which may determine the choice technique for a multi-staining application. Chapter 6.10 Immunoluorescence Alternative dyes Immunoluorescence (IF) is a common laboratory technique Alternatives to the conventional chromogenic dyes are colloidal used in almost all aspects of biology. This technique, based gold-labeled antibodies that can be used with bright ield micros- on pioneering work by Coons and Kaplan (26, 27), and later copy, with silver enhancement, Green Fluorescent Proteins (GFP by Osborne (28), has been widely both in research and clinical and their variants), and Quantum dots. The latter, especially, has diagnostics. Applications include the evaluation of cells in sus- been found to be superior to traditional organic dyes on several pension, cultured cells, frozen tissue, FFPE tissue, beads, and counts, such as brightness (owing to the high-quantum yield), as microarrays for the detection of speciic proteins. In IF tech- well as their higher stability (owing to less photodestruction). They niques, antibodies are chemically conjugated to luorescent can be linked to antibodies or streptavidin as an alternative to dyes such as luorescein isothiocyanate (FITC) or tetramethyl luorochromes (24). However, the size of these conjugates poses rhodamine isothiocyanate (TRITC). As in the enzymatic meth- problems of steric interference and diffusion, in terms of getting ods these labeled antibodies can be use directly or indirectly these inorganic particles into cells or organelles. to bind to the antigen of interest, which allows for antigen detection through luorescence techniques. The degree of luorescence can then be quantiied using a low cytometer, array Chapter 6.9 Automated Image Acquisition and Analysis in Multiple Staining scanner, or automated imaging instrument, or visualized using luorescence or confocal microscopy. IF techniques can be used on both fresh and ixed tissue samples, though the latter Digital image analysis will increase the number of usable dyes present problems of autoluorescence. because it does not rely on the human eye for detection and differentiation. A digital image is acquired at excitation wavelengths relevant for the dyes applied, and separate detectors re- Table 6.3 Advantages and disadvantages of direct and indirect immunoluorescence. cord individual colors. Thus, digital image analysis will allow the combination of both luorescent and immunoenzyme dyes (25). Pros Detectors, however, have biased color vision. They amplify colors differently than does the human eye. Therefore, dyes Direct Immunoluorescence Indirect Immunoluorescence Simpler Higher signal (amplified) Antibodies from the same species Flexibility (array of fluorescent colored secondary antibodies) used in image analysis should be optimized for the best it possible with the detector’s ilter properties. Image analysis systems incorporate algorithms that allow Low costs Cons Lower signal More steps Higher costs Antibodies from the same species cannot be used together Less flexibility Background may be amplified compensation for overlapping emission spectra, comparable to low cytometry. They also allow signal gating within a range of wavelengths of interest, enabling users to see only signals 87 Chapter 6 | Detection Methods Principle of Fluorescence Fluorescence and phosphorescence are both types of lu- Molecule goes into higher energy state minescence. When molecules with luminescent properties absorb light, they emit light of a different wavelength. With luorescence the emission of light occurs extremely rapidly after the absorption of excitation light, whereas with phosphorescence emission continues for milliseconds to minutes after Excitation decay energy (heat) lost towards semi-stable state High energy Excitation by external light source Light emission when the molecule goes back to ground state the energy source has been removed. Fluorescent materials give off light because of their atomic structure. Electrons are arranged in discrete energy levels surrounding the atom’s nucleus, with each level having a predetermined amount of en- Low energy ergy. When an electron absorbs the energy from a photon of light (Figure 6.10) it becomes ‘excited’ and jumps to a higher, less stable, energy level. The excited state does not last long. Figure 6.10 Principle of luorescence. The half-life of the excited state is generally less than 10 seconds. The electron loses a small amount of energy as heat, and the remainder of the extra energy is given off in the form A range of wavelengths of light can excite the electrons of a of a photon. The emitted luorescence has a lower energy luorochrome. For example, luorescein will luoresce when hit than the absorbed light, so the wavelength of the emitted light by light with any wavelength between 450 nm and 520 nm. is longer than that of the excitation light. However, the closer the excitation wavelength is to 495 nm, the Figure 6.9 Cultured pulmonary artery endothelial cells stained for tubulin (red), actin (green) and DNA (blue). The dual immunoluorescence procedure used rabbit-anti-actin and mouse-anti-alpha tubulin as primary antibodies. The secondary antibodies used were Texas Red-conjugated goat, anti-rabbit IgG and FITC-conjugated goat, anti-mouse IgG. The sample was also stained with the DNA-speciic dye Hoechst 33342. Scale bar is equal to 20 microns. 88 Detection Methods | Chapter 6 more luorescence will be produced. This optimal wavelength lem with immunoluorescence is photobleaching. Photobleaching is called the excitation peak. Similarly, the light produced by is when the luorophore looses its ability to luoresce. This pho- luorochromes has a range of wavelengths. The emission of tochemical destruction is due to the generation of reactive ox- light from luorescein ranges from 490 nm to 630 nm, and the ygen species in the specimen as a byproduct of luorescence emission peak is approximately 515 nm. Since the phenome- excitation (Figure 6.12). Photobleaching can be minimized by: (a) non of luorescence was irst explained in 1852 by a British sci- decreasing the excitation light in both intensity and duration, (b) entist, Sir George Stokes, the shift in wavelength from short to reducing the availability of singlet oxygen (1O2) by the addition of long during luorescence is called “Stokes shift” (Figure 6.11). singlet oxygen scavengers (= antifade reagents), and (c) using a low concentration of a luorochrome with high-quantum eficiency. Some luorochromes have a small Stokes shift, while other luorescent compounds have large Stokes shifts. For example, Autoluorescence the luorochrome luorescein can be excited by blue-green Biological autoluorescence in mammalian cells due to lavin light, and its Stokes shift is only about 20 nm, which means coenzymes (FAD and FMN: absorption, 450 nm; emission, 515 that the light emitted is green. This contrasts with another lu- nm) and reduced pyridine nucleotides (NADH: absorption, 340 orochrome, phycoerythrin, which also can be excited by blue- nm; emission, 460 nm) can be problematic in the detection green light, but has a large Stokes shift and thus the light will of luorescence probes in tissues and cells. Fixation with al- be emitted in a different color (yellow). dehydes, particularly glutaraldehyde, can result in high levels of autoluorescence. This can be minimized in ixed cells by Photobleaching washing with 0.1% sodium borohydride in phosphate-buff- As with most luorescence-based techniques, a signiicant prob- ered saline (29) prior to antibody incubation. Problems due to Excitation source, 470 nm Peak FITC excitation, 495 nm Intensity (%) Stroke shift 100 Energy Peak FITC emision, 520 nm S3 80 S2 Intersystem crossing 60 S1 T1 Actual FITC emision, 520 nm 40 Ground energy state S0 20 Absorption (blue) 0 400 Phophoresence Green fluorescence emission 425 450 475 500 525 550 575 600 625 650 675 700 Wavelength (nm) Figure 6.11 Excitation and emission spectrum of luorescein. When luorescein is excited at a wavelength other than its peak excitation (470 nm in this example), the shape of the emission curve (darker green) remains the same, but the relative intensity is reduced. The eficiency of the excitation at 470 nm is 45% of peak excitation. Figure 6.12 Illustration of how a singlet-excited state can convert to a triplet-excited state. Photobleaching is the irreversible decomposition of the luorescent molecules in the exited state because of their interaction with molecular oxygen prior to emission. 89 Chapter 6 | Detection Methods autoluorescence can be minimized by selecting probes and small amount of FITC luorescence within the PE band. These optical ilters that maximize the luorescence signal relative unwanted signals must be electronically removed or the meas- to the autoluorescence. Other factors that limit IF include the urement for each detector will overestimate the actual signal. performance of the detection instrument (i.e. how well the mi- This process is called luorescence compensation and can be croscope has been calibrated and set), the speciicity of the automatically calculated in many detection systems using sin- antibodies, and the specimen preparation. gle positive controls. Fluorescence Overlap Applications of IF in Pathology One of the problems that must be dealt with when measuring Some practical applications of IF in diagnostic pathology are: luorescence of more than one color is the possibility that the „„ Analysis of protein antigens in fresh, frozen or, less often, emission signals overlap. It is necessary to remove the over- ixed tissues; sub-cellular localization of protein antigens lapping signal or it will give a false level for one or more colors. in tissue culture monolayers; and observation of bacterial For example, as shown in igure 6.14, there is signiicant over- or parasitic organisms. Immunoluorescence is primarily lap when using FITC and PE. A range of wavelengths will be used in the research setting, or in clinical research setting, collected for each detection channel. In the igure, these are on frozen tissue. In particular where antibodies compatible identiied as the FITC detector bandwidth and the PE detector with formalin ixation and parafin embedding have not bandwidth. These band-pass optical ilters will allow photons been developed. within this wavelength range to reach the detector. However, „„ A major practical use is for luorescence in situ hybri- as can be seen in Figure 6.14, there is a very small amount of dization (FISH), luorescent labeled DNA is used to detect PE luorescence, which is within the FITC band, and similarly a gene aberrations in cells. FITC detector PE detector bandwidth bandwidth Intensity (%) 100 Excitation source 80 60 40 Overlap 20 FITC emision 0 400 425 450 475 PE emision 500 525 550 575 600 Wavelength (nm) Figure 6.13 NADH autoluorescence in a human colon carcinoma cell line (HCT116). Ultra-violet excitation at 363 nm was used and the emitted luorescence greater than 440 nm was collected. Scale bar is 10 microns. Courtesy of Giselle M. Knudsen, Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University, West Lafayette, IN, USA. 90 Figure 6.14 Fluorescence overlap of FITC and PE. 625 650 675 700 Detection Methods | Chapter 6 „„ Immunoluorescence potentially has a wider dynamic ran- ge than immunoenzyme staining, as there is no enzymatic ampliication involved and thus the dynamic range is determined solely by the sensitivity of the detectors (25). Quantitative immunoluorescence staining coupled with digital scanning of slides and image analysis algorithms have been utilized to create an automated quantitative immunoluorescence technique which has been applied in various studies (30). „„ Multi-staining (see multi-staining section) „„ Visualization of cell structures by super resolution microscopy Chapter 6.11 Future Perspectives The IHC technique continues to undergo evolution and improvement, driven by ongoing demands of reproducibility, sensitivity and quantiication. Today, automated systems enable standardized visualization of targets in tissue with increased sensitivity and improved signal to background ratio. Chromogen, luorescence and multistain technologies are being employed. Increasingly, stained slides are submitted for digital scanning and signals quantiied using image analysis algorithms. The demand for more information from each slide, to conserve available tissue, will inevitably lead to increasing use of multistaining technologies in the pathology laboratories. In addition, targeted therapies have created a need for more quantitative biomarker information, launching a rapidly growing range of new types of IHC tests, variously termed ‘prognostic markers’, ‘predictive markers’, ‘companion diagnostics’ or ‘advanced personalized diagnostics’ (Chapter 11). Thus, future IHC-based tests will increasingly rely upon standardized, approved kits and reagents, in combination with an automated image analysis system for the evolution into quantitative pathology. Acknowledgements Sections, in whole or parts thereof, from the previous editions of this Guidebook are used in the 6th edition. We sincerely thank and acknowledge the contribution of the authors. Special acknowledgements to: Mark Key, J. Paul Robinson, Jennifer Sturgis, References 1. Coons A, Creech HJ, Jones RN. Immunological properties of an antibody containing a luorescent group. Proc Soc Exp Biol Med 1941;47: 200-2.; Coons A, et al. J Immunol 1942; 45, 159-170. 2. Graham R, Karnovsky M. The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: Ultrastructural cytochemistry by a new technique. J Histochem Cytochem 1966; 14: 291-302. 3. Nakane PK. Simultaneous localization of multiple tissue antigens using the peroxidase labeled antibody method: A study of pituitary glands of the rat. J Histochem Cytochem 1968;16:557-60. 4. Taylor CR, Burns J. The demonstration of plasma cells and other immunoglobulin- containing cells in formalin-ixed, parafin-embedded tissues using peroxidase-labelled antibody. J Clin Pathol 1974; 27:14-20. 5. Taylor CR, Mason DY. The Immunohistological detection of intra cellular immunoglobulin in formalin-parafin sections from multiple myeloma and related conditions using the immunoperoxidase technique. Clin Exp Immunol 1974; 18:417-29. 6. Mason TE, Phifer, RF, Spicer SS. An immunoglobulin-enzyme bridge method for localizing tissue antigens. J Histochem Cytochem 1969;17:563-9. 7. Sternberger LA, Hardy PH Jr., Cuculis JJ, Meyer HG. The unlabeled antibody-enzyme method of immunohistochemistry. Preparation and properties of soluble antigen-antibody complex (horseradish peroxidase-antihorse-radish peroxidase) and its use in identiication of spirochetes. J Histochem Cytochem 1970;18:315. 8. Hsu SM, Raine L, and Fanger H. Use of avidin-biotin peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J Histochem Cytochem 1981;29:577-80. 9. Heras A, Roach CM, Key ME. Enhanced polymer detection system for immunohistochemistry. Lab Invest 1995;72:165 (Abstract). 10. Chilosi M, Lestani M, Pedron S, Montagna L, Benedetti A, Pizzolo G, et al. A rapid immunostaining method for frozen sections. Biotech Histochem 1994;69:235. 11. Sabattini E, Bisgaard K, Ascani S, Poggi S, Piccioli M, Ceccarelli C. The EnVision™ system: a new immunohistochemical method for diagnostics and research. Critical comparison with the APAAP, ChemMateTM, CSA, LABC, and SABC techniques. J Clin Pathol 1998;51:506-11. 12. Gross AJ, Sizer IW. The oxidation of tyramine, tyrosine, and related compounds by peroxidase. J Biol Chem 1959;234:1622-1614. 13. Bobrow MN, Harris TD, Shaughnessy KJ, Litt GJ. Catalyzed re porter deposition, a novel method of signal ampliication. Application to immunoassays. J Immunol Methods 1989;125:279-285. George L. Kumar, Nanna K. Kristensen and Lars Winther. 91 Chapter 6 | Detection Methods 14. Adams JC. Biotin ampliication of biotin and horseradish peroxidase signals in histochemical stains. J Histochem Cytochem 1992;40:1457-63. 15. Merz H, Malisius R, Mann-Weiler S, Zhjow R, Hartmann W, Or scheschek K, Moubayed P, Feller AC. Methods in laboratory in vestigation immunoMax. A maximized immunohistochemical method for the retrieval and enhancement of hidden antigens. Lab Invest 1995;73:149-56. 16. Molinie V, Fromont G, Sibony M, Vieillefond A, Vassiliu V, Cochand-Priollet B, et al. Diagnostic utility of a p63/alpha-methylCoA-racemase (p504s) cocktail in atypical foci in the prostate. Mod Path 2004; 17:1180-90. 17. Taylor CR, Levenson RM. Quantiication of Immunohistochemistry – issues concerning methods, utility and semiquantitative assessment. Histopathol 2006; 49:411-424. 18. van der Loos CM. Multiple immunoenzyme staining: methods and visualizations for the observation with spectral imaging. J Histochem Cytochem 2008; 56:313-28. 19. Van der Loos CM. Immunoenzyme Multiple Staining Methods: BIOS Scientiic Ltd.; 1999. 20. Chaubert P, Bertholet MM, Correvon M, Laurini S, Bosman FT. Si multaneous double immunoenzymatic labeling: a new procedure for the histopathologic routine. Mod Path 1997; 10:585-91. 21. van der Loos CM, Gobel H. The animal research kit (ARK) can be used in a multistep double staining method for human tissue specimens. J Histochem Cytochem 2000;48:1431-8. 22. Mason DY, Micklem K, Jones M. Double immunoluorescence labelling of routinely processed parafin sections. J Pathology 2000;191:452-61. 23. Dickinson ME, Bearman G, Tille S, Lansford R, Fraser SE. Multispectral imaging and linear unmixing add a whole new dimension to laser scanning luorescence microscopy. BioTechn 2001; 31:1272-8. 24. Wu X, Liu H, Liu J, Haley KN, Treadway JA, Larson JP, et al. Immu noluorescent labeling of cancer marker Her2 and other cellular targets with semiconductor quantum dots. Nat Biotech 2003; 21:41-6. 25. Jubb AM, Landon TH, Burwick J, Pham TQ, Frantz GD, Cairns B, et al. Quantitative analysis of colorectal tissue microarrays by im munoluorescence and in situ hybridization. J Pathol 2003; 200:577-88. 26. Coons AH ea. Proc Soc Exp Biol Med 1941;47:200-2. 27. Coons AH and Kaplan MH. Localization of antigen in tissue cells. II. Improvements in a method for the detection of antigen by means of luorescent antibody. J Exp Med 1950; 91:1-13. 28. Weber K, Bibring T, Osborn M. Speciic visualization of tubulincontaining structures in tissue culture cells by immunoluorescence. Cytoplasmic microtubules, vinblastine-induced paracrystals, and mitotic igures. Exp Cell Res 1975; 95:111-20. 92 29. Bacallao R, Sohrab S, Phillips C. Guiding Principles of Specimen Preservation for Confocal Fluorescence Microscopy. In: Pawley JB, editor. Handbook Of Biological Confocal Microscopy: Springer US; 2006. p. 368-80. 30. Camp RL, Chung GG, Rimm DL. Automated subcellular localization and quantiication of protein expression in tissue microarrays. Nat Med 2002; 8:1323-8. 93 Part I: The Staining Process Chapter 7 Digital Pathology Joachim Schmid, PhD Clive R. Taylor, MD, DPhil Dig•i•tal Pa•thol•o•gy (n.) A dynamic, image-based environment that enables the acquisition, management and interpretation of pathology information generated from a digitized glass slide. The Digital Pathology Association web site Digital Pathology | Chapter 7 Chapter 7.1 Microscopy – From Analog to Digital diology converted to a digital format over a remarkably short period of time. The change was technology driven. It radically Early in the 19th century, improvements in the manufacture of transformed the way in which radiology is practiced today, and lenses allowed production of affordable, high quality, compound as such provides a glimpse into the future of digital slides in microscopes (Figure 7.1). The impact on the biological sciences pathology. To date, digital imaging has found only limited ap- and on medicine was enormous. Pathology, a new branch of plication in pathology, primarily for education and research, medical sciences was born, with surgical pathologists effective- with diagnostic use limited to isolated telepathology applica- ly practicing ‘image analysis’ on a personal, subjective basis. tions, and a few immunohistochemical ‘quantitative analyses’. A crucial difference from radiology is that pathologists begin, not with a digital ile, but with a piece of tissue, that must be ixed, embedded, cut and stained, prior to obtaining a digital image. Current pathology practice is that glass slides are distributed to the pathologist, and diagnosed by the usual light microscopic approach, as they have been for a hundred and more years. Selected slides may then be digitally scanned on an elective basis, for research or educational purpose, or for performance of quantitative algorithms. Until recently, the scanning time has been prohibitively slow using digital imaging methods, adding greatly to the time required for a pathologist to complete a case. But all this is changing in the face of rapid improvements in scanning technology, both hardware and software. Chapter 7.2 Automation of Microscopy Figure 7.1 Early image analysis device. Automated manual microscopy has been evolving since the irst demonstration of telepathology in 1968. The concept has One hundred and ifty years later, the accumulated literature and evolved along two distinct pathways, one driven by technolo- experience of surgical pathology has grown to such an extent gy, the other by need. The technology ield focused on adapt- that it is the ‘gold standard’ for diagnosis of many diseases. Even ing new knowledge developed for other disciplines in an effort in this molecular age, cancer treatment is rarely initiated without to prove that automated microscopy was possible. Improve- a tissue diagnosis. However, the methods of microscopic exam- ments in digital camera resolution, speed and idelity, the in- ination have changed relatively little in more than a century (1). vention of the robotic microscope and stage, the development Hematoxylin and eosin, and other stains in common use today, of the Internet and the exponential advancements in computer were irst introduced in the 1850s. Immunohistochemistry was technology, including processor speed, memory and storage, utilized in routinely processed FFPE tissues in 1974, and en- all have contributed to eliminating the hurdles that delayed de- tered general use in the next two decades. In 2013, pathologists velopment of a viable automated microscopy system. The sec- are still performing cancer diagnosis by examining glass slides ond group focused on need, with the goal of responding to de- at a microscope. However, new computer-based technologies mands in anatomic pathology, such as rapid second opinion, offer the real prospect of radical change. pathology stafing of remote frozen sections, cost reduction for reference laboratories, medical student and resident teaching, Radiology and pathology share a foundation of interpreting continuing medical education, and improved storage and re- images for diagnostic purposes. A little over a decade ago, ra- trieval of slides (2-8). 95 Chapter 7 | Digital Pathology The concept of developing a functional robotic telepathology net- Chapter 7.3 Slide Scanning work, with the aim of providing real-time expert opinions for frozen sections and dificult cases was put forth by Ronald Weinstein in The irst step in digital pathology is to obtain a digital rep- the mid 1980s (9, 10). Performance studies supported the feasi- resentation of a pathology slide. The scanned image or images bility of telepathology, and a demonstration project for the U.S. are saved as two-dimensional digital iles. However, patholo- Department of Defense showed how slides prepared in El Paso, gy slides and the tissue on pathology slides are three-dimen- Texas could be read remotely via satellite, at the Armed Forces sional structures, albeit that the third dimension is only a few Institute of Pathology in Washington, D.C. Although the prototype microns. With a low power magniication lens (such as a 5x was successful, there were many limitations to commercialization, lens), scanning a single focal plane is often suficient for cap- most involving a lack of telecommunication infrastructure and turing this three-dimensional tissue with acceptable focus. standards. Since it was not cost effective to purchase a satellite, With higher magniication (such as a 40x) lenses, the depth of cheaper alternatives were sought, and the concept of “static tele- focus is shallow, necessitating that the scanning system has pathology” was introduced as an alternative. In static telepathol- the capability to automatically focus on, select and capture a ogy, a pathologist captures and saves a digital image (photomi- single focal plane. In addition, more sophisticated systems in crograph), using a camera mounted on a microscope, and then fact capture several single focal planes, a process known as forwards one or more images to a remote computer, where they Z-stacking, that allows the viewing pathologist a good replicate may be reviewed by a second pathologist. E-mail and servers of focusing up and down through the tissue section. were available to facilitate the transfer of images, and standards in image formats evolved rapidly. However, the static nature of All modern systems have the ability to autofocus; however, they the images and the inability of the viewing pathologist to examine all do it in a slightly different manner. A second camera can be other ields and to change magniication, severely limited clinical used to monitor and adjust the focus continuously as the slide is use. In practice, thousands of static images or photomicrographs scanned. Or a map of the tissue can be made available on the are required to fully represent a standard pathology slide, and a slide to be scanned, and ‘points of focus’ can be created. Then, method was required to organize and display them, so as to rep- as the slide is scanned, the system performs an autofocus at each resent a facsimile of the original slide. calculated focus point. In this way, systems may overcome problems such as variations in thickness of the tissue or tissue folds. As a result, only few pathologists used these systems diagnostically, although many used digital photomicrographs for tumor The process of scanning a slide also differs between various boards, teaching and other educational purposes. systems. All processes involve acquiring multiple images in some fashion and then stitching them together to create a rep- The next major technical step was the creation of ‘stitching’ soft- resentation of the slide. Some systems acquire images as tiles, ware, which allowed the digital representation of an entire micro- while other systems use methods such as line scanning. scopic slide by digitizing individual microscopic ields and then 96 stitching them together to create a virtual slide. This process was Scanning systems generally are judged by two criteria: speed laborious and time consuming and the computer processing and and resolution. The total speed of acquisition involves not only storage requirements pushed up against the limits of then avail- acquiring the image (scan time), but also entails stitching the im- able technology. As a result of these limitations, the next systems ages together, image compression and storing the resulting im- created were hybrids, containing elements of both static and age on a computer for viewing. Since the purpose of acquiring a dynamic systems. These systems digitized an entire slide at low virtual slide is to view or analyze it in the future, some additional power magniication, creating a tissue map for the pathologist to processing is often implemented to facilitate this viewing process. select areas of interest, which would then be re-digitized at higher For example, to implement the full functionality of a standard light magniication and forwarded as a series of static images. microscope, the scanned image must be viewable at similar mag- Digital Pathology | Chapter 7 niications found on a standard microscope, namely, 1x, 2.5x, 5x, Chapter 7.4 Digital Slide Review 10x, 20x and/or 40x. Assuming the slide is scanned, stitched and stored at the equivalent of a 40x objective, viewing the 40x image A digital slide is a digital representation of a glass slide that can be does not require additional processing. However, to view the im- viewed on a screen at any location. Different systems have differ- age at the equivalent of a 5x objective, the 5x image must be de- ent approaches to display information. Not only is the digital slide rived from the 40x image before it can be displayed. This process information displayed, but patient and case information can be dis- results in a signiicant delay between the time an area is selected played, as well as multiple images and/or thumbnails of multiple and the time it is displayed on a monitor. To reduce this delay, the slides. These interfaces have been termed ‘pathology cockpits’ as acquired image is processed to a format where the intermediate they allow the pathologist to work in one location on all aspects of magniication levels have already been calculated and stored. a case. The user interface allows to perform similar tasks such as While this format, known as a pyramidal format, decreases the changing the magniication, moving the slide in any direction, sav- time to load intermediate magniication views, it increases the im- ing regions as image iles, annotating speciic areas of the slide, age processing (time) necessary following image acquisition and launching image analysis applications and generating reports. produces a larger ile to be stored (11). An image format for digital pathology has been developed as a DICOM format (Digital Imaging and Communications in Medicine), a standard for distributing Chapter 7.5 Applications in Digital Pathology and viewing medical images. Like applications (apps) for smartphones, new software for imThe ability of the image acquisition system to resolve features pres- age analysis will be utilized. Routine use of these apps (mostly ent in the microscopic slide is known as the absolute or “point-to- algorithms) will extend the capabilities of pathologists beyond point resolution” and is dependent on the microscope objectives, subjective morphologic criteria, and beyond the present com- the camera lens and the analog to digital conversion process. pass of the human eye (1). What is more important is the actual resolution, as appreciated by the observer, which is also dependent on hardware and software Applications for digital pathology today include, but are not compression techniques and on the characteristics of the monitor limited to: used to display the image. In radiology these monitors are subject to standards, not yet established and accepted for pathology. Figure 7.2 Philips Ultra Fast Scanner, an example of a digital pathology scanner with a time to display of less than one minute at 40x magniication. „„ Quality assurance of routine and IHC slides, using objective measurements of color and intensity Figure 7.3 Pathology cockpit by Omnyx, LLC, A GE Healthcare and UPMC Joint Venture. An example of a digital pathology interface. 97 Chapter 7 | Digital Pathology „„Image analysis and quantiication „„Telepathology; remote viewing and sharing of digital slides „„Obtaining a second opinion in dificult cases (consults) forms well at selecting relevant areas on the slide using morphology; an algorithm performs well at quantifying intensities and counts. „„Obtaining a rapid second opinion (as for frozen sections) „„ Reports including images for the diagnostic report (archiving Image Quality and recovery) Quantiication/ Algorithm Result „„Tumor boards „„ Internal and external quality assurance at the interpretative Algorithmn Staining Quality Region Selection level (diagnostic review) „„Research; drug development, clinical trials, 3D pathology „„Education; medical students, residents and CME This growing range of applications will be considered here pri- Figure 7.4 Quantiication/Algorithm results are dependent on staining quality, image quality, the algorithm, and region selection. marily with reference to immunohistochemistry. Additionally, the principle of garbage-in, garbage-out cannot Chapter 7.6 Image Analysis and Quantification be overstated. There is no point in attempting to use image analysis to quantify IHC slides unless one is sure that the im- Collecting an image for subsequent analysis has more rigorous munohistochemical procedure has been performed in a repro- requirements than the capturing of images for viewing only. Since ducible fashion. Automation (Chapter 9), standardization and the image will be ‘analyzed’, procedures must be established to the proper use of controls (Chapters 4, 5 and 14) in the pro- ensure that the image is captured reproducibly, and that the system cess of preparing the IHC slides are essential prerequisites for is operating in the detectable range for the image that is analyzed. successful image analysis. For example, if an image is acquired for quantifying the amount of HER2/neu protein, then the exact same image must be produced As shown in Figure 7.4, staining quality, image quality, algorithm whether the slide is scanned today, tomorrow or six months from and region selection are important components to achieve good now. The reproducibility of such an image depends on a number of quantiication/algorithm results. Staining quality can be opti- factors such as stage and camera drift over time, luctuation/varia- mized by using ready-to-use reagents and automated staining tion in the light source, and variation in the analog-to-digital conver- equipment with deined staining protocols. Additionally, linking sion process. In addition, further complexity may be introduced by speciic pre-treatment, antibodies, and staining protocols to an the use of different objective magniications. To avoid these com- algorithm is key for a consistent result. Having consistent, stand- plications, calibration must be performed on a regular basis (most ardized staining is necessary for successful image analysis be- systems today perform system calibration automatically). cause differences in staining lead to differences in acquired digital images, thereby affecting image analysis. The image quality 98 Image analysis represents a signiicant step in standardizing the of a digital slide must meet rigorous standards in order to apply interpretation of slides in pathology (12). The creation of an image image analysis methods for the purpose of quantiication, or to analysis application is not an easy task. Just as it takes consider- differentiate between components of the morphology of the tis- able time and effort to create and validate a pharmacodiagnostic sue. Clearly the algorithm must also be developed to produce assay, it takes signiicant effort to create and validate a clinically results that are relevant for the clinician. Areas to be evaluated useful image analysis algorithm. These algorithms are meant to (such as the tumor regions) must be selected for the algorithm complement pathologists, not replace them. Accordingly, algo- to be included in the analysis. Selection of the region(s) for anal- rithms should not attempt to emulate those skills and capabilities ysis should be made based upon predetermined guidelines, in which pathologists excel, but rather should aid pathologists in relying on the pathologist’s expertise. Thus, image analysis is an tasks where the human eye does not excel. The pathologist per- aid to the pathologist. Digital Pathology | Chapter 7 „„HER2 quantiication Chapter 7.7 Analysis of IHC stains – ’Eye vs Algorithm’ „„DNA ploidy quantiication With the exception of these few semi-quantitative IHC tests and Algorithms for the following areas are in use today: „„ER/PR quantiication „„Ki-67 quantiication algorithms, the great majority of IHC stains performed in the sur- „„p53 quantiication gical pathology laboratory are performed for the purposes of assisting in the identiication of cell types, and ultimately tumor In these quantitative algorithms the image analysis process is diagnosis and classiication. In this context, the question ‘asked’ intended to mimic the pathologist, producing semi-quantitative of the IHC stain is whether or not there is a signiicant positive results on an arbitrary scoring system, of 0 to 3+ , or similar. reaction (in relation to the controls), and whether it is positive A number of the software programs designed for this purpose or negative in the context of cell identiication, not ‘how much have been cleared as algorithms by the FDA, as Class II devic- positivity’ is present? As noted above, the experienced trained es. ASCO/CAP guidelines for HER2 and ER/PR recommend the human eye is relatively good at distinguishing patterns as ‘nor- use of image analysis for the evaluation of these stains (13, 14). mal’ or not, and grouping the ‘abnormals’ into known disease The use of digital pathology for reading of H&E slides is current- categories. The human eye is much less reliable at counting by ly under discussion and will probably be a Class III device. number and calculating percentage values, and is quite poor at judging variations in intensity, unless major. In many respects the computer, given the appropriate digital slides, is almost the converse, good at assessing intensity, reproducibly, good at counting, good at event detection, but less good at categorizing complex patterns. Given the high-quality, high-resolution images, and standards against which the computer (software) can be calibrated, including ‘training’ in pattern recognition, a great number of these shortfalls in digital image analysis are diminishing, and diminishing rapidly. So much so that the use of ‘intelligent’ trained software to assist the pathologist in interpretation of IHC images (as well as perhaps H&E stains) is now a reasonable expectation in the very near future. For accurate quantiication, the use of computer driven algorithms will be indispensible, but at this point in time still awaits the general availability of IHC staining methods that deliver a known and deined amount of signal (stain) to target (antigen), as well as the development of quantii- Figure 7.5 Pathologists can now use a digital system from Philips to help them score Her2-neu stained slides to assess treatment options for breast cancer patients. able internal reference or calibration controls (Chapter 1). Chapter 7.8 Quality Assurance in IHC Dako HercepTest™ Region Score Slide Score Staining intensity 3 3 Presently, the results of IHC staining, including control slides, are HercepTest™ Score 3+ 3+ assessed by technologists and/or pathologists, on an individual subjective basis, without direct comparison to prior control slides Results Strongly Positive Figure 7.6 Dako HercepTest™ algorithm result displayed on the Omnyx system. from earlier runs or slides from validation studies. Digital images of control slides allow for direct and accurate comparison of positive (and negative) control slides for spectral pattern and in- 99 Chapter 7 | Digital Pathology tensity, providing a more precise means of quality assurance of Additionally, the acquisition and assembly of multiple regis- automated staining platforms (or manual methods). It is likely that tered parallel slides from single tissue blocks may provide a these types of programs will become an intrinsic component of ‘virtual 3D tissue block’, that will provide for levels of morpho- the monitoring process in future generations of IHC autostainers. logic examination and manipulation far beyond current capabilities, all of course contingent upon ability to manage huge While reproducibility of stain color and intensity is important in IHC amounts of data and control costs. With the recent speed of in general, it is absolutely critical that precise measurement of per- progress in these areas almost anything seems possible! formance occurs if IHC is to be used for true quantiication, extending beyond current semi-quantitative visual or automated scoring New methods to combine morphologic analysis and staining ex- approaches (0, 1+, 2+, 3+, etc.). Under these circumstances, the pression evaluation will also support the integrated analysis of human eye might not appreciate subtle changes in intensity that different markers. Especially for cases where ‘multiplex staining’, easily may affect whether a slide is scored as 1+ or 2+ (and hence combining IHC, ISH and luorescence methods will be used to negative or positive), whereas a digital slide can be evaluated for review and analyze slides that have been stained with multiple intensity by appropriate algorithms with great accuracy. markers, where the evaluation of patterns of staining of single cells and colocalization information will become critical. Chapter 7.9 Future Trends in Digital Pathology In summary, digital pathology is an important new technology that already has penetrated pathology education signif- As already noted the experience of ‘digitized radiology’ provides icantly. With recent improvements in the technology, other insight into future developments in pathology. Most of the current uses are sure to follow quickly. applications described above will continue and will expand and new, yet unimagined applications will emerge. Acknowledgements As has occurred in radiology, digital pathology will be used to Sections, in whole or parts thereof, from the previous editions of review complete cases including H&E, special stains and IHC this Guidebook are used in the 6th edition. We sincerely thank and ISH slides. In scenarios where the complete worklow in and acknowledge the contribution of the authors. Special ac- the pathology laboratory is digitized the information will all be knowledgements to: Kenneth J. Bloom and Mark Verardo. collected in the pathologist cockpit and used to generate comprehensive patient reports. The impact upon worklow (Chapter 10) in pathology laboratories will be profound. Result LIS Evaluation/ Algorithm Staining Scanning Figure 7.7 Complete digital worklow. 100 Digital Pathology | Chapter 7 References 1. Taylor C. From Microscopy to whole Slide Digital Images: A Century and a half of Image Analysis. Appl Immunohistochem Mol Morphol 2011; Vol 19, 6:491- 493. 2. Demichelis F, Della Mea V, Forti S, Dalla Palma P, Beltrami Ca. Digital storage of glass slides for quality assurance in histopathology and cytopathology. J Telemed Telecare 2002; 8:138-142. 3. Harris T, Leaven T, Heidger P, Kreiter C, Duncan J, Dick F. Comparison of a virtual microscope laboratory to a regular microscope laboratory for teaching histology. Anat Rec 2001; 26 5:10-14. 4. Heidger PM Jr, Dee F, Consoer D, Leaven T, Duncan J, Kreiter C. Integrated approach to teaching and testing in histology with real and virtual imaging. Anat Rec 2002; 26 9:107-112. 5. Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Van der Kwast T, Isola J. Webbased virtual microscopy in teaching and standardizing gleason grading. Hum Pathol 2005; 36:381-386. 6. Kumar RK, Velan GM, Korell SO, Kandara M, Dee FR, Wakeield D. Virtual microscopy for learning and assessment in pathology. J Pathol 2004; 204:613-618. 7. Marchevsky AM, Wan Y, Thomas P, Krishnan L, Evans-Simon H, Haber H. Virtual microscopy as a tool for proiciency testing in cytopathology: a model using multiple digital images of Papanicolaou tests. Arch Pathol Lab Med 2003; 127:1320-1324. 8. Okada DH, Binder SW, Felten CL, Strauss JS, Marchevsky AM. “Virtual microscopy” and the internet as telepathology consultation tools: diagnostic accuracy in evaluating melanocytic skin lesions. Am J Dermatopathol 1999; 21:525-531. 9. Weinstein RS: Prospects for telepathology. Hum Pathol 1986; 17:433-434. 10. Weinstein RS, Bloom KJ, Rozek LS. Telepathology and the net working of pathology diagnostic services. Arch Pathol Lab Med 1987; 111:646-652. 11. Glatz-Krieger K, Glatz D, Mihatsch MJ. Virtual slides: high-quality demand, physical limitations, and affordability. Hum Pathol 2003; 34:968-974. 12. Bloom K, Harrington D. Enhanced accuracy and reliability of Her-2/neu Immunohistochemical scoring using digital microscopy. Am. J. Clin Pathol. 2004; Vol. 121, 620-630. 13. Wolff AC, Hammond ME, Schwartz JN, et al. American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med 131:18–43, 2007 14. Wolff AC, et al. ASCO/CAP HER2 Testing Guideline Update. Arch Pathol Lab Med. 2013 in press. 101 Part II: The Potentials and Pitfalls Chapter 8 Optimization of Immunohistochemical Reactions Gale E. Pace, MT (ASCP), BSc Op•ti•mi•za•tion (n.) An act, process, or methodology of making something (as a design, system, or decision) as fully perfect, functional, or effective as possible. Merriam-Webster Online Dictionary Optimization of Immunohistochemical Reactions | Chapter 8 Chapter 8.1 Introduction This chapter will discuss the formulation and use of several basic immunohistochemical reagents, including proteolytic The process of implementing immunohistochemical tests enzymes for tissue pre-treatment, antibody diluents, blocking in the diagnostic laboratory has been greatly simpliied by and enhancing solutions, and wash solutions as they relate to the availability of standardized reagents, instruments, and assay optimization. Complementary information on the topics assay protocols from commercial manufacturers. However, of tissue preparation, antigen retrieval, background suppres- researchers and diagnosticians who wish to develop new sion, and other aspects of assay optimization may also be immunohistochemical assays or to explore new applications found in the related chapters in this Guidebook. for existing tests must carefully consider all of the different steps in the process, including the methods of tissue preparation and the reaction conditions for each assay step in order to obtain clear, speciic antigen signals and to minimize Chapter 8.2 Tissue Digestion using Heat-induced Epitope Retrieval non-speciic (background) reactions, as well as interpretation and reporting (see Chapter 1). Antigen retrieval, often referred to as the heat-induced epitope retrieval technique (HIER) is predominantly based on heating The complex nature of tissue specimens presents both an slide-mounted specimens in a buffer solution, followed by a aid and a challenge to the development of any new immuno- cooling-off period. histochemical test. Tissue morphology can be tremendously helpful for interpreting the speciicity of staining signals in a High-temperature heating of tissues is used as a non-enzy- new immunohistochemical reaction; on the other hand, tis- matic pretreatment for immunohistochemical staining of for- sues are notoriously prone to non-speciic binding artifacts, malin-ixed, parafin-embedded tissue sections. HIER is com- and are also sources of enzymatic and other biochemical ac- monly used to overcome the effects of covalent cross-links that tivities that can generate false signals from assay reagents are formed in tissues during formalin ixation (see Chapter 3). and confound even the most expert interpretation. The goal of While controlling the temperature and incubation time in retriev- assay optimization is to enhance the strength and speciicity al solution is vital in this process, pH is also very important in of the signal generated by the immunological and enzymat- maintaining optimal morphology and consistent immunoreac- ic staining reactions while suppressing noise and artifacts. tivity. Common pH’s in clinical IHC applications range between Figure 8.1 A) Anti-BCL2 1:200 with pH 6.1 retrieval. Figure 8.1 B) Anti-BCL2 1:200 with pH 9.0 retrieval. 103 Chapter 8 | Optimization of Immunohistochemical Reactions 6.0-9.0. Low pH solutions have been observed to preserve tis- teolytic digestion; however, glycoproteins that contain carbo- sue morphology while demonstrating a more distinct nuclear hydrate epitopes may beneit from proteolysis if the epitope is pattern staining intensity. High pH retrieval solutions depend- ‘unmasked’ by cleaving and de-constraining the peptide back- ing on the chemical composition tend to increase staining in- bone, allowing the antibody access to the epitope for binding. tensity of most antibodies, with target epitope in membrane Non-protein antigens may also beneit from the improved rea- and cytoplasm structures. For IVD-labeled antibodies the gent penetration resulting from tissue digestion. speciication sheet from the manufacturer provides the appropriate HIER protocol which demonstrates the most consistent Preliminary experiments should be conducted to determine result. For example BCL2 demonstrates a weak staining in lym- the appropriate choice of enzyme, incubation time, tempera- phoid tissues when pH 6.1 citrate retrieval solution is used, but ture, and concentration for optimal digestion. Most of the pro- very strong and speciic appropriate immunoreactivity results teolytic enzymes commonly used for IHC display their highest when using pH 9.0 EDTA retrieval solution (Figure 8.1). activities near 37 °C, and many protocols use this reaction temperature in order to achieve the digestive effect in the short- For more information on antigen retrieval or HIER, please refer est possible time. In some cases, however, the use of a lower to Chapter 3. temperature may be preferable. By reducing the rate of the digestion reaction, one may extend the reaction time and thus exert more control over the digestion process – a method that Chapter 8.3 Tissue Digestion using Proteolytic Enzymes is especially useful for situations requiring very mild digestion. Table 8.1 lists several commonly used enzymatic reagents and the typical incubation conditions used in IHC. Proteolytic digestion is commonly used to overcome the effects of covalent cross-links that are formed in tissues during forma- In general comparison studies by Quality Assurance Organi- lin ixation. Controlled proteolysis can improve the penetration zations such as UK NEQAS and NordiQC have revealed that of reagents into the tissue structures and restore the immuno- enzymatic methods perform less well than antigen retrieval or dominant conformation of epitopes of interest, thus allowing the HIER and are much less reproducible. So much so that their primary antibody to reach and bind its intended target. If the use is recommended only in those rare circumstances where enzymatic cleavage points for the protease are in proximity to HIER does not give satisfactory results. a formalin cross-link, digestion may induce a relaxation of the rigid protein structure and facilitate contact between the primary Table 8.1 Proteolytic enzymes and typical incubation conditions. antibody and the corresponding antigenic determinant. Enzyme Approximate activation temperature Incubation time Proteolytic enzymes cleave speciic amino acid sequences Proteinase K 25-37 °C 5 min within the peptide chain of the digested protein; since proteas- Trypsin 37 °C 10 min different effects upon tissues depending upon the type of ix- Pepsin 37 °C 5-20 min ative used for processing the tissue, the antigenic target, and Protease XXIV 37 °C 5-20 min Pronase 25-37 °C 30 min es differ in their cleavage speciicities, they can have markedly the epitope recognized by the antibody. Because the effectiveness of proteolytic digestion is dependent upon multiple factors, optimal digestion conditions must be determined empirically for each tissue and antigen/antibody combination. Carbohydrate epitopes, for instance, being non-proteinaceous, may be expected to be unaffected by pro- 104 NOTE: Formalin does not preserve tissue proteins by coagulation but it is thought to form cross-links with basic amino acids. Ethanol and mercuric chloride-based ixatives are based on coagulation. With few exceptions retrieval should not be performed on ethanol-ixed tissues. It should only be conducted with limited controlled protocols in mercuric–chloride-based ixatives. Optimization of Immunohistochemical Reactions | Chapter 8 Chapter 8.4 Endogenous Enzyme Blockers before the application of the enzyme-labeled immunohistochemical detection reagent. This principle is used onboard the The two most common enzyme activities that are used to gen- Dako Omnis instrument. erate chromogenic signals in immunohistochemistry, horseradish peroxidase and alkaline phosphatase, are also both encountered as endogenous activities in a variety of cellular Table 8.3 Common endogenous enzyme blocking reagents for horseradish peroxidase and alkaline phosphatase systems. and tissue specimens. If the endogenous enzymes are similar Blocker Enzyme in speciicity to the enzymes used in the immunohistochem- Dual endogenous enzyme block Horseradish peroxidase and alkaline phosphatase labels false-positive signals that interfere with, and even overwhelm, Hydrogen peroxide Horseradish peroxidase label the signals from the immunohistochemical reactions. Enzyme Levamisole + chromogen except intestinal alkaline phosphatase Alkaline phosphatase label Weak acid (0.3 N HCl), including intestinal alkaline phosphatase Alkaline phosphatase label ical detection system, the endogenous activity can produce blockers are used to inhibit the activity of such endogenous enzymes within cells and tissue specimens. Table 8.2 lists some of the common sources of these endogenous activities. Simple reagents may be used to completely block these endogenous enzymes by either competitive inhibition or acid inactivation (see Chapter 15, for further information on enzyme inhibitors) Table 8.2 Endogenous enzymes found in a variety of cells and tissue types. Enzyme: Peroxidase Enzyme: Alkaline Phosphatase* Red blood cells Placenta Intestine - situated between cellular components of mucosa Granulocytes Proximal tubules of kidney Eosinophils Osteoblast in bone Hepatocytes Arterial & capillary endothelial cell surfaces Muscle Stromal reticulum cells Kidney Neutrophils Monocytes Follicle and mantle zones in most lymphoid tissue Figure 8.2 Example of endogenous peroxidase in red blood cells of kidney stained with DAB. *Alkaline phosphatase is destroyed by routine ixation and parafinembedding procedures. Generally speaking, enzyme blockers are applied prior to the addition of antibody reagents in the staining protocol. However, the enzyme blocking reagent may interfere with the immunohistochemical reaction, which may alter sensitive epitopes to the extent that they are no longer recognized by their cognate antibody. In such cases the blocking reaction should be attempted after the primary antibody has been applied, but Figure 8.3 Example of endogenous alkaline phosphatase in ileum stained with Permanent Red. 105 Chapter 8 | Optimization of Immunohistochemical Reactions Chapter 8.5 Protein Blocking Reagent Chapter 8.7 Antibody Diluents These protein-containing reagents are used to reduce non-spe- Antibody diluents are buffered solutions that are used to for- ciic reactions that may result from the binding of antibodies and mulate working solutions of antibodies for use in IHC assays. other assay reagents to various tissue components. Antibodies Because antibody conformations are highly dependent upon and detection systems used in immunohistochemistry are primar- the aqueous environment, diluent formulations can signiicantly ily protein-based reagents, and they are susceptible to non-spe- alter the stability and binding properties of antibodies, affecting ciic binding. Protein blockers act to minimize non-speciic protein both epitope speciicity and “non-speciic” interactions with the adsorbance by competing for the non-speciic protein binding Fc region of the antibody molecule. Variations in charge, hydro- sites on the specimen. By occupying the tissue binding sites phobicity, glycosylation, and other physico-chemical properties with blocking protein prior to (or during) the incubations with the between antibodies can make it dificult to accurately predict other immunohistochemical reagents, the non-speciic signals the behavior of antibody-diluent combinations, but the availa- that would otherwise develop can be greatly reduced. A more bility of several effective commercial diluents has simpliied the thorough discussion of the various non-speciic protein-protein empirical process of diluent selection. interactions that occur between immunohistochemical reagents and tissues, and the protein blockers that may be used to prevent Ionic interactions are one of the primary forces controlling the such interactions, is given in Chapter 15. immunochemical binding of antigens with antibodies, and these interactions are pH-dependent. Generally speaking, buffers that approximate pH (7.3-7.4) are often suitable for dilution of prima- Chapter 8.6 Endogenous Biotin ry antibodies, and are a good starting point for optimization. The isoelectric point, or pI (the pH at which the net electric charge Biotin, may be protein-bound to tissue and can interfere with prop- of a molecule is zero), for immunoglobulins can range from 5.8 er interpretation of staining patterns, when using a labeled (strept) to 8.5 for a given antibody. If the pH of the diluent is too close to avidin-biotin (LSAB) or ABC complex-based visualization systems. the pI of the antibody, solubility can be diminished, with negative Kidney and liver contain high amounts of endogenous biotin, while effects upon both signal and background. If low signal and high lesser amounts may be found in the GI tract, lung, spleen, and background are observed at the intial pH tested, try raising or cells grown in culture media containing biotin (RPMI). Biotin block- lowering the pH of the diluents buffer by 0.5 pH units. Antibody ing is based on the extreme afinity of (strept)avidin for biotin. Once interactions are also affected by ionic strength; most diluents the avidin stage is saturated, free biotin is added the system is es- contain millimolar quantities of sodium chloride or other salts sentially closed to further interaction with subsequent free or con- which serve to reduce non-speciic interactions among charged jugated (strept)avidin or biotin stages. Excess avidin is applied to molecules. Excessive ionic strength should be avoided, since it the tissue. Each biotin molecule is capable of binding to only one may overcome the speciic charge interactions required for high avidin molecule. Once the avidin stage is saturated, free biotin is avidity antibody-epitope binding. added. The system is essentially closed to further interaction with 106 subsequent free or conjugated (strept)avidin or biotin stages. The Several pre-formulated antibody diluents are now available need for biotin blocking solutions has decreased due to the devel- from commercial manufacturers; these diluents can simplify opment of labeled polymer visualizations. Polymer visualizations the optimization process when selecting diluents for a new an- provide the technician with fewer protocol steps due to the conju- tibody or assay, and may also confer increased stability to the gation of the enzyme and secondary antibody on a dextran back- working primary antibody solution. Many of the commercially bone. Polymer systems have the potential for greater numbers of available antibody diluents are based upon Tris-HCl buffers horseradish peroxidase enzymes to produce increased staining containing a detergent and proprietary stabilizers. Some dil- intensity. For these reasons the use of traditional ABC systems to- uents also contain protein-based background-reducing com- day is limited to speciic research and clinical needs. ponents such as bovine serum albumin or serum proteins. Se- Optimization of Immunohistochemical Reactions | Chapter 8 rum-containing diluents should be used with caution; serum binding to the primary antibody can cause a reduction in sensitivity, while binding to the secondary antibodies used in IHC detection systems can result in false positive results. Antibody O CH2 Hydrogen bonds C CH2CH2 OH Ionic interactions C CH2 O CH3 in which they will be used, and be aware that refrigeration and heating can cause changes in the pH of temperatureroom temperature before use. „„Phosphate buffered saline (PBS) should generally not be Hydrophobic interactions CH CH2 CH3 CH3 CH2 CH2 CH CH3 CH3 changes. Always prepare buffers at the same temperature manufacturer for a particular antibody CH3 CH3 more than a few days unless validated stability data is available „„The pH’s of Tris-based buffers are sensitive to temperature used as a diluent unless speciically recommended by the Antigen CH2CH stable the working solution „„Highly diluted antibody solutions should not be used for sensitive buffers. Always allow reagents to equilibrate to NH2 -O CH2CH2CH2CH2NH3+ „„Generally speaking, the more dilute the antibody, the less CH Van der Waals forces Chapter 8.8 Antibody Concentrations CH3 Antibodies belong to a group of proteins called immunoglobulins (Ig) which are found in the plasma or serum. Five major classes of immunoglobulins exist – IgM, IgG, IgA, IgD and IgE. Most antibodies in IHC are from the IgG class. Optimal reactivity can Figure 8.5 Antibodies are attracted to antigens initially through electrostatic interactions, and subsequently through Van der Waals and hydrophobic interactions. be achieved when using the appropriate diluents. The optimization process includes diluting the antibody at the recommended dilution and one serial dilution above the recommended dilution and one below to ensure the appropriate staining reaction. More Antibody stability in solution cannot be predicted without tho- on this topic can be found in Chapter 4 and Chapter 5. rough stability studies; technicians are advised to follow proper quality control procedures for stability validation if primary antibodies are to be diluted in the laboratory and utilized for Chapter 8.9 Incubation times extended periods of time. An advantage to using commercially diluted primary antibodies is the built-in customer protec- All other factors held constant, incubation times affect every tion provided by the regulatory mandates that govern reagent step in immunohistochemistry protocol. Insuficient primary manufacturers. Manufacturers must demonstrate the stability antibody or visualization incubation times have a direct effect of commercially produced reagents for deined periods to es- on the colorimetric expression. Shorter incubation times results tablish a predictable shelf life for their antibody products. in partial staining or weak stain intensity, while longer incubation times can increase staining intensity. Optimum is usually Important Points to Consider Regarding Antibody Diluents: „„ Diluent pH can be an important determinant of effectiveness „„High concentrations of sodium chloride and azides are not achieved during incubation times less than 20 minutes. Typically, manufactures’ speciications provide recommended incubation ranges for tissue demonstrating a high level of tar- used frequently as preservatives in commercial preparations, geted epitope with high afinity antibodies. However, optimal but these components can reduce antibody reactivity incubation time, i.e. saturated incubation reactions should give „„Excessive ionic strength can decrease speciic staining by interfering with antibody-antigen binding suficient staining intensity to identify high expression and, especially, low expression of immunoreactivity. The importance of 107 Chapter 8 | Optimization of Immunohistochemical Reactions identiication of low expression cells and structures is covered Table 8.4 Effects of temperature on pH of Tris-buffered saline. in more detail in Chapter 4 and Chapter 5. Chapter 8.10 Wash Buffers Temperature pH 5 °C 8.18 25 °C 7.6 37 °C 7.30 Wash buffers are used to remove excess reagents from the specimen after each incubation step in the assay protocol. Commonly used wash buffers, including those that are com- Wash buffers can be employed to counteract the effects of mercially available, include Tris-buffered saline (TBS) and non-speciic reagent binding to tissue specimens. When con- phosphate-buffered saline (PBS). ditions require very high speciicity reactions, such as when using highly sensitive detection methods, an increase in the 108 Tris-Buffered Saline saline and detergent content of the wash buffer can be used to Tris(hydroxymethyl)aminomethane-based wash buffers are of- minimize non-speciic binding. The addition of high concentra- ten utilized in combination with the non-ionic determent Tween tions of salts (e.g. sodium chloride) and detergents (e.g. Tween 20. Commercial wash buffer preparations may also include a 20) to wash buffers will signiicantly reduce the non-speciic preservative, such as 0.01% sodium azide, to prevent the growth binding of many immunohistochemical reagents, including an- of microorganisms and extend the shelf life of the reagent. tibodies and detection reagents such as labeled polymers. Tris-buffered solutions are pH-sensitive; as the solution decreas- Phosphate-Buffered Saline (PBS) es in temperature, pH increases at a rate of approximately 0.03 PBS is also commonly used as a wash buffer for IHC. PBS’s ad- units per degree centigrade, so care should be taken to equili- vantages are reduced auto luorescence in immunoluorescent brate buffer solutions to the proper working temperature in order assays, and it is relatively inexpensive compared to Tris-based to maintain consistent pH. This effect is illustrated in Table 8.4. buffers. However, in some cases PBS can cause higher levels Figure 8.6 A) Anti-Cytokeratin, Clone 34βE12, Dako Code M0630, on human prostate tissue, rinsed with 150 mM NaCl, 0.05% Tween 20, Tris-Buffered Saline, Dako Code S3006, pH 7.6 at 25 °C. Non-speciic staining is evident in the lumen and the connective tissue. Figure 8.6 B) Anti-Cytokeratin, Clone 34βE12, Dako Code M0630, on human prostate tissue, rinsed with 300 mM NaCl, 0.1% Tween 20 Tris-Buffered Saline, Dako Code S3306, pH 7.6 at 25 °C. Staining is conined to speciic signals in the cells of the prostatic epithelium. Optimization of Immunohistochemical Reactions | Chapter 8 of nonspeciic staining, and it has been observed to reduce product will appear darker and more intense than the non-en- the speciic binding abilities of certain monoclonal antibodies hanced DAB product. Since the enhancer metals work only (Anti-CD30, for example). upon the reduced chromogen that has been deposited by the peroxidase reaction, the enhancer reaction generally does not Suggestions for Making Wash Buffers „„Always replace buffers after expiry, and do not combine old contribute signiicantly to background. Incubation times for enhancement should be determined by the individual laboratory, buffers with new based on the desired hue of the chromogen. mendations; buffering ability may be compromised, and The use of enhancers adds yet another uncontrolled variable to ionic strength will no longer be optimal if the recommended the entire staining process, and their use should be reserved for dilution is exceeded those cases where a satisfactory intensity cannot be achieved „„Do not dilute reagents beyond the manufacturer’s recom- „„Clearly label all new solutions with the date when they are by adjusting antibody or label concentrations or incubation times. opened and made „„Do not mix different types of buffers with one another „„Use distilled, organically iltered deionized, or high quality reagent grade water only to make and dilute buffers. Do not use tap water, as the presence of trace quantities of metals, inorganic ions, and other contaminants in tap water may interfere with immunohistochemical reactions Chapter 8.11 Chromogen Enhancers for DAB The DAB chromogen is a popular choice for signal generation A B Figure 8.7 A) Standard DAB chromogen without enhancement. B) DAB enhanced with Dako DAB Enhancer. with peroxidase-based immunohistochemical detection systems. The DAB reaction product, which normally appears as a light brown precipitate, may be enhanced by reaction with any Chapter 8.12 Type of Glass Slides one or a combination of several types of metals. The metals enhance the reduction of the DAB reaction product and darken FFPE tissue sections bind to positive charged glass slides by its appearance. Copper, silver, nickel, gold and cobalt have all the negatively charged carboxyl end of cellular proteins. There been used as enhancers. are a number of glass slides on the market that address tissue adherence for immunohistochemistry and insitu-hybridization. Metal enhancers may be added directly to the chromogenic Poly-L-Lysine, Silanized slides, and plus slides have the com- DAB reaction, or they may be applied in a separate incubation mon characteristic of binding tissue sections. There are also a step following the development of the DAB reaction product number of additives which can be used in the water bath during and a rinse in high-purity water. The latter approach is rec- sectioning that cause tissue to adhere to plain or non-charged ommended when it is desirable to ine-tune the degree of en- slides. It should be mentioned that using the combination of tis- hancement; the optimal incubation time for the enhancement sue adhesives and charged slides results in tissue falling off the reaction may be determined empirically by monitoring the de- slide. Other causes of tissue loss include hand lotion contami- velopment of the enhanced chromogen under the microscope. nating the waterbath, insuficient section drying temperature or The hue and intensity of the inal DAB reaction product will vary time where water is trapped between the glass and the tissue, depending upon the type and concentration of the metal(s) extended heat induced retrieval time, and tissue with high adi- used in the enhancing reaction, but in general the enhanced pose content where there is low cellular protein content. 109 Part II: The Potentials and Pitfalls Chapter 9 Automation in IHC Ole Feldballe Rasmussen, PhD, MSc Au•to•ma•tion (n.) Automatically controlled operation of an apparatus, process, or system by mechanical or electronic devices that take the place of human labor. Merriam-Webster Online Dictionary Automation in IHC | Chapter 9 Chapter 9.1 History of IHC Automation for manual staining that before beginning the staining process the specimens should be circled with a delimiting pen that The irst automated devices for immunohistochemistry (IHC) serves as a guide when wiping away excess liquid to prevent appeared in mid-late 1980s and were based on differing the specimen from being accidentally wiped off; that to avoid technologies (1-3). One device used totally enclosed slide evaporation of solution, the slides should be laid lat with the chambers with a computer-controlled microluid transport specimen facing upward in a humidity chamber; that during system and reagents placed on a carousel (3), and another buffer wash excess liquid should be wiped off using an absor- was based on capillary action between two slides with tissues bent tissue and consequently that when processing a large facing each other (1). A key driver for implementation of au- number of specimens, only 3 to 5 slides should be wiped at tomation was to avoid the labor-intensive and thus expensive one time before applying the appropriate solution.” manual staining. It has been estimated that during the manual staining procedure each slide must be manipulated over 100 times, and, depending on the skill set, one technician can The quote also serves as an illustrative background to highlight some of the key advantages of IHC automation: manually handle up to 50 slides per day (4). The irst automat- „„ Reduced hands-on time that may compensate for increa- ed device capable of both IHC and in situ hybridization (ISH) sed staining volume, frees up skilled resources to perform was described in 1990 (5). The staining was performed in a other essential tasks such as cutting and embedding or capillary gap-based reaction chamber with the slides placed slide review, and may compensate for increased dificulties specimen-side down. in recruiting skilled staff as seen in many regions of the world. „„ Standardization, which ensures that staining is repeatedly Automation of IHC quickly caught on – one might compare that performed in the same way and does not depend on development with what we currently see with the entrance of differences in personal skills. Furthermore, standardization Next Generation Sequencing into cancer diagnostics – and combined with ease-of-use facilitates use of multiple techni- already in 1995, several IHC instruments were commercially cians at multiple shifts, thereby increasing laboratory lexibility. available. A total of ive systems were compared by Le Neel „„ High reproducibility by ensuring that any given step of the et al. (6) who concluded that automation is possible in routine staining procedure is continuously performed as speciied laboratory setting, and that the different systems served different laboratory needs. These stainers were all semi-automated, with deparafinization and antigen retrieval performed off in- by any given protocol „„ Optimized use of reagents „„ Improved error control via process control (e.g. control struments, and all had limited user interface. of slides to be stained and correct selection of reagents Some instrument generations later, today’s most advanced ting of correct liquid level or process temperature), plus IHC staining instruments are now fully automated, handling all alarm notiication if action is needed. via barcode tracking) and process monitoring (e.g. repor- steps from baking to counterstaining. Some instruments can perform ISH, the graphical user interface is designed for user-friendliness, and they can be fully integrated with the hospi- Chapter 9.3 Staining Technologies tals’ laboratory information systems. There are currently at least four different IHC staining principles being used on commercially available instrument systems. Chapter 9.2 Key Advantages of IHC Automation Open Individual Slide System Today, automated IHC staining is commonplace. However in The slides are arranged horizontally with reagents being ap- the 1990s it was a major step forward with manual IHC being plied using a dispenser that typically directs reagents to one or highly complex as described by Moreau in 1998 (7): “outline more zones on the slides depending on the location and size 111 Chapter 9 | Automation in IHC of the tissue. These types of systems are the closest to mimic the entire slide. The capillary gap system exists in a vertical manual staining. version where reagents are applied by placing the slides in a reagent jar and reagents are drawn out by blotting, with washing being performed by repeated reagent draw up and out. Capillary gap staining is also used in a horizontal version using a replaceable covertile that can be moved back and forth by the instrument and in this way create a pseudo-capillary gap. Slide Reagent Tissue 50 µm gap Slide Figure 9.1 Open individual slide system. Reagents are applied on top of the tissue by an automated dispenser system. The principle has been and is used on a range of slide stainers, including Dako Autostainer Link 48 and intelliPATH from Biocare Medical. Reagent Slide (with or without tissue) Tissue Liquid Overlay Technology With this technology, an inert luid is deposited over the entire slide. Reagents are either overlaid or deposited into the overlay luid. This technology allows reagent mixing on the slide as well as reactions at elevated temperature as the overlay luid restricts evaporation. Completed slides must be cleaned off the overlay luid before coverslipping. Figure 9.3 Vertical capillary gap staining. A deinite and very small spacing between the two surfaces creates capillary forces to draw and/or keep liquid between the two planar units. This principle has been and is being used on a range of stainers including the Dako TechMate instruments (vertical orientation of slides) and Bond from Leica Biosystems (horizontal orientation of slide using an overlay covertile). Reagent mixing by air Inert fluid Dynamic Gap Staining The dynamic gap staining technology uses capillary forces to secure homogeneous spreading of reagent throughout the Slide Reagent Tissue staining area during reagent application. During reagent incubation, dynamic movement of the coverlid ensures continuous Figure 9.2 Liquid overlay technology. Reagents are either overlaid with an inert overlay luid or dispensed into/under the overlay luid. This principle is currently being used by Ventana Medical BenchMark systems. movement of reagents across the staining area. In this way the dynamic gap staining technology provides homogeneous reaction conditions across the entire staining area throughout the full incubation time. In addition, the technology decreases Capillary Gap Staining incubation time and a tight humidity control of the system facil- The capillary gap technology utilizes capillary forces to draw itates use of elevated temperature to further decrease incuba- and/or keep liquid between two planar units that may be either tion time. Moreover, no slide cleaning is required post staining. two microscope slides with tissue facing each other, or a slide 112 and a cover plate. This requires a deinite spacing between the An intermediate technology between capillary and dynamic two units to ensure equal and correct capillary forces across gap staining is the wave-based system from Celerus where Automation in IHC | Chapter 9 Dispenser Movable Coverlid Tissue Slide Reagent is applied Coverlid moves forward and opens Coverlid slides back and closes Dynamic gap cycle completed Figure 9.4 Dynamic gap staining. The dynamic gap uses two motions. The lid is moved back and forth along the glass slide, and the upper end of the glass lid is moved up and down. When the glass lid and glass slide ends are aligned, the dynamic gap is created. Each cycle takes 16 seconds.The dynamic gap principle is implemented in the IHC staining module of the Dako Omnis staining instrument. paired slides are repeatedly being opened and closed during There is obviously no strict line between open and closed sys- reagent incubation. tems. Fully open systems are particularly relevant in research settings to investigate the expression of the growing number of Open vs. Closed Systems new biomarkers for biomedical research. In contrast, there is a Automated IHC staining systems are commonly referred to strong tendency to move towards more closed systems in clin- as being 'open' or 'closed'. In open IHC staining systems, the ical settings where standardized and high-quality performance users may select any reagent (incl. target retrieval buffer, an- for a deined number of antibodies is required. tibody, and visualization system) and staining protocol (incl. temperature, incubation time, wash time) that they prefer and thus fully design the staining according to speciic needs. A Table 9.1 The advantages and disadvantages of open and closed systems. primary advantage of open systems is that they offer a very high degree of lexibility. Any lab, pathologist or scientist may Pros design an assay with any antibody, visualization system or protocol to fulill their speciic needs. However, it is critical to note Open systems Closed systems Free choice of visualization system High level of standardization through use of RTU reagents High antibody flexibility High level of staining consistency; many available RTU antibodies Research application friendly Reduced hands-on time Increased need for internal protocol optimization Reduced staining protocol options Require highly skilled technicians or increased risk of manual errors Antibody restrictions; some may not work optimally that in order to ensure consistent and high staining quality, such systems require highly skilled technicians. In closed systems, the visualization reagents and protocols are typically locked (often termed Ready-To-Use (RTU)). In addition, other protocol steps, typically antigen retrieval, may also be locked or restricted. The primary antibody and antibody incubation time may also be available in RTU format, but also lexibility and use of other antibodies is allowed, as it is not realistic that one vendor can supply all antibodies required Cons Reduced staining consistency in any given clinical laboratory. Closed systems offer a much higher degree of standardization and lexibility in use of staff with varying skill sets. 113 Chapter 9 | Automation in IHC Sequential vs. Parallel Staining Processes Batch vs. Continuous Loading For most automated systems, the staining is fully sequential in Previously, all automated IHC stainers were 'batch stainers'. the sense that any given slide stays at the same slide position Here, a large number of slides are loaded and off-loaded at in the instrument and all staining processes, in principle from the same time with no manual intervention, and typically with baking to coverslipping, are carried out at that position. Such a a pre-set completion time deined by the instrument according system puts huge demands on the capability of each staining to the number of slides and complexity of the selected staining unit, as it has to perform very different tasks. The advantage is protocols. The primary advantage is that a large number of that once the slide is placed in the instrument, all subsequent slides can be completed in one run within a deined time; thus staining actions are “hands off”. enabling staff to perform other laboratory tasks during the ixed run-time. Another advantage is that these instruments typically An alternative is found in 'parallel processing' where one or more offer signiicant capacity for overnight runs to ‘catch up’ on of the staining protocol steps are carried out on separate instru- days with high slide volume. ments or instrument modules. The very basic example is baking done in a separate oven or in a seperate module onboard the In recent years, automated stainers with options for continuous instrument. Another example is a separate unit for deparafin- loading and unloading have appeared. In this context the term ization, antigen retrieval and rehydration. A newly introduced 'continuous' covers a spectrum of options for independently load- staining system beneits from having a number of modules, each ing smaller batches; typically 5-10 slides that will be completed at performing one or more of the process steps. Overall, the ben- different times. The completion time will depend on loading time, eits of parallel processing are two-fold: 1) Each module can be protocol complexity and potentially also priority. In the extreme optimized to perform the speciic process step(s) and 2) It allows case, single slides are loaded and unloaded independently. The a higher degree of lexibility for initiating staining of a new set of primary advantages of continuous loading include reduced time slides. Furthermore, by having specialized units, new ways of to irst result, the possibilities to continuously prioritize patient cas- handling process steps are facilitated. One such example is the es, and increased lexibility in the lab worklow. Potential draw- pre-treatment unit that combines antigen retrieval and rehydration backs include increased hands-on time per slide or patient case, into one single step. The primary drawback of parallel processing and less overall eficient use of the instrument. – when separated on different instruments – is the addition of an extra manual step compared to full automation. Some continuous IHC stainers depend on “instrument free time”, so introduction of new slides – and possible new reagents required for new protocols – does not interfere with planned protocol steps for slides already in process on the instrument. One option to circumvent this limitation is to include designated slide loading and unloading stations, which creates full lexibility for slide loading – as long as there are open positions in the loading station. It is still possible to apply priority to speciic patient cases and the instrument will continuously identify the best solution for effective staining according to the priorities selected. Likewise, there is increased lexibility in unloading, either removal as soon as a small batch has been completed to facilitate rapid assessment, or removal in larger batches Figure 9.5 Inside Dako Omnis, a robotic arm moves the slide rack from one module to the next allowing for staining protocol steps to be carried out in separate instrument modules. 114 to reduce overall hands-on time. A system with loading and unloading stations requires some level of parallel processing capability of the instrument. Automation in IHC | Chapter 9 for 4 out of 12 antibodies compared to standard baking for 1 hour at 60 °C. Reagents and Staining Protocols First of all, the reagents must be of high quality with high speciicity and suficient sensitivity to give correct staining. Also, the staining protocols must be optimized to ensure that high-qualLoading station ity staining results are obtained consistently and reproducibly, see details in Chapter 5. In this context, it is important to note that staining protocols must be optimized for a speciic set of reagents together with a speciic instrument type. The same protocol might not be optimal for two different instrument types, even though they may be using the same staining principle. Moreover, we have seen examples where speciic antibodies perform very well on one instrument system, but less optimally to very poorly on another when different staining principles apply. One factor may be staining temperature, but other factors may be of importance as well. Unloading station The Instrument From the above it is obvious that automation by itself does not Figure 9.6 Separate loading and unloading stations enable true continuous loading of slides. guarantee high staining quality, many other factors must be in place as well. However, it is critical that the instrument is able to perform each protocol step in a precise and reproducible manner. Important elements include: Chapter 9.4 Requirements for Effective Automated Staining „„ Correct reagent application: there must be suficient rea- gent to cover the tissue, both to ensure that the whole tissue is indeed stained and to avoid drying out which will Today’s IHC staining instruments are generally of high standard and capable of delivering high-quality staining results. Nonetheless, there are a number of requirements that must be fulilled to realize the full potential of automated IHC staining. give detrimental artifacts. Application of too much reagent will negatively impact the number of tests per reagent vial. „„ Correct incubation times: the instrument must be able to control the incubation time precisely; too short a time may give insuficient staining, while too long a time may give High Quality Tissue An absolute requirement for good staining quality is that staining that is too intense, impacting readability and increasing the risk of background staining. the pre-analytical process has been correctly performed, „„ Temperature control: this is a key parameter, particularly including short ischemic time, controlled transport to the for antigen retrieval, and to some degree also for reagent pathology lab and, not least, that the tissue has been ixed incubation. The extent of temperature control on an instru- correctly. The importance of ixation on the staining quality ment has direct impact on staining protocol optimization. has been documented in several reports (8, 9, and references therein). Furthermore, correct baking is also important. An important instrument feature with major inluence on instru- Williams et al. (10) reported that baking for 30 min on a 70 ment eficiency is its ‘scheduler’. A scheduler is the software °C hot plate had a negative impact on the staining quality that in detail organizes the sequence of each individual step 115 Chapter 9 | Automation in IHC Figure 9.7 Correct reagent application: there must be suficient material to cover the tissue to ensure that the whole tissue is indeed stained. Figure 9.8 The scheduler software organizes the sequence of all individual steps an instrument must perform from start to end of the staining process for the given load of slides. For some instruments, the scheduler also calculates when slides will be completed. that an instrument must perform for each slide from start to end the risk of introducing false negative or false positive results. of the staining process. The more complex the staining proto- As noted elsewhere (Chapters 1, 4, 5 and 14), it is vital that col is, the greater the diversity of protocols in the same run, revalidation is performed when any changes are made to and the more slides to be stained simultaneously, the great- reagents or protocols. er are the demands placed on the scheduler. The eficiency and dynamics of a scheduler have a direct impact on the to- Last, but not least, the instrument must be maintained ac- tal staining time; both for batch staining and when slides are cording to the manufacturer’s instructions, both with respect loaded continuously. In the latter case, there are two options: to daily/weekly/monthly internal maintenance tasks and the either the staining sequence for the new slides is added to the regular service intervals. In this regard, it is important always queue, or the scheduler creates a new plan if some of the new to be aware of potential irregularities, e.g. extensive or re- slides have high priority, which often prolongs the completion duced buffer use, deviation in quality of control slides and time of current slides. In order to give the scheduler as good a change in instrument noise. Following major repair, or if in- basis for an as effective action as possible, it is important that struments are moved, it is necessary to again establish with the incubation and wash times allow the maximum lexibility controls that the instrument is working properly. For example, that is consistent with good staining quality. it is not uncommon to identify incorrect instrument leveling as the root cause for incorrect staining results. Some vendors Some of today’s staining instruments are offered as part of offer service contracts to ensure timely response and minimal a ‘staining solution’ that consists of an instrument, the soft- downtime with ixed costs. ware, and a large portfolio of ready-to-use (RTU) reagents (antigen retrieval buffer, antibodies, visualization reagents) that comes with optimized staining protocols. This not only Chapter 9.5 Automation vs. Worklow gives high staining quality, but also provides short overall 116 staining time (turnaround time). Many laboratories make Worklow describes the method and sequence in which an modiications to staining protocols to obtain a staining pat- activity is performed. Today, there is a lot of focus on estab- tern according to in-house material (tissue and reagents), lishing eficient worklow in the pathology laboratory from internal needs or personal preferences. However, when do- sample receipt, through processing and analysis, to inal re- ing internal modiications it is important to consider potential porting and storage of slides and data. Creating an eficient impact on the overall staining time – and most importantly worklow is highly dependent on the situation in each labora- Automation in IHC | Chapter 9 tory. It will differ among laboratories depending on a number of factors including type of samples received, sample load, Chapter 9.6 Key Features for Complete Staining Solutions number of staff and staff competence, and not least variation in working hours. A critical element in an eficient worklow is When looking for a new IHC staining instrument, there is a selection of the type and level of automation that best supports large range of features to consider; some of which have been the worklow. Choice of instrumentation is thus one of several mentioned above, e.g. the ability to perform batch and/or important factors for an eficient laboratory. continuous staining. In this context, it should be noted that it is important to compare instruments in the context of the There are many examples of how instrumentation and/or laboratory in which they will be used. Also, it is useful to con- changed use of instrumentation have improved the worklow sider the ‘complete staining solution’ provided by the vendor, in pathology laboratories. Examples include the use of rapid including reagents, protocols, software, service. Each of the processing instruments that have the potential to signiicantly separate elements – not least how well they work together – reduce time to result, and the use of IHC staining instruments impacts the overall performance in terms of quality, eficiency in a continuous loading mode rather than a batch mode, fa- and cost. Below, a range of important solution features are cilitating both shorter time to result and a balanced workload. briely discussed. The level of importance of each feature Also, automation of special stains can free up laboratory re- may differ among laboratories; what is really important is that sources signiicantly. In short, automation has great potential to the combination of features has the best it to the needs of the improve the laboratory worklow provided it increases eficien- individual laboratory. cy in key parameters including: „„ reduction of time to analysis „„ reduction of hands-on time from data entry to slide storage „„ reduction of human errors „„ increase of information from a specimen – the actual High Quality Staining The most important feature for an IHC staining solution is that it consistently can deliver high-quality staining, with high sensitivity and speciicity for all relevant combinations of staining protocols and tissues. This outcome can only be achieved parameters being highly dependent on the focus of and via effective interaction between high-quality reagents and requirements to the laboratory staining protocols that work seamlessly with the instrument. Turnaround Time The turnaround time (TAT), the time from loading of slides to completion of staining, has great impact on the laboratory worklow, including when the pathologist receives the stained slides for evaluation. Consequently, a short TAT facilitates a rapid answer to the oncologist, and by that may have direct impact on when patient treatment can be initiated. The TAT will also have a signiicant impact on the total slide throughput. A small, but signiicant, detail is the scheduler’s ability to accurately predict when a slide or slide run is completed, to facilitate rapid post-processing of the slide(s) and timely preparation and loading of the next slide(s). Figure 9.9 An overview of all slides being processed on all instruments in the laboratory can improve the worklow by increased information, control and overview of the processes. The TAT of a speciic staining solution is inluenced by multiple parameters, including how many slides are loaded onto a given instrument. Depending on the instrument design and 117 Chapter 9 | Automation in IHC scheduler (see section 9.4), even a small change in the num- tors that are not isolated to the instrument alone, e.g. remote ber of slides per batch may signiicantly alter how eficient- access to instrument status, report generation, and need for ly the instrument performs. Likewise, changes in incubation cleaning of stained slides or instrument accessories. Another times are likely to impact the overall staining time, directly in important parameter to assess is the total reagent package, case of prolonged incubation, or indirectly by affecting the including handling of reagent vials and bulk luid bottles, as scheduler. Manufacturers of complete staining solutions will well as the ease of use of the software (user-friendliness), e.g. typically aim to optimize their instrument, software and rea- when setting up new staining protocols or requesting new gents to give a low TAT; thus any in-house laboratory modii- slides for staining. cation of protocols may increase the TAT. Capacity With the constantly increasing number of slides in the routine diagnostic laboratory, without concomitant increase in staff, the capacity of the staining solution is a major parameter. The capacity can be measured in several ways. It is important that capacity is assessed according to laboratory needs such as the length of the workday, whether the solution is used throughout the 24 hours of the day, average daily workload, and the slide number peaks and distribution. Special consideration should be made on need for overnight capacity, as that may be a very important option for management of increasing slide volume and heavily luc- Figure 9.10 More onboard capacity for reagents and bulk luid bottles can reduce hands-on time due to little or no reagent handling during the day. tuating slide volumes between days. Hands-On Time Functionality With increasing workload, laboratories are under pressure to Today, all IHC staining instruments can perform the complete reduce the hands-on time in order to process the daily volume staining process from peroxidase blocking to counterstain- of slides. Factors such as duration of daily start-up procedures, ing. Many instruments include deparafinization and antigen time spent on slide loading and un-loading, as well as rea- retrieval whereas other approaches have separated the two gent, bulk luid and waste handling times are important. Each processes. Onboard deparafinization and antigen retrieval of these factors may include several important components, reduce the number of times that slides need to be handled according to the speciic needs of the laboratory; one example and the risk of manual errors. A split of these functionalities being the reagent capacity of the instrument. The lower the allows for parallel processing so that deparafinization/an- reagent capacity, the greater the need to change reagents be- tigen retrieval can be performed in parallel on a dedicated tween runs, whereas a high reagent capacity may require little instrument and thus be ready for immediate staining when or no reagent handling during the day. An additional important the IHC staining instrument has completed its run. Eficient instrument feature is related to maintenance, including daily, use of this approach has a positive impact on total eficien- weekly and long term maintenance. cy and throughput. Furthermore, some instruments have the capability to bake slides, which reduces the number of slide 118 Regarding hands-on time, it is truly a question of assess- handling steps. This advantage must be measured against ing not only the instrument itself but the complete staining the quality of baking, the eficiency/throughput of the instru- solution, including how instrument, software and reagents ment, and not least which method has the best it in overall interact. Important parameters include a wide variety of fac- laboratory worklow. Automation in IHC | Chapter 9 The inal functionality to be mentioned is the possible appli- Chapter 9.7 Next Steps in Automation cation of the instrument to luorescence in situ hybridization (FISH), or a bright ield microscope version, such as chromo- Looking 5-10 years ahead, new technologies and new types genic in situ hybridization (CISH). As hybridization assays are of biomarkers will be commonplace in routine pathology signiicantly different from IHC assays, dual capability puts laboratories. Examples of new types of biomarkers include increased demands on instruments and integrated stain- mRNA, miRNA and other non-coding RNAs, DNA methyla- ing solutions. It is therefore important to carefully evaluate tion variants, and post-translationally modiied proteins. With the quality of both IHC and ISH stainings and whether the regard to new technologies, digital imaging is certain to oc- instrument can effectively run both types of assays in paral- cupy a central position, fully integrated with the IHC stain- lel without signiicantly impacting TAT and capacity. Further, ing solutions and other slide-based staining solutions (see the overall it of dual IHC and ISH stainers into the laboratory Chapter 7). New, non-slide-based tests, e.g. next generation worklow must also be evaluated. Important elements in this sequencing, are in the process of entering routine diagnostic context are the distribution and volume of IHC and ISH slides application providing new important information for cancer and the need for speciic report times (e.g. can either assay diagnostics, but at the cost of morphological information. types be run overnight, or can ISH slides be batched and run Thus, IHC and other slide-based staining techniques will in on set days only?). 10 years time continue to play an important role in cancer diagnostics, however, in an environment that is quite different Laboratory Accreditation Support from today’s pathology laboratory. More and more clinical laboratories are subject to accreditation requirements, and in some countries laboratories must be accredited to perform clinical testing. In the USA, accreditation is also a requirement for reimbursement through the Medicare system. Accreditation is a mechanism of value to verify that laboratories have an appropriate quality management system, can properly perform speciic assays, and are able to properly document test results. In this context automated IHC stainers, that incorporate complete staining solutions, represent an approach to monitoring important assay parameters, which at the same time can include a reporting format that suits the laboratory needs according to its quality management system. Required accreditation data vary according to the country in which the laboratory is situated, as well as the quality management system of the individual laboratory. Requirements may be quite detailed, down to the level of speciic information of the IHC staining on a per slide basis, or the speciication of which lot of wash buffer was used. Likewise, instrument maintenance and service must typically Figure 9.11 New technologies and new types of biomarkers will be commonplace in routine pathology laboratories in the next 5-10 years. This will put new demands on future instrumentation. The image shows RNA FISH demonstrating differential RNA expression within a cell population. Courtesy of Robert A. Arch, Agilent Laboratories. be carefully logged, often with an ‘on instrument’ record. 119 Chapter 9 | Automation in IHC The future staining solutions will need to be able to generate or units. Finally, it will be very important that the new staining much more information from less sample material, in a more ef- solutions will be able to work seamlessly with whole slide digital fective and faster manner without compromising the requirement scanners and laboratory/hospital information systems. for quality; that surely will increase over the next 10 years. „„ More information will be derived from new biomarkers and biomarker types, with growing use of dual- and multiplex assays with several antibodies being applied to the same slide as well as antibody and F/CISH combinations. 2. Tubbs RR, Bauer TW. Automation of immunohistology. Arch Pathol Lab Med 1989;113:653-7. ongoing, perhaps growing, demand for quantiication 3. MaWhinney WHB, Warford A, Rae MJL, Lauder I. Automated immunohistochemistry. J Clin Pathol 1990; 43:591–6. „„ There will be demand for accurate quantiication, with at an individual cell level within a heterogeneous tumor cell population. „„ A drive towards less invasive sampling methods will reduce the amount of sample material, at least for some cancer types, and the solutions must provide more information from less sample material. „„ Complete staining solutions will support new multiplex and quantitative assays and become signiicantly more effective with a reduction in hands-on time per slide. Eficiency may not be measured on a per slide basis but rather on a per patient case basis. Increased eficiency will be implemented via an increase in the functionality of complete staining solutions as well as of the whole laboratory We will probably see improved alignment between tissue cutting and slide loading, as well as integrated slide scanning and distribution of Whole Slide Images (Chapter 7). „„ There will be a continued drive towards reduced time to diagnosis that will be translated to the staining part of the worklow, as reduced time to result or TAT. This goal may be affected by adoption of rapid staining protocols, employing new visualization chemistry that can also handle several biomarkers simultaneously. Again, it is important to stress the concept of a complete staining solution, incorporating instrumentation, reagents and protocols in an optimized system. In order to be able to live up to the new customer needs, it will be imperative that instruments, reagents and software are developed to work optimally together. For example the introduction of more rapid, more sensitive staining methods will not only require new visualization technology, but also instruments and software schedulers that support short incubation times for eficient use of the instrument modules 120 References 1. Stark E, Faltinat D, Von der Fecht R. An automated device for immunocytochemistry. J Immunol Methods 1988;107:89-92. 4. Haines DM, Chelack BJ. Technical considerations for developing enzyme immunohistochemical staining procedures on formalin-ixed parafin-embedded tissue for diagnostic pathology. J Vet Diagn Invest 1991; 3:101-12. 5. Takahashi T, Ishiguro K. Development of an automatic machine for in situ hybridization and immunohistochemistry. Analyt Biochem 1991; 196:390-402. 6. Le Neel T, Moreau A, Laboisse C, Truchaud A. Comparative evaluation of automated systems in immunohistochemistry. Clin Chim Acta 1998; 278:185-192. 7. Moreau A, Neel TE, Joubert M. Approach to automation in immunohistochemistry. J Clin Chim Acta 1998; 278:177-87. 8. Boenish T. Effect of heat-induced antigen retrieval following inconsistent formalin ixation. Appl Immun Mol Morphol 2005; 13:283-86. 9. Shi S-R, Liu C, Taylor CR. Standardization of Immunohistochemistry for formalin-ixed, parafin-embedded tissue sections based on the antigen-retrieval technique: From experiments to hypothesis. J Histochem Cytochem 2007; 55:105-9. 10. Williams JH, Mapham BL, Wright DH. Tissue preparation for immunocytochemistry. J Clin Pathol 1997; 50:422-8. 121 Part II: The Potentials and Pitfalls Chapter 10 Optimizing Laboratory Worklow David McDonald, BSc, MLT, LSSBB Work•low (n.) 1. The low or progress of work done by a company, industry, department, or person. 2. The rate at which such low or progress takes place. The American Heritage® Science Dictionary Optimizing Laboratory Worklow | Chapter 10 Chapter 10.1 Introduction to Worklow „„ Slide load interval – how frequently can slides be loaded The term “Worklow” has been deined in numerous ways and „„ Batch size – does the current process enable the maximum onto a stainer without compromising throughput consistency? scope of activity over time and across the evolution of sophisticated observation and statistical analysis techniques. Ulti- number of slides to be loaded with each load event? „„ Process complexity – to what extent do manual processes mately any deinition will have merit, based upon the ability such as documentation, written tracking of slides and stains, to demonstrate the information required to make an educated post stain labeling of slides, case assignment rules and and fact-based decision as to future direction. That direction details of protocol impede and delay the overall process? must be aligned with the deined criteria. Worklow must be designed around a series of individual solutions that respect and exploit the unique characteristics and Surgery and Sampling Patient Record Pathologist’s Interpretation needs of each operation. After all, no two labs are the same. Accessioning and Grossing Chapter 10.2 Specimen Tracking IHC Staining Embedding Within the scope of daily workload processing, one of the sin- Potential Manual Processes Antigen Retrieval Sectioning gle most pervasive forms of delay incorporated into an overall process is the need to identify and track individual specimens/ blocks/slides as they advance through each subsequent pro- Microscopy Patient ID on slide H&E Staining Tissue Sample on Slide Deparaffinization cess step, i.e. initial accessioning, embedding, microtomy, labeling of slides, drying, staining, cover slipping and delivery to the pathologists. Many of these steps are still manually performed in most laboratories today, greatly increasing the risk for mislabeling errors at one or more of these checkpoints, thereby leading Figure 10.1 No two labs are the same. Manual handing in one or more of the illustrated steps is frequent in most labs. to potential misidentiication of patient samples. In the US, both the College of American Pathologists (CAP) (1) Worklow is based upon the simple principle of deriving a com- and The Joint Commission (JCAHO) (2) have issued directives plete and thorough understanding of the entire current pro- that address the importance of a specimen identiication sys- cess, the multiplicity of steps inherent within that process and tem throughout the clinical laboratory. Across the world, hospi- all of the factors which impact the current achievable outcome. tal labs have implemented various Quality and Safety protocols An important rule of worklow resides in the belief that only by as part of their accreditation systems. However, the direct con- having a complete understanding of the current processes can nection between patient identiication in the hospitals’ labora- an alternative, more effective process be implemented. The tory information system and the IHC staining process steps are understanding of an IHC pathology ‘facility speciic’ worklow not yet fully implemented in all labs. resides in the assessment of the following parameters: „„ Slide availability – how do current processes in the laboratory Historically, sample tracking is accomplished through the affect at what point in the shift slides become accessible for use of manual documentation logs. In recent years, vendors staining; and subsequently how do they impact at what time slid- have developed speciic software and hardware solutions to es become available for delivery to the pathologist for review? address and attempt to eliminate the requirement of manual „„ Stainer load capacity – how many slides can be loaded without compromising stainer throughput capacity and consistency? documentation logs within the scope of Anatomic Pathology. When investigating such a solution for a speciic laboratory, it 123 Chapter 10 | Optimizing Laboratory Worklow is important to consider the following criteria: „„ Overall system lexibility – ease of connectivity „„ Individual process step points of contact – where in the Chapter 10.3 Specimen Collection and Identiication process is it important and/or desired to capture specimen/ Specimens are collected and delivered to the anatomic block/slide handling and transfer advance pathology lab through a variety of mechanisms and can „„ Scope of system implementation – single site versus multiple site take the form of wet specimens, blocks and slides. These mechanisms will vary in complexity based upon the catch- The tracking solution usually consists of software that links the ment area of the speciic testing facility and may include facility’s LIS to a server-software that collects and feeds label in- in-house, multi-site afiliated and non-afiliated referral spec- formation to slide label printers and assay information to instru- imens. Requisitions may be of either a standardized format, ments. This connection enables tracking of multiple interaction or a combination of multiple form and requisition types, add- points within the process, such as accessioning, grossing, tissue ing a further layer of complexity. processing, embedding, microtomy, staining, slide distribution and archiving; thus capturing the entire scope of daily operations. LIS connectivity among sites offers a distinct advantage in The tracking capability is enabled through a speciic barcode the process of specimen collection and identiication through attributed to specimens, blocks and slides. It is a huge beneit site-speciic preixes in the specimen identiication number. for an optimized tracking system to be deployed, such that the The lack of integrated LIS connectivity among testing referral total number of access points into the system is virtually unlimit- sites, however, need not be a barrier to connectivity for staining ed. Only in this manner can clerical staff, technical staff and pa- purposes. Through wide area networks (WAN), some systems thologists have the ability to perform search functions on speciic are able to connect separate multiple sites to enable remote specimens, blocks and slides. Tracking systems can ultimately ordering, such that incoming slide orders are able to populate eliminate the need for manual documentation, streamline the pro- instrument management software. cess low and provide a means of error-prooing the entire system. Obstacles/Challenges Manual processes Automated processes Request form or phone Electronic request from LIS with instrument integration Slide labeled (with pencil or label) Slide label printed and scanned There is a perceived limitation that only through LIS interfaces can multiple sites achieve the level of connectivity required to transmit orders to an individual testing site facility. There may also be the misconception that all sites are required to use the Slide information entered on an instrument same LIS and version software. Improvements/Solutions In the absence of LIS interface connectivity, WAN solution software can enable multiple sites to achieve the level of commu- Label printed and applied to slide nication required in order to permit direct transfer of slide order requests into a single test site instrument software. Staining performed Staining performed Data is re-entered for storage purposes Slide scanned for electronic storage Label printed and applied to slide Chapter 10.4 Specimen Accessioning Reducing the complexity of specimen accessioning resides in the ability to translate a physical specimen order request into a visible Figure 10.2 Example of worklow challenges involving slide labeling and data entry. LIS integration with IHC instrumentation can potentially save many manual labeling steps. 124 slide protocol that is recognized by the staining instrumentation in use. By utilizing a suitable interface connectivity between the Optimizing Laboratory Worklow | Chapter 10 Figure 10.3 A busy lab can beneit from systems that are able to connect separate multiple sites to enable remote ordering. facility LIS and the instrument server software, users can place Improvements/Solutions orders into the LIS that can then be assigned to individual instru- Through the advent of direct download capabilities into the ment staining runs directly from workstations in the laboratory. instrument software, with accompanying order request information and notations, pending slide runs can be planned and Orders are typically placed within the LIS directly by the executed much sooner in the daily routine. The result is that pathologist, or additionally can be submitted via separate slides become available for review sooner into the shift, and requisitions, sometimes hand written in a separate order log overall turnaround time is reduced. book. One speciic advantage of interface connectivity between the facility LIS and the instrument server software is elimination of the delays associated with multiple layers of Chapter 10.5 Grossing specimen accessioning, at the same time reducing opportunities for error. At the point of grossing, as in the many other aspects of specimen processing, delays and errors may occur as a result of Obstacles/Challenges inherent complexity of multiple sequential steps within that par- Translation of order requests into the instrument stain protocol ticular process. Ideally, the generation of specimen cassettes can also be a source for delay in the daily initiation of stain- is optimized when each cassette is printed/etched automati- ing runs. Historically, Histology departments have developed cally as a direct result of data entry at the requisitioning step. time speciic cut-offs, after which staining runs are initiated with the intent of capturing all of the pending orders up to Most tracking software enables the generation of cassettes with that point in the day. Outstanding slides are conirmed by case and specimen speciic identiication within a unique bar- pulling pending logs and/or checking separate request logs, code assignment, as a parallel activity directly associated with and cutting/retrieving the associated slides, all of which may requisition and ordering. This capability eliminates the need to require up front manual activity. switch from specimen requisitioning to either a separate soft- 125 Chapter 10 | Optimizing Laboratory Worklow ware application to generate speciic cassettes, or a manual daily activity. One manner by which optimization of worklow keying in of each individual cassette at the cassette printer. The may be addressed is covered by the worklow term “trigger optimization of specimen entry, transfer to grossing, and then point transfer” – namely the rapid shift from one activity to the on to histology, augments and maximizes the ability for effective next based upon accumulation of inventory (work), such as use of load capacity and short tissue processing cycles. embedded blocks or cut slides. In contrast, standard batch event processing is a function of arbitrary individual shift start/ Obstacles/Challenges stop times, established hours of operation and current process Typically, specimen requisitioning involves the batch registra- capability limitations. Common scenarios can be character- tion of multiple specimens, either delivered to the laboratory, ized by the following practices: or collected during a sweep of associated specimen generating locations within the facility (operating rooms, day surgery Tissue Processing sites, specialty clinics). Once these specimens have been „„ Specimens are grossed as they come in throughout the registered, the resulting batches transition from the gross lab day and tissue processors are all loaded at the same time into properly labeled cassettes, matched with the specimen. for an overnight cycle, up to a speciied ‘cut off’ time This crucial step requires planning and dedicated lead time in advance of grossing. Additionally, specimen registration is not always localized adjacent to the grossing stations, a situation that can impede the progress of specimens to the stage of cutting, as well as limit accessibility for the request and generation of additional cassettes as required. Batch Event Obstacles „„ Delays in loading processor baskets into appropriate short cycle processors for certain tissues (rapid biopsies) „„ All cassettes become available at the same time the fol- lowing day, the large batch of cassettes for sorting, embedding and microtomy creating a bottleneck effect Improvements/Solutions The design of a process whereby cassettes are printed in as- „„ Loading subsets of appropriate tissue cassettes into short cycle processors, as they become available, removes a sociation with, and as a direct result of, specimen registration portion of the workload from the remaining process batch expedites the progression of specimens through grossing as and improves throughput of the overall process single integrated low, such that individual specimens can be accessed in real time, rather than as part of a batch transfer event between sequential steps in the process. Embedding/Microtomy „„ Typically multiple staff members embed all cassettes from the overnight processing cycles and once all embedding Additional direct beneit results from the placement, wherever pos- is completed, then microtomy begins sible, of specimen registration directly adjacent to grossing, and „„ Slides are hand-written or etched at the ‘front end’ of mi- the provision of at least one supplemental cassette printer within crotomy, in advance of cutting, and re-labeled at the back the immediate grossing area. This arrangement allows for addition- end, post routine or special staining al cassette requests without a separate request to the registration clerk, or interruption of the registration and grossing processes. Batch Event Obstacles „„ By waiting until embedding is complete, microtomy, which is an inherently long and labor intensive process, becomes Chapter 10.6 Tissue Processing and Microtomy the next rate limiting step, impacting the time available to load and stain slides. By identifying and setting an avail- 126 The optimization of worklow within tissue processing and mi- able block count, which triggers some staff members to crotomy is a factor of batch size and sequential versus parallel begin moving from embedding to microtomy, the overall activity. Batch event or “all-or-nothing” processing is a non-op- process is shortened and stain throughput, case assembly timal activity that often becomes the rate limiting step of all and distribution are optimized. Optimizing Laboratory Worklow | Chapter 10 „„ ‘Hand writing’ slides at the front end and labeling at the back end means that slides are handled and identiied twice. With tracking software capabilities, labeled slides can be generated in real time prior to microtomy, thereby eliminating one of these slide labeling interactions, reducing time, while simultaneously removing an opportunity for error. Manual transfer processes are subject to the rate-limiting factors of: „„ The number of staff manning the workstation „„ Other duties which intermittently take staff away from the workstation „„ Slide load intervals, which can be inconsistent and dis- continuous „„ Time required for ‘off-line baking’ and transfer to the stai- ning step „„ Slide rack capacity and the inherent time required to cut suficient slides to ill a rack „„ Sequence timing limitations in the X-Y-Z motion of the robotic rack transfer arm, affecting overall throughput time Integrated Staining Platforms Integrated staining platforms remove the requirement of having trained laboratory staff immediately available at the appropriate times for transfer between subsequent steps. Eficiency remains dependent upon; „„ The ability to take advantage of all online features, while Figure 10.4 Exchanging hand-written slides with barcoded labels and software with tracking capabilities reduce time and error opportunity. ensuring that they perform as intended (for example, does the oven actually perform an adequate baking process) „„ Making optimal use of inherent process-balanced design features, such that bottlenecks are avoided on the unit Chapter 10.7 Routine Staining and Coverslipping in any of the process steps (does baking or staining take The staining of routine slides generally is accomplished either move through the remaining processes and dip tanks?) a disproportionate time relative to how slide racks can by a manual transfer staining process, or by use of an inte- „„ Manual oversight and intervention may still be required grated staining platform. With integrated staining methods, the (how many coverslips can be loaded at once to alleviate processes of slide baking, staining and coverslipping are performed on a ‘single load’ instrument platform. the constant need for reloading?) „„ Performing the required frequencies of onboard reagent iltering, swap-out exchange and/or position transfer Manual Transfer Processes (replacing the irst xylene or alcohol dip tank and moving Manual transfer processes rely on the transfer of slide racks the others in sequence to the next position as required). across each of the individual and separate steps of baking in a stand-alone oven, deparafinization, staining and dehydration (often on an automated stainer), followed the manual load and Chapter 10.8 Slide/Block Reconciliation unload activity associated with a stand-alone coverslipping device. In each of these scenarios, worklow optimization be- Case assembly for veriication and distribution is another time comes a function of reducing the number of steps, load inter- consuming process and a source of imparted delay into overall vals and batch size. turnaround time. It is therefore important to design a functional system that facilitates case assembly with a minimum of manual activity and time commitment. 127 Chapter 10 | Optimizing Laboratory Worklow Obstacles/Challenges The organization and structure of the slide sorting and case Chapter 10.9 Laboratory Asset Tracking and Worklow Management assembly process presents a signiicant obstacle to worklow. Typically sorting, assembly and distribution is an end Within any anatomical pathology environment, the ultimate of shift batch event, and not a continuous process. Also the goal should be to achieve maximum access to all the relevant number of people brought to the task requires a large labora- information in the course of managing daily worklow. tory footprint to accommodate multiple racks of ‘routine’ H&E slides for sorting to multiple case folders, accompanied by an additional degree of disassembly or reorganization of cut The ultimate management of worklow is a function of three speciic factors; slides in order to carry out further staining processes (such „„ Tracking capabilities – the ability to trace specimens, blocks as IHC and special stains, grouped by antibody, pre-treat- and slides throughout the entire process, from requisitioning ment protocol or stain group). to archival. Tracking should ideally also provide for attribution of process-speciic activities to individual staff Improvements/Solutions within the scope of the entire process. Introducing more frequent (ideally continuous) checking of „„ Connectivity – the ability to place orders locally as well as blocks and sorting of routine slides minimizes occurrence of remotely, and to generate a barcode that identiies and “all-or-nothing” batch events during the shift, or at the end of marries specimen, cassette and slide. A slide barcode the shift, and expedites the low of completed slides to the pa- should be speciic down to the individual stain and proto- thologist for review. In addition, reducing disassembly of sets col level. There should be an unlimited number of access of cut slides for IHC and Special Stains, by advance planning entry points to enable the monitoring of workload sta- and organization not only facilitates staining by case, but also tus. Connectivity should also remain lexible in order to minimizes hands-on manipulation on the part of the technolo- accommodate third party ancillary devices as required, gist. This process can be accomplished by choosing staining (barcode label printers, for example). equipment with high slide and onboard reagent capacity, plus „„ Reporting – Reports should be easily conigurable and the ability to combine slides with differing protocol require- custom tailored to the needs of the facility for capture and ments side-by-side within the same run. organization of information as required Request for special stains LIS Server Request for IHC stains Request for IHC/ISH stains Middleware (ULISA) Figure 10.5 Integration from a Laboratory Information System (LIS) to the stainer instruments can be accomplished through a single point of connection using an integration module, e.g. the Universal LIS Agent (ULISA). The ULISA manages communication with a LIS server and provides the various instruments with the data received. 128 Optimizing Laboratory Worklow | Chapter 10 „„ Facility for full installation on any user PC, wherever it Obstacles/Challenges Historical challenges include the ability to interface fully with the makes sense to track workload freeing technologists from facility LIS provider, proprietary limitations with respect to all asso- instrument speciic workstations ciated ancillary devices and standard versus conigurable reporting structures, which may or may not meet the needs of the facility. Improvements/Solutions „„ Provide for workload monitoring at multiple sites in an integrated delivery network „„ Facilitate the ability to generate orders remotely and stain locally Ideally, the most lexible solution will enable a facility to achieve all of the following; „„ Provide an independent test order module for pathologists „„ Facilitate a single barcode label throughout the entire process „„ Enable the coniguration of facility speciic report format „„ Enable the generation of barcode labels at the microtome station, to cut and label slides in real time and as single Chapter 10.10 One Worklow Does Not Fit All piece low „„ Enable deployment at the reagent fridge for log in and management of reagent inventory Doctor Grossing When it comes to addressing worklow, it is important to acknowledge that each pathology laboratory, while performing Processing Embedding Surgery Sectioning Staining Specimen Receiving and Data Entry Slide Viewing LIS connectivity Hospital Pathologist Figure 10.6 Some laboratories are disconnected; the process steps are all critical to the inal result, yet they are not integrated with one another. 129 Chapter 10 | Optimizing Laboratory Worklow similar activities, differs in speciic workload processes, with respect to absolute slide volumes, stain menu diversity, hours „„ Have a second staff member review and verify what was sent to the laboratory of operation, stafing levels, catchment area, and interface with „„ Have only the tissue processing cassettes for the case one or more institutions for which service is provided. The most being grossed at the prosection station during that cases effective strategy to approach worklow evaluation will incorpo- dissection. Immediately remove from the work bench and rate the following capabilities: „„ Assessing and cataloguing those functions that perform well versus those that need improvement „„ Fully understanding the current worklow, identifying speciic obstacles to success and deriving criteria by which alternative process capabilities may be assessed „„ Deriving the proper key performance indicators to measure process activity and capability „„ Facilitating statistical analysis to identify historical growth discard all cassettes that are not utilized for the processing of the case. „„ When grossing, only have the tissue cassettes for the current case on the cutting board at the time of dissection. All other tissue cassettes should be with their appropriate specimen and not on or near the prosection station. „„ When embedding, be certain to wipe off forceps after every use to insure that any residual tissue is removed thereby preventing carryover from one cassette to another rate, current achievable throughput against demand, ven- „„ Work in tandem with all those who come in contact dor-related capabilities with respect to instrument produc- with the specimen including the Operating Room Team, tivity, slide capacity, waste generation, and the ability to Transporters, Specimen Accessioning staff, Pathologists’ accommodate incremental slide volume over time „„ Methods for aligning input from all key stakeholders, from Assistants, Residents, Pathologists, Histologists, Secretaries, Transcriptionists, slide and block iling room staff, etc. bench Technologist, department Managers, Pathology Director, and Pathologists, and together with hospital administrators and clinical users as necessary Do not „„ Do not pre-label requisitions, specimen containers or tissue cassettes „„ Do not accession similar case types (tissue) in sequence Chapter 10.11 General Sample Labeling and Tracking (by James Happel, DLM (ASCP) HTL, from IHC Staining Meth- „„ When embedding, do not open more than one cassette at one time „„ When cutting on the microtome, never cut a second parafin block until the irst block has been cut to completion ods, 5th Edition, published by Dako.) „„ Do not leave ribbons on the water bath after picking up Do „„ Do not leave printed labels in the work area. Discard all sections on slides „„ Have clear and concise procedures based on CAP and JCAHO specimen identiication recommendations „„ Remove waste baskets from the specimen collection areas „„ Retain all trash for a minimum of three working days „„ Label only one case at a time „„ Keep all materials from a single case together (requisition, specimen containers, cassettes, etc.) „„ Have the staff responsible for placing the specimen in the transport vessel (i.e. specimens collected during a surgical procedure, cultures to be sent for microbiological analysis, tubes of blood, etc.) be the same individual who labels all materials generated for that case 130 unused labels as soon as it is identiied that they are not to be used. 131 Part II: The Potentials and Pitfalls Chapter 11 Companion Diagnostics Hans Christian Pedersen, MSc Jan Trøst Jørgensen, PhD, MSc Com•pan•ion di•ag•nos•tics (n.) A companion diagnostic device is an in vitro diagnostic (IVD) device that provides information that is essential for the safe and effective use of a corresponding therapeutic product. also known as Predictive Markers, Theranostics, or Advanced Personalized Diagnostics. www.fda.gov Companion Diagnostics | Chapter 11 Chapter 11.1 Introduction In the 1990s, the US-based biotechnology company, GenenPredicting the response to a pharmacological intervention is tech, ‘humanized’ the monoclonal antibody trastuzumab (Her- an optimal goal for any healthcare professional working with ceptin®, Roche/Genentech) and placed it under development pharmacotherapy, especially when it comes to serious and life- for treatment of women with metastatic breast cancer. Trastu- threatening diseases such as cancer. Early correct diagnosis zumab is a drug that targets the external domain of the HER2 and effective intervention are two elements of key importance receptor. Genentech very early on realized that only patients in the treatment of cancer. In the event of a wrong treatment with cancer cells that had a high expression of HER2 receptor decision the disease may become disseminated, with little or no protein responded to treatment with trastuzumab, pointing to chance of cure. Companion diagnostics (CDx) hold the promise the need for a companion diagnostic selecting such patients of improving predictability of outcome of a speciied pharmaco- for therapy. Data from the subsequent phase III study with tras- logical intervention, thereby constituting an important tool for the tuzumab in HER2-positive patients with breast cancer showed oncologist in relation to choice of treatment. In a related sphere that this was a wise decision (5). Furthermore, this decision there is an urgent need to improve the development pathway later led directly to the development of the CDx HercepTest™, of new drugs, with better predictability of the outcomes of R&D designed to detect HER2 in a semi-quantitative manner (4). in terms of approved and useful product. CDxs have already been shown to be useful tools in this respect, especially when A few years after the inalization of the phase III trial, a retrospec- it comes to clinical development (1). The way that many drugs tive statistical analysis of the data showed some rather remark- are being developed today is far from optimal and, in general, able results with respect to the design of this study. Based on this the productivity of the R&D departments in the pharmaceutical analysis, it was concluded that the use of the HER2 IHC assay as industry and biotech companies seems to be decreasing. Re- a pre-selection tool had played a key role. If this assay had not cently, it has been estimated that around USD 60 billions, of the been available it would have required enrolment of several thou- USD 85 billions spent on R&D globally every year, are wasted sand patients in order to reach a conclusive result, instead of the due to failures during the drug development process (2). 469 patients that made up the total phase III study population (5, 6). It has also been said that if the HER2 IHC assay had not existed, we would likely not have access to trastuzumab today, due to the Chapter 11.2 History of Companion Diagnostics fact that it would have been discarded during clinical development because of lack of eficacy in an unselected patient population (7). The idea of combining drugs and diagnostics is not new. When ® The development of trastuzumab showed for the irst time how cru- the selective estrogen receptor modulator tamoxifen (Nolvadex , cial a CDx assay can be in relation to the development of a target- AstraZeneca) was developed in the 1970s for the treatment of ed anti-cancer drug. The HercepTest™ assay was approved for advanced breast cancer, data on estrogen receptor status was routine clinical use in September 1998 simultaneously with trastu- correlated with the treatment outcome. Based on a phase II study zumab in a novel coordinated process between two divisions of the performed in patients with metastatic breast cancer, published in Food and Drug Administration (FDA) (4). The parallel development 1976, the investigators concluded: “A high degree of correlation of Herceptin® and HercepTest™ is outlined in Figure 11.1. between response and positive estrogen-receptor assay suggests the value of the diagnostic test as a means to select patients Within the last decade, HER2 has proven not only to be an for tamoxifen treatment” (3). Despite the fact that this conclusion important target in the treatment of breast cancer, but more re- was reached as early as the mid seventies, more than 20 years cently also gastric cancer (8, 9). In 2010 the FDA approved use would pass before the next predictive or selective diagnostic as- of HercepTest™ and the HER2 FISH pharmDx™ Kit (Dako) for say saw the light of day, namely the HercepTest™ (Dako), an im- the assessment of gastric cancer patients for whom trastuzu- munohistochemical (IHC) assay for detection of overexpression mab treatment is being considered. Further, within the last of the human epidermal growth factor receptor 2 (HER2) protein. couple of years the HER dimerization inhibitor pertuzumab 133 Chapter 11 | Companion Diagnostics HercepTestTM Genentech develops the CTA PMA submitted for HercepTestTM Genentech/Dako cooperation starts CE marking of HercepTestTM FDA approves HercepTestTM 1992 1993 Phase II starts 1994 1995 1996 Phase III starts 1997 1998 Phase III enrollment closed 1999 2000 2003 CPMP approves Herceptin® IND submitted for Herceptin® BLA submitted for Herceptin® Phase I starts FDA approves Herceptin® Trastuzumab (Herceptin®) Figure 11.1 Co-development of the drug Herceptin® and the companion diagnostic, HercepTest™. CTA: Clinical Trial Assay; PMA: Premarket Approval; CE: Conformité Européene; IND: Investigational New Drug; BLA: Biologic License Application; CPMP: Committee for Proprietary Medicinal Products. (Perjeta™, Roche/Genentech), and the antibody-drug conju- names are used in the literature and by the regulatory authorities, gate ado-trastuzumab emtansine (Kadcyla™, Roche/Genen- such as pharmacodiagnostics, theranostics, pharmacogenomic tech) have both been approved for treatment of breast cancer. biomarkers, advanced personalized diagnostics, and compan- During the clinical development of these two compounds both ion diagnostics. Within the last few years ‘companion diagnos- HercepTest™ and the HER2 FISH pharmDx™ Kit were used tics’ has been used more and more frequently and this name has for patient selection, resulting in the FDA approval of updates also been adapted by the FDA; however, theranostics is still used to the intended use of these assays, so as also to include se- quite frequently especially in the academic medical literature (1). lection of patients under consideration for treatment with pertuzumab or ado-trastuzumab emtansine. For this book chapter, we will adhere to the deinition of CDx that the FDA recently published in a draft guideline (12). According As shown in Table 11.1 only two assays have obtained FDA to this deinition a CDx assay is an in vitro diagnostics device approval for use in relation to gastric cancer and the two new that provides information that is essential for the safe and ef- HER2 targeting compounds; pertuzumab and ado-trastu- fective use of a corresponding therapeutic product. The FDA zumab emtansine. further speciied three areas where a CDx assay is essential: „„ To identify patients who are most likely to beneit from a particular therapeutic product Chapter 11.3 Companion Diagnostics and ‘Personalized Medicine’ „„ To identify patients likely to be at increased risk of serious adverse reactions as a result of treatment with a particular therapeutic product There is lack of consensus regarding the terminology of a di- 134 „„ To monitor response to treatment for the purpose of ad- agnostics assay that is developed in parallel to a targeted drug justing treatment (e.g., schedule, dose, discontinuation) to and used to guide the treatment decision. A number of different achieve improved safety or effectiveness. Companion Diagnostics | Chapter 11 According to the FDA, a CDx can be used both to predict out- on clinical development (1). A deinition that focuses on the come (eficacy and safety) and to monitor the response. predictive or selective characteristics of the CDx assay and makes a link to personalized medicine is: ”A pre-treatment test The predictive or selective characteristics of CDxs have at- performed in order to determine whether or not a patient is tracted the most attention so far. Use of a CDx assay facilitates likely to respond to a given therapy. This type of test is clas- the design of clinical trials with a smaller number of subjects, siied as a predictive or selective test and is a prerequisite for which in turn has a clear positive effect on the resources spent implementation of personalized and stratiied medicine” (13). Table 11.1 Companion diagnostics assays approved through the premarket approval (PMA) process by the FDA. For further information see (10). Companion Diagnostic Type of Assay Indication(s) mentioned in the Indications for Use Drug(s) mentioned in the Indications for Use PathVysion HER-2 DNA Probe Kit (Abbott Molecular) FISH Adjuvant treatment of breast cancer Trastuzumab (Herceptin) Vysis ALK Break Apart FISH Probe Kit (Abbott Molecular) FISH Non-small cell lung cancer (NSCLC) Crizotinib (Xalkori) InSite HER-2/neu Kit (Biogenex Laboratories) IHC Breast cancer Trastuzumab (Herceptin) HercepTest™ (Dako) IHC Breast cancer Metastatic gastric or gastroesophageal junction adenocarcinoma Trastuzumab (Herceptin) Pertuzumab (Perjeta)1 Ado-trastuzumab emtansine (Kadcyla)1, 2 FISH Breast cancer Metastatic gastric or gastroesophageal junction adenocarcinoma Trastuzumab (Herceptin) Pertuzumab (Perjeta)1 Ado-trastuzumab emtansine (Kadcyla)1, 2 CISH Breast cancer Trastuzumab (Herceptin) EGFR pharmDx™ Kit (Dako) IHC Colorectal cancer Cetuximab (Erbitux) Panitumumab (Vectibix) c-Kit pharmDx™ Kit (Dako) IHC Gastrointestinal stromal tumors (GIST) Imagines (Gleevec/Glivec) Bond Oracle Her2 IHC System (Leica Biosystems) IHC Breast cancer Trastuzumab (Herceptin) SPOT-Light HER2 CISH Kit (Life Technologies) CISH Breast cancer Trastuzumab (Herceptin) Therascreen KRAS RGQ PCR Kit (Qiagen) RT-PCR Colorectal cancer Cetuximab (Erbitux) COBAS 4800 BRAF V600 Mutation Test (Roche Molecular Systems) RT-PCR Melanoma Vemurafenib (Zelboraf) Inform Her-2/Neu (Ventana Medical Systems) FISH Breast cancer None (Prognostic) INFORM HER2 Dual ISH DNA Probe Cocktail (Ventana Medical Systems) CISH Breast cancer Trastuzumab (Herceptin) PATHWAY Her2 (Ventana Medical Systems) IHC Breast cancer Trastuzumab (Herceptin) 1 Breast cancer only, 2Approved March 2013, in US only 135 Chapter 11 | Companion Diagnostics Chapter 11.4 Co-Development of Drug and Companion Diagnostics the Critical Path to New Medical Products, plus the FDA’s release in April 2005 of its Drug-Diagnostic Co-Development Concept Paper and Draft Guidance paper on companion di- Companion diagnostic development requires long term com- agnostics (12, 18). The main activities involved in each stage mitment from the parties involved. In an ideal scenario a diag- are reviewed in the following pages. The order and timing nostics manufacturer and a drug developer engage in a long for each of these stages depends on the biomarker, the clin- term partnership that if successful leads to the launch of the ical trial design and the extent to which the diagnostic/phar- drug and the companion diagnostic assay. The diagnostics maceutical companies are willing to invest, prior to having partner brings the assay forward in parallel to the drug devel- proof of concept of their predictive biomarker hypothesis in opment stages, ensuring that the assay has the appropriate patients. In accordance with the theme of this book we will as- level of evidence at each stage, in order to support required sume that the selected biomarker for companion diagnostic testing during each phase of the clinical trial. The combina- development is an immunohistochemistry-based assay. tion of a relevant biomarker and the clinical outcome data into a predictive or selective assay demands close collabora- Feasibility Stage tion between the diagnostic industry and the pharmaceutical Feasibility has two parts, identifying a relevant biomarker for industry during the development of companion diagnostics predicting eficacy of the drug in patients and establishing an (Figure 11.2). This process is also supported by the FDA’s assay for the relevant predictive biomarker. The irst part in- Critical Path Initiative, launched in March 2004 with the white volves many years of work understanding the biology of the paper Innovation-stagnation: Challenge and Opportunity on disease and how the drug target interacts with the cancer cell, Pharmaceutical Development Stages Pre-Clinical Clinical Phase II Phase I Demonstrate Feasibility Develop Prototype Create design specifications for indications Finalize work on prototype and protocols Confirm performance for FFPE tissues Train test center Conduct market assessment Align marketing and regulatory strategy Phase III FDA Filing/Approval Launch Analytical Validation Clinical Validation Drug-diagnostic Registration Commercial Phase Finalize design specifications and verify assay IDE submission Align and complete regulatory co-filings Scale up production and stock finished kits Determine clinical cut-off Manufacture and validate finished kits Conduct clinical trials Align launch and reimbursement plans Compile clinical data for submission Obtain FDA NDA/BLA and PMA approvals Global registration Conduct training and education Initiate co-marketing activities Launch clinically validated kits Companion Diagnostic Development Stages Product lifecycle management Figure 11.2 Collaboration between the diagnostic industry and the pharmaceutical industry during the development of companion diagnostics. FFPE: Formalin-Fixed, Parafin-Embedded; FDA: Food and Drug Administration; NDA: New Drug Application; BLA: Biologic License Application; PMA: Pre Market Approval. 136 Companion Diagnostics | Chapter 11 identifying the critical determinants that dictate if a cell is like- In addition, in the prototype assay stage relevant positive ly to respond to the drug. The second part aims to establish and negative control material should be established, controls whether it will be possible to develop a companion diagnos- which can be used by laboratories running the test in order tic assay as desired. Feasibility studies cover development or to accept or reject an assay run. This control material can be identiication of suitable antibodies for the target protein that either normal histologic structures in the tissue being stained, work on clinically relevant tissues, e.g. if the drug targets a pro- that are positive/negative for the protein (internal control), or tein deregulated in non-small cell lung cancer (NSCLC) the an- screened normal/cancer tissue with known protein status (ex- tibody must be able to detect the protein of interest in clinical ternal control). The development and inclusion of positive and NSCLC samples. Important points to consider are that the as- negative control material by the laboratory is a requirement for say should be functional within the relevant clinical entities, so testing in order to assure the validity of the testing procedure that if the assay is used for prospective or retrospective testing (36). Thus, laboratories that will deploy testing using the proto- in early stage clinical trials, it will work effectively in all disease type assay have to be appropriately trained in performing the entities being tested (lung, melanoma, kidney, etc), and will assay procedure, establishing controls and reporting results. be compatible with the sample types and sample processing methods present in the patient population, i.e. formalin-ixed, During the early clinical trials the prototype assay can be em- parafin-embedded surgical resection specimen or biopsies ployed retrospectively or prospectively depending on the level from the tumors. Another important aspect at this stage is to of evidence for association between biomarker and response establish the speciicity of the assay and sensitivity for detec- to therapy. A ‘go/no go’ decision is made based on the study of tion of the relevant protein in the tissue. It should be noted that both the drug and the biomarker in question. Prototype assays the terms ‘speciicity’ and ‘sensitivity’ in the above context refer are typically labeled RUO as the prototype assay application in to technical aspects of the test and are not the same as when early stage drug development is only used as a research tool for used to describe the ability of the validated assay to detect exploratory purposes in terms of investigating the relationship ‘non-responders’ and ‘responders’ in the patient population between biomarker status and response to therapy (14, 15, 18). (clinical speciicity and sensitivity – see below). Analytical Validation Stage Prototype Stage In order to deploy an assay for prospective selection of patients Once the points above have been covered in the prototype as- in a phase II/III it is recommended that the assay must previously say stage, a feasibility assay/kit can be developed. The assay have been validated (18). Based on earlier stage clinical trials, a description (protocol) should encompass how the entire im- cut-off or threshold for positivity for the target protein is established munohistochemical staining is performed from obtaining and and if possible linked to the eficacy of the drug. The cut-off value preparing the patient sample, through staining, interpretation needs to be established before the inale clinical validation of CDx and reporting of assay results. Under design control, the as- assay in phase III. This cut-off is built into inal coniguration of the say design speciications should include deinitions sensitivi- device, and validation studies are performed using kits that have ty and speciicity, interpretation rules and criteria, and clinical been manufactured in accordance with inal production proce- indications. These speciications are typically setup and test- dures in a GLP controlled setting. In addition, if control materials, ed during this stage. In respect to the interpretation system, a such as parafin embedded cell lines that express known levels data report form is established to make sure all relevant inter- of target protein, are included as part of the analytically validat- pretation data are collected when the prototype assay is later ed assay, then these have to be developed at this stage. Final applied to clinical samples. A limited number of relevant assay device design and coniguration are irst veriied, including test- robustness studies may also be performed during this stage; ing of accuracy, technical sensitivity, and speciicity, robustness however, the majority of studies establishing assay robustness (tolerance), and precision (intra-assay run, inter-assay run, inter-lot and reproducibility are performed following development of variability, inter-reader variability, inter-instrumentation variability). the analytically validated assay. Finally, external analytical validation studies are performed to doc- 137 Chapter 11 | Companion Diagnostics Reproducibility Inter laboratory variation Inter observer variation Day to day variation { { { Site 1 Site 2 Site 3 Observer 1 Observer 2 Observer 1 Observer 2 Observer 1 Observer 2 Run 1, 2, 3, 4, 5 Run 1, 2, 3, 4, 5 Run 1, 2, 3, 4, 5 Run 1, 2, 3, 4, 5 Run 1, 2, 3, 4, 5 Run 1, 2, 3, 4, 5 Figure 11.3 Example of reproducibility study. ument reproducibility (day to day, inter-observer, inter-laboratory). as deined by the cut-off. If testing of a ‘significant risk device’ See Figure 11.3 for example of reproducibility study. is to be deployed on US patients, an Investigational Device Exemption (IDE) is required in order to ensure that the device is The sensitivity of the assay to detect the target protein is estab- safe and effective for patient selection in the clinical trial setting lished (limit of detection), together with pattern of reactivity in and does not put patients participating in the clinical trial at risk. patient material; non-pathological as well as pathological. The The FDA will inspect the sponsor to conduct the trial under full speciicity of the assay is demonstrated, along with showing IDE regulations. An IDE application is submitted by the sponsor the absence of cross-reactivity to closely related targets. Test- responsible for deployment of the companion diagnostic dur- ing will deine the stability of the actual reagents comprising a ing the clinical trial and typically contains: investigational plan, kit, but will also demonstrate working stability of the assay as report of prior investigations, device description, device perfor- well as onboard stability of the kit. Today, the FDA only recog- mance, investigators’ information, monitoring, Institutional Re- nizes real time stability data when approving companion diag- view Board information, and informed consent documents. An nostic assays to be released on the IVD market (18). IDE may be required during the prototype phase if prospective selection of patients is based on the test result. The purpose of the activities performed during the analytically validated assay stage is to ensure that a robust assay with Upon IDE approval, the analytically validated assay will be stringent requirements can be deployed in a safe and effective used in conjunction with phase III/pivotal clinical trials for the manner in clinical trial testing. The analytically validated assay medical drug, to establish clinical validation of the device, and is manufactured in compliance with the applicable FDA Quality to establish the assay as an IVD companion diagnostics. System Regulations (QSR) and labeled IUO, supplemented by a text that clearly deines that ‘‘The performance characteristics of this product have not been established.’’ Laboratories that will perform testing during clinical trial must be trained in deploying Chapter 11.5 Clinical Validation of the Companion Diagnostics the assay according to the instructions for use, and patholo- 138 gist(s) must be trained in performing interpretation of staining The traditional way to evaluate safety and eficacy of a new pattern and reporting of results. Training includes criteria for re- drug is to perform a randomized clinical trial, in which the per- porting if a patient is positive or negative for the target protein formance of the new drug is compared to standard treatment. Companion Diagnostics | Chapter 11 If the result of such a comparative trial shows superiority of versus non-responders, and in this respect the clinical diag- the new drug over standard treatment, the usual interpretation nostic accuracy of the assay is important. Data on clinical sen- would be that the new drug is the preferred choice for patients sitivity, clinical speciicity, positive predictive value (PPV) and with the disease. However, a new drug with 10-15% superior- negative predictive value (NPV) for the assay are important di- ity over the current standard treatment may still be the wrong agnostic metrics to consider in this regard. Several designs for choice for many patients. Diseases are heterogeneous and the clinical drug-diagnostic co-development have been proposed traditional randomized clinical trial only takes this variation into and they can be broadly categorized into three groups (17-19): account to a limited extent. This type of study does not answer the question about the eficacy of the new drug in the individual patient, and extrapolation of the average study result to all „„ Untargeted/all-comers „„ Targeted/enrichment „„ Stratiication/marker-by-treatment-interaction patients may often be a wrong decision (16, 17). The drug-diagnostic co-development model is a way to obtain insight into In the untargeted or all-comers design, all patients meeting disease heterogeneity, thereby making any extrapolation from the eligibility criteria of the study protocol will be entered into clinical research to clinical practice much more accurate. the trial. Despite the patients being tested at study entry, the treatment assignment is independent of the test result from the In the drug-diagnostic co-development model the clinical tri- CDx assay and the patients are either randomized to the new als should be designed in such a way that both safety and treatment or to the standard treatment (Figure 11.4). eficacy of the drug, and the performance of the CDx assay are assessed at the same time. As described in Figure 11.2 it is desirable that the CDx development starts early during pre- New clinical development of the drug, so an analytical validated assay can be ready before start of phase III. Further, the clinical cut-off value for the assay needs also to be established, which requires detailed access to both the diagnostic test results as CDx + Patients R CDx Testing CDx − Standard well as clinical outcome data from the patients treated with the drug in question. The clinical outcome data that are collected from the phase II studies, which, for oncology studies, typically will be objective response rate, and sometimes time to progression and progression-free survival. If a relationship can be established between the assay result and the clinical outcome data, the next step will be to select the cut-off value for the as- Figure 11.4 The untargeted/all-comers design. After inclusion in the study all the patients are randomized (R) to either the new treatment (New) or the current standard treatment (Standard) independent of the CDx assay result. CDx+ indicates test positive patients; CDx– indicates test negative patients. Despite the patients being tested at study entry, the treatment assignment is independent of the test result from the CDx assay. say that deines a result as test positive (CDx+) or test negative (CDx−). The ability of an assay to discriminate between CDx+ and CDx− results at a given cut-off value will depend both on If the relationship between known pathophysiology and the the analytical performance of the assay and the strength of mechanism of the drug is unclear, this type of design address- the clinical response to treatment with the drug under devel- es the question of whether the new treatment is beneicial for opment. The clinical cut-off value for the assay should be se- all patients, or only a subset of patients, especially the CDx+ lected prior to performing the pivotal phase III trials, which will group (17, 19, 20). However, an untargeted design will not be provide the inal evidence of safety and eficacy for the new very useful if the prevalence of the CDx+ patients is low, be- drug as well as the clinical validation of the CDx assay (17). cause the treatment effect in the overall population will be diluted (21). If the prevalence of the CDx+ patients is suficiently A CDx assay will only be useful if it provides information that high the untargeted design has the advantage that it allows for can discriminate between patients who are likely responders a preliminary calculation of the clinical diagnostic sensitivity, 139 Chapter 11 | Companion Diagnostics speciicity, PPV and NPV for the assay. In general the untarget- (12, 17, 19). With this design the patients are stratiied based on ed design must be regarded as explorative, and is often used the result of the CDx assay and subsequently ran-domized to during early clinical development in phase I/II. either the new treatment or the standard treatment (Figure 11.6). If there is clear evidence of a strong relationship between a CDx+ status and treatment-outcome with the drug (e.g. from New previous phase I and/or II studies), a targeted or enrichment de- CDx + R sign can be used (12, 17). With this design, all the patients are tested by means of a CDx assay, but only the CDx+ patients are Standard enrolled in the study and subsequently randomized to either the new treatment or to the standard treatment groups (Figure 11.5). Patients CDx Testing New New CDx + Patients R CDx Testing R CDx − Standard Standard CDx − Figure 11.5 The targeted/enrichment design. With this design only patients that have a positive CDx+ result are randomized (R) to either the new treatment (New) or standard treatment (Standard). CDx+ indicates test positive patients; CDx– indicates test negative patients. Figure 11.6 The stratiication/marker-by-treatment-interaction design. With this the patients are stratiied based on the result of the CDx testing and subsequently randomized (R) to either the new treatment (New) or standard treatment (Standard). CDx+ indicates test positive patients; CDx– indicates test negative patients. In fact, this design is to some extent similar to conducting two This design have been used more often recently, for example to independent randomized trials, one in CDx+ patients and one in demonstrate safety and eficacy of anti-cancer drugs, such as CDx− patients. However, with the stratiied design these trials are vemurafenib (Zelboraf™, Roche/Genentech), crizotinib (Xalko- conducted under one umbrella, and differ from one large single ri®, Pizer), pertuzumab and ado-trastuzumab emtansine, and randomized study in a couple of ways; in that both sample size to validate clinically the corresponding CDx assay (22-25). A calculation, and randomization are stratiied based on results targeted design study was also used when trastuzumab went from the same CDx assay in the stratiied design approach. The through inal phase III testing in women with advanced breast advantage of this design is that it allows the calculation of the cancer in the 1990’s (5). This design generally requires ran- clinical diagnostic sensitivity, speciicity, PPV and NPV (17). So domization of a smaller number of patients as compared with far very few trials have been conducted using this design, but the untargeted designs, due to the fact that only CDx+ patients one of these is the Marker Validation for Erlotinib in Lung Can- are enrolled in the trial, which makes the study population more cer (MARVEL) study. This trial was designed to evaluate whether homogeneous. However, this design allows only PPV to be cal- epidermal growth factor receptor (EGFR) gene status, measured culated and not sensitivity, speciicity and NPV. by luorescence in situ hybridization (FISH), could be used to predict the response to treatment with erlotinib (Tarceva®, Roche/ 140 The targeted design approach has one drawback as it does not Genentech). Based on the FISH testing result patients with ad- provide insight into the eficacy of the new treatment in CDx− vanced non-small cell lung cancer (NSCLC) were divided into patients. However, this aspect is included in the stratiication two groups, EGFR+ and EGFR-, and subsequently randomized design, also called the marker-by-treatment-interaction design to either erlotinib or to pemetrexed (Alimta®, Lilly) (26). Companion Diagnostics | Chapter 11 Chapter 11.6 Companion Diagnostics and Regulatory Aspects risk-based classiication system will also be part of the new in vitro diagnostic regulation to be introduced in the EU, and CDx assays will likely be classiied as high risk products (27). If a diagnostic CDx assay is developed in conjunction with a targeted cancer drug, the CDx assay may later end up deter- When it comes to the regulations for CDx assays, the FDA has mining the conditions for use of the drug after approval. The been at the forefront compared to other national health agen- CDx assay then becomes a kind of gatekeeper in relation to cies. Already in 2005 the FDA issued a draft concept paper which patients to treat with the drug in question. This gate- on drug-diagnostic co-development, where both the analytical keeper role requires that the CDx assay must live up to certain and the clinical requirements for in vitro diagnostics assays standards during development, both in relation to analytical and developed in conjunction with a speciic drug were discussed clinical performance. This aspect has long been recognized by (18). In this draft concept paper a model for the parallel de- the FDA in the US, where most CDx assays must go through velopment of a drug and a diagnostic assay was suggested. the PMA process. In the European Union (EU) these types of This model was very much inluenced by the process through assays are not subject to any premarket approval process by which the HER2 IHC assay and trastuzumab were developed the regulatory authorities, but only a conformity assessment and by Genentech in the 1990’s. On several occasions the FDA has CE marking, which normally is performed by the manufacturers indicated that a draft guideline would be issued, based on this themselves. However, the growing number of CDx assays under drug-diagnostic co-development concept paper from 2005, development, and their increasing importance in relation to the and input from the different stakeholders, but at the time of the clinical use of targeted anti-cancer drugs, has raised questions present writing such guidance has not yet been released. as to whether the regulatory framework for these types of assays ought to be changed in the EU (1). In fact, very recently the Eu- Overall, within the last few years the regulatory framework for de- ropean Parliament has suggested changes in the regulation for velopment of CDx assays has gradually been changed to relect medical devices. With this newly proposed regulation, which is the important role that they play in relation to both drug develop- expected gradually to come into force from 2015 and onwards, ment and care of the individual patient. For a number of targeted it is anticipated that a type of premarket approval process will be anti-cancer drugs the CDx assay result has now become the de- introduced, where an independent notiied body will be involved cisive factor in relation to the treatment decision in the clinic. This in conformity assessment relating to in vitro diagnostic medical key role of the CDx assays has not only been recognized by the devices, which of course will include CDx assays (27). FDA, but also now by other regulatory health agencies worldwide. From a medical and regulatory point of view CDx assays are considered high risk medical devices, because clinical deci- Chapter 11.7 Learnings from EGFR sions and actions are taken based directly on the test results. For example, an incorrect diagnostic result from a CDx assay Upon the success of trastuzumab in clinical trials it was antici- could lead to a wrong treatment decision, which for a cancer pated that other targeted agents against growth factor receptors patient could have serious consequences. In general, in US, would follow and that similar correlations of overexpression of the classiication of a diagnostic device depends on the in- a growth factor and beneit of treatment would be found. Epi- tended use of the assay and its associated risk to the patient. dermal growth factor receptor (EGFR) was shown to be overex- An intended use with a higher risk will direct the classiication pressed, acting as an oncogene in several cancers, rendering of the device towards a high risk class III device. Intended it an attractive target for therapeutic intervention (32). Several uses judged to carry moderate risk allow the lower risk clas- clinical trials were initiated with EGFR inhibitors such as cetuxi- siication, of class II, while devices of low risk are classiied mab (Erbitux®, Bristol-Myers Squibb / Eli Lilly) and panitumumab as class I devices (28). In the US, the FDA normally classiies (Vectibix®, Amgen). Patients were tested with IHC-based assays CDx assays as high risk class III diagnostic devices. A similar for detection of EGFR expression by tumor, however, a clear cor- 141 Chapter 11 | Companion Diagnostics relation between overexpression of protein and the patient’s re- be introduced for clinical use before 2018, and that most of these sponse to treatment was not found (30-33). Instead, exploratory will likely have a CDx linked to their use (11). testing for mutations in patients from the clinical trials revealed a linkage between K-RAS mutation and resistance to EGFR in- Immunohistochemistry-based companion diagnostics may be hibitors, suggesting that inhibition of EGFR expression was by- limited in their precision to measure the target protein, for reasons passed by downstream signaling (34). Subsequently, speciic that include non-standard sample preparation, assay variation, mutations in the EGFR receptor have been linked to response to or poor observer-to-observer reproducibility, all of which point to- therapy with erlotinib and been used in clinical trials as predic- wards a need to move into more standardized controlled quanti- tive biomarkers. Consequently, the COBAS EGFR Mutation Test tative assay methods, that also incorporate digital image analysis has recently received FDA approval as a companion diagnostic to measure accurately the amount of target protein in the tissue. for erlotinib (35). This example illustrates the need for thorough understanding of the biology of the disease and the pathway(s) Technological advances in sequencing methods have also made being targeted in order to devise the best strategy for a com- signiicant progress in the past decade and comprehensive se- panion diagnostic, prior to engaging in clinical trials. quencing efforts using next generation sequencing methods have revealed the genomic landscapes of the a number of common human cancers (37). These indings already have had pro- Chapter 11.8 Conclusion and Future Perspectives found effects on our understanding on the genetic background of different cancers types and will inevitably lead to new therapeutic IHC or FISH-based assays are commonly used as companion development, and thereby companion diagnostic development. diagnostic tests in routine pathology laboratories, making them It is likely that sequencing of critical genes performed in combi- the current methods of choice for an increasing number of onco- nation with measurement of the functional output of the deregulat- logical drug development programs. The use of a CDx assay to ed pathways by immunohistochemistry and other protein based select patients for treatment during clinical development is clearly technologies will lead to increasingly complex companion diag- a way to fulill the promise of personalized medicine, from both nostic testing schemes that in turn will result in new and innovative a patient and regulatory perspective. In this chapter we have trial designs and approval procedures. tried to summarize the efforts necessary for development and regulatory approval of an IHC-based companion diagnostic. In order to successfully employ a companion diagnostics strategy, a References concerted effort from both the pharmaceutical company and the 1. Jørgensen JT. Companion diagnostics in oncology – Current status and future aspects. Oncology 2013; 85:59-68. diagnostic manufacturer is needed, which clearly requires major commitments of time and resources by both parties. 2. The Lancet. Where will new drugs come from? Lancet 2011; 377:97. The development of trastuzumab is often held up as an example 3. Israel L, Leung BS. Phase II study of tamoxifen: report of 74 patients with stage IV breast cancer. Cancer Treat Rep 1976; 60:1431-35. of successful development of a targeted anti-cancer drug and has served as the main inspiration for the current drug-diagnostic co-development model. It is now 15 years ago since the approval of HercepTest™. Looking at the list of CDx assays that have gone through the premarket approval process at the FDA, reveals a relatively short list, with modest diversity. Ten of the 15 assays on the list are measuring HER2, either as protein overexpression or as gene ampliication (10) (see Table 11.1). However, this situation is very likely to change dramatically in the years to come, as it has been estimated that as many as 80 targeted cancer drugs could 142 4. Jørgensen JT, Winther H. The Development of the HercepTest – From Bench to Bedside. In: Molecular Diagnostics – The Key Driver of Personalized Cancer Medicine (Jørgensen JT and Winther H eds.). Pan Stanford Publishing, Singapore, 2010; 43-60. 5. Slamon DJ, Leyland-Jones B, Shak S et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001; 344: 783-92. 6. Simon R, Maitournam A. Evaluating the eficiency of targeted designs for randomized clinical trials. Clin Cancer Res 2004; 10:6759-63. 7. Hortobagyi GN. Opportunities and challenges in the development of targeted therapies. Semin Oncol 2004; 31 (1 Suppl 3):21-7. Companion Diagnostics | Chapter 11 8. Ross JS, Slodkowska EA, Symmans WF, et al. The HER-2 receptor and breast cancer: ten years of targeted anti–HER-2 therapy and personalized medicine. Oncologist 2009; 14:320-368. 9. Jørgensen JT. Targeted HER2 treatment in advanced gastric cancer. Oncology 2010; 78:26-33. 10. FDA. In Vitro Companion Diagnostic Devices. Updated: November 21, 2012. (www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/InVitroDiagnostics/ucm301431.htm). 11. Thomas J, Stratton E, Keppens M. Companion diagnostics: emerging strategies and issues in pharmaceutical development. Expert Rev Mol Diagn 2012; 12:561-3. 12. FDA. In Vitro Companion Diagnostic Devices. Draft Guidance. FDA, July 14, 2011. (http://www.fda.gov/downloads/Medical Devices/DeviceRegulationandGuidance/GuidanceDocuments/ UCM262327.pdf). Accessed April 18, 2013. 13. Jørgensen JT. Companion Diagnostics and the Drug-Diagnostic Co-Development Model. Drug Dev Res 2012; 73:390-97. 14. Swanson J. Companion diagnostics take off (online). Available from URL:http://www.genomeweb.com/dxpgx/companion-diagnostics-take 15. Metcalfe TA. Development of novel IVD assays: a manufacturers perspective. Scand J Clin Lab Invest 2010; 70:23-6. 25. Verma S, Miles D, Gianni L, et al. Trastuzumab emtansine for HER2 -positive advanced breast cancer. N Engl J Med. 2012; 367:1783-91. 26. Freidlin B, McShane LM, Korn EL. Randomized clinical trials with biomarkers: design issues. J Natl Cancer Inst 2010; 102:152-60. 27. Proposal for a Regulation of the European Parliament and of the Council on medical devices, and amending Directive 2001/83/ EC, Regulation (EC) No 178/2002 and Regulation (EC) No 1223/2009. (http://ec.europa.eu/health/medical-devices/iles/ revision_docs/proposal_2012_542_en.pdf). Accessed May 1, 2013. 28. Scherf U, Becker R, Chan M, et al. Approval of novel biomarkers: FDA's perspective and major requests. Scand J Clin Lab Invest Suppl. 2010; 242:96-102. 29. Vincenzi B, Santini D, Tonini G. New issues on cetuximab mechanism of action in epidermal growth factor receptor – negative colorectal cancer: The role of vascular endothelial growth factor. J Clin Oncol 2006, 24 (12), Apr. 20 30. Campaglio M, Locatelli A, Olgiati C, Normanno N, Somenzi G, Vigano, L, et al. Inhibition of proliferation and induction of apoptosis in breast cancer cells by the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor ZD1839 (‘Iressa’) is independent of EGFR expression level. J Cell Phys 2004; 196:259-28. 16. Woodcock J. The prospects for “personalized medicine” in drug development and drug therapy. Clin Pharmacol Ther 2007; 81:164-9. 31. Chung KY, Shia J, Kemeny NE, Shah M, Schwartz GK, Tse A, et al. Cetuximab shows activity in colorectal cancer patients with tumors that do not express the epidermal growth factor receptor by immunohistochemistry. J Clin Oncol 2005; 23:1803-10. 17. Winther H, Jørgensen JT. Drug-Diagnostic Co-Development in Cancer. Pharm Med 2010; 24:363-75. 32. Scaltriti M, Baselga J. The epidermal growth factor receptor pathway: a model for targeted therapy. Clin Cancer Res 2006; 12:5268-72. 18. FDA. Drug-diagnostic co-development concept paper. Draft. April 2005. (www.fda.gov/downloads/drugs/scienceresearch/researchareas/pharmacogenetics/ucm116689.pdf). Accessed April 23, 2013. 33. Johnston JB, Navaratnam S, Pitz MW, Maniate JM, Wiechec E, Baust H, et al. Targeting the EGFR pathway for cancer therapy. Curr Med Chem 2006; 13:3483-92. 19. Mandrekar SJ, Sargent DJ. Clinical trial designs for predictive biomarker validation: theoretical considerations and practical challenges. J Clin Oncol 2009; 27:4027-34. 34. Benvenuti S, Sartore-Bianchi A, Di Nicolantonio F, Zanon C, Moroni M, Veronese S, et al. Oncogenic activation of the RAS/RAF signaling pathway impairs the response of metastatic colorectal cancers to anti-epidermal growth factor receptor antibody therapies. Cancer Research 2007; 67:2643-8. 20. Simon R M, Paik S, Hayes DF. Use of archived specimens in evaluation of prognostic and predictive biomarkers. J Natl Cancer Inst 2009; 101:1446-52. 21. Mandrekar SJ, Sargent DJ. All-comers versus enrichment design strategy in phase II trials. J Thorac Oncol 2011; 6:658-60. 22. Chapman PB, Hauschild A, Robert C, Haanen JB, Ascierto P, Larkin J. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med 2011; 364:2507-16. 23. Kwak EL, Bang YJ, Camidge DR, et al. Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. N Engl J Med 2010; 363:1693-703. 24. Baselga J, Cortés J, Kim SB, et al. Pertuzumab plus trastuzumab plus docetaxel for metastatic breast cancer. N Engl J Med 2012; 366:109-19. 35. Halait H, DeMartin K, Shah S , Soviero S, Langland R, Cheng S, et al. Analytical performance of a real-time PCR-based assay for V600 mutations in the BRAF gene, used as the companion diagnostic test for the novel BRAF inhibitor Vemurafenib in metastatic melanoma. Diagn Mol Pathol 2012; 21:1-8. 36. Clinical and Laboratory Standards Institute (formerly NCCLS). Quality Assurance for Design Control and Implememtation of Immunohistochemistry Assays; Approved guideline-Second Edition. CLSI document I/LA28-A2 (1-56238-745-6) Clinical and Laboratory Standard Institute, 940 West Valley Road, Suite 1400, Wayne, Pennsylvania 19087 USA, 2011 37. Vogelstein B, Papadopoulos N, Velculescu VE, et al. Cancer genome landscapes. Science 2013; 339: 1546-58. 143 Part II: The Potentials and Pitfalls Chapter 12 Tissue Microarray – Construction and Quality Assurance Rashmil Saxena, BFA Sunil Badve, MD Mi•cro•ar•ray (n.) A supporting material (as a glass or plastic slide) onto which numerous molecules or fragments usually of DNA or protein are attached in a regular pattern for use in biochemical or genetic analysis. Merriam-Webster Dictionary of Medical Terms Tissue Microarray – Construction and Quality Assurance | Chapter 12 Chapter 12.1 Introduction The origin of TMAs can be attributed to Dr Hector Battifora’s humble ‘sausage’ blocks (1), in which a number of tissues, typ- Recent times have seen the advent of high throughput as- ically from different organs, were thrown together in the same says such as array comparative genomic hybridization, cDNA block and the tissue distribution of a particular antigen/protein microarray and Next Generation Sequencing techniques, was assessed. A signiicant disadvantage of this technique was which have led to the rapid discovery of thousands of poten- that when tumors or tissues from the same site were put togeth- tial biomarkers. However, these need to be validated in tis- er it was dificult, if not impossible, to trace them back to the sue-based studies in large datasets to prove their potential patient. This prevented meaningful analysis of prognostic mark- utility. As these datasets are typically present in the form of ers. However, many laboratories, small and large (including our formalin-ixed, parafin-processed tissue blocks, immunohis- own), adopted this technology to generate multi-tissue or mul- tochemical (IHC) methods are ideal for validation. However, ti-organ tissue blocks. The next step in the development of TMA performing whole-section IHC on hundreds to thousands of was described by Wan et al. (2) who used a 16-gauge needle blocks requires a lot of resources in terms of reagents and to manually bore cores from tissue blocks and array them in a time. Analysis needs to be batched; batch to batch variability multi-tissue straw in a recognizable pattern. This method was could interfere with the analyses. In addition, an average block further modiied by Kononen et al. (3) using a 4 mm skin biopsy will yield less than 300 slides of 5 µm each. The tissue microar- punch. They used a cast of a small amount of melted parafin ray (TMA) technique circumvents some of these problems. to record the position of each punch specimen. This landmark study lead to the development of a TMA precision microarray instrument with an x-y guide by Beecher Instruments (Sun Prairie, WI). This device enabled real high-throughput analysis, with arraying of up to 1,000 cores in the same block. Chapter 12.2 Advantages and Disadvantages of TMAs Advantages The major advantage of TMAs is that they allow the performance of tissue-based assays (immunohistochemistry, histochemistry, in situ hybridization, etc.) on a large number of patient samples in an eficient and cost-effective manner. With TMA technology, several hundred representative cores from several hundred patients may be included on a single glass slide for assay. Thus, signiicantly more tissue can be conserved than if the blocks were to be sectioned serially. TMAs have been generated from all tissue types, including decalciied bone and core biopsies. Figure 12.1 A) Principle of tissue microarray (TMA) analysis. Cylindrical cores are obtained from a number (up to 1,000) of individual formalin-ixed, parafin-embedded tissue blocks. These are transferred to a recipient TMA block. Each TMA block can be sectioned up to 300 times. All resulting TMA slides have the same tissues in the same coordinate positions. The individual slides can be used for a variety of analyses saving labor and reagent costs while maintaining uniformity of assay. Typically a minimum of three cores for each case are used for 0.6 mm cores. The latter are usually rotated 90 degrees and embedded vertically to ensure presence of tissue of interest in multiple cuts. In addition, methods for generating TMAs from fresh frozen tissue using blocks made from either optimal cutting temperature compound, or from a mix of gelatin-sucrose, have been described (4). More recently, a technique for developing ‘patch’ TMAs from unstained slides has been described (5). 145 Chapter 12 | Tissue Microarray – Construction and Quality Assurance Disadvantages cells in tumors. It is also dificult to study certain facets of tu- The major disadvantage of TMAs is that each core (or set of mor biology, such as interactions between the tumor and it’s cores) represents a fraction of the lesion. This was considered stroma, as these stromal components may not be adequately a major weakness, particularly in the early days of the TMA. represented in the cores. The use of large cores (2 mm) has However, multiple studies in different organ systems have now been advocated for these types of studies. demonstrated that consistent and comparable results can be obtained using TMA cores as with whole sections. In order to obtain comparable results two main strategies have been used. Chapter 12.3 Types of TMAs The irst is increasing the number of cores from each case. It is typical in breast cancer to use at least two cores from each case The type of TMA to be generated depends on the question when using a 1 mm core; a minimum of three cores from each being asked within the study. The following are the commonly case for 0.6 mm cores. The number of cores may vary accord- used types: ing to the disease site; for example, it is typical to use ive cores Cell Line Arrays in colon cancer. Although a 2 mm core might be theoretically These arrays consist of normal or cancer cell lines that are grown considered better than multiple smaller cores, in practice this is in culture. The major function of these arrays is to survey the pres- usually not true. Smaller cores permit sampling of different tumor ence of proteins that are known to be present in one or more of the areas and are, therefore, more likely to be representative of the cell lines. In addition, cell line arrays can be used to analyze the entire tumor. In addition, smaller cores tend to inlict a lesser de- utility (plus sensitivity and speciicity) of an antibody in detecting gree of damage on the original tissue blocks. The second strat- proteins. The most common example of this type of array is the egy consists of increasing the number of tumors included in the 3-cell line control that is used with HER2 testing in breast cancer. study. This method averages out the errors that might result from tumor heterogeneity associated with the use of tissue cores. Random Tissue/Tumor Arrays These arrays contain tissues from multiple sites and contain tumor and/or non-tumor tissues. Small arrays of this kind can be used for quality control measures, such as monitoring of existing reagents/antibodies, as well as work-up of novel re-agents. In addition, they can also be used as discovery tools. For example a survey of CD10 in tumors arising in multiple tissue sites lead to the discovery of its utility in diagnosing uterine stromal tumors (6). Consecutive Case Array This type of array is constructed using consecutive cases belonging to a single tissue site. These types of arrays are extremely useful for quality control purposes, including identifying shifts and drifts in reagent quality. They are also useful in studyFigure 12.2 Photograph of low density TMA block. ing the prevalence of a protein/antigen in a given tumor type, and analyzing the relationships between different biomarkers. When Not to Use TMAs 146 TMAs are not recommended for certain types of studies. In cer- Tumor Characteristic-based Array tain tumors such as glioblastoma, there is such marked heter- This is a special type constructed solely on the basis of a given ogeneity within tumors that this feature may not be adequately characteristic, such as patient age or tumor grade. The latter is captured in TMA studies. In addition, TMAs are also not very useful for evaluating the frequency of a marker throughout the useful to study rare or focal events, such as number of immune spectrum of tumor differentiation. Similarly TMAs can be gen- Tissue Microarray – Construction and Quality Assurance | Chapter 12 erated based on the expression of a biomarker, such as es- how they should be distributed across the array to avoid bias. In trogen receptor or HER2/neu positive or triple negative breast general the number of cases required in a TMA is dependent on cancers. These types of biomarkers are useful in analyzing in- the size of the difference expected in the outcome and the degree terrelationships between different cellular pathways. of variance in the samples; this sample size calculation is best done with the assistance of a statistician. Outcome-based TMAs Progression Arrays may need input from, or participation of, treating physicians/on- These types of arrays are used to analyze the role of protein(s) in cologists. As an example, it took two years for the design and cancer progression and consist of normal tissues from patients construction of TMAs in a study that involved 200 cases at six without cancer, normal tissue from patients with cancer, pre-inva- sites in the United States and Canada. The study was conducted sive lesions and tumor (from local and metastatic sites). The addi- to validate deinitively tissue biomarkers of prostate cancer recur- tion of normal tissue from close to the tumor and those much fur- rence after radical prostatectomy. The scientiic team involved ther away from the tumor site might enable study of ‘ield effect’. pathologists, clinicians, statisticians and cancer researchers (7). Outcome-based Arrays The following steps are recommended for breast TMA construction: These special arrays are the most valuable and most dificult to Step 1: Deine the question generate, as they involve collation of tissues from patients that As described above TMAs are created to answer speciic ques- have the same disease and have been more or less similarly treat- tions. It is important to deine clearly this question at the outset. ed and followed up for a signiicant period of time. The period of The question will help deine the number of cases and cores that follow-up depends on the type of disease or tumor being studied. need to be used in the generation of the TMA. For example, a These types of arrays are mostly used to evaluate prognostic or TMA containing 20 cases might be suficient for routine quality predictive biomarkers. The presence of biomarkers in tumor sub- control/ assessment, but is not enough for biomarker assessment. types might then be used to design novel therapeutic strategies. Step 2: Review the cases to be included in the TMA Other Special Types Pull all the cases to be included in the TMA together. If the blocks TMAs can be generated based on speciic question being asked, have been previously cut into for other clinical or research purpos- whether it be race (Caucasians versus African Americans), sex es, it is prudent to review a fresh H&E slide to ensure that the slide is (male versus female) or more tissue-oriented questions such as representative of the block. Review all the slides and mark areas of center of the tumor versus invasive edge of the tumor. interest. It is useful to mark multiple areas from more than one block, as blocks may be depleted or misplaced. Areas to be sampled (tumor, normal, and pre-malignant tissues) should be identiied. Chapter 12.4 Team Required for TMA Construction Step 3: TMA core size and number of cores Size of the cores: The typical core sizes used for TMA constructs Construction of TMA is a team effort (7). The irst and foremost are 0.6 mm, 1.0 mm, 1.5 mm and 2.0 mm. Many workers consider question that needs to be answered is: why is the TMA being the small 0.6 mm cores as the standard of practice. Use of smaller constructed? This will decide the composition of the team. For the core diameters, however, allows for a greater number of cores to generation of the simplest of TMA, a technologist might be all that be extracted from the lesion and a greater number of cores that is required. However, most TMA synthesis will require close collab- can it into the TMA block. In addition, they tend to inlict little dam- oration with the pathologist, who will identify the areas of interest age on the donor and recipient blocks and the cores are easier and mark them for the technologists to obtain cores. It is a good to remove and replace from these blocks. The larger core sizes idea to involve biostatisticians from the onset rather than just ask- have the advantages of being more robust and the cores are ing them to do the data analysis. They can help with deciding the more resistant to damage during handling. However, these larger number of cores from each case, total number of cases needed, sizes can lead to increased likelihood of dificulty in extracting the 147 Chapter 12 | Tissue Microarray – Construction and Quality Assurance cores from the blocks, as well as greater chance of the blocks being broken or cracked during the TMA generation process. Number of Cores The optimal number of cores, to be included in the TMA, is trol and to address tumor heterogeneity. Three types of control tissues may be used: „„ Tissue-speciic controls: Normal tissues and cell lines from the organ site can help in comparative analysis of the marker expression status, in addition to helping ensure standardization. marker dependent and can vary depending on the degree of „„ Biology-associated controls: It is useful to insert ‘path- tumor size and heterogeneity. In general, the greater the de- way associated’ controls to ensure that the reagents are gree of intratumoral heterogeneity for any given marker, the working well; thus functioning as good internal controls greater is the number of cores that will be required. When within the TMA block. Common examples include endo- using 0.6 mm sized cores, it is typical to use a minimum of metrium for hormone receptor, testes or lymph node or three cores per case. Three 0.6 mm cores are still better than tonsils, for proliferation. one 1.0 mm core, even though the tissue surface area is es- „„ Organ system controls: examples include adrenal gland, sentially identical. For small tumors, three 1.0 mm cores could brain, breast, colon, kidney, liver, lung, pancreas, placenta, result in destruction of the donor block; so tumor size would prostate, testes, salivary gland, uterine myometrium (smooth also drive number/size of cores to be taken. Studies that have muscle). These controls are particularly useful when the used 1 mm core punches have tended to use two cores (8). TMA is being used to analyze novel markers, as one or more of these tissues can serve as internal controls. Nor- Density mal tissue TMAs, at a minimum, should contain: liver, The maximum number of cores that should be placed on a sin- kidney, endometrium, lymph node, colon, and testis. gle block will vary depending on core size, block size, and IHC methodology, among other factors. It is best to avoid placing so many cores on a TMA that the surface section of cores becomes larger than the antibody coverage area on the slide programmed by the autostainer (e.g. Dako Autostainer). Similarly, too many cores diminish the amount of parafin at the edge of the block creating dificulties in sectioning. Cores should start at least 3 mm away from the block edges, to prevent the parafin from cracking. Maximum number of cores per block should therefore depend on the comfort level of the technician, as well as the pathologist, who is ultimately going to read the slides. For these reason, it is typical for most workers to put somewhere between 100 and 300 of 0.6 mm cores in a TMA block. Distance The distance between cores should NOT exceed the core diameter. It is easier for the microscopist to follow the rows and columns if he/she can “lead” from one core to another. If the distance between cores is large, it dificult to follow the chain of cores and may result in skipping of lanes and false recording of data when performing manual interpretation. Step 4: Identify control tissues to be included in the block Controls should be placed on each TMA block, for quality con- 148 Figure 12.3 TMA map and block design: TMA layout should be asymmetric and irregular so that it is relatively easy to orient the TMA block. This irregularity should be obvious to the histotechnician who is cutting the block, so that all the cuts from the block are taken on the slides in an identical manner. In addition, locating the controls in an asymmetric manner is also helpful when reading the slides. For example, the following features may be included: 1) Blank rows and columns that do not run down the center lines of the TMA blocks, but to one side, so that the block is cut into two-third and one-third grids. 2) A blank corner for orientation or a tail coming out from close to one of the corners. 3) Asymmetric distribution of control cell lines and tissue controls. Placing stained cores of control tissues at the edge of the grid can be useful to mark orientation. Tissue Microarray – Construction and Quality Assurance | Chapter 12 Step 5: Make a TMA map depicting the layout devices consist of a stand, in addition to a positioning apparatus, The TMA map may consist of a simple Excel sheet, or may be and ensure vertical punching of the blocks and proper placement a more sophisticated datasheet made using one of the TMA within the grid. Fully automated devices additionally have inte- generation programs. This map also serves as a guide in or- grated computers that can be programmed to select the donor der to arrange blocks in the sequence in which they need to sites from different blocks and transfer them in the recipient block. be arrayed. Thus, the TMA map will contain the exact location of each case, including the duplicate samples, and where Donor Block controls are located. Mini-arrays (“City Blocks”) of the cores The block from which a core will be taken is referred to as the (3x5, 4x5, 5x5, 6x5) can be spaced for easy orientation, with donor block. The area of the donor block to be cored for TMA control tissue in the rows between the mini-arrays. should be selected by a pathologist. Although it is intuitive, it must be stated that the donor blocks should be optimally pro- Issues Related to Layout „„ TMA layout should be asymmetric and irregular to assist orientation (see Figure 12.3) „„ If multiple TMA blocks are being made for the same project, cessed and should not contain any poorly processed areas. Similarly, cores should be obtained from the block before the block gets depleted. The thicker the donor blocks the greater the number of useful sections that can obtained from the TMA. one consideration is to carry a small proportion of cases onto Core punches should be pushed gently into the TMA block, other blocks (e.g., 10%) and not too deeply as this can damage the needle as well as „„ Cores from the same case: Ideally, if same-patient cores are the block. When using semi-automated devices it is easier to to be placed on the same block, they should be dispersed mark the depth of the punch to the level of the plastic of the on the block. This will decrease the risk of interpretation bias. cassette. It has been suggested that heating the tissue core However, some researchers prefer this arrangement since it for 10 minutes, before inserting it into the recipient block, al- permits immediate ‘normalization’ or ‘conirmation’ of the lows better fusion of the parafin within the core and that sur- results of the different cores from the same patient. If same rounding the core. Fusion avoids loss and folding of tissue patient cores are to be dispersed across multiple blocks, it cores during sectioning of the TMA (9). is better to place them in different regions of the array (outer and inner); (outer in one block and inner in the other), with Recipient Block random placement, rather than placing them in the same The block into which the cores are placed is referred to as location in each block. This is done to prevent similar artifacts the recipient block. It is best to place the cores towards the affecting all cores. center of this block in order to prevent cracking of the block. After the cores are inserted, place the TMA at 37 °C degree Step 6: Creating the TMA itself Instrumentation overnight, and then on the cold plate of the tissue embedding The need for specialized instrumentation for creating TMAs is en- to temper the array. Multiple sections from the block should tirely based on the number of cores and value of the tissue being be cut at the same time to prevent wastage of tissue. Incom- inserted in these TMAs. For TMAs being constructed for quality plete sections should not be discarded; these can be used for control/quality assessment or work-up of new reagents, the num- standardization of staining technique (see below). station, with subsequent two to three 1-hour cycles of hot/cold ber of cores inserted is relatively low. This enables use of larger cores and diminishes the need for specialized instrumentation. Staining TMAs However, for TMAs to be made from valuable cases with scant When performing staining of the TMA, the step is to ensure materials, it is necessary to use these instruments. The simplest of that the staining procedure actually works in the laboratory and these devices consists of hand-held punches and is generally not the procedure has been standardized. If the TMA has been very useful for a serious TMA project, where it is necessary to use obtained from an outside institution, it is important, if possible, at least an intermediate grade device. These intermediate grade to get other tissue processed in that laboratory, or alternatively 149 Chapter 12 | Tissue Microarray – Construction and Quality Assurance poor-quality sections (incomplete or discards) from the TMA, selection) helps to prevent the creation of biased TMAs. It is for practice and standardization; following standardization useful to perform common biomarker analysis on sections from good quality TMA sections should be used for analysis. As the created TMA to conirm the “normal” distribution of known TMA sections are usually larger, they require special care to parameters. Comparison of this data with prior clinical data (e.g. ensure that the entire section (the whole array) is covered with ER analysis) obtained from whole section analysis is particularly reagents, otherwise uneven staining will be observed. useful to validate utility of the TMA. Alternatively the incidence of expression of a number of biomarkers in the TMA should be One of the limitations of the TMA is that the tissues in the tumor compared to that in published literature (using whole sections). cores have been processed at different times and often with different protocols. This will lead to optimal staining of some tumors, Quality Assurance Measures but also sub-optimal staining (over or under-staining) of quite a It is critical to perform and analyze H&E sections from the TMA few tumor cores. However, the large number of cases included in to conirm the presence of tissue of interest (usually tumor) in the the TMA can to some extent compensate for this limitation. TMA sections. In addition, H&E should be performed at regular interval (e.g. on every 25th slide) from the TMA blocks. The above Step 7: Validation and quality assurance tests should be reviewed by a pathologist familiar with the study. Measures for the TMA should include the following: Validation The use of TMAs enables analysis of large datasets, however Chapter 12.5 TMA Analysis this ability does not by any means suggest that the dataset is not skewed (10). This skewing may be the result of the institution’s The analysis of the TMA has two components. The irst involves location (population distributions with regards to race, ethnici- analysis of the slides and recording of the data. The second ty, and access to health care), or type of practice (community involves data analysis. hospital versus referral center). These biases collectively might 150 inluence the tumor size, grade and subtype composition of the Slide Analysis cases in the dataset. Such abnormalities of the dataset need The TMA slides can be analyzed manually; alternatively auto- to be recognized and allowed for in interpretation of indings; mated image analysis programs that can assist with the ana- the involvement of a biostatistician from the start (i.e. at case lysis are also available. The need for these programs is based Figure 12.4 A) Low power photograph of TMA stain. B) High power photograph of TMA stain. Tissue Microarray – Construction and Quality Assurance | Chapter 12 on the work volume as well as density of the TMAs. Some References programs generate a virtual slide (Whole Slide Image – WS) 1. Battifora H. The multitumor (sausage) tissue block: Novel method for immunohistochemical antibody testing. Lab Invest 1986; 55:244-8. of the TMA and further analysis can be done using a computer screen. Use of a WSI has the advantage of avoiding burn to the slides and all the cores can be analyzed at the same “optical and illumination” conditions. It additionally permits electronic storage of the fresh images and later re-analysis if 2. Wan WH, Fortuna MB, Furmanski P. A rapid and eficient method for testing immunohistochemical reactivity of monoclonal antibodies against multiple tissue samples simultaneously. J Immunol Methods 1987; 103:121-9. tions, which fade with viewing and time. 3. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, et al. Tissue microarrays for high-throughput molecular proiling of tumor specimens. Nat Med 1998; 4:844-7. Data Analysis Step 1: Data cleaning 4. Hoos A, Cordon-Cardo C. Tissue microarray proiling of cancer specimens and cell lines: Opportunities and limitations. Lab Invest 2001; 81:1331-8. required. This capability is particularly beneicial for FISH sec- Given the large number of samples in a typical TMA study, analysis of the data can become quite a challenge. One needs to exclude the cases that are not informative; it is not unusual to lose up to 10% of cases due to insuficient representation of tissue of interest. On the informative cases, strategies for conversion of multiple values (one per core) for each case into a single data point have to be devised. The commonly used strategies include using the highest value or a numerical mean of the values obtained (for review see (11). Each method used for normalization has its own advantages, as well as limitations. Step 2: Statistical analysis The tests used to determine the p value will be dependent on the type of data (i.e. nominal or categorical), as well as the degree of variance within the data. For simple analyses of relationships, contingency tables, and chi-squared tests are used. For demonstration of survival distributions, most researchers use the Kaplan-Meir plot and then apply Log-rank analysis to test survival differences between groups. The most frequently used analytical strategy is to subdivide patient material into highand low-risk groups, based on the expression of novel biomarkers. Some commercially available computer programs, such as X-tile program (11), may assist the selection of the best cut-off point. This cut-off point needs to be conirmed in a separate series of cases to validate its utility. The NCI – EORTC group has developed the REMARK (Reporting recommendations for tumor marker prognostic studies) (12) guidelines 5. Deng FM, Zhao Y, Kong X, Lee P, Melamed J. Construction of tissue microarrays using pre-existing slides as source of tissue when parafin blocks are unavailable. J Clin Pathol 2013; 66:627-9. 6. Chu P, Arber DA. Parafin-section detection of cd10 in 505 nonhematopoietic neoplasms. Frequent expression in renal cell carcino ma and endometrial stromal sarcoma. Am J Clin Pathol 2000; 113:374-82. 7. Hawley S, Fazli L, McKenney JK, Simko J, Troyer D, Nicolas M, et al. A model for the design and construction of a resource for the validation of prognostic prostate cancer biomarkers: The canary prostate cancer tissue microarray. Adv Anat Pathol 2013; 20:39-44. 8. Badve SS, Baehner FL, Gray RP, Childs BH, Maddala T, Liu ML, et al. Estrogen- and progesterone-receptor status in ecog 2197: Comparison of immunohistochemistry by local and central laboratories and quantitative reverse transcription polymerase chain reaction by central laboratory. J Clin Oncol 2008; 26:2473-81. 9. Tsao SC, Wu CC, Wen CH, Chai CY, Chen YT. Improved technique for manually constructing tissue microarrays for large-core arrays. Appl Immunohistochem Mol Morphol 2013; 21:85-9. 10. Rimm DL, Nielsen TO, Jewell SD, Rohrer DC, Broadwater G, Waldman F, et al. Cancer and leukemia group b pathology committee guidelines for tissue microarray construction representing multicenter prospective clinical trial tissues. J Clin Oncol 2011; 29:2282-90. 11. Camp RL, Dolled-Filhart M, Rimm DL. X-tile: A new bio-informatics tool for biomarker assessment and outcome-based cut-point optimization. Clin Cancer Res 2004; 10:7252-9. 12. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark GM. Reporting recommendations for tumor marker prognostic studies (remark). J Natl Cancer Inst 2005; 97:1180-4. which should be followed whenever possible. 151 Part II: The Potentials and Pitfalls Chapter 13 Immunohistochemical Visualization of Molecular Tests Jens Mollerup, PhD, MSc Jan Trøst Jørgensen, PhD, MSc Mo•le•cu•lar test (n.) The analysis of human DNA, RNA, and chromosomes to detect heritable or acquired disease-related genotypes, mutations, phenotypes, or karyotypes for clinical purposes, which include predicting risk of disease, identifying carriers, and establishing prenatal or clinical diagnoses or prognoses. McGraw-Hill Concise Dictionary of Modern Medicine Immunohistochemical Visualization of Molecular Tests | Chapter 13 Chapter 13.1 Introduction and visualization of the signals by deposition of chromogen in the tissue. Finally, counterstaining with hematoxylin is performed to Molecular tests are increasingly used in the diagnosis of can- enhance and visualize the nuclear borders. cers. The information gained from looking at the genetic characteristics of the cancer can be of importance for the treatment Haptens can be luorochromes such as FITC, Texas Red or decision. Many of the molecular tests, involving PCR (polymer- Cy-3 or they can be non-luorescent such as dinitrophenyl, ase chain reaction) or sequencing techniques, are performed biotin or digoxygenin. The enzymes most often used to gene- by specialized genetic laboratories or cytology laboratories. rate the deposition of chromogens are horseradish peroxidase However, with the introduction of the chromogenic technique (HRP) or alkaline phosphatase (AP). In silver-enhanced in situ that transforms the luorescent signals to signals that are visi- hybridization (SISH) the probe location is visualized as a black ble in a standard bright ield microscope, some of these tests coloration due to silver precipitation (4). have moved back to the pathology laboratory. This chapter will focus on the immunohistochemical (IHC) aspects of molecular testing for genetic aberrations in cancer. CISH FISH Target DNA (gene) Target DNA (reference) Target DNA (gene) Target DNA (reference) Fluorescent probe Fluorescent probe Fluorescent probe Fluorescent probe Enzyme labeled antibody Enzyme labeled antibody Chromogen Chromogen ISH (in situ hybridization) is the common name for a sensitive and robust technique that is mostly used to evaluate gene ampliications, deletions, and translocations, as well as chromosomal copy number changes, in tissue or cell material. The ability to visualize genetic rearrangements and quantify copy numbers of speciic genes in the morphological context of the tissue has widespread importance. ISH has been used as a research tool for more than 40 years (see 1) and the method has also gained Mount and read slides popularity as a diagnostic tool in clinical oncology, exempliied by determination of epidermal growth factor receptor 2 (ERBB2 /HER2) status in breast and gastric cancer (2, 3). ISH methods allow the speciic recognition of DNA target sequences in the nuclei of target cells using luorescence- or hapten-labeled sequence pairing probes. The techniques are referred to as FISH, CISH or SISH depending on whether luorochromes, chromogens or silver is used to visualize the probe. In FISH, luorescent probes are directly visible using luorescence Mount and read slides microscopy. In CISH, luorochromes or other labels attached to the probes are used as haptens for antibody-directed enzymatic deposition of chromogen that can be viewed using bright ield Texas Red-labeled DNA probe FITC-labeled FISH PNA probe Anti-Texas Red conjugated with alkaline phosphatase (AP) Anti-FITC conjugated with horse radish peroxidase (HRP) Red chromogenic detection Blue chromogenic detection microscopy. The principle is illustrated in Figure 13.1. In practice, the FISH procedure should be followed until just prior to the inal dehydration step. Instead of proceeding with dehydration, the slides are immersed in CISH wash buffer and are ready for CISH staining. CISH staining is a standard IHC staining with initial blocking of endogenous peroxidase activity, addition of a primary antibody mix containing antibodies for the luorescent labels Figure 13.1 A schematic illustration of the procedures for FISH and CISH. 153 Chapter 13 | Immunohistochemical Visualization of Molecular Tests Chapter 13.2 FISH versus CISH dure, as properly processed nuclei should expose the red to blue signal ratio expected of normal diploid cells. The results FISH has been used as a sensitive and reliable technique of a CISH test is an almost complete 1:1 conversion ratio of for evaluating gene status since 1986 (5). The publication of FISH signals to CISH signals (Figure 13.2), resulting in a high a chromogenic HER2 ISH procedure in 2000 was the irst to level of concordance between the two analyses. show that quantiication of HER2 copy numbers was possible in formalin-ixed, parafin-embedded breast cancer tissue using CISH (6). To obtain information regarding tumor aneuploidy at that time, a reference marker was visualized on a serial slide from the same specimen. Today, it is possible to get information regarding both HER2 and a reference marker (which typically is the centromere of chromosome 17) on the same slide using dual color CISH. This simpliies interpretation of cases with borderline gene copy number by offering simultaneous visualization of the two probes, e.g. the ability to distinguish between true gene ampliication/deletion and chromosomal aneuploidy on the same slide. The combination of a dual color procedure and bright ield evaluation is attractive, because it allows for easy interpretation of staining without the use of luorescence microscopes. Studies have shown that inter-laboratory and A B Figure 13.2 The images illustrate a “dot-to-dot” conversion of FISH signals (A) to CISH signals (B). The red lorescent signals are converted to red chromogenic signals and the green luorescent signals to blue chromogenic signals (Dako DuoCISH™). In practice the sample was stained in the FISH procedure, evaluated by luorescence microscopy and subsequently converted to CISH. Due to the nature of the image capture, i.e. by using an already mounted FISH slide to create the CISH stain, the CISH staining quality is compromised (12). observer-to-observer variability is reduced for interpretation of HER2 status using CISH, when compared to FISH; this advantage also applies when it comes to inexperienced observers (7-9). Compared to FISH, CISH has several other advantages: „„ Bright ield microscopes used for CISH are generally available in diagnostic laboratories, whereas luorescence microscopes are more complex, more expensive and require the use of dark rooms/light-restricted rooms „„ Bright ield CISH allows for good visualization of tissue structures with good ability to distinguish appropriate tumor areas „„ CISH signals do not fade over time and stained slides may be archived and re-evaluated, for retro-spective studies or educational purposes at a later point in time „„ Documentation by image acquisition of bright ield images is simpler compared to luorescence images Furthermore, the evaluation time for CISH specimens have been shown to be considerably shorter than for identical specimens evaluated by FISH (10, 11). This observation is likely to be due to the good visualization of tissue morphology in the bright ield microscope. Normal cells within the tissue may serve as an internal control for success of the staining proce- 154 Figure 13.3 A graphical morphing of two sections stained with HER2 FISH pharmDx™ and HER2 CISH pharmDx™ kits, respectively. Immunohistochemical Visualization of Molecular Tests | Chapter 13 Chapter 13.3 Principle of the CISH Procedure Red luorochrome and green for CEN-17 when using FITC CISH is considered an easy way to do molecular testing, rescent probe signals are converted to chromogenic signals but it should be kept in mind that the staining procedure has by an immunohistochemical staining procedure. First, en- several reagent and protocol steps that are critically impor- dogenous peroxidase is inactivated and then slides are in- tant for robust and correct staining results. The procedure cubated with an antibody mix, comprising HRP-labeled an- will vary depending on the vendor and the speciic marker ti-FITC and AP-labeled anti-Texas Red. Following antibody luorochrome. After a stringency wash and rinses, the luo- kit. The main steps in the procedure for Dako’s HER2 CISH incubation, red and blue signals are generated by incuba- pharmDx™ Kit are briely explained below, serving as an tion with the red and blue chromogen solutions. The enzy- example of the general principles of converting FISH signals matic activity and the presence of the reactive chromogens into CISH signals. result in formation of visible red and blue end-products at the target sites of the Texas Red- and FITC-labeled probes. After deparafinization and rehydration, the sections are Thus, red luorescent signals are converted to red chromo- heated in pre-treatment solution. The next step involves pro- genic signals and green luorescent signals are converted teolytic digestion using pepsin either at room temperature to blue chromogenic signals (13). The chromogenic signals or at 37 °C. The optimal pepsin incubation time depends on are clear and distinct and are supported by a hematoxylin the ixation history of the tissue and should be determined counterstain to enhance morphological features. The red by the user. Following the proteolytic step, sections are de- and the blue signals are insoluble in water and are mount- hydrated and the probes applied. Following co-denaturation ed in a water-based permanent mounting media. In Figure of section and probes, the hybridization is carried out. The 13.4, representative images of two breast cancer specimens speciic hybridization to the two target regions result in for- are shown. The signals emanating from a CISH staining are mation of a distinct luorescent signals at each HER2 gene readily apparent using low power objectives (10x and 20x) locus and distinct luorescent signals at each chromosome or can be clearly visualized and counted using high power 17 centromere, for example red for HER2 when using Texas objectives (40x, 60x or 100x). A B Figure 13.4 Examples of HER2 ampliied (A) and non-ampliied (B) FFPE breast cancer sections visualized with a 60x objective. Stained with Dako HER2 CISH pharmDx™ Kit. 155 Chapter 13 | Immunohistochemical Visualization of Molecular Tests To determine the HER2 gene status, the invasive component of the tumor cells is located in the bright ield microscope, and enumeration of the red (HER2) and blue (CEN-17) signals is The use of validated products and procedures are important conducted. The HER2/CEN-17 ratio is calculated based on the to provide the best possible assay result and subsequent counting of at least 20 nuclei. Normal cells within the analyzed treatment guidance in clinical decisions. Following the intro- tissue section serve as an internal positive control of the stain- duction of molecular testing using CISH, a large number of ing process and the pre-analytical treatment of the specimen. studies have reported on the agreement between FISH and When performing the staining reaction described above, a CISH/SISH with respect to HER2 gene status in breast and complete one to one conversion of FISH signals to CISH gastric cancer tissue. However, many of these studies are signals takes place as illustrated in Figure 13.2. In this par- small and include a relatively low number of tissue samples. ticular staining, a FISH staining was irst completed, luo- In order to identify the larger published concordance studies rescence pictures taken, coverslip removed and the section between FISH and CISH/SISH a PubMed search was under- was washed prior to performing the CISH immunohisto- taken. Studies that contained 200 or more breast or gastric chemical staining reaction. Then, new bright ield pictures cancer tissue samples were selected if the publication was were taken to allow comparison of identical tissue areas as in English. Since the main purpose of the study was a direct illustrated. Many publications have shown a very high level comparison between FISH and CISH/SISH, the percent over- of agreement between the ratios for HER2/CEN-17 by FISH all agreement should have been calculated. In Table 13.1 the and CISH (see Table 13.1). result of the literature search is presented. Ten concordance studies were indentiied that fulilled the search criteria and Table 13.1 Concordance between FISH and CISH/SISH. All the studies listed in the table have evaluated for overall agreement with respect to HER2 gene status between the two ISH methods. CISH: Chromogenic in situ hybridization, SISH: Silver-enhanced in situ hybridization. Study 156 Sample size (n) Tissue Assay Concordance (% Overall Agreement) the number of tissue samples in these studies ranged from 200 to 588 with the overall agreement percentage ranging from 93.0% to 100.0%. The literature survey showed that the overall agreement between the FISH and CISH/SISH assays is generally very high, with 7 out of 10 studies having an Park YS, et al (14) 588 Gastric SISH 98.3 agreement percentage of 95% or higher. Lee Y, et al (15) 543 Breast SISH 96.7 For many years, HER2 testing in the clinic has been done us- Riethdorf S, et al (16) 399 Breast CISH 94.9 Mollerup J, et al (10) 348 Breast CISH 98.3 Hyun CL, et al (17) 309 Breast CISH 96.8 Hanna WM, et al (18) 254 Breast CISH 95.3 Hwang CC, et al (19) 227 Breast CISH 100.0 Gong Y, et al (20) 226 Breast CISH 98.5 Brügmann A, et al (21) 201 Breast CISH 93.0 Sàez A, et al (22) 200 Breast CISH 94.8 ing IHC with relex testing using FISH, or with FISH as the initial test and most of the important clinical studies done with HER2-targeted compounds in HER2-positive cancer patients have used the traditional IHC and/or FISH assays for patient selection (23-28). However, correlation between HER2 status measured by CISH and response to trastuzumab (Herceptin®, Roche/Genentech) has now been demonstrated in clinical studies (29, 30). One of these studies was the FinHer Study in primary breast cancer, which showed an improved three-year, recurrence-free survival for the HER2-positive patients treated with trastuzumab. Currently, four different ISH products are approved by the FDA as in vitro companion diagnostic devices that can be used to provide predictive information prior to the use of a particular therapeutic product (see Table 13.2). Immunohistochemical Visualization of Molecular Tests | Chapter 13 Table 13.2 Overview of dual color in situ hybridization products currently assigned as companion diagnostic devices by the FDA. Accessed May 31, 2013. (www.fda.gov/MedicalDevices/ProductsandMedicalProcedures/InVitroDiagnostics/ucm301431.htm) 1 Trade Name Manufacturer Assay type Tissue Drug tradename HER2 CISH pharmDx™ Kit Dako, Agilent Technologies CISH Dual color Breast Herceptin® (trastuzumab) INFORM HER2 DUAL ISH DNA Probe Cocktail Ventana Medical Systems, Inc. SISH/CISH Dual color Breast Herceptin® (trastuzumab) HER2 IQFISH pharmDx™ Dako, Agilent Technologies FISH Dual color Breast, Gastric Herceptin® (trastuzumab); Perjeta™ (pertuzumab)1; Kadcyla™ (ado-trastuzumab emtansine)1,2 PathVysion HER-2 DNA Probe Kit Abbott Molecular Inc. FISH Dual color Breast Herceptin® (trastuzumab) Breast cancer only, 2Approved March 2013, in US only Finally, from an analytical and a clinical viewpoint, HER2 CISH assays have shown to provide results similar or identical to FISH assays and it must be concluded that CISH is a reliable alternative to FISH for HER2 testing. Acknowledgements Sections, in whole or parts thereof, from the previous editions of this Guidebook are used in the 6th edition. We sincerely thank and acknowledge the contribution of the authors. Special acknowledgements to: Ulla Henriksen, Sven Müller and Andreas Schönau. References 1. Holer H. What's new in "in situ hybridization". Pathol Res Pract 1987; 182:421-30. 2. Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ, et al. American society of clinical oncology/college of american pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol 2007; 25:118-45. 3. Jørgensen JT. Targeted her2 treatment in advanced gastric cancer. Oncology 2010; 18:26-33. 4. Powell RD, Pettay JD, Powell WC, Roche PC, Grogan TM, Hainfeld JF, et al. Metallographic in situ hybridization. Hum Pathol 2007; 38:1145-59. 5. Pinkel D, Straume T, Gray JW. Cytogenetic analysis using quantitative, high-sensitivity, luorescence hybridization. Proc Natl Acad Sci U S A 1986; 83:2934-8. 6. Tanner M, Gancberg D, Di LA, Larsimont D, Rouas G, Piccart MJ, et al. Chromogenic in situ hybridization: A practical alternative for luorescence in situ hybridization to detect her-2/neu oncogene ampliication in archival breast cancer samples. Am J Pathol 2000; 157:1467-72. 7. van de Vijver M, Bilous M, Hanna W, Hofmann M, Kristel P, Penault-Llorca F, et al. Chromogenic in situ hybridisation for the assessment of her2 status in breast cancer: An international validation ring study. Breast Cancer Res 2007;9:R68. 8. Di Palma S, Collins N, Bilous M, Sapino A, Mottolese M, Kapranos N, et al. A quality assurance exercise to evaluate the accuracy and reproducibility of cish for her2 analysis in breast cancer. J Clin Pathol 2008; 61:757-60. 9. Bilous M, Morey A, Armes J, Cummings M, Francis G. Chromogenic in situ hybridisation testing for her2 gene ampliication in breast cancer produces highly reproducible results concordant with luorescence in situ hybridisation and immunohistochemistry. Pathology 2006; 38:120-4. 10. Mollerup J, Henriksen U, Muller S, Schonau A. Dual color chromogenic in situ hybridization for determination of her2 status in breast cancer: A large comparative study to current state of the art luorescence in situ hybridization. BMC Clin Pathol 2012; 12:3. 11. Bhargava R, Lal P, Chen B. Chromogenic in situ hybridization for the detection of her-2/neu gene ampliication in breast cancer with an emphasis on tumors with borderline and low-level ampliication: Does it measure up to luorescence in situ hybridization? Am J Clin Pathol 2005;123:237-43. 12. Bang Y, Chung H, Xu J, Lordick F, Sawaki A, Al-Sakaff N, et al. Pathological features of advanced gastric cancer (gc): Relationship to human epidermal growth factor receptor 2 (her2) positivity in the global screening programme of the toga trial. J Clin Oncol (Meeting Abstracts). 2009;27:4556. 157 Chapter 13 | Immunohistochemical Visualization of Molecular Tests 13. Petersen KH. Novel horseradish peroxidase substrates for use in immunohistochemistry. J Immunol Methods 2009; 340:86-9. 14. Park YS, Hwang HS, Park HJ, Ryu MH, Chang HM, Yook JH, et al. Comprehensive analysis of her2 expression and gene ampliication in gastric cancers using immunohistochemistry and in situ hy bridization: Which scoring system should we use? Hum Pathol 2012; 43:413-22. 15. Lee Y, Ryu Y, Jeong H, Chang H, Kim Y, Kim A. Effectiveness of silver-enhanced in situ hybridization for evaluating her2 gene status in invasive breast carcinoma: A comparative study. Arch Med Res 2012; 43:139-44. 16. Riethdorf S, Hoegel B, John B, Ott G, Fritz P, Thon S, et al. Prospective multi-centre study to validate chromogenic in situ hybridisation for the assessment of her2 gene ampliication in specimens from adjuvant and metastatic breast cancer patients. J Cancer Res Clin Oncol 2011;137:261-9. 17. Hyun CL, Lee HE, Kim KS, Kim SW, Kim JH, Choe G, et al. The effect of chromosome 17 polysomy on her-2/neu status in breast cancer. J Clin Pathol 2008;61:317-21. 18. Hanna WM, Kwok K. Chromogenic in-situ hybridization: A viable alternative to luorescence in-situ hybridization in the her2 testing algorithm. Mod Pathol 2006;19:481-7. 19. Hwang CC, Pintye M, Chang LC, Chen HY, Yeh KY, Chein HP, et al. Dual-colour chromogenic in-situ hybridization is a potential alternative to luorescence in-situ hybridization in her2 testing. Histopathology 2011;59:984-92. 20. Gong Y, Sweet W, Duh YJ, Greenield L, Fang Y, Zhao J, et al. Chromogenic in situ hybridization is a reliable method for detecting her2 gene status in breast cancer: A multicenter study using conventional scoring criteria and the new asco/cap recommendations. Am J Clin Pathol 2009;131:490-7. 21. Brugmann A, Lelkaitis G, Nielsen S, Jensen KG, Jensen V. Testing her2 in breast cancer: A comparative study on brish, ish, and ihc. Appl Immunohistochem Mol Morphol 2011;19:203-11. 22. Saez A, Andreu FJ, Segui MA, Bare ML, Fernandez S, Dinares C, et al. Her-2 gene ampliication by chromogenic in situ hybridisation (cish) compared with luorescence in situ hybridisation (ish) in breast cancer-a study of two hundred cases. Breast 2006;15:519-27. 23. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Ba jamonde A, et al. Use of chemotherapy plus a monoclonal antibody against her2 for metastatic breast cancer that overexpresses her2. N Eng J Med 2001;344:783-92. 24. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, Goldhirsch A, Untch M, Smith I, et al. Trastuzumab after adjuvant chemotherapy in her2-positive breast cancer. N Engl J Med 2005;353:1659-72. 25. Romond EH, Perez EA, Bryant J, Suman VJ, Geyer CE, Jr., Davidson NE, et al. Trastuzumab plus adjuvant chemotherapy for operable her2-positive breast cancer. N Engl J Med 2005;353:1673-84. 158 26. Bang YJ, Van Cutsem E, Feyereislova A, Chung HC, Shen L, Sawaki A, et al. Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of her2-positive advanced gastric or gastro-oesophageal junction cancer (toga): A phase 3, open-label, randomised controlled trial. Lancet 2010;376:687-97. 27. Baselga J, Cortes J, Kim SB, Im SA, Hegg R, Im YH, et al. Pertuzumab plus trastuzumab plus docetaxel for metastatic breast cancer. N Engl J Med 2012;366:109-19. 28. Verma S, Miles D, Gianni L, Krop IE, Welslau M, Baselga J, et al. Trastuzumab emtansine for her2-positive advanced breast cancer. N Engl J Med 2012;367:1783-91. 29. Joensuu H, Kellokumpu-Lehtinen PL, Bono P, Alanko T, Kataja V, Asola R, et al. Adjuvant docetaxel or vinorelbine with or without trastuzumab for breast cancer. N Engl J Med 2006;354:809-20. 30. Francis GD, Jones MA, Beadle GF, Stein SR. Bright-ield in situ hybridization for her2 gene ampliication in breast cancer using tissue microarrays: Correlation between chromogenic (cish) and automated silver-enhanced (sish) methods with patient outcome. Diag Mol Pathology 2009;18:88-95. 159 Part II: The Potentials and Pitfalls Chapter 14 Controls Ole Feldballe Rasmussen, PhD, MSc Rikke Jørgensen, PhD, MSc Con•trol (n.) A standard of comparison for checking or verifying the results of a scientiic experiment. The American Heritage® Dictionary of the English Language Controls | Chapter 14 Chapter 14.1 Introduction nosis. The controls will show whether all reagents are in good order, and very importantly, they will also show if the staining system Many factors may introduce variations in immunohistochem- is working correctly and thus demonstrate potential day-to-day, istry including ischemic time, tissue ixative and ixation time, operator-to-operator and/or instrument-to-instrument variations. tissue processing, eficiency of epitope retrieval, selection of antibody/antibody clone, detection system, instrumentation – For many of the potential variations there are other measures in and not least interaction among the many different sources of place to safeguard; unfortunately, it is not always possible to ac- potential variation. count for extreme situations or equipment defects. Reagent expiry is one example. Reagent expiry dates are determined exper- All major suppliers of diagnostic systems to the pathology lab- imentally and extensively documented during development and oratories have implemented measures to safeguard the quality veriication; testing includes extensive transportation scenarios, of their systems. These measures are implemented in develop- with changes in temperature and prolonged storage at increased ment and validation as well as during manufacturing and suppli- temperature. However, such investigation will never be able to er quality control. However, with the many potential factors that encompass all situations, e.g. that a package is lost for weeks may inluence an immunohistochemical staining, it cannot be as- during shipment. Likewise, a defective refrigerator can impact re- sumed that any given IHC staining is correct. Consequently, it is agent quality. Sub-optimal reagent quality should be detected by important to include controls for veriication of results for in vitro appropriate controls as part of the staining process. diagnostic use. It is also important to understand what information a given control can provide or not provide. Many countries have regulatory requirements for diagnostic methods, reagents, instruments and software, as well as vendor and laboratory responsibilities. For each laboratory it is important to understand and implement these requirements accordingly. For reference, a thorough description on Quality Management and Regulation in the US is given in (1). Vendor responsibilities In addition to country-speciic requirements, the site of manufacture may also impact vendor responsibilities. In the EU, vendor responsibilities are determined by a speciic IVD directive (98/79/EC). Basically, there are requirements for development and validation, as well as for manufacturing and quality control of IVD products. Key parameters that must be documented are included in the data/speciication sheet or corresponding product speciic documentation. For example, antibody vendors Figure 14.1 This chapter will describe the range of controls that should be used in a diagnostic laboratory. must ensure appropriate speciicity using a range of immunochemical techniques, including immunoelectrophoresis, Western blot, double diffusion immunoassays and ELISA. Testing on transgenic cells expressing the speciic, as well as closely Chapter 14.2 Purpose of Controls related, antigens may also be performed. It is, however, imperative also to include testing on tissue; both cancer tissue and In short, IHC staining controls are required for validation of stain- normal tissue, for high and low expression structures (see also ing results to document and/or ensure that each and every IHC Chapter 5). For primary antibodies, manufacturers typically irst test is performed correctly and can be used to support the diag- test antibodies on a range of positive tissues to identify the op- 161 Chapter 14 | Controls timal antibody dilution in combination with the chosen staining For every immunohistochemistry staining run, at least one control protocol. Next, immunohistochemistry testing is extended to an should be included. This is straightforward when staining is car- expanded panel of additional tissues known to either contain, ried out in batch mode (see Chapter 9), however with continuous or not contain, targeted antigens. For new antibody lots, manu- slide loading, “a staining run” must be deined in a different way, facturers typically perform quality control to verify the speciicity e.g. once per day, or once per x number of slides. The staining and sensitivity that were documented during development. run control may be expanded to at least one per antibody, and over the past years there has been an increasing use of control Laboratory responsibilities material on every slide (see below for further discussion). Laboratories performing IHC assays hold key responsibilities relating to controls: „„ Laboratory validation of new antibodies, protocols and systems „„ New reagent lot veriication „„ Performance control for each run The saying, “if it isn’t documented, it didn’t happen”, directly applies to the staining control mentioned. It is important that all procedures and results – as well as procedural changes, are documented and proper record-keeping procedures are applied. In addition, it is advisable to implement a trend-tracking procedure to The exact responsibilities vary depending on the country and capture small changes over time, e.g. running a standard at reg- accreditation program. If the laboratory is enrolled in the US ular intervals. Many labs track the relative percentages of HER2 CAP’s (College of American Pathologists) laboratory accred- scores (0, 1+, 2+, 3+) to use as a indirect measure of assay per- itation program, the laboratory must comply with a speciic formance over time. Changes in the relative percentage of HER2 checklist (2). Here it is speciied that “The performance char- scores may warrant further investigation of the cause. If results acteristics of each assay in the immunohistochemistry labo- may be inluenced by changing environmental conditions, it is also ratory must be appropriately validated before being placed advisable to keep track of the relative humidity and temperature. into clinical use”. The goal is to verify vendor speciications and the optimal staining protocol, as part of validating overall clinical performance. It is important to test a panel of positive and negative tissues to determine assay sensitivity and spec- Chapter 14.3 Categories of Controls and Control Material iicity. With respect to the number of tissues to be included for validation of an antibody, 10 positive and 10 negative tissues The terminology for the various control types are in some parts are suggested for well-characterized antibodies with a limited of the literature not clearly deined. This chapter uses the ter- number of targets/applications. A recent study investigated minology emphasized by Torlakovic (4). the use of immunohistochemistry validation procedures and practices for non-predictive assays (3). The study was based on 727 responders to a questionnaire composed of 32 items. Key results were that 68% of the laboratories had written vali- Control material can be categorized in different ways: „„ Reagent controls – Positive reagent control is the actual reagent; e.g. the dation procedures, and 86 % had validated the most recently primary antibody when tested on control material. Testing introduced antibody. With respect to the number of different is done during development and product validation to ensure reagent speciicity and sensitivity. tissues included in the validation, 75% used 21 or fewer cases for the validation. – Negative reagent control is a reagent substitute for the primary antibody used to assess potential speciicity New reagent lot veriication is also addressed in the CAP Anatomic Pathology Checklist, where it is stated that “the performance of new lots of antibody and detection system reagents 162 issues/false positive staining reaction „„ Tissue controls – Positive tissue control is tissue with the speciic antigen is compared with old lots before or concurrently with being at known, relevant and stable level. Purpose is to docu- placed into service”. ment correct staining. Controls | Chapter 14 – – Negative tissue control is tissue without the speciic an- impractical to devise and run controls for every combination tigen present – or not present in speciic regions. Purpose of pre-analytical and analytical factors, data on controls in di- is to document speciicity of the staining. agnostic immunohistochemistry, as well as in research, must Tissue blocks/tissue microarrays, consisting of a few to be included for valid and reproducible test results. Correctly several different tissues that may serve as control mate- used controls can reveal if the staining protocols were run rial for a range of antibodies (see Chapter 12) „„ Internal tissue controls – correctly, and help determine day-to-day and operator-to-operator variations. The presence of the target antigen (protein) within normal elements of the tissue under investigation is an internal positive control „„ Tissue process control – Constitutively expressed proteins generally expressed in all tissues at the same level. Such a protein may serve as control for not only the staining process but also for the pre-analytical steps. „„ Cell line controls – Control material based on cultured cells with speciic antigenic characteristics. This type of control typically is made for very speciic purposes, in particular predictive assays. Chapter 14.4 Use of Controls in Daily Routine Testing The many sources of variations, as well as the potentially critical consequences thereof, call for standardization, assay validation and use of controls. Although awareness of the com- A B C Figure 14.2 On-slide positive tissue controls, where speciic controls are placed on the same slide as the test specimen are strongly preferred. A) Illustration of the principle of multi-core blocks. Tissue controls are the cores in red, yellow, gray and green. Patient sample is placed below B) Example of on-slide tissue controls. C) Example of cell line controls. plexity of immunohistochemical assays and the inevitable variation is increasing, a meta-analysis of 100 peer-reviewed papers on immunohistochemistry revealed that only 13% of the Reagent controls articles described positive and negative controls run on identi- Negative reagent control for polymer-based detection systems cally prepared samples. More than half of the articles either did Since the start of IHC, it has been good practice to include not mention controls, or did not run controls as separate speci- a negative reagent control to check for speciicity of the rea- mens (5). The papers included in the meta-analysis have both gents. In recent years, there has been a signiicant improve- diagnostic and research focus, underlining the importance of ment in both sensitivity and speciicity of detection systems. A paying proper attention to inclusion of controls in both types of key element has been development of polymer-based, biotin- IHC applications. free visualization systems, which has reduced the need for negative reagent controls for such detection systems. In fact, This section will elaborate on the different types of controls recently, the US CAP organization changed its recommenda- used in IHC and their applications in the diagnostic labo- tion for US laboratories, now stating that for polymer-based (bi- ratory. Controls are a necessity for proper interpretation of otin-free) systems, negative reagent controls may be omitted staining, and omitting them will entail a risk. Although it is at the discretion of the laboratory director (2). 163 Chapter 14 | Controls Currently, most antibody speciication sheets from major vendors NOTE: A special situation worthy of note is when a panel of two include suggestions for appropriate negative reagent controls, or more antibodies (of the same species) is applied to serial sec- with no reference to choice of detection system; and as stated tions, using the same detection system. In this instance, negative- in the US CAP recommendation, the negative reagent control ly stained areas of one slide may serve as the negative/non-spe- should be omitted only when given appropriate consideration. ciic binding background control for other antibodies in the panel. Thus, laboratories must still pay attention to potential artifacts In cases where it is necessary to evaluate non-speciic bind- or special cases, for example occurrence of non-speciic DAB ing, potentially caused by sources other than the primary an- precipitates when performing staining of small bacteria or vi- tibody, additional patient tissue sections may be stained with ruses. DAB precipitates may resemble small stained elements, selected reagents. For example, tissues may be stained with which could complicate interpretation of the staining result. A just the secondary antibody and/or the enzyme, followed by negative control would help determine if the staining seen in application of the substrate/chromogen. In cases where the the patient sample is a false positive result. suspected non-speciic staining may be the result of endogenous enzyme present within the tissue, this possibility can be Negative reagent control for biotin-based detection systems conirmed by application of the substrate/enzyme only. For tests using biotin-based detedtion systems, a negative reagent control test should still be included in all runs, due to Tissue Controls the inherent possibility of endogenous reactivity. This test has As noted in above, both positive and negative tissue controls to be done on a separate section of patient tissue, using the provide important information and must be included in daily same staining protocol with same target retrieval and detection routine regardless of type of detection system used. Typical- system, except that the primary antibody is replaced by the ly, the tissues being used for controls are carefully selected negative reagent control. normal or cancer tissues previously analyzed in the laboratory. However, in some cases the actual patient tissue being For monoclonal mouse primary antibodies, the negative re- tested can be used as tissue control. For all tissue controls, agent controls may be developed by different methods. The if the staining does not perform as expected, results from the optimal method is an antibody of the same isotype, present respective test specimen should be considered invalid. in the same immunoglobulin concentration, using the same diluent and exhibiting no speciic reactivity with the given Positive tissue control human tissues tested. A less optimal alternative is to use mix- The positive tissue control must contain the target antigen at rel- tures of antibodies representing all or most relevant IgG sub- evant, known and stable expression level. It serves to document types. Finally, the diluent itself may also be used as an alterna- that proper staining has been performed and conirms that the tive which, however, is neither eficient nor desirable. target retrieval procedure has been carried out correctly. Thus, a positive tissue control assesses correct staining protocol perfor- For polyclonal antibodies, negative reagent controls should be mance (temperature, time and correct application of reagents). a dilution of immunoglobulin fractions, or whole normal (non-immune) serum of the same animal source. Again, the negative As positive tissue controls are indicative of properly prepared reagent control should be applied in the same concentration as tissue, they should be as accurate as possible in the same the test antibody, and the same diluent should be used. manner as the patient samples. Optimally, autopsy/biopsy/surgical specimens should be ixed, processed and embedded 164 Using the same protocol as the primary antibody, the nega- as soon as possible for best preservation of antigens. Gener- tive reagent control should be applied to a sequential section ally, autopsy tissue is least preferred because of inevitable de- of each patient specimen, to evaluate nonspeciic staining in lays before ixation (cold ischemia, with degradation of some that particular tissue. antigens, see also Chapter 2). Controls | Chapter 14 Furthermore, the positive controls should be properly docu- separate slide - as a batch control or daily control, or they may mented, including sectioning date and storage conditions for be included on the same slides as the test specimens. Using the control material. Special attention should be given to the sta- batch run controls, with the control section on a separate slide, bility of the speciic antigen(s) to be controlled, as antigen deg- the control slide will be indicative that the staining process has radation occurs in cut, unstained slides, at different rates de- run correctly. However, batch controls will not be able to iden- pendent on the antigen (6). A different challenge may be access tify a missing reagent (false negative), wrong antibody (false to optimal tissue control material; in particular in cases where positive) or inappropriate protocol (false negative or false posi- normal tissue may not be used, e.g. for certain viral infections. tive) on any other slides run on the instrument in that batch oth- Finally, the use of patient material for control tissue must comply er than on the control slide itself. Consequently, on-slide pos- with local or regional regulations. itive controls, where speciic controls are placed on the same slide as the test specimen are strongly preferred (Figure 14.2). Ideally these controls should contain a spectrum of weak to strongly positive staining reaction. If such tissue is not avail- More and more laboratories are using TMA cores as on-slide able, at least a weakly positive tissue should be used, as this controls. It is possible to build a few basic multi-tissue control provides the best basis to evaluate whether a particular stain- blocks each containing a small number of control tissues to ing reaction is optimal (see also Chapter 4 and Chapter 5). cover most of the markers used in clinical diagnostics. NordiQC, an independent scientiic organization arranging quality On-slide positive tissue controls schemes for pathology laboratories, has made a suggestion In the daily routine, positive tissue controls may be run on a for set of four multi-tissue control blocks each containing four Colon, HE Tongue, LE Case 3 A Case 2 B Case 1 Figure 14.3 Normal colon from three different cases stained with Anti-Actin, clone HHF35. Smooth muscle cells in vessel wall, muscle layers and lamina muscularis mucosa, deined as high expression (HE) structure, show comparable staining intensity in all three samples. B) Normal tongue from three different cases stained with Anti-Actin, clone HHF35. The myoepithelial cells of the mucous/salivary glands, deined as low expression (LE) structure, show comparable staining intensity in all three samples. 165 Chapter 14 | Controls different – and preferably normal – tissues that can be used when selecting the normal tissue that is optimal as control for most of the markers used in clinical diagnostics. For each tissue, by using a validated protocol that is able to identify block, the speciic antibodies covered are speciied, together variations in antigen expression. with the reaction pattern on each control tissue, as well as quality indicators for diagnostic use if available (7). As an example, Negative Tissue Control one of the multi-tissue control blocks contains appendix, tonsil, A negative tissue control is tissue that lacks the speciic antigen pancreas and liver, and the reaction patterns are described for – or where the antigen is not present in speciic regions. It serves approximately 100 different antibodies. to ascertain speciicity of the staining. If using the ‘on-slide TMA method’, one of the cores should be negative, or contain struc- Dako has created an Atlas of Stains book that for each an- tures/cells that are expected to be negative for the antibody that tibody in the FLEX RTU (Ready-To-Use) series speciies rec- is being used. The negative tissue control must be included ommended control tissues with low expression (LE) and high to identify the correct speciicity of the antibody; showing no expression (HE) quality indicator structures to use. In addi- staining of structures or cells that are known to lack the anti- tion,matching descriptions and images of the expected stain- gen. The negative tissue control also serves for identiication ing patterns are included. of sub-optimal protocol (high background). Just as for positive controls, tissue used for negative controls should be prepared in Example of positive tissue controls the same manner as the patient sample. The shown example is for Anti-Actin, clone HHF35, where the deined HE and LE structures are present in different tissue Internal Tissue Controls types; in colon all smooth muscle cells in vessel wall, muscle Positive internal controls, also known as ‘built-in’ or intrinsic layers and lamina muscularis mucosa are deined as HE (Fig- controls, contain the target antigen within normal tissue ele- ure 14.3A), whereas LE is the myoepithelial cells of the mu- ments, in addition to the tissue elements to be evaluated. This cous/salivary glands in tongue (Figure 14.3B). circumstance appears ideal, as the tissue elements to be evaluated have been treated exactly as the positive control. However, the level of target in the internal positive control is not pre- where antigen expression varies too much to be deined as determined, and may or may not be as stable as the external stable. Using the same Anti-Actin as example, normal liver tissue control. Thus, when analyzing the slide, careful assess- tissue shows staining of perisinusoidal smooth muscle cells ment of the internal positive control is important. Obviously, if – but only in some samples (Figure 14.4). This variability em- the test slide only contains tumor tissue, an internal control is phasizes that all normal tissues may not, by default, be suita- not an option. One example of a positive internal tissue control ble as control tissue. Detailed analysis should be carried out is weak expression of the estrogen receptor in normal breast Liver However, evaluation of various normal tissues can show cases Figure 14.4 Normal liver from three different cases. The staining intensity of perisinusoidal smooth muscle cells varies from weak to negative, and is consequently a poor control tissue due to the variable antigen expression between tissue samples. 166 Controls | Chapter 14 ducts. Thus, if the section to be tested is selected to include of controls that makes tissue process control effective in daily normal benign duct structures, these cells will be an excellent routine exists. positive control. Another example is presence of S-100 protein in both melanoma and in normal tissue, such as peripheral nerves and dendritic reticulum cells (Figure 14.5). Internal Chapter 14.6 Cell Line Controls controls can of course also function as negative controls when cells known to lack the antigen in question are used. Several FDA-approved companion diagnostics contain cell line controls as part of the kit, or sold separately. These cell lines are speciically developed to monitor staining of the antigen of interest and should be included in all staining runs as an additional protocol control. Just as with tissue controls, cell line controls may be positive or negative. Positive cell line controls monitor staining performance by assessing target retrieval, blocking, antibody incubation and visualization. Negative cell line controls assess speciicity and, depending on the characteristics of the chosen cell line, may also provide information on performance. It must be emphasized that because processing differs for these cell lines, they do not serve as control for pre-analytic variables; only internal controls do that. Figure 14.5 Normal appendix stained with Anti-S-100. The satellite cells and the Schwann cells of the peripheral nerves show a moderate to strong staining reaction. An ideal negative cell line control will contain an amount of target antigen that is suficiently low to produce no staining if the procedure has been performed correctly. At the same Quantiied, ‘built in’, or internal, proteins may potentially serve time, the amount may be suficiently high to produce a weakly as calibrators for quantiication in IHC (Quantiiable Internal positive stain if the run has been performed under conditions Reference Standard – see Chapter 1, and (9), in much the that produce an excessively strong staining. same way as housekeeping genes, such as actin, serve as internal reference controls for expression studies (8, 9). An ideal positive cell line control would contain a number of target antigens producing staining of medium intensity. This would allow the control to assess both stains that are too weak Chapter 14.5 Tissue Process Control and stains that are too strong. The controls described in section 14.4 all serve to document An example of the way in which cell line controls can be used is the staining process. However, the pre-analytic steps, in par- illustrated by Dako’s HercepTest™ kit, which contains three cell ticularly the ischemic time and ixation, can have signiicant line controls: 0 (negative), a 1+ (weak staining) and a 3+ (strong impact on the capability of the tissue to provide good stain- staining). All are placed on the same microscope slide. The scale ing. Unfortunately, it has not yet been possible to develop ef- used for HER2 scoring (0, 1+, 2+, 3+) is arbitrary, and does not icient controls that are optimal for the pre-analytic process. It represent the proportional difference in target molecules per has been speculated to use constitutively expressed proteins cell. The control cell line with 3+ score has an estimated num- present in the test sample (8, 9). Battifora (8) suggested the ber of 1,400,000 to 2,390,000 receptors per cell, up to a 30-fold use of vimentin for this purpose. However, no broad system increase over the number of receptors expressed by the 1+ cell 167 Chapter 14 | Controls line (12, 13) (Table 14.2). It is important to note that the staining ciic slides. In addition, should the reading of control or test slides intensity of the control slides should not be used to interpret the give any reason for concern, the staining process data are likely staining intensity in patient samples by comparison. The cell lines to indicate which step may have introduced an error. Furthermore, are used only to indicate the validity of the staining run. the staining process data may be used to assess the stability of the instrument and indicate if one parameter may be drifting, so Table 14.2 The correlation between the number of HER2 receptors per cell and the resulting IHC score (9, 10). Cell line HER2 receptors IHC Score Staining Pattern MDA-231 21,600 0 No discernible membrane staining MDA-175 92,400 1+ Faint, discontinuous membrane staining SK-BR-3 1,400,0002,390,000 3+ Moderate/strong, complete membrane staining that service can be applied if necessary (see also Chapter 16). Chapter 14.8 External Quality Assurance Programs Most routine IHC staining results have no common quantitative measures. Instead, diagnostic IHC results are typically qualitative (yes/no answer) and are based on subjective interpretation by pathologists of varying experience (11). Quality control and assurance therefore remain crucial, and warrant high attention Negative Cell Line (MDA-231) 1+ Cell Line (MDA-175) 3+ Cell Line (SK-BR-3) by both manufacturers and the pathology laboratory. Internal quality control procedures in the individual laboratory are important for the reproducibility of the IHC performance, i.e. they will determine if the protocols, reagents and interpretations are working the same way, from day to day, and from individual to individual. However, internal quality control procedures will not necessarily identify if the staining is sub-optimal and if the IHC test Figure 14.6 Cell line controls for Dako HercepTest™. is sub-optimally calibrated an IHC result may be obtained, it just may not be correct (12). Consequently, it is important for laboratories to participate in proiciency testing through an external quality Chapter 14.7 Monitoring the Staining Process assessment (EQA) scheme. The EQA organization independently assess staining results from a range of laboratories, and identify New staining instruments from several manufacturers, exem- good staining quality as well as insuficient or poor staining, plus pliied by the Dako Omnis and Autostainer Link 48, collect in- inappropriate protocols and possible interpretation problems. formation concerning a range of the key staining parameters; EQA schemes assess staining results by sending out tissue sam- either on a batch, slide rack, or single slide basis. The type of ples to be stained using the laboratory’s routine IHC staining pro- information collected may include onboard reagent tempera- cedures. Laboratories then return their results, and in some cases ture and time, to monitor reagent stability, target retrieval time examples of the stained slides, which are compared with all other and temperature. The data include temperature ramp-up and participating laboratories and summarized in a inal report. ramp-down time, staining incubation time, and temperature at each individual step, as well as reagent use. Today, monitoring of the staining process cannot replace the controls discussed above; however, it can provide information prior to inspection of control slides that might indicate invalidity of the staining run, or indicate that special attention must be paid to spe- 168 Selected EQA schemes „„ Canadian Immunohistochemistry Quality Control (Canada) – www.ciqc.ca „„ College of American Pathologists (CAP) (United States) – www.cap.org „„ NordiQC (Based in Denmark) – www.nordiqc.org Controls | Chapter 14 „„ UK NEQAS (United Kingdom) – www.ukneqas.org.uk ries for the burden of both identifying proper tissue and internal manufacture of the control slides. However, such systems serve Similar programs are introduced in Australia, Austria, Germa- to validate only the analytic phase of IHC, and do not serve to ny, and other countries or regions around the world. control for pre-analytic error such as internal reference controls – if developed to routine use may do. A inal note is that availability of such systems is likely to speed up the adoption of on-slide Chapter 14.9 Future Aspects controls – one critical factor, of course, being the price. To ensure diagnostic certainty, IHC quality control will remain a References very important and integrated part of the daily routine. Across 1. Prichard J. Quality management and regulation. In: Handbook of practical immunohistochemistry. New York: Springer Science; 2011. the world, there is an increasing requirement for laboratories to be accredited, and often accreditation includes a need for participation in proiciency testing. Thus the EQA schemes will see increased participation – and importance – becoming more widespread across the world. We also see that on-slide controls will be commonplace to provide a slide speciic control, the challenge here being the workload and associated cost in 2. CAP. Anatomic pathology checklist. www.cap.org: College of American Pathologists; 2013. p. 34-8. 3. Hardy LB, Fitzgibbons PL, Goldsmith JD, Eisen RN, Beasley MB, Souers RJ, et al. Immunohistochemistry validation procedures and practices: A college of american pathologists survey of 727 laboratories. Arch Pathol Lab Med 2012;137:19-25. making and placing the on-slide controls. 4. Torlakovic EE. Quality control by tissue microarray in immunohisto chemistry. J Clin Pathol 2012 [Letter, 15 June 2012]. As discussed in Chapter 7, there is an increasing use of digital 5. Colasacco C, Mount S, Leiman G. Documentation of immunocytochemistry controls in the cytopathologic literature: A meta-analysis of 100 journal articles. Diag Cytopathol 2011;39:245-50. whole slide imaging, and it is expected that this will become a widespread tool in the routine laboratory. In the context of this chapter, it is important to point out that both manufacturers and laboratories must comply with the requirements described in section 14.2, and ensure proper validation, both from the vendor side and at the laboratory, of the intended use of the system. Regulations for digital pathology are more extensively discussed in Chapter 7 (13). Moving further along, it is interesting to speculate to which extent digital imaging will be used in the process of automating the reading of control material, namely the tissue controls, cell block controls or protein spots. New technologies on the horizon will likely facilitate more eficient – and standardized – means of generating tissue control substitutes, one example being peptide IHC controls, where formalin-ixed peptide epitopes resembling the target protein are covalently attached to the glass microscope slides (14). Another possibility that may approach a more complete control of the IHC ‘total test’ (Chapter 1) is the ‘faux tissue’ or ‘Histoid’, which if fully developed has the potential also to control ixation variables, if not cold ischemia time (18, 19). When the performance of such systems has been documented to be comparable to tissue controls, they could be an effective tool relieving the laborato- 6. Xie R, Chung J-Y, Ylaya K, Williams RL, Guerrero N, Nakatsuka N, et al. Factors inluencing the degradation of archival formalinixed parafin-embedded tissue sections. J Histochem & Cytochem 2011;59:356-65. 7. NordiQC. Recommended control tissue - multitissue control block. www.nordiqc.org/Techniques/Recommended_control_tissue.htm2005. 8. Battifora H. Assessment of antigen damage in immunohistochemistry. The vimentin internal control. Am J Clin Pathol. 1991;96:669-71. 9. Mass R. Erb-b2 as a therapeutic target. Totowa, NJ: Humana Press Inc.; 2006. 10. DeFazio-Eli L, Strommen K, Dao-Pick T, Parry G, Goodman L, Winslow J. Quantitative assays for the measurement of her1-her2 heterodimerization and phosphorylation in cell lines and breast tumors: Applications for diagnostics and targeted drug mechanism of action. Breast Cancer Res. 2011;13:R44. 11. Taylor CR. The total test approach to standardization of immunohistochemistry. Arch Pathol Lab Med. 2000;124:945-51. 12. Vyberg M, Torlakovic E, Seidal T, Risberg B, Helin H, Nielsen S. Nordic immunohistochemical quality control. Croat Med J. 2005;46:368-71. 13. Lange H. Digital pathology: A regulatory overview. LabMedicine. 2011;42:587-91. 14. Bogen SA, Vani K, McGraw B, Federico V, Habib I, Zeheb R, et al. Experimental validation of peptide immunohistochemistry controls. Appl Immunohistochem Mol Morphol. 2009;17:239-46. 169 Part II: The Potentials and Pitfalls Chapter 15 Background Helle Grann Wendelboe, MSc Ba•ck•gro•und (n.) Extraneous signals that can be confused with the phenomenon to be observed or measured. The Free Online Dictionary Background | Chapter 15 Chapter 15.1 Introduction mer conjugates, the diffusion rate of these molecules is lower than for low-molecular weight conjugates. In addition, polymer As immunostaining of histological tissue specimens becomes conjugates based on a hydrophobic backbone have a tenden- more diversiied in methodology and more sensitive in detec- cy to be “sticky”. This can be resolved by applying multiple tion, background staining has develop into one of the most com- wash steps, adding detergent to the wash buffer, and by pro- mon problems in immunohistochemistry. Background staining in longing washing time. tissue sections may be due to several factors, some of which are speciic to the antigen and antibody reaction or detection meth- The latest guidelines from CAP (1) leave it up to the individual od, and others, which are of a more general character. The ter- laboratory to evaluate whether the selected detection meth- minology used in this chapter uses the term “unwanted speciic od poses a negligible risk for non-speciic staining, such that staining”, if the staining is mediated by interactions between the negative reagent control for the polymer-based detection any antibodies and their respective epitopes, and “non-speciic methods can be eliminated (Chapter 14). However, some staining” for all other interactions. The following description will staining situations still justify the use of negative reagent con- cover the major causes of background staining related to anti- trol for polymer-based detection methods, e.g. if instrumen- bodies, detection methods, and other general factors and will tation is used where a reaction between the polymer-based offer possible solutions to these problems. detection system and the substrate-chromogen can occur in an error situation, or where the substrate-chromogen by itself can be captured by foreign items like graphite or dust. Chapter 15.2 Detection Methods Capturing of chromogens such as DAB pose a challenge when the target structure is small, for example helicobactor Polymer-Based Detection Methods pylori. For small structures it is therefore necessary always to Non-speciic staining caused by endogenous avidin/biotin can include a negative reagent control. be completely avoided by use of polymer detection systems. However, polymer-based detection systems may induce gen- Vendors of IVD reagents for IHC will continue to recommend eral non-speciic staining if insuficient washing is performed the use of negative reagent control for polymer-based detec- after polymer application. Due to the large size of some poly- tion methods to comply with FDA regulations. A B Figure 15.1 Red blood cells showing endogenous peroxidase activity. A) before, and B) after blocking with 3% hydrogen peroxide. 171 Chapter 15 | Background Horseradish Peroxidase-Based Detection Methods tions and frozen sections, routine quenching of endogenous For practical purposes in immunohistochemistry, endogen- peroxidase is also advisable. ous peroxidase activity can be deined as any activity that results in the decomposition of hydrogen peroxide (H202). Specimens rich in endogenous peroxidase activity may be pro- Such activity is a common property of all hemoproteins, cessed using an alkaline phosphatase detection method instead such as hemoglobin (red cells), myoglobin (muscle cells), of a peroxidase method, eliminating this type of background. cytochrome (granulocytes, monocytes), and catalases (liver and kidney). Peroxidase activity may also be encountered in tissue adjacent to vascularized areas, due to the diffusion of Alkaline Phosphatase-Based Detection Methods blood cells prior to ixation. Endogenous alkaline phosphatase activity is frequently encountered in intestine, kidney, osteoblasts, endothelial cell sur- The most commonly used procedure for suppressing endo- faces, neutrophis, stromal reticulum cells, lymphoid tissues, genous peroxidase activity in formalin-ixed tissue is the incu- and placenta. In frozen tissue, where endogenous alkaline bation of sections in 3% H2O2 for 5-10 minutes (Figure 15.1). phosphatase activity is most pronounced, routine quenching of endogenous alkaline phosphatase is recommended. In Methanolic H2O2 treatment (11 parts 3% H2O2 plus 4 parts most formalin-ixed tissue sections, interpretation is not usually absolute methanol) for 20 minutes is also used, but is not impaired by endogenous alkaline phosphatase, which makes recommended for specimens where cell surface markers are quenching an optional choice. Most forms of endogenous al- to be stained. Methanolic treatment may also detach frozen kaline phosphatase can be quenched by including 5 mM le- sections from their carrier glass. Endogenous peroxidase ac- vamisole in the chromogen substrate solution (Figure 15.2). tivity can also be suppressed by a mixture of sodium azide The intestinal form of alkaline phosphatase is the exception and H2O2 (1). However, in most work with formalin-ixed tissue and resists this treatment, but it can be quenched by treating sections, the interpretation of speciic staining is not impaired the tissue sections with a weak acid wash prior to the applica- by endogenous peroxidase activity. If the formalin-ixed tis- tion of the primary antibody. sue is rich in blood-containing elements then it is a good idea to quench endogenous peroxidase activity. In cell prepara- A B Figure 15.2 Placenta showing endogenous alkaline phosphatase activity. A) before, and B) after blocking with levamisole. 172 Background | Chapter 15 Combined Peroxidase and Alkaline Phosphatase (Double Staining) Double staining using these enzymes requires quenching of both endogenous activities. To achieve this, use the H2O2 method for endogenous peroxidase and the weak acid method for endogenous alkaline phosphatase. The sequence of blocking endogenous activities is optional: the procedure will work effectively conducting either step irst. Also, reagents that block both endogenous peroxidase and alkaline phosphatase in one step are available. Biotin/Streptavidin-Based Detection Methods Endogenous avidin-binding activity (EABA) has been observed with all biotin-based techniques, due to the presence Figure 15.3 Endogenous biotin expression in the kidney. of biotin in a wide variety of tissues. Biotin is bound to enzymes and other protein especially in the liver (hepatic nodules), kidney (tubular epithelia) and lymphoid tissue (paracortical histio- EABA is best suppressed by sequential incubations of 10 to 20 cytes) (Figure 15.3). minutes of tissue sections, irst with 0.01% to 0.1% avidin, followed by 0.001% to 0.01% biotin prior to the staining protocol (6). EABA is usually observed within cytoplasm and is most pro- Avidin has four binding sites for biotin while each biotin molecule nounced when using frozen tissue sections. Parafin-embed- can bind to only one avidin molecule. The irst incubation with avi- ded tissues also retain substantial endogenous biotin. Other din effectively blocks endogenous biotin but simultaneously adds examples of EABA include the nonimmunochemical staining three more potential biotin-binding sites to the specimen. This of myelin (3) and mast cells (Figure 15.4), in both frozen and means there are extra biotin binding sites open to link antibodies parafin-embedded tissue (4). Guesdon et al (5) found EABA or detection systems that can give background staining. There- in granulocytes from mouse spleen. fore it is important to block these extra biotin binding sites with A B Figure 15.4 Avidin-biotin-complex (ABC) binding to mast cells in submucosa. A) before, and B) after blocking for endogenous avidin binding activity (EABA). 173 Chapter 15 | Background subsequent biotin incubation. Before blocking products became be extracted or displaced by the subsequent tissue processing commercially available, egg white (avidin) and skimmed milk steps. Extracellular antigens, or those of low molecular weight, (biotin) could be used, at the risk of inconsistent performance. are more likely to diffuse than high molecular weight antigens. Because avidin is a glycoprotein containing 10% carbohy- A typical example is the diffusion of thyroglobulin from thyroid drates and has a pI of 10, it tends to bind nonspeciically to follicular epithelium and colloid into surrounding stromal tis- lectin-like and negatively charged tissue components at phys- sue. Similarly, speciic background may result when the tissue iological pH. Streptavidin contains no carbohydrates and has marker is also present in high concentrations in blood plasma, a pI of ive. Its introduction to IHC has largely eliminated these and has diffused in the tissue prior to ixation. This effect can problems. A sugar solution can block the lectin-like elements. be seen when tonsil tissue is stained for immunoglobulin heavy Many commercially available avidin detection systems contain and light chains (Figure 15.5), particularly when ixation was modiied avidin to minimize non-speciic avidin binding. Poly- not performed promptly, and when antisera were not diluted mer-based detection systems, as noted, circumvent problems suficiently. Ingestion of target antigens by phagocytes may of background caused by avidin-biotin binding. also produce speciic background staining, resulting in staining patterns not normally seen in such cells. Chapter 15.3 Antigen Retrieval (Heat-Induced Epitope Retrieval) Antigen retrieval has been reported both to eliminate and introduce cytoplasmic and nuclear background in immunohistochemical procedures (7) (see also Chapter 3). A possible explanation is that antigen retrieval inluences antigen-antibody binding activity, and thereby affects binding of the antibody to tissue proteins. Different types of antigen retrieval solutions, with different buffer compositions, pH and chelating abilities exist. Citrate pH 6.0, Tris/EDTA pH 9.0, and TRS pH 6.0 retrieval solutions vary in the way they inluence antigen-antibody binding. Retrieval time and heating temperature can also inluence antigen-antibody binding, so for new antibodies it is advisable to investigate which antigen retrieval solution and time are best to Figure 15.5 Undesirable staining of plasma proteins with anti-kappa light chains. Plasma cells stain speciically. optimize signal and minimize background (Chapter 3). Chapter 15.4 General Factors Natural and Contaminating Antibodies in Polyclonal Antibodies (Antisera) Natural Antibodies 174 Antigen Diffusion Low-level natural antibodies present in a polyclonal antibody Unwanted speciic background staining may occur when the or antiserum, as a result of prior environmental antigenic stim- tissue marker to be stained has diffused from its sites of synthe- ulation, may increase in titer during immunization with use of sis or storage into the surrounding tissue. Because many ixa- adjuvants. As a consequence, they can give rise to non-spe- tives penetrate tissues slowly, it is important to keep tissue spec- ciic staining. In 1979, Osborn et al (8) reported that sera from imens as small as possible and to ix immediately (Chapter 2). non-immunized rabbits and goats, but not from guinea pigs, Otherwise the antigens may not be adequately ixed and may contained environmental antibodies to keratins. This inding Background | Chapter 15 may be an example of ‘speciic epithelial background’ stain- Problems stemming from natural and contaminating antibod- ing caused by natural antibodies. Although also observed by ies, of course, do not occur with monoclonal antibodies pro- others, attempts to isolate or remove these antibodies from the duced in tissue culture, but may be present in monoclonal an- antiserum were not successful (9). tibodies prepared from ascites luid. Most natural antibodies are of the non-precipitating type and Cross-Reactivity occur only in relatively low concentrations. Binding of these anti- Background staining due to antibody cross-reactivity may result bodies usually is not detected on tissue if the antiserum is used when target tissue antigen epitopes are shared with other pro- at a suficiently high dilution, or with the incubation periods. teins. A typical example is the use of unabsorbed antiserum to carcinoembryonic antigen (CEA). Because CEA shares epitopes Contaminating Antibodies with some normal tissue proteins and blood group antigens, Isolated antigens used for immunization are rarely pure. If a non-speciic staining may result. Careful absorption of such anti- host’s immune system reacts to the impurities, antibody rea- sera, or in the case of monoclonal antibodies careful screening of gents containing immunoglobulins of unwanted speciicity clones, will eliminate this type of background staining. (contaminating antibodies) will result. Usually these contaminating antibodies are present in low concentration and will not Non-speciic antibody cross-reactivity with similar or dissimilar detract from the functional immunohistochemical speciicity of epitopes on different antigens may also be the cause of con- high-titer antisera, provided they are diluted suficiently. fusing background staining. This effect is rare however, and can be avoided by using antibodies from hyper-immunized Contaminating antibodies may be related to infectious agents; animals, or carefully selected clones. other animal species kept in the same facilities, or carrier proteins used for immunization. These antibodies may be of Cross-reactivity of antigens from related species is a common special concern when dealing with antisera against synthetic problem in multi-staining methods. This dificulty can often be peptide. Small peptides are not antigenic, and must therefore avoided by using afinity-puriied antibodies, subtype speciic an- be coupled to carrier proteins prior immunization. The antisera tibodies or site/region speciic antibodies (see also Appendix A). produced will therefore contain antibodies against the carrier protein and the peptide. Fc Receptors Fc receptors (FcR) are a family of detergent-soluble membrane However, if contaminating antibodies potentially may interfere glycoproteins with approximate molecular weights of 50-70 kDa. with speciicity, afinity absorption of the antiserum is usual- They comprise less than one percent of the total membrane ly performed. ‘Batch-absorbed’ antisera almost always con- proteins and are most frequently present on macrophages and tain residual levels of contaminating antibodies (mostly of the granulocytes. They have also been reported on B cells and non-precipitating type), and will cause nonspeciic staining of some T cells. The intrinsic afinity of the FcR for monomeric IgG tissue if used at excessively high concentration (9). is approximately 1x106 to 1x108 M–1, but is higher for polymers and immune complexes of IgG. There is considerable class/ Monitoring and evaluating the results of absorption by use of subclass and species speciicity among different FcR’s. For ex- such techniques as immunodiffusion, immunoelectrophoresis ample, the FcR on some human cells was found to bind mouse and rocket immunoelectrophoresis can only be used to de- monoclonal IgG2a and IgG3, but not other IgG subclasses (10). termine non-speciicity. This monitoring cannot establish the Goat sera do not react with FcR’s of human leucocytes (11). speciicity of an antiserum in a tissue section environment, where a multiplicity of antigens are present. Ultimate mo- Background staining due to FcR is more common in frozen no-speciicity must be demonstrated by use of the designated sections, smears, and in lightly ixed tissues, than in tissues technique, and by extensive use of tissues. ixed by harsher procedures. It can be avoided by use of 175 Chapter 15 | Background F(ab')2 fragments instead of whole IgG molecules, and by careful screening of monoclonal antibodies. subclass IgG1 and IgG3 are more hydrophobic than those belonging to subclasses IgG2 and IgG4. Furthermore, some isolation procedures for IgG class antibodies promote the formation Hydrophobic Interaction of aggregates, thereby further increasing their hydrophobicity. In aqueous media, hydrophobic interactions between macro- Storage of immunoglobulins (antibody reagents) may also in- molecules occur when surface tensions are lower than that of crease their hydrophobicity and lead to aggregation and po- water (called van der Waals forces). These interactions can lymerization. This effect frequently leads to a diminution in, or be interatomic as well as intermolecular, and originate through loss of, immune reactivity. Attendant increase in non-speciic the luctuating dipolar structure within these macromolecules. background staining by use of a polyclonal IgG fraction, when compared to that obtained by use of the original whole antise- Hydrophobicity is a property shared to varying degrees by most rum, has been observed (13). proteins and is imparted primarily through the side chains of neutral aromatic amino acids phenylalanine, tyrosine and tryptophan. The diluent buffer’s formulation can also inluence hydrophobic By their lower attraction for water molecules, these amino acids binding between monoclonal IgG and tissue proteins: the great- tend to link to one another, thus expelling water from the molecule. er the proximity of diluent pH and the isoelectric point (pI) of an- While hydrophobicity is one of the natural forces that confer sta- tibodies, the stronger hydrophobic interaction will be. The lower bility on the tertiary structure of peptides, it also imparts stability the ionic strength of the diluent, the weaker will be the strength to formed immune complexes, and depending on environmental of hydrophobic attraction. The following anions and cations are factors, can exist also between different protein molecules. arranged in order of their diminishing effect on hydrophobicity: Anions: PO43-, SO42-, Cl-, NO3-, SCN- Tissue Proteins Cations: NH4+, K+, Na+, Ca2+ In tissue, proteins are rendered more hydrophobic by ixation with aldehyde-containing reagents, such as formalin and glutaralde- Other possible methods to reduce hydrophobic interactions be- hyde. Increased hydrophobicity often results from cross-linking tween tissue and reagent proteins include adding detergent, for reactive epsilon- and alpha-amino acids within and between ad- example Tween 20, or ethylene glycol to the diluent, or raising jacent tissue proteins. The extent of this hydrophobic cross-linking the pH of the diluent used for polyclonal antibodies only. during ixation is primarily a function of time, temperature and pH. 176 Changes in these factors will likely result in variable hydropho- The most widely practiced measure to reduce background due bicity due to variable cross-linking of tissue proteins. Therefore, to hydrophobic interaction is to use a protein blocking solution, once optimized, ixation procedures must be maintained and either in a separate step, or by adding it to the antibody dilu- controlled. Tissues that commonly have the most background ent. However this strategy will only be successful if the block- staining as a result of hydrophobic, as well as ionic, interactions ing protein is a type that can compete effectively with IgG or are connective tissue: collagen laminin, elastin, proteoglycans its aggregates or conjugates, for hydrophobic binding sites. and others, squamous epithelium (keratins), and adipocytes (due Separate incubation with a solution containing blocking protein to lipids, if incompletely removed during processing with xylene). is best carried out immediately prior to application of the pri- Excessive background staining due to overixation with formalin mary antibody. The solution should contain proteins identical to may sometimes be remedied by postixation with Bouin’s, Zenk- those present in the secondary link or labeled antibody, but not er’s or B5 ixative (12), but at a potential cost of decreased anti- to those in the primary antibody, in order to prevent non-speciic genicity, and poor performance of antigen retrieval methods. binding of the secondary antibody. Antibodies The addition to the primary antibody diluent of one percent bo- Of the major serum proteins, immunoglobulins unfortunately vine serum albumin (BSA) is probably the most widely practiced are particularly hydrophobic. In general, mouse antibodies of step for reducing non-speciic binding due to hydrophobic inter- Background | Chapter 15 action. Use of non-fat dry milk (14) or of casein (15) for reducing Chapter 15.5 Miscellaneous Sources background staining is also recommended. Casein, when used as a blocking agent, an antibody diluent and in the wash buff- Physical injury to tissue, drying out prior to ixation, or incomplete er, was found to result in signiicantly less background staining penetration of ixative, may cause diffuse staining of all or most tis- compared with normal swine and sheep sera (15). sue elements within an affected area. Similar diffuse background staining of both the section and the glass slide, usually limited to Because of the different uses of biotinylated antibodies today, the area of antibody incubation, has been observed and may be it should be of interest to note that biotinylation can change the due to residual embedding medium. Sections mounted routinely pI of the antibody in excess of three units, for example from in water baths containing protein additives, such as Knox gelatin a pI of 8 for the antibody to less than pI 5 for the conjugate or Elmer’s glue, may also show this type of diffuse background, (16). This change may have a marked effect on the solubility especially in procedures of high staining sensitivity. Water baths of these conjugates, possibly due to increased hydrophobicity. should be free of bacterial or yeast contamination. Ionic and Electrostatic Interactions Non-speciic staining due to undissolved chromogen granules Ionic interactions are among the prime forces that control im- may on occasion also be encountered. munochemical interaction between antigens and their corresponding antibodies. They may, however, also contribute to Non-immunologic binding of horseradish peroxidase (either in non-speciic background. free form or as a conjugate) to HbsAg in hepatocytes was reported by Omata et al (19). The precise nature of this binding The pI of the majority of polyclonal IgG ranges from approxi- is not known. mately 5.8 to 8.5. At physiological pH and at the pH commonly used for diluents, antibodies can have either net negative or Necrotic areas of tissue may stain with all reagents. Nadji and positive surface charges. Ionic interaction of some antibodies Morales (20) provide an excellent collection of color plates illus- with tissue proteins can be expected if the latter possess op- trating background staining, with accompanying explanations. posite net surface charges. Negatively charged sites on endothelia and collagen ibers have been reported to interact Excessive counterstaining may compromise visualization of with cationic conjugates composed of rabbit Fab fragments the speciic staining signal. and horseradish peroxidase type VI (pl 10.0) (17). In general, interactions of the ionic type can be reduced by use of diluent Sub-optimal operation and maintenance of automated platforms buffers with higher ionic strength. Addition of NaCl to the dil- can compromise the effectiveness of the wash between reagent uent buffer can reduce background staining stemming from incubations, or of the reagent-handling components, e.g. the ionic interactions, but its routine use in diluents for monoclonal dispensing probe. It can potentially lead to unwanted non-spe- antibodies is not recommended (18). ciic staining reaction, if two reagents react at the wrong step in the IHC staining reaction due to ineficient washing of the slide, Unfortunately most diffuse background staining results from a or the reagent-handling probe. It is, therefore, important to ad- combination of ionic and hydrophobic interactions. Remedies here to the recommended service schedule and maintenance for one type of interaction may aggravate the other. checks for any automated platform performing IHC staining. Complement-Mediated Binding Complement-mediated binding may occasionally be a cause of Chapter 15.6 General Aspects background in frozen tissue when whole antisera are used. However by the time large pools of antisera have been prepared for While it is clear that background staining can be caused by use, several of the complement factors are usually inactivated. the factors outlined above, it is also important to work with 177 Chapter 15 | Background well-characterized reagents and established protocols, in order to avoid background, or the necessity to troubleshoot background staining. Many vendors of IHC reagent offer “system solutions,” which are IHC product lines with carefully optimized buffers, target retrieval reagents, primary antibodies, detection reagents and substrates to be run on an automated platform (the RTU approach – Chapters 4 and 5). These products are designed to provide users with consistent, optimal staining. Several countries have established national quality programs, such as United Kingdom National External Quality Assessment Service (UK NEQAS) (21), Nordic Immunohisto- 9. Conway de Macario E, Macario AJ, Jovell RJ. Slide immunoenzymatic assay for human IgE (SIA-IgE). J Immunol Meth 1986; 90:137-41. 10. Gadd SJ, Ashman LK. Binding of mouse monoclonal antibodies to human leukaemic cells via the Fc receptor: a possible source of 'false positive' reactions in speciicity screening. Clin Exp Immunol 1983; 54:811-8. 11. Alexander EL and Sanders SK. F(ab')2 reagents are not required if goat, rather than rabbit, antibodies are used to detect human surface immunoglobulin. J lmmunol 1977; 119:1084-8. chemical Quality Control (NordiQC) (22), the Canadian Immu- 12. Caron BL and Banks PM. Lab Investig 1979; 40:244-5. nohistochemistry Quality Control (CIQC) (23) and the College 13. Boenisch T. Unpublished observations. of American Pathologists (CAP) (24). These organizations, 14. Duhamel RC and Johnson DA. Use of nonfat dry milk to block nonspeciic nuclear and membrane staining by avidin conjugates. J Histochem Cytochem 1985; 7:711-4. formed in response to the need for raised awareness of quality control and best practices in IHC laboratories, have fostered higher performance in many laboratories, including reduction of background staining. See Chapter 14 for further discussion. Information on individual national programs can be found on each program’s Web site. References 1. College of American Pathologists. Anatomic Pathology Checklist 2013; p. 34-38. 2. Li C-Y, Li CY, Ziesmer SC, Lazcano-Villareal O. Use of azide and hydrogen peroxide as an inhibitor for endogenous peroxidase in the immunoperoxidase method. J Histochem Cytochem 1987; 35:1457-60. 3. Sternberger LA and Sternberger NH. The unlabeled antibody method: comparison of peroxidase-antiperoxidase with avidin-biotin complex by a new method of quantiication. J Histochem Cytochem 1986; 34:599-605. 4. Hsu S-M and Raine L. In DeLellis RA (ed). Advances in Immunohistochemistry. New York: Masson 1984:31-42. 5. Guesdon J-L et al. The use of avidin-biotin interaction in immunoenzymatic techniques. J Histochem Cytochem 1979; 27:1131-9. 6. Wood GS and Warnke R. Suppression of endogenous avidin-binding activity in tissues and its relevance to biotin-avidin detection systems. J Histochem Cytochem 1981; 29:1196-204. 7. Wieczorek R, Stover R, Sebenik M. Nonspeciic Nuclear Immuno reactivity After Antigen Retrieval Using Acidic and Basic Solutions. J Histotech 1997; 20:139-43. 178 8. Osborn M, Franke WW, Weber K. Visualization of a system of ilaments 7-10 nm thick in cultured cells of an epithelioid line (Pt K2) by immunoluorescence microscopy. Proc Natl Acad Sci (USA) 1977; 74:2490-4. 15. Tacha DE and McKinney LA. Casein Reduces Nonspeciic Background Staining in Immunolabeling Techniques. J Histotech 1992;15: 127-32. 16. Wadsley JJ and Watt RM. The effect of pH on the aggregation of biotinylated antibodies and on the signal-to-noise observed in immunoassays utilizing biotinylated antibodies. J Immunol Meth 1987; 103:1-7. 17. Pino RM. Binding of Fab-horseradish peroxidase conjugates by charge and not by immunospeciicity. J Histochem Cytochem 1985; 33:55-8. 18. Boenisch T. Diluent buffer ions and pH: their inluence on the performance of monoclonal antibodies in immunohistochemistry. Appl Immunohistochem 1999; 7:300-6. 19. Omata M, Liew CT, Ashcavai M, Peters RL. Nonimmunologic binding of horseradish peroxidase to hepatitis B surface antigen. A possible source of error in immunohistochemistry. Am J Clin Pathol 1980; 5:626-32. 20. Nadji M and Morales AR Immunoperoxidase Techniques: A Practical Approach to Tumor Diagnosis. Chicago: American Society of Clinical Pathologists, 1986. 21. United Kingdom National External Quality Assessment Service (UK NEQAS): www.ukneqas.org.uk. 22. Nordic Immunohistochemical Quality Control (NordiQC): www.nordiqc.org. 23. Canadian Immunohistochemistry Quality Control : www.ciqc.ca 24. College of American Pathologists: www.cap.org 179 Part II: The Potentials and Pitfalls Chapter 16 Troubleshooting Helle Grann Wendelboe, MSc Anette Lykke, MLS, Dip Gale Pace, MT(ASCP), BSc Trou•ble•shoot•ing (n.) Discovering why something does not work effectively and making suggestions about how to improve it. Cambridge Advanced Learner's Dictionary Troubleshooting | Chapter 16 Chapter 16.1 Introduction Section 2 presents a method of systematically adding one IHC reagent at a time to determine at which stage in a staining Immunohistochemistry (IHC) is a multi-step process that requires protocol non-speciic or undesired staining may be occurring. specialized training in the processing of tissue, the selection of appropriate reagents and interpretation of the stained tissue sec- Section 3 is a simple chart used to deine the type of tissue tions. In general, IHC staining techniques allow for the visualiza- specimen, the IHC reagents, and the staining protocol already tion of antigens by sequential application of a speciic antibody in use by the laboratory personnel. The user is encouraged to to the antigen, a secondary antibody to the primary antibody, an copy this chart and use it to help troubleshoot any problems enzyme complex and a chromogenic substrate. The enzymatic that may be encountered in the IHC staining process. activation of the chromogen results in a visible reaction product at the antigen site. Because of its highly complex nature, the causes Section 4 is a guide to reading a manufacturers’ speciication sheet of unexpected negative reactions, undesired speciic staining, or for IVD (in vitro diagnostic) antibodies. The guide includes general undesired background may be dificult to isolate. The information information for use in immunohistochemistry, including ixation, rec- contained in this chapter should enable the user rapidly to pinpoint ommended visualization systems, recommended titer and diluent, and resolve problems encountered during the staining procedure. pre-treatment methods, and selection of required controls. Section 1 is a compilation of common problems encountered Section 5 is a guide to check that the automated platform used when using IHC staining reagents, including the underlying to perform the staining has operated correctly during the stain- causes of staining failure and the recommended corrective ing process. actions. The chart is divided into sections describing inadequate staining, general background staining, and limited background staining. Section 1 Common Problems Inadequate Staining (little or no speciic staining) Possible cause of poor staining Solution A) Both the positive controls and the specimen tissue show little or no speciic staining, except for counterstain. The tissue section may show little or no background staining. Primary antibody or labeled reagent omitted. Reagents used in wrong order. Repeat the procedure using the manufacturer’s staining system specification sheet, or the standard operating procedure reagent checklist as established by the individual laboratory. Excessively diluted or excessively concentrated reagents; inappropriate incubation time and or temperature. Determine correct concentration for each reagent (see Chapter 4 and Chapter 5). Depending on the degree of staining obtained, if any, a 2- to 5- fold increase in concentration may be needed. Incubation temperature and incubation time are inversely proportional and will affect results. To determine optimal incubation protocol, vary either the time or temperature for each reagent in the IHC staining system. Generally, incubation times can be extended if little or no background is detected. Overnight incubation at higher dilution may also be effective. Primary antibody diluted with inappropriate buffer. Use of PBS or TBS as an antibody diluent. Lack of stabilizing or carrier protein. Detergent in diluent. Check formula and compatibility of antibody diluent. A change of ion content and/or pH of the antibody diluent can cause a diminution in the sensitivity of the antibody. Addition of NaN3 should be avoided. This problem is primarily seen with monoclonal antibodies. 181 Chapter 16 | Troubleshooting Possible cause of poor staining Solution Primary antibody defective; one or several secondary or ancillary reagents defective. Do NOT use product after expiration date stamped on vial. Replace defective or expired antibody; repeat staining protocol, replacing one reagent at a time with fresh, in-date reagents. Store products according to each product specification sheet or package insert If using a neat or concentrated antibody, and directed by the manufacturer to store frozen, the reagent may be aliquoted to avoid repeated freezing and thawing Do not freeze ready-to-use (RTU) or customer diluted products Follow manufacturer recommendations on product specification sheets, package inserts, and reagent labels Dissociation of primary antibody during washing or incubation with link antibodies. Particularly a feature of low affinity antibodies: Polyclonal primary antiserum: Attempt staining at lower dilutions (higher concentrations) Monoclonal primary antibody: Replace with higher affinity antibody of identical specificity Re-optimize incubation times for washing buffer and link antibody Use of alcohol-based counterstain and/or alcohol-based mounting media will remove aqueous– based chromogens. Excessive counterstaining may compromise proper interpretation of results. Incorrect preparation of substrate-chromogen mixture. Incompatible buffer used for preparation of enzyme and substrate-chromogen reagents: Use of PBS wash buffer with an alkaline phosphatase staining system. Sodium azide in reagent diluent or buffer baths for immunoperoxidase methodologies. Antigen levels are too low for detection by the employed visualization system. May be due to loss of antigenic differentiation in some tumors or loss of antigenicity due to sub-optimal tissue ixation. 182 Repeat staining, using water-based counterstain and mounting media Use a permanent chromogen, such as DAB/DAB+, that is not affected by organic solvents Use a counterstain that: Will not excessively stain tissue sections Can be diluted so as not to obliterate the specific signal Reduce incubation time of the counterstain Repeat substrate-chromogen treatment with correctly prepared reagent Staining intensity may be decreased when excess DAB/DAB+ is present in the working reagent Check compatibility of buffer ingredients with enzyme and substrate-chromogen reagents. Repeat staining: Commercial phosphate buffers may contain additives that will inhibit alkaline phosphates activity Avoid sodium azide in diluents and buffers. A concentration of 15 mM/L sodium azide, which is routinely added to IHC reagents to inhibit bacterial growth, will not impair HRP conjugated labels Utilize a higher sensitivity staining system Prolong incubation time of primary antibody Re-optimize incubation times and concentrations of ancillary reagents Perform antigen retrieval, if applicable, using a range of pH buffers (see Chapter 3) Steric hindrance due to high antigen level and possible prozone effect. Re-optimize concentration of the primary antibody and ancillary reagents. Antibody concentration of the primary antibody may be too high. Use of inappropriate ixative. Use of certain ixatives may damage or destroy antigens or epitopes in the tissue specimen. Use of non-cross linking ixatives may allow the elution of antigens soluble in IHC reagents. Different ixatives may affect standardization of cells. Check manufacturer’s specifications regarding recommended fixatives known to be effective with antibody and protocol in use. Immunoreactivity diminished or destroyed during embedding process. Use a paraffin wax with a melting temperature ~55-58 °C. Wax used for embedding should not exceed 60 °C. Immunoreactivity diminished or destroyed during dewaxing at high oven temperature. Oven temperature not to exceed 60 °C. NOTE: The intensity of immunostaining may be diminished when tissue is exposed to prolonged heat at this stage in the protocol. Paradoxically heating of the section in aqueous solution is used to recover antigenicity in the AR process (See Chapter 3). Refer to the primary antibody specification sheet for additional information. Immunoreactivity diminished or destroyed on pre-cut tissue sections. The intensity of immunostaining may be diminished when pre-cut tissue sections are exposed to air. Use freshly cut sections and re-seal paraffin-embedded blocks. Immunoreactivity diminished or destroyed by the enzyme blocking reagent altering a speciic epitope. More common on frozen sections: apply the primary antibody prior to the enzymatic block to ensure its reaction. In such cases the blocking reagent can be applied at any point after the primary and before the enzyme-labeled components. Troubleshooting | Chapter 16 Possible cause of poor staining Solution Excessive wash buffer or blocking serum remaining on tissue section prior to application of IHC reagents. Excess residual reagent will dilute the next consecutive reagent. Repeat staining, making sure to wipe away excess washing buffer and blocking serum. Antigen retrieval protocol is inappropriate or has been omitted. Many tissue antigens require proteolytic enzyme digestion or heat-induced antigen retrieval performed prior to staining (Chapter 3). The need for pre-treatment depends on the type and extent of fixation, specific characteristics of the antigen and the type of antibody used. Use the pretreatment method recommended by the manufacturer. No single pre-treatment is suitable for all applications. Repeated reuse of antigen retrieval buffer. Do not reuse buffer. Sections incorrectly dewaxed. Prepare new sections and deparaffinize according to standard laboratory protocol, using fresh xylene or xylene substitute. Failure to achieve the optimal temperature required for heat induced antigen retrieval. When using a waterbath or steamer, allow sufficient time for the retrieval buffer to equilibrate to a temperature range of 95-99 °C At high altitude (greater than ~4,500 feet), the buffer will boil at less than 95 °C Utilize a closed heating system such as a pressure cooker, autoclave or Pascal, or utilize a low temperature protocol if standardization of the validated procedure is not affected Excessive or incomplete counterstaining. Re-optimize concentration of counterstain and incubation time. Instrument malfunction. Ensure automated stainer is programmed correctly and is running to manufacturer’s specifications. B) Positive control tissue shows adequate speciic staining with little or no background staining. Specimen tissue shows little or no speciic staining with variable background staining. Specimen held for too long in a cross-linking ixative, usually in formalin, causing “masking” of antigenic determinants. Control appropriately ixed. Standardize routine fixation, matching test specimens to control tissues. Proteolytic digestion or antigen retrieval will break down cross-linking and render some tissue antigens reactive (Chapter 3). Refer to the primary antibody specification sheet for additional information. Sectioned portion contains crush artifact caused by grossing tissue with dull scalpel or razor. Serum proteins diffuse through tissue and are fixed in place. Cut new tissue block if available, using sharp blade. Sectioned portion of specimen contains necrotic or otherwise damaged elements. Ignore physically damaged portions of stained tissue sections. Uneven ixation of section; portion of specimen not penetrated by ixative. Loss of antigenicity in unixed tissue. Fix tissue biopsy for longer period of time or fix smaller pieces to ensure complete penetration. Unfixed tissue tends to bind all reagents non-specifically. General Background Background is seen in both control tissue and specimen tissue. Background staining may affect several tissue elements, such as connective tissue, adipose tissue and epithelium. Excessive incubation with substrate-chromogen reagent. Reduce incubation time. Substrate-chromogen reagent prepared incorrectly. Repeat incubation with correctly prepared chromogen reagent. Secondary or link antibody cross-reacts with antigens from tissue specimen. Absorb link antibody with tissue protein extract or species-specific normal serum from tissue donor. Secondary or link antibody and/or tertiary reagents too concentrated. Repeat staining. Determine correct concentration for each reagent. Incubation temperature and incubation time will affect results. To determine optimal incubation protocol, vary both the time and temperature for each reagent in the IHC staining protocol. Substrate-chromogen reagent prepared incorrectly. Repeat incubation with correctly prepared chromogen reagent. 183 Chapter 16 | Troubleshooting Possible cause of poor staining Solution Slides inadequately rinsed. Gently rinse slide with wash buffer bottle and place in wash bath for 5 minutes. Gentle agitation of the wash bath may increase effectiveness when used with cytoplasmic or nuclear staining protocols. Insufficient saline or detergent in wash buffer. High-sensitivity staining systems may require higher concentrations of saline or detergent in the wash buffer. Refer to the staining system specification sheet for optimal formulation. Blocking serum or wrong blocking serum used. Block with serum from the host of the secondary or link antibody. Avoid serum that contains auto-immune immunoglobulins. Alternatively, a serum-free protein block, lacking immunoglobulins, may be substituted for the serum block. Sections incorrectly dewaxed. Prepare new sections and deparaffinize according to standard laboratory protocol using fresh xylene or xylene substitute. Non-speciic binding of the secondary antibody with an animal tissue specimen. Use a secondary antibody that has been absorbed against a species specimen, or use a secondary antibody produced in a host that exhibits little or no cross-reactivity with the tissue source. Instrument malfunction. Ensure automated stainer is programmed correctly and is running to manufacturer’s specification. Specimen tissue and negative reagent control slides show background staining. Positive and negative control tissue show appropriate speciic staining. May involve several tissue elements such as connective tissue, adipose tissue and epithelium. 184 Specimen held for too long in a cross-linking ixative, usually in formalin, causing ‘masking’ of antigenic determinants due to aldehydes cross-linking and increased hydrophobicity of tissue. Standardize routine fixation. Proteolytic digestion or antigen retrieval will break down crosslinking and render some tissue antigens reactive. Refer to the primary antibody specification sheet for additional information. Sectioned portion of specimen not penetrated by ixative. Loss of antigenicity in unixed tissue. Unixed tissue tends to bind all reagents nonspeciically. Fix tissue biopsy for longer period of time or fix smaller pieces to ensure complete penetration. Sectioned portion contains crush artifact caused by grossing tissue with dull scalpel or razor. Serum proteins diffuse through tissue and are ixed in place. Serum proteins diffuse through tissue and are fixed in place. Re-cut tissue using sharp blade. Sectioned portion of specimen contains necrotic or otherwise damaged elements. Ignore physically damaged portions of stained tissue sections. Excessive or unevenly applied subbing agent on poly-Llysine, charged, or silanized slides. Some IHC reagents may bind to these products, resulting in a light stain over the entire slide surface. Some slides may be unevenly coated, and will exhibit the above problems on only a portion of the tissue or glass. Antigen diffusion prior to ixation causing speciic background outside the expected antigen site. Avoid delays in fixation of the tissue. Tissue sections too thick. Cut tissue sections thinner. Formalin-fixed paraffin-embedded tissue sections should be approximately 4-6 µm; cryostat section 4-6 µm or less. Incomplete permeabilization of tissue sections. Seen in frozen sections, cell smears and non-paraffin embedded tissue: incomplete permeabilization of cells allows unattached reagents to become trapped within the cells and resistant to removal by wash buffer. Troubleshooting | Chapter 16 Possible cause of poor staining Solution Negative reagent control slide shows background. Positive control tissue, negative control tissue and specimen tissue show expected speciic staining. Negative control serum insufficiently diluted. Use properly diluted negative reagent control serum: For polyclonal antibodies, dilute the negative reagent control serum until the protein concentration is equal to that of the primary antibody For monoclonal antibodies, dilute the negative reagent control serum until the Ig concentration is equal to that of the primary antibody Contaminating antibodies in the negative control serum are cross-reacting with proteins from the specimen tissue. Replace the negative reagent control serum; repeat staining protocol. Negative reagent control serum contaminated with bacterial or fungal growth. Replace product with non-contaminated serum. Limited Background Areas of inconsistent staining on controls, specimens and glass slides. Protein trapped beneath the tissue during the mounting process will allow partial lifting of the section. Pooling of IHC reagents beneath the section, or partial detachment of the tissue from the slide may occur. Avoid the use of commercial adhesives, glue starch or gelatin in water baths when mounting tissue sections. Avoid allowing water from an initial section mounting to flow over an area where additional sections will be mounted. This is particularly important when using charged or silanized slides. Undissolved granules of chromogen. Ensure that chromogen in tablet or powder form is completely dissolved, or switch to a liquid chromogen. Incomplete dezenkerization of tissue ixed with B5 or mercury containing reagents. Remove embedding medium thoroughly, using fresh reagents. Incomplete dezenkerization of tissue ixed with B5 or mercury containing reagents. Perform dezenkerization with fresh reagents. Bacterial or yeast contamination from mounting waterbath. Clean and refill waterbath. Partial drying of tissue prior to ixation. Unaffected areas show normal staining. Instrument malfunction. Immerse tissue promptly in fixative or holding reagent Keep moist during the entire staining process Use a humidity or moist chamber during incubation steps When using an automated staining instrument, addition of wet towels to the sink may prevent drying of slides Ensure automated stainer is programmed correctly and is running to manufacturer’s specification. Adipose or connective tissue in specimen, negative control tissue, positive control tissue and negative reagent control slides. Background in connective and epithelial tissue. Hydrophobic and ionic interactions between immunoglobulins and lipoid substances in fatty tissue. Non-specific staining of fatty tissue rarely interferes with interpretation of specific staining and can usually be disregarded. Primary antibody and negative reagent control serum are insufficiently diluted. Reoptimize the dilution of the primary antibody and negative control serum. Epithelial tissue in specimen, negative control tissue, positive control tissue and negative reagent control slides. Staining is moderate to marked, especially in epidermal epithelium. Background in epithelia accompanies background in connective tissue. Both the primary antibody and negative control serum contain contaminating antibodies to epithelial elements, possibly cytokeratins. Excessive formalin ixation of tissues may increase protein cross-linking, resulting in tissue hydrophobicity. Use a higher dilution of the primary antibody and negative control serum Increase the incubation time Replace the antibody Proteolytic digestion or antigen retrieval will break down cross-linking and render some tissue antigens reactive. Refer to the primary antibody and/or the negative reagent control specification sheet for appropriate pre-treatment. 185 Chapter 16 | Troubleshooting Possible cause of poor staining Solution Focal cytoplasmic staining observed in epithelium in the specimen tissue. Focal cytoplasmic staining is seen, particularly in intermediate and supericial layers of the epidermis. May be caused by passive absorption of plasma proteins into degenerating epidermal cells. This observation is rare and should not interfere with interpretation of specific staining. Background seen in all control and specimen tissue when using an immunoperoxidase staining system. Unquenched endogenous peroxidase activity may be seen in all hemoprotein-containing specimens, including hemoglobin in erythrocytes, myoglobin in muscle cells, cytochrome in granulocytes and monocytes and catalases in liver and kidney. Use alternate or prolonged peroxidase blocks or use another enzyme label such as alkaline phosphatase Eosinophils and mast cells are particularly resistant to peroxidase quenching. Use a peroxidase blocker Use special stains: eosin will stain eosinophils a bright red-orange Background seen in all control and specimen tissue when using an alkaline phosphatase staining system. Unquenched endogenous alkaline phosphatase activity may be seen in leucocytes, kidney, liver, bone, ovary bladder, salivary glands, placenta and gastro-intestinal tissue. Add levamisole to the alkaline phosphatase chromogen reagent or use another enzyme label such as horseradish peroxidase. Intestinal alkaline phosphatase is not quenched by the addition of levamisole. Pretreat the tissue with 0.03 N HCl. Background seen in all control and specimen tissue when using a biotin-streptavidin staining system. Endogenous protein-bound biotin (water-soluble B vitamin). High amounts of biotin are found in adrenal, liver, and kidney. Lesser amounts are found in the GI tract, lung, spleen, pancreas, brain, mammary gland, adipose tissue, lymphoid tissue, and cells grown in culture media containing biotin as a nutrient. Use a biotin block or chose another non-biotin based staining system. Background of skeletal or smooth muscle tissue in positive control tissue, negative control tissue, specimen tissue and negative reagent control. Cause is not understood. It is possibly due to antibodies to muscle antigens in primary and negative reagent control serum. Should not interfere with interpretation of specific staining. Undesired ‘Speciic’ Staining. Positive staining of leucocyte membranes in specimen tissue, positive control, negative tissue control and negative reagent control. Binding of the Fc portion of Ig by Fc receptors on the cell membrane of macrophages, monocytes, granulocytes and some lymphocytes. Use F(ab’)2 or F(ab) fragments for the primary and secondary antibodies rather than intact antibodies Add detergent to the wash buffer Positive staining of histiocytes and granulocytes in the specimen tissue only, with a marker not normally reactive with these cells. Phagocytosis of antigens may render phagocytes positive for the same. Rare. Should not interfere with interpretation of specific staining. Positive membrane staining of specimen tissue and negative reagent control tissue when using a horseradish peroxidase staining system. Tissue from persons infected with Hepatitis B virus and expressing Hepatitis B surface antigen may exhibit undesired staining. Utilize a non-peroxidase staining system. Miscellaneous Loss of viability of cell cultures. Some manufacturers produce antibodies and reagents for in vitro use only. These products may contain preservatives, usually sodium azide, which is a known poison. 186 Utilize an in vivo product for application on viable cells. For use on cell cultures only: sodium azide may be dialyzed out of some reagents. Contact Dako Technical Support for additional information. Troubleshooting | Chapter 16 Section 2 Systematical Approach Using a Troubleshooting Flow Chart This low chart can be used to determine source(s) of non-specif- tissue is only counterstained with hematoxylin. If the described ic staining that has been encountered when using an IHC staining result (to the right) does not match the observed staining pattern protocol. Each step (Slide box to the left) is a suggestion for rea- when using the suggested setup, proceed to the next step in the gents to be tested on the indicated tissue type with known posi- low chart. In the next step (Slide #2), a chromogen is added to tive or negative expression pattern. In the irst step (Slide #1) the the staining protocol, before counterstaining and so forth. Background Staining Encountered with Peroxidase Reagents Reagents Result/Action Slide 1 Brown endogenous pigment (such as melanin) observed: Positive Control Tissue: Counterstain with hematoxylin „„To distinguish melanin pigment from DAB chromogen, Azure B can be used as a counterstain. The melanin stains blue-green, while the DAB remains brown. „„An alternate method is to use AEC as the chromogen. If Result/Action does not match the observed staining: However, if high levels of pigment exist in the tissue, the Go to next step red chromogen may be partially obscured. Since bleaching protocols to remove melanin may compromise tissue antigenicity, it should be avoided if at all possible. Slide 2 Brown/red color observed: Positive Control Tissue: DAB/AEC chromogen and counterstain „„Indicates endogenous peroxidase activity in the tissue sections. It is present in all hemoprotein containing tissue including erythrocytes, muscle, liver, kidney, granulocytes and monocytes. If Result/Action does not match the observed staining: Go to next step „„Block with 3% hydrogen peroxide or other peroxidase blocking reagent. Using a new bottle of hydrogen peroxide, perform a 3% H202 peroxidase block, followed by DAB Slide 3 and an appropriate counterstain. Positive Control Tissue: Peroxidase Block + Secondary Antibody + Streptavidin-HRP DAB/AEC Counterstain Brown/red color observed: „„Indicates endogenous biotin activity in the tissue sections. Protein-bound biotin may be found in adrenal, liver, kidney, GI tract, lung, spleen, brain, mammary gland, adipose tissue, lymphoid tissue and cell grown in culture media If Result/Action does not match the observed staining: Go to next step containing biotin (RPMI, NCTC, MEME). „„Block with a biotin block or switch to a staining system that is not dependent on the streptavidin/biotin reaction 187 Chapter 16 | Troubleshooting Slide 4 Reagents Result/Action Positive Control Tissue: Peroxidase Block + Biotin Block (if required) + Secondary Antibody + Streptavidin-HRP + DAB/AEC + Counterstain Brown/red color observed: „„ Indicates non-speciic or undesired binding of the secondary antibody to the tissue sections. This primarily occurs when the secondary antiserum has not been prepared for use on a speciic species tissue. If Result/Action does not match the observed staining: Go to next step „„To determine if this is the problem, absorb out non-speciic proteins by adding 2, 5 or 10 µL of normal serum (from the species of tissue to be stained) per 100 µL of the secondary Slide 5 antibody Positive Control Tissue: Peroxidase Block + Biotin Block (if required) + Negative Reagent Control + Secondary Antibody + Streptavidin-HRP + DAB/AEC Brown/red color observed: „„ May indicate non-speciic binding of the primary antibody carrier-protein. Perform a protein block with normal serum from the host of the link antibody. Add 0.05-0.1% Tween 20 to wash buffer to decrease protein attachment. If Result/Action does not match the observed staining: Go to next step „„ Antigen retrieval lipofusion-artifact may appear as granule staining in liver and cardiac tissue, or as speciic staining in pancreatic sections Slide 6 Brown/red color observed on Negative Control Tissue: Negative Control Tissue: Perform complete staining protocol „„ Monoclonal antibody: Possible contamination „„ Polyclonal antibody: Possible contamination or undesired antibody in the host Ig fraction „„ Antigen retrieval lipofusion-artifact may appear as granule staining in liver and cardiac tissue, or as speciic staining in pancreatic sections Background Staining Encountered with Alkaline Phosphatase Slide 1 Red/blue color observed: Positive Control Tissue: Fast Red, Fuchsin or BCIP/NBT + Counterstain „„Indicates endogenous alkaline phosphatase activity in the tissue sections. It is present in liver, kidney, GI tract, bone, bladder, ovary, salivary gland, placenta, leukemic, necrotic or degenerated cells. If Result/Action does not match the observed staining: Go to next step „„Block with levamisole (Intestinal alkaline phosphatase may be quenched by the addition of 0.03 N HCl prior to the addition of the alkaline phosphatase) 188 Troubleshooting | Chapter 16 Result/Action Reagents Slide2 Red/Blue color observed: Positive Control Tissue: Streptavidin-AP + Fast Red, Fuchsin or BCIP/NBT + Counterstain „„ Indicates endogenous biotin activity in the tissue sections. Protein-bound biotin may be found in adrenal, liver, kidney, GI tract, lung, spleen, brain, mammary gland, adipose tissue, lymphoid tissue and cells grown in culture media containing If Result/Action does not match the observed staining: Go to next step biotin (RPMI, NCTC, MEME). „„ Block with a biotin block or switch to a staining system that is Slide 3 not dependent on the streptavidin/biotin reaction Positive Control Tissue: Biotin Block (if required) + Secondary Antibody + Streptavidin-AP + Fast Red, Fuchsin or BCIP/NBT + Counterstain Red/blue color observed: „„ Indicates non-speciic or undesired binding of the secondary antibody to the tissue sections. This primarily occurs when the secondary antiserum has not been prepared for use on a speciic species tissue. If Result/Action does not match the observed staining: Go to next step „„ To determine if this is the problem, absorb out non-speciic proteins by adding 2, 5 or 10 µL of normal serum (from the species of tissue to be stained) per 100 µL of the secondary Slide 4 antibody Positive Control Tissue: Biotin Block (if required) + Negative Reagent Control + Secondary Antibody + Streptavidin-AP + Fast Red, Fuchsin or BCIP/NBT + Counterstain. Red/blue color observed: „„ May indicate non-speciic binding of the primary antibody carrier-protein. Perform a protein block with normal serum from the host of the link antibody or a protein block; add 0.05- If Result/Action does not match the observed staining: Go to next step 0.1% TWEEN 20 to wash buffer to decrease protein attachment. „„ Antigen retrieval lipofusion-artifact may appear as granule staining in liver and cardiac tissue or as speciic staining in pancreatic sections Slide 5 Red/blue color observed on Negative Control Tissue: Negative Control Tissue: Perform complete staining protocol „„Monoclonal antibody: Possible contamination „„Polyclonal antibody: Possible contamination or undesired antibody in the host Ig fraction „„Antigen retrieval lipofusion-artifact may appear as granule staining in liver and cardiac tissue, or as speciic staining in pancreatic sections 189 Chapter 16 | Troubleshooting Negative Reagent Control Reagents Result/Action „„ (Human tissue) Perform the peroxidase blocking protocol Negative Control Reagent: Perform complete staining protocol. from Slide #2 under “Background Staining Encountered with Peroxidase Reagents” „„ Perform a biotin block if required, protein block if required, apply the appropriate negative reagent control (see below), apply biotinylated secondary antibody, apply streptavidin/ HRP reagent and DAB „„ Prepare a negative reagent control – Polyclonal: non-immunized sera from the same species, diluted to the same protein concentration as the primary antibody – Monoclonal: negative reagent control that matches the isotype as the primary antibody. Additionally, the diluent used to manufacture a monoclonal primary antibody and isotypic negative control should contain the same ions. Diluents containing sodium or phosphate ions may change the sensitivity of some monoclonal antibodies. – Calculation: – Ig or total protein concentration of primary antibody divided by dilution factor of primary antibody = x – Ig or total protein concentration of negative reagent control divided by x = dilution factor of negative reagent control 190 Troubleshooting | Chapter 16 Section 3 Troubleshooting Chart Background staining is deined as unexpected or undesirable Tissue Specimen: Successful staining of tissue with an IHC sent the target antigen. Frequent causes of background staining marker is dependent on the type and preparation of the spec- are endogenous enzyme activity and endogenous biotin. staining seen on the test or control tissue, which does not repre- imen. The chart below provides a convenient check list, as to [procedure at each step in the ‘Total Test’. Peroxidase is an enzyme of the oxido-reductase class that reacts with a substrate containing hydrogen peroxide as the elec- Species: tron acceptor. To block this activity, a variety of hydrogen per- (important to note in research studies). oxide reagents can be applied to cells producing this enzyme. Organ/tissue source: Alkaline phosphatase is an enzyme having various isoforms, Collection: which are produced in the leukocytes, liver, bone, intestine, Surgical specimen/biopsy placenta and Regan (carcinoma). Addition of levamisole to the Post-mortem specimen chromogen/substrate will inhibit endogenous alkaline phos- Fine needle aspirate phatase activity, with the exception of the intestinal isoform. If Peripheral blood (include anti-coagulant) necessary, this can be blocked with a weak acid wash, such Brushing as 0.03-0.5 N HCl. Biologic luid Cell culture Biotin, a B vitamin, may be protein-bound to tissue and can Other interfere with proper interpretation of staining patterns when Tissue preparation: Parafin embedded using a streptavidin or avidin reagent. To block this binding, a biotin/avidin block. Plastic embedded Cryostat section Cytospin Peroxidase block: 3% H2O2 Cell smear Methanol/H2O2 Mono-layer cultured cells Sodium azide Other Peroxidase Block (e.g. Dako Code S2003) Tissue ixation: Type of ixative Other Alkaline phosphatase block: Total length of time in ixative, including during transport, Levamisole grossing and on the tissue processor 0.03 N HCl (not for use on cryostat tissue) Size of specimen ; size of block; wheterh additional blocks Other are available if needed Tissue mounting: Slide mount Tissue thickness Biotin block: Biotin Block (e.g. Dako Code X0590) Other (e.g. skimmed milk) Protein block: Gelatin, glue commercial adhesive or starch in the water bath Protein Block (e.g. Dako Code X0909) Other Normal sera or Ig from host species of the secondary antibody Other Blocking of endogenous components that may produce spurious staining. 191 Chapter 16 | Troubleshooting Section 4 Speciication Sheets Below is an example of the information supplied in a typical Dako package insert for an IVD (in vitro diagnostic) concentrated antibody. The information and placement in the package insert will vary. Information You Need to Know Information Located on the Speciication Sheet/Package Insert * Regulatory Status of the Primary Antibody Intended use For in vitro diagnostic use Tissue Preparation Specimen preparation Paraffin sections: The antibody can be used for labeling formalin-fixed, paraffin-embedded tissue sections. Tissue specimens should be cut into sections of approximately 4 µm. Pre-treatment: Pre-treatment of formalin-fixed, paraffin-embedded tissue sections with heat-induced epitope retrieval (HIER) is required. Optimal results are obtained by pretreating tissues with HIER using diluted EnVision™ FLEX Target Retrieval Solution, High pH (50x) (Codes K8000/K8004). Deparaffinization, rehydration and epitope retrieval can be performed in Dako PT Link (Code PT100/PT101). For details, please refer to PT Link User Guide. The tissue sections should not dry out during the treatment or during the following immunohistochemical staining procedure. For greater adherence of tissue sections to glass slides, the use of FLEX IHC Microscope Slides (Code K8020) is recommended. After staining the sections must be dehydrated, cleared and mounted using permanent mounting medium. Indicates the type of specimen that was used during validation studies. In many cases this would include formalin-fixed tissue and frozen sections. Use of other fixatives requires validation by each individual laboratory. Staining procedure Visualization: The recommended visualization system is EnVision™ FLEX, High pH (Code K8000/K8010) using a 20 minute incubation at room temperature. Follow the procedure enclosed with the selected visualization system(s). Automation: The antibody is well-suited for immunohistochemical staining using automated platforms, such as Dako Autostainer, Autostainer Plus and Autostainer Link as well as PT Link for pre-treatment. Indicates the recommended visualization system to be used with the antibody. Conditions will differ if other detection systems are used. It also indicates that the antibody can be used for automated staining. Staining procedure Dilution: The recommended dilution of Monoclonal Mouse Anti-Human PSMA, Clone 3E6, Code M3620, is 1:50. Dilute the antibody in Dako Antibody Diluent (Code S0809). Incubate pretreated tissue sections for 20 minutes at room temperature. These are guidelines only. Optimal conditions may vary depending on specimen and preparation method, and should be validated individually by each laboratory. Includes a suggested dilution for the antibody and the recommended diluent. The dilution is a suggested starting point, but may require further optimization depending on specimen, preparation method, temperature of the laboratory or automated instrumentation. Choosing the Visualization System Diluting the Primary Antibody * Comments Indicates that a product meets the FDA requirements as a clinical diagnostic product. Likewise, a icon indicates the reagent meets European Union requirements. This section also indicates the optimal epitope retrieval procedure and warns against procedures that may destroy the epitope. Specimen preparation and staining procedure sections can and will change periodically, to reflect changes in technology. So remember to retain copies of each version of the reagent specification sheet. Version numbers are usually found on each page. NOTE: If your state regulatory agency requires written documentation that a reagent can be used for automated staining and this indication is not listed on the specification sheet, you may wish to contact the manufacturer’s technical support group for further information. * Examples included. The information given will vary depending on the individual antibodies. Always refer to the actual package insert for speciic information 192 Troubleshooting | Chapter 16 Information You Need to Know Information Located on the Speciication Sheet/Package Insert * Comments Controls Staining procedure Controls: Positive and negative control tissues should be run simultaneously using the same protocol as the patient specimens. The positive control tissue should include prostate and the cells/structures should display reaction patterns as described for this tissue in the “Performance characteristics” section. Negative control: The recommended negative control reagent is Dako Negative Control, Mouse IgG1 (Code X0931), diluted to the same Ig concentration as the primary antibody. Unless the stability of the diluted antibody and negative control has been established in the actual staining procedure, dilute these reagents immediately prior to use. Positive and negative controls should be run simultaneously with patient specimens. Use of a negative reagent control (ANP.22570) is no longer required by the College of American Pathologists (CAP), based on Clinical Laboratory Improvement Amendments (revised July 31 CLIA 2012), for each patient or patient block in a staining run when using polymer detection systems. The latest guidelines from CAP (5) leave it up to the individual laboratory to evaluate whether the selected detection method poses a negligible risk for non-specific staining that the negative reagent control for the polymer-based detection methods can be eliminated. Positive Control Tissue Performance characteristics Normal tissues: In prostate, glandular epithelial cells show a moderate to strong cytoplasmic and/or membranous staining reaction. Abnormal tissues: In 92/102 prostate adenocarcinoma, glandular epithelial cells showed a moderate to strong cytoplasmic and/or membranous staining reaction. CLIA 2003 Sec. 493.1273 (3) Mandates that fluorescent and immunohistochemical stains must be checked for appropriate positive and negative reactivity each time they are used. Most IVD antibody specification sheet will list tissue that will exhibit positive and negative staining patterns in the Performance Characteristics section. NOTE: abnormal tissue will not necessarily be labeled. Both negative and positive tissue controls should be processed using the same fixation, embedding, mounting, drying, epitope retrieval and immunostaining protocols as the patient tissue. Section 5 Automated Platform Performance Checks Information You Need to Know Information Located at the Instrument Comments The right protocol has been applied to the slide Right protocol Find the location for the completed slides in the automation software and look up the slide ID of the slide in question and check that the applied protocol is the correct one The information can be located in different places based on your automated solution. If the location is not known then contact your automated platform supplier 193 Chapter 16 | Troubleshooting 194 Information You Need to Know Information Located at the Instrument Comments Target retrieval has been performed under the right conditions Target retrieval Target retrieval solution: Find the location of the specific data for the target retrieval procedure – this information can be a part of the slides’ log file under completed slides Check that the right target retrieval solution has been used. Sub-optimal results can be seen if a high pH target retrieval solution is used for an antibody that according to the specification sheet requires a low pH target retrieval solution. Check that the target retrieval solution has been within the expiration limits when used for the slide Target retrieval temperature: Check that the temperature has been held within the limits of the target retrieval equipment throughout the course of the target retrieval process. A too high temperature can lead to impaired morphology and “over retrieval” of the antigen epitopes. Low temperature can lead to inadequate retrieval of the epitopes and thereby reduced staining intensity or lack of stained epitopes. The temperature data for the slide can be located at different places in the instrument software based on which automated solution is used. It is recommended to consult the user guides for the automated platform or to contact the supplier. High altitude installations need to provide information in the datalog that appropriate temperature was achieved Target retrieval time: Find the location of the specific data for the target retrieval time. The target retrieval time for the slide can be located at different places in the automated platforms software based on which automated solution is used. It is recommended to consult the user guides for the automated platform to find the location or to contact the supplier. Check that the time the slide actually received target retrieval is within the allowed limit for the assay performed If positive control has been run for the assay on the slide in question an evaluation of the effect of any deviations can be made based on the positive control. Based on this evaluation it can be determined if the slide can be used even though not processed inside the allowed limits. Special care shall be taken if the time has been reduced because the level of positivity in the samples is not known and can differ from the positive control and thereby potentially result in a false negative result. If any of the checks performed for the target retrieval step show irregularities it is recommended to re-run the sample and/or to get the automated equipment serviced by the manufacturer. LIS Protocols align with workstation test Protocols from the server Verify that correct protocols are received from the server Find the appropriate test name for the protocol. Verify that this name is mapped in the list of IHC or ISH test protocols. Contact the vendor technical support or bioinformatics department for assistance. Important to check this whenever new protocols or tests have been added to the server of the system. LIS may not automatically update and map new test protocols from the server. Staining process has been performed under the right conditions Staining process Reagents: Find the location of the specific data for the executed protocol – this can be separate or a part of the slides’ log file under completed slides. The location of the information can be looked up in the user guides for the automated platform or provided by the supplier. Check that the right protocol including the right reagents have been used to stain the slide in question. Sub-optimal results may be seen if another reagent than the validated for the assay either by the laboratory itself or the supplier and the laboratory in combination is used. Check that the right reagent volume has been applied according to the protocol Check that all the reagents used have been within the expiration limits when used for the slide Check that the label of the reagents used is actually in agreement with liquid in the bottles used on the automated platform. This can be done by looking at what the bottle previously has been used for in the bottle history (can be located different places dependent of automated platform used). If the bottle history shows successful use of the reagent prior to this then the right reagent is in the bottle. If you use reagents that you dilute from concentrate then it is important to check that the dilution has been done correctly. This can be done by validating the new dilution against a previous dilution still within the expiry limits. It is recommended that these checks are done to eliminate other variables like different automated platform or other reagents in the process. If a failure of either use of the target retrieval reagents or the target retrieval platform has been identified, remember to search for other slides which potentially have been submitted to the same failure and perform a quality check of these related slides. Check of the instrument overall performance is not based on a single slide. However, if a given automated platform is the main denominator between failing slides then it should be considered to make a check of the automated platform and potentially get it serviced by the supplier. It is recommended that controls are applied on every slide to ensure that it has received the exact same treatment as the sample being evaluated which supports the trouble shooting for the samples to a higher degree than a separate run or daily control will do. Troubleshooting | Chapter 16 Information You Need to Know Information Located at the Instrument Comments Check whether the reagent has been stored according to recommendation from the supplier Check that the reagent volume usages have been adequate compared to the calculated use. This can be done by checking the bottle history together with the weight of an empty container and compare this to the container used making it possible to estimate the consumption. Staining temperature: If temperature control is used during the staining process then check that the temperature has been held within the limits of the staining process. The location of the information can be found in the user guides for the automated platform or provided by the supplier. Check that the operating conditions for the automated platform are within specifications. The specified operating conditions for automated platforms are listed in the user guides. Incubation time: Find the location of the executed incubation times for the slide in question. This location can vary based on which automated platform is used. The location of this information is accessible either from the user guides or can be provided by the supplier. Check that the incubation times the slide actually received are within the allowed limits for the assay performed. Variation in incubation time can result in variation in staining results. Too long incubation time can potentially lead to increased level of background staining which can influence the interpretation of the slide. Too short incubation times can lead to false negative results. Wash: Find the location of the executed wash incubation times for the slide in question. This location can vary based on which automated platform has been used. The location of this information is accessible either from the user guides or can be provided by the supplier. Check that the incubation time the slide actually received in the wash steps are within the allowed limits for the assay performed. Wide variation in wash times can be allowed for most assays. However, too short wash time potentially can result in unwanted specific as well as non-specific background due to inadequate wash. Extensive wash times can for special assays lead to reduced intensity. Check that the wash buffer volume has been within the acceptable limits. How this is done can be different based on the automated platform used. Most platforms have estimated buffer consumption per run which can be checked against the actual usages. Furthermore, description of how to check the wash volume per slide may be provided in the user guides or can be requested from the supplier. Check that the wash buffer used is the buffer specified for the assay or has been validated for the assay by the laboratory Check that the wash buffer used for the slide has been within the expiry limits when used for the slide in question General: If a positive control has been run for the assay on the slide in question an evaluation of the effect of any deviations can be made based on the positive control. Based on this evaluation it can be determined if the slide can be used even though not processed inside the allowed limits. Special care shall be taken if the time has been reduced because the level of positivity in the samples is not known and can differ from the positive control and thereby potentially result in a false negative result. Check of leveling of the slide during the staining process. How this is done will vary based on the automated platform used. The information can be obtained either from the user guides for the automated platform or requested from the the supplier. It is important to check the level of the slide as a slide which is not in level during staining can lead to the area of interest not receiving the necessary reagents. This will result in inconsistent staining. If problems with the automated platform is indentified it is recommended to check other samples processed in the same period on the same platform or platforms in order to identify potential relations. To confirm or discharge the automated platform as the reason to problems with the staining quality it is a good idea to search for related samples e.g. other samples treated on the same automated platform run or in the same slide position, this can help lead in the right direction when doing the troubleshooting. If problems with the automated platform is identified, a service of the instrument is recommended 195 Chapter 16 | Troubleshooting Information You Need to Know Information Located at the Instrument Automated IHC staining platform has been performing as expected Automated IHC staining platform Reagent volume: Make sure that the volume your automated platform is set to use is adequate to cover the entire area of your sample over the total duration of the incubation time. It is important to take into account the potential evaporation of the reagent over time. Drying out of the tissue during the staining process can result in staining effects including no staining, inconsistent staining, extensive background and other staining artifacts Contamination: Contamination of the probe on the automated platform can lead to mixing of reagents in the probe and/or on the slide. This can lead to false positive staining or background on the slide. The inclusion of negative tissue controls can help identify if a contamination has taken place. A contamination of the visualization system in the substratechromogen will not be identified by the negative control as it will be present both on the positive control and the negative control. In order to identify this contamination it is necessary to make an investigational test where the first dispense of substrate chromogen can be removed and examined by mass spectrometry or ELISA for small quantities of visualization component. Contamination with bacterial and/or fungal growth can be seen in automated IHC platforms when the recommended maintenance schedule is not followed and/or some parts has been defective which increase the risk of bacterial and/ or fungal growth. Inspection of the visual parts of the instrument as well as keeping the maintenance schedule can prevent the contamination from happening. If growth is expected normal microbiological methods can be used to determine the presence of both bacteria and fungus. After identification the automated platform has to be cleaned according to specifications listed in the user guides or the supplier can be contacted for advice on how to clean. Contamination caused by inadequate wash of the automated platforms probe. After the probe aspirates reagent it has to be washed before transferred to another reagent bottle. If this wash is not adequate then the second bottle can be contaminated with reagent from the previous bottle. This will be identified as an unspecific reaction for a given marker e.g. CD20 staining in the nuclei if the first aspiration wash from Ki67. This contamination can be confirmed by having the probe go into a bottle of wash buffer or other neutral fluids after the first aspiration and then measure the content of the previous reagent by either mass spectrometry or ELISA. Contamination of the tissue by the probe dropping a drop of unrelated reagent onto the slide. This will be recognized as a false positive reaction either being in the wrong structure or wrong location. Use of positive tissue controls including multiple organ types on each slide will make it possible to identify contamination of this origin. Waste separation: Automated platform separates hazardous from non-hazardous waste. Failure of this separation does not impact the staining process and thereby should not influence the staining result. Bulk luid supply: The automated platforms normally have a function which makes it possible to check whether the supply of bulk fluids is working adequately. This check can be a prime of the bulk fluid trough the system securing that there are no leakages or clots preventing the fluid from flowing. Check of the bulk fluid supply is described in the user guides or can be requested from the supplier. General: If in doubt about the performance of the instrument, thorough observation of the automated platform during the staining process can give an indication of whether the individual steps is being executed as expected. Check that the instrument is closed correctly as the performance depends on the surrounding conditions for optimal staining. Improper closing can result in inability to start or create staining conditions which are sub-optimal do to the air getting into the instrument through the open cabinet. Instrumentation should be installed away from direct sunlight. Sunlight can affect the staining conditions and viability of reagents. Acknowledgements References Sections, in whole or parts thereof, from the previous editions 1. Wood G, et al. Suppression of endogenous avidin-binding activity in tissues and its relevance to biotin- avidin detection systems. J Histochem Cytochem 1981;29:1196-204. of this Guidebook are used in the 6th edition. We sincerely thank and acknowledge the contribution of the authors. Special acknowledgements to: Karen N. Atwood and Dako Technical Support Group 2. Sayaki H, et al. Azure B as a counterstain in the immunohistological evaluation of heavily pigmented nevomelanocytic lesions. Appl Immunohistochem 1995; 3:268-71. 3. Federal Register: January 24, 2003;68(12) 42CFR Part 493. 4. College of American Pathology; Anatomic Pathology Checklist, October 2005. 5. College of American Pathology; Anatomic Pathology Checklist, July 2013; p. 34-38. 196 197 Appendix A Antibodies Revised by: Sussie S. Jensen, MSc Original version by: Thomas Boenisch An •ti •bod •y (n.) A Y-shaped protein on the surface of B cells that is secreted into the blood or lymph in response to an antigenic stimulus, such as a bacterium, virus, parasite, or transplanted organ, and that neutralizes the antigen by binding speciically to it; an immunoglobulin. The American Heritage® Dictionary of the English Language Antibodies | Appendix A The central reagent common to all immunohistochemical techniques is the antibody. The availability of new monoclonal antibodies, antisera and their immunoglobulin fractions to an ever-increasing number of clinically useful tissue biomarkers (antigens) has enormously expanded the quantity and quality of the immunohistologic repertoire. This Appendix A briely covers the structural and biochemical features of antibodies and antigens and is as such not a necessity in order to understand the immunohistochemical staining procedure. However, to better comprehend the potential of immunohistochemical as used in clinical research and clinical pathology as well as associated problems, it is helpful to have a basic knowledge of antibodies, their potentials and their limitations. Appendix A.1 Immunoglobulins Antibodies belong to a group of proteins called immunoglobulins (Ig) that are present in the blood of immunized individuals. Figure A.1 Diagram showing the structure of an immunoglobulin molecule. It comprises two identical heavy (H) chains and two identical light (L) chains. The heavy chains of IgG are denoted as gamma (γ) chains. The two L chains are either of type kappa (κ) or lambda (λ). Inter- and intrachain disulphide bonds (──♦──) contribute to the structure and stability of the molecule. One molecule of IgG has a molecular weight of ~150 kDa. The removal of cells and ibrin from blood is used to collect the serum fraction frequently referred to as antiserum. Listed in order of decreasing quantity found in plasma or serum, im- Much of the structure of the IgG molecule has been determined munoglobulins comprise ive major classes: immunoglobulin G in part by proteolytic digestions and chemical dissociation of the (IgG), IgA, IgM, IgD and IgE. Each is composed of two identi- molecule (Figure A.2). When the IgG molecule is digested by cal heavy chains (H) and two identical light chains (L). The H the enzyme papain it results in the cleavage of a covalent bond chains differ in antigenic and structural properties, and deter- on the N-terminal side of the inter-heavy chain disulide bridg- mine the class and subclass of the molecule. The two L chains es. This yields two antigen-binding fragments (Fab, monovalent) are either of type kappa (κ) or lambda (λ). Distribution of κ and and one crystalline fragment (Fc). The digestive enzyme pepsin λ chains differs in all Ig classes and subclasses, as well as be- cleaves the γ chains on the C-terminal side of the inter-heavy tween different species. Covalent bounds join L to H and H to chain disulide bridges, resulting in one antigen-binding fragment H chains. By participating in the tertiary structure, they confer (F(ab')2, bivalent). In this case, the Fc fragments are destroyed. greater stability to the immunoglobulin molecule. Using chemical reductive dissociation of an IgG molecule will split interchain disulide bridges, and if the free sulfhydryl groups Of the ive classes of immunoglobulins, IgG and IgM will be con- are blocked, it results in the formation of two H chains (molecular sidered in more detail here, as these are by far the most frequent- weight of 50 kDa each) and two L chains (25 kDa each). ly utilized antibodies in immunohistochemistry. Unless otherwise noted, most of what is described of the IgG structure in this text The IgG molecule can be further divided into so-called do- was learned from studies with human IgG of subclass IgG1. mains, namely the variable domains (V) and the constant domains (C). Each domain contains 110 to 120 amino acids and IgG one intrachain disulide bond. On the variable domain of the IgG has the general formula of one molecule of IgG is com- light chain (VL), and on the variable domain of the heavy chain posed of two γ heavy chains, and two light chains of either type κ or type λ (γ2 κ2 or γ2 λ2) (Figure A.1). (VH), the amino terminals of the immunoglobulin molecule are located. Together, VL and VH form the antigen-combining site. 199 Appendix A | Antibodies Several hypervariable (HV) regions are located within the VL and VH domains of the antibody. It is these hypervariable regions that recognize the antigen. During the formation of the antibody/antigen complex, the HV regions are brought into close proximity to the antigenic determinant (epitope). The distance between the antigen and HV regions of the antibody is approximately 0.2 to 0.3 nm. In this region, unique structural speciicities called idiotypic determinants are located. Each antibody clone expresses its own idiotype. Each L chain also has one constant domain (CL) in addition to the VL domain. The H chain has three constant domains (CH1, CH2 and CH3). The hinge regions are located between the CH1 and CH2 domains of the H chains. Minor differences within these hinge regions contribute to the subclass speciicity of immunoglobulin G. The same are designated by subscripts as in IgG1, IgG2a, IgG2b, IgG2C and IgG3 for mice and IgG1, IgG2, IgG3 and IgG4 for humans. Whereas in human IgG the overall Figure A.2 Diagram showing the structure of rabbit IgG (which exists as a single major subclass). The heavy (H) and light (L) chains are composed of variable (V) and constant (C) domains and are linked by inter- and intrachain disulide bonds (──♦──). Proteolytic digestion with papain (• • • • •) yields two antigen-binding fragments (Fab) and one crystalline fragment (Fc), whereas digestion with pepsin (– – – –) yields one F(ab')2 fragment. ratio of κ to λ is 2:1, mice have approximately 95% κ chains, and therefore most monoclonal IgG antibodies from this species have κ chains. The number of disulide bridges linking the heavy Whereas IgG is the most abundant antibody in the hyperimmunized chains also varies among the IgG subclasses. IgG1 and IgG4 host, in the newly immunized animal, IgM is the irst humoral anti- each have two, while IgG2 and IgG3 have four and ive, respec- body detectable. The primary antibody formation proceeds in sev- tively. Because of the lexibility of the hinge region, the angle eral major stages. The antigen irst reaches equilibrium between that both Fab fragments form can vary to accommodate varying extra- and intravascular spaces, then undergoes catabolism result- distances between epitopes. ing in smaller fragments, and inally is eliminated from the intravascular spaces by the newly formed antibodies. The period from the Rabbits have only one IgG subclass which, like the human and introduction of an antigen until the irst appearance of humoral IgM mouse IgG, also has a molecular weight of ~150 kDa, with two antibodies is called the latent period and may last approximately heavy chains (~50 kDa each) and two light chains (~25 kDa one week. Within two weeks, or in response to a second antigen each) under non-reducing conditions. challenge, IgG class antibodies usually predominate. Like all proteins, antibodies are subject to catabolism. Whereas antibodies of IgM class IgM have a relatively short half-life of only four to six days, IgG IgM is a pentamer (MW ~900 kDa) consisting of ive subunits antibodies have a mean survival of approximately three weeks. The of ~180 kDa each (Figure A.3). The general formula can be ex- serum antibody level will decrease after this period. pressed as (μ2 κ2) or (μ2 λ 2)5. Each subunit is linked by a sulfhydryl-rich peptide, the J chain (15 kDa), and consists of two Antibody formation on the molecular level is a complex pro- heavy chains μ and two light chains of type κ or λ. The J-chains cess, and a detailed account of it is beyond the scope of contribute to the integrity and stability of the pentamer. As with this guidebook. IgG, IgM subunits can be fragmented by enzymatic and reduc- 200 tive cleavage into F(ab')2, Fab and Fc portions, as well as heavy Immunoglobulins in different species and light chains, respectively. Human, mouse and rabbit have different sets of immuloglobulins Antibodies | Appendix A A B Figure A.3 Diagram showing A) the ive subunits of mouse IgM linked by disulide bridges (──♦──) and the J chain to form a pentameric ring structure. B) Each subunit comprises two µ heavy (H) chains and two light (L) chains each composed of constant (C) and variable (V) domains. and subtypes within each immunoglobulin class. Humans and its ability activate an antibody generation response. In immuno- mice have ive antibody classes (IgA, IgD, IgE, IgG, and IgM), but histochemistry, the biomarkers of interest are proteins, and full in rabbits only four classes have been identiied (IgA, IgE, IgG, and length or fragments of proteins are often used as these are the IgM). The most commonly used immunoglobulin for immunohisto- most potent antigens. Shorter peptide sequences (10-20 amino chemistry, IgG, exists in ive subclasses in mouse but only one in acids) can be used with advantage in terms of higher speciic- rabbit. Humans have two IgA subclasses, mice have only one IgA ity and reduced likelihood of cross-reactivity, but requires that subclass, but 14 IgA subclasses have been found in rabbits. the peptide is coupled to a carrier protein (e.g. keyhole limpet Table A.1 List of immunoglobulin subclasses in human, mouse and rabbit. hemocyanin (KLH), bovine serum albumin (BSA) or ovalbumin (OVA)). The requirement of a carrier protein also applies to Ig Class Human Mouse Rabbit IgA IgA1, IgA2 IgA IgA1, IgA2, IgA3, IgA4, IgA5, IgA6, IgA7, IgA8, IgA9, IgA10, IgA11, IgA12, IgA13, IgA14 (DNP) and biotin. These smaller substances – often referred to carbohydrates, DNA and smaller molecules like dinitrophenol as haptens - can also elicit an immune response, but are less potent antigens compared with protein and their derivatives. IgD IgD IgD IgE IgE IgE IgE The part of the antigen that is speciically recognized by the IgG IgG1, IgG2, IgG3, IgG4 IgG1, IgG2a, IgG2b, IgG2c*, IgG3 IgG immune system is called an epitope; this is the antigenic de- IgM IgM IgM IgM terminant that is bound by the antibody. One protein can have multiple epitopes, but a monoclonal antibody only recognize *IgG2c is found in some inbred mice strains one of these epitopes in the protein. The epitopes of proteins are divided into two categories, conformational epitopes and linear epitopes, based on their structure and interaction with the antibody (1). Linear epitopes are deined by the primary Appendix A.2 Antigens amino acid sequence of a particular region of a protein and vary typically from 5 to 20 amino acids in length. Conforma- When a biomarker of interest is used for immunizing an animal, tional epitopes are deined by the spatial structure of the native the biomarker is referred to as an antigen, a name derived from protein and are situated on disparate parts of in the protein 201 Appendix A | Antibodies sequence, but are brought close to each other in the folded ammonium sulfate precipitation and ion exchange chromatog- native structure. In IHC, the conformation of the proteins in a raphy, Protein A or G puriication or afinity isolation. tissue sample will be affected by the formalin ixation and the antigen retrieval process, which may be the reason that some antibodies directed against linear epitopes perform better (2). Polyclonal antibody Epitopes To determine which part of the biomarker is most suited for generating a speciic antibody, the biomarker should be thoroughly investigated. Features like the amino acid sequence, Antigen structure and spatial position of the epitope in the protein can be examined by integrated computational and microarray-based analysis to ind the most immunogenic epitopes and to prevent cross-reactivity (3). Appendix A.3 Polyclonal Antibodies Polyclonal antibodies are a heterogeneous mixture of antibodies directed against various epitopes of the same antigen Figure A.4 Schematic diagram of polyclonal antibodies binding to various epitopes on an antigen. (Figure A.4). The antibodies are generated by different B-cell clones of the animal and as a consequence are immunochemically dissimilar. The antibodies in a polyclonal mixture can Appendix A.4 Monoclonal Antibodies have slightly different speciicities and afinities (see Antibody Afinity). Polyclonal antibodies are most frequently produced Monoclonal antibodies are a homogeneous population of im- in rabbits but also made in other mammals including goat, munoglobulin directed against a single epitope. The antibod- swine, guinea pig and cow. Rabbits are frequently the species ies are generated by a single B-cell clone from one animal and of choice in polyclonal antibody production due to the appreci- are therefore immunochemically similar. Monoclonal antibod- able amount of anti-serum and relative rarity of human antibod- ies are most commonly produced in mice and rabbits, but may ies to rabbit proteins when compared to other species such as also be produced in other species such as rat and camel. goats. Additionally, rabbit antibodies precipitate human proteins over a wider range of either antigen or antibody excess. Mouse Monoclonal Antibodies The mice are immunized and boosted every two weeks over a 202 Polyclonal antibodies are produced by immunizing with an- period of two months. The animal’s immune response is moni- tigen (also known as immunogen) using doses ranging from tored through periodic testing of the serum. Upon achieving an 10 ug to 200 ug. Antigen can be prepared with or without an acceptable immune response, the B lymphocytes are isolated adjuvant such as Freund’s Complete or Incomplete Adjuvant from the spleen and fused with an immortal cell line (myeloma which can boost the immune response. For smaller or less im- cell line/fusion partner). The B lymphocytes confer the capaci- munogenic proteins or peptides, the immunogen should be ty to produce speciic immunoglobulin while the fusion partner coupled to carrier proteins. The immunization period lasts from cell line enables immortality and indeinite growth in culture. The 3 to 8 months with biweekly immunogenic boots. The polyclon- fused and immortalized cell line is called a hybridoma. The hy- al antibody preparation can be used in the form of stabilized bridoma cell line is cultured, selected and sub-cultured by limit- antisera or further puriied. Puriication of the immunoglobulin; ing dilution to isolate a stable clone with a high antibody produc- to eliminate other serum proteins can be performed through tion capacity. For production of tissue culture supernatant, the Antibodies | Appendix A hybridoma cell line is cultured in multiple tissue culture lasks. sized epitopes. There is also a tendency with rabbits to Bioreactors can be used for large scale hybridoma growth and generate antibodies with higher afinity and overall avidity. antibody generation. A bioreactor is a system that continually The resulting rabbit antibodies also have enhanced binding replenishes cells with fresh media and promotes growth for cul- properties due to heavy glycosylation. Mouse hybridomas tivation of concentrated amounts of antibody. In mice, ascites however tend to generate a higher yield of immunoglobulin luid has also been used for production due to a very high con- than rabbit hybridomas and the mouse hybridoma cell lines centration of antibody compared with tissue culture superna- are typically more stable in culture. tant; however contaminating antibodies will be present in ascites preparations. Most monoclonal antibody preparations are from stabilized tissue culture supernatant. Further puriication of immunoglobulin can be performed through the use of protein A Appendix A.5 Polyclonal Antibodies versus Monoclonal Antibodies or G chromatography or afinity puriication. When comparing the advantages and disadvantages of poly- Rabbit Monoclonal Antibodies clonal and monoclonal antibody preparations, there are ben- While the vast majority of monoclonal antibodies are pro- eits to both. Due to their multiclonality, polyclonal antibodies duced in mice, a growing number of rabbit monoclonal anti- are typically more robust reagents when used on routinely-pro- bodies are being manufactured. The process for generating cessed tissue specimens. The ability of a polyclonal reagent a rabbit hybridoma is comparable to the mouse hybridoma to recognize multiple epitopes on a single molecule means generation described in Figure A.6. The use of rabbits for that the reagent is not as subject to the deleterious effects on monoclonal antibody production confers some advantag- epitopes of pre-analytic processing of specimens, as are mon- es over mouse monoclonals. These advantages are also oclonal antibodies which are directed against a single epitope. shared by rabbit polyclonals. Rabbits are reported to have The presence of antibodies to multiple epitopes however more diverse epitope recognition than mice (less immuno- can increase the chance for cross-reactivity (see Antibody dominance) and an improved immune response to small- Cross-Reactivity) with other proteins. Monoclonal antibody Epitopes Antigen Figure A.5 A given monoclonal antibody clone reacts with only one speciic epitope on an antigen. Figure A.6 The process of monoclonal antibody production. 203 Appendix A | Antibodies Monoclonal antibodies have the advantage of lot-to-lot con- On the other hand, monoclonal antibodies are of uniform afinity sistency and lack the inherent variability of polyclonal antibod- and loss of staining may be due to the dissociation of the antibody ies due to the immunological state of the animal. The use of a from its epitope if the afinity is low. Therefore monoclonal antibod- hybridoma in monoclonal antibody production enables a sus- ies of high afinity should be selected, if possible. As indicated tained production of antibody. above, factors that weaken the antigen-antibody bond such as high salt concentrations, high temperature and very low pH during the washing of the specimens should be avoided. Experience in Appendix A.6 Antibody Afinity the handling of antibodies in immunohistochemistry has shown that the washing and incubation in buffer baths can be safely reduced Antibodies from hyperimmunized animals not only differ with re- and that gentle agitation helps to reduce background staining (7). gard to the determinants they recognize on multivalent antigens, but also differ in their afinities for the same. The term “afinity” has The afinity of antibodies is also related to their capacity to form been used to describe both intrinsic and functional afinities (4). insoluble immune complexes. Generally, the higher the afinity of an antibody, the greater its tendency to form a precipitate. The intrinsic afinity of an antibody resides in the HV region and Precipitation proceeds through a rapid stage in which soluble is determined by the same sequence of amino acids that deter- antigen-antibody complexes form, followed by slower aggre- mines speciicity. In immunohistochemistry, the functional afinity gation and, eventually, precipitation. Non-precipitating anti- of an antibody or an antiserum can be very loosely deined by bodies are mostly of lower afinity and are incapable of forming the time required to reach equilibrium with the tissue antigen. If the lattice required for precipitation to occur. equal aliquots of two antibodies or antisera of identical titer are incubated for increasing periods of time with the antigen on the Monoclonal antibodies, regardless of whether they are of high tissue, the antibody that reaches a plateau of maximum staining or low afinity, do not form a lattice with antigen, and, hence intensity irst is of a higher functional afinity. The term “avidity” only rarely form insoluble precipitates. However, in immunohis- has been used synonymously to describe functional afinity (5), tochemistry, the capability of a primary antibody to form pre- but has also been used to denote the strength of the binding cipitating immune complexes is of little importance because reached between antibody and its antigen (6). The term avidity reaction with immobilized tissue antigen entails antibody cap- has also been used to describe the sum total of all intrinsic afin- ture onto tissue rather than precipitation. ities found in a polyclonal antibody population. Because antigen-antibody reactions are reversible, the simple Appendix A.7 Antibody Cross-Reactivity immune complexes formed on tissue may occasionally dissociate during the washing cycles used in immunohistochemistry. The term “cross-reactivity” denotes an immunochemical activity The ease and extent of this dissociation vary from antibody to that can occur either between an antibody and two or more anti- antibody, and low salt concentrations as well as low temperatures gens or vice versa, when an antigen reacts with several different will reduce the likelihood of weak staining due to dissociation of antibodies. Typical examples are when anti-λ (or -κ) chain anti- an already formed immune complex. Thus, high afinity antibod- bodies interact with all ive Ig classes or when carcinoembryonic ies are desirable and have the advantage that during washing, antigen (CEA) reacts with antibodies against CEA, blood group dissociation is less likely to occur than with low-afinity antibodies. antigens and normal tissue proteins, respectively. The common As mentioned before, a polyclonal population of antibodies con- denominator in each case is the sharing of at least one common tains a more or less continuous spectrum of low to high afinities epitope between several antigens. against several epitopes on a given antigen. Therefore after incu- 204 bation with a primary antibody of this type, excessive washing is Another valid use of the term cross-reactivity denotes the ex- unlikely to result in any appreciable loss of staining. perimentally or accidentally induced changes within one or Antibodies | Appendix A several epitopes, through antigen retrieval (8), leading to a Appendix A.9 Antibody Stability and Storage possible loss of speciicity by a given monoclonal antibody for this antigen. The term cross-reactivity also describes the inter- Polyclonal antibodies, when stored unfrozen and used sub- action of an antibody with similar or dissimilar epitopes on un- sequently in immunohistochemistry, are somewhat less stable related antigens. This latter phenomenon however is frequently as immunoglobulin fraction compared to whole antiserum (7). a property of low afinity antibodies, and is usually subject to However, this reduced stability was found to depend largely on change because of afinity maturation during immunization. the method of puriication and storage as well as on the method of application. Exposure of antibodies to extreme pH, as Cross-reactivity of antibodies to human antigens with identical well as high or very low concentrations of salts during puriica- or similar antigens of other species, or “cross-species cross-re- tion tends to decrease their stability more than does exposure activity,” can be of interest to the researcher and veterinarian be- to mild conditions such as ion exchange chromatography. cause of the scarcity of animal-speciic antibodies. To overcome this, two publications reported the results of cross-species reac- Monoclonal antibodies also have been shown to be inluenced tivity studies using commercially available antihuman polyclonal in their performance by methods of puriication and storage; and monoclonal antibodies (9, 10). It was demonstrated that the 42% of monoclonal antibodies investigated by Underwood majority of animal antigens selected showed strong reactivity and Bean showed changes in speciicity, afinity and cross-re- with antihuman antibodies. However, for more technical detail activity (11). Antibodies of class IgM and subclass IgG2b were on the use of a given mouse primary antibody on animal tissues, especially sensitive. the reader is referred to animal research kit products. Antibody stability in commercially produced reagents is deterThe terminology of cross-reactivity however is misplaced when mined best by real-time and real-temperature testing by each describing any observed staining by the same antibody of manufacturer. Most manufacturers demonstrate stability by different cells or tissue components, regardless whether they testing during a pre-determined period of time, i.e. the “shelf contain common antigens, as this would distort the strict immu- life”. While many antibodies may retain activity longer, the only nochemical deinition of the term. regulatory requirement for the manufacturer is to certify the period of time that the antibody has been tested. There is no requirement to continue testing until the antibody loses activity. Appendix A.8 Antibody Reaction Rates Sometimes, the conditions for the storage of reagents in the user’s laboratory are not identical to those that prevailed during Although under ideal conditions antibodies react with their antigens the manufacturer’s shelf life studies. Because of the possibility very rapidly, in immunohistochemistry the conditions are rarely ide- of adverse storage conditions after the purchase of the prod- al. However, very short incubation periods are made feasible by the uct, the manufacturer can only offer a limited liability instead of relatively rapid reaction rates that occur when higher concentra- predicting the actual demise of a reagent. tions of high-afinity primary and link antibodies are used. Generally, the size and shape of the antibody molecule and its conjugates The only possible corollary to these requirements is to allow lab- or complexes appear to be of little consequence in immunohisto- oratories to document the activity of the product until the loss of chemistry. Insuficient tissue penetration, even when staining nu- the same. Alternatively, laboratories may aliquot and freeze un- clear or cytoplasmic antigens, has never been observed, regard- diluted antibody at -20 °C for later use. At this time, laboratories less of whether primary antibodies of class IgM (900 kDa), large must conirm activity prior to the use of the antibody in any test. complexes like PAP (400-430 kDa) or APAAP (~560 kDa) or dextran-polymer-linked reagents were used. However, it is reasonable Finally, expiration dating as practiced today also serves the to assume that gross overixation of tissue may make penetration purpose of conforming to regulatory requirements. Regulato- more dificult for antibodies and their complexes. ry guidelines in place in the US for clinical laboratories have 205 Appendix A | Antibodies been mandated by the Clinical Laboratory Improvement Act Store most RTU antibodies and their conjugates solutions at of 1988 and by the College of American Pathologists. These 2-8 °C, because freezing and thawing is known to have a del- regulations mandate that expired reagents cannot be used in eterious effect on their performance. This also applies to entire the clinical diagnostic laboratory on human tissue. kits that contain ready-to-use reagents, including monoclonal antibodies. Store concentrated protein solutions such as an- Handling of Antibodies tisera and immunoglobulin fractions in aliquots and frozen at In order to achieve optimal performance from reagents used in -20 °C or below, in order to prevent cycles of repeated freezing immunohistochemistry, it is imperative to observe basic rules and thawing. Bring frozen protein solutions to room tempera- for their handling and storage. If properly maintained, most ture slowly, and avoid temperatures above 25 °C. reagents will remain stable for months or even years. Recommendations given by the manufacturer on speciication sheets Use and Care and on vial labels should always be followed. Proper reagent care can reduce problems stemming from contamination, heat or excessive light exposure. Reagent contamina- Receiving tion can be avoided by the use of clean pipette tips. Prompt return Although many commercially produced immunochemicals are of reagents to proper storage conditions will prolong their shelf life. guaranteed to be stable for up to several years, ready-to-use (RTU) Familiarity with the nature of antibodies, their capabilities and antibodies have a shorter shelf life. Upon receipt, immunochem- limitations, will allow the user to better utilize these reagents icals should be promptly stored according to the manufacturer's and to more eficiently solve problems, if they occur. recommendations. The users should log reagents by entering the manufacturer’s lot numbers, expiration date, date of receipt and in- References voice number. These entries provide valuable information for the 1. Van Regenmortel MH. What is a B-cell epitope? Methods Mol Biol 2009; 524:3-20. user, especially if later reclamations should become necessary. Storage Containers and Temperature Perhaps the two most important considerations when storing antibodies are the storage container and the temperature. Ideally, preferred materials for storage containers of protein solutions should have negligible protein adsorptivity. Polypropylene, polycarbonate or borosilicate glass are recommended and are used widely. Solutions containing very low concentrations of protein (i.e. less than 10-100 µg/mL), should receive an addition of immunochemically inert protein. Generally, 0.1% to 1.0% bovine albumin is used to reduce loss through polymerization and adsorption onto the container. Containers made of clear and colorless materials are preferred, as these will allow ready inspection of contents. Container labels also should allow access for inspection. 2. Fowler CB, Evers DL, O'Leary TJ, Mason JT. Antigen retrieval causes protein unfolding: evidence for a linear epitope model of recovered immunoreactivity. J Histochem Cytochem 2011; 59:366. 3. Peri C, Gagni P, Combi F, Gori A, Chiari M, Longhi R, et al. Rational epitope design for protein targeting. ACS Chem Biol 2013; 8:397-404. 4. Hornick CL and Karush F. Antibody afinity. 3. The role of multivalance. Immunochemistry 1979;9:325-40. 5. Steward MW and Steensgaard J. Antibody Afinity: Thermodynamic Aspects and Biological Signiicance. Boca Raton: CRC Press, 1983. 6. Herschowitz HID Immunophysiology: Cell function and cellular interactions in antibody formation. In Bellanti JA. Immunology III. Philadelphia: Saunders, 1985. 7. Boenisch T. Formalin-ixed and heat-retrieved tissue antigens: a comparison of their immunoreactivity in experimental antibody diluents. Appl Immunohistochem 2001; 9:176-9. 8. Alexander J and Dayal Y. Appl Immunohistochem 1997; 5:252-3. Probably more than any other factor, observe proper storage temperature as recommended by the manufacturer. Monitor refrigerators and freezers used for storage of immunochemicals for accurate and consistent temperatures. Store valuable or large quantities of immunochemical reagents in equipment with temperature alarm and emergency back-up power systems. 206 9. Smith RA. J Histotech 1990; 13:255-69. 10. Martin CA and Badrán AF. Evaluation of antibody crossreactivity of mouse tissues: study of a panel of antibody reacting to human antigens. Appl Immunohistochem 1998; 6:84-8. 11. Underwood PA and Bean PA. The inluence of methods of production, puriication and storage of monoclonal antibodies upon their observed speciicities. J Immunol Meth 1985; 80:189-97. 207 Appendix B Basic Immunochemistry Revised by: Sussie S. Jensen, MSc Original version by: Thomas Boenisch Im •mu •no •chem•is •try (n.) A method for the detection and localization of proteins and other cellular components using antibodies that speciically label the materials. Collins English Dictionary – Complete and Unabridged Basic Immunochemistry | Appendix B In immunohistochemistry (IHC), antibody titer and dilutions as of monoclonal antibodies in cell culture supernatants. These well as incubation time, temperature and pre-treatment of tis- dilutions may likely be exceeded in the future due to ever-in- sue samples are tightly interwoven in their effect on staining creasing sensitivities of newer detection methods, including quality. These factors can be changed independently, or as is the use of more effective antigen retrieval procedures. more often the case, in complementary fashion to bring about positive differences. The predominant goal of an IHC staining is to achieve optimal speciic staining accompanied by mini- Appendix B.2 Antibody Dilution mal interference from background staining. This Appendix B will briely describe how changes to titer, dilution, incubation Correct dilutions will contribute to the quality of staining if they time and temperature may inluence the staining reaction. See are prepared accurately and consistently. Often a manufactur- Chapters 4 and 5 for detailed discussion of the implications er offers ready-to-use (RTU) reagents ready for use, or recom- when changing these parameters in an IHC staining protocol. mends dilution ranges compatible with other variables such as method, incubation time and temperature. If this information is not provided, optimal working dilutions of immunochemical rea- Appendix B.1 Antibody Titer gents must be determined by titration. Correct dilutions are best determined by irst selecting a ixed incubation time and then Optimum antibody titer may be deined as the highest dilution by making small volumes of a series of experimental dilutions. of an antibody (monoclonal or polyclonal) that results in maxi- Depending on specimen size, applications of 0.1 0.4 mL of solu- mum speciic staining with the least amount of background un- tion per section is generally adequate. It should be noted that at der speciic test conditions. This highest dilution is determined least on parafin sections optimal dilutions of primary antibodies primarily by the absolute amount of speciic antibodies present. are not only signaled by a peak in staining intensity, but also by the presence of minimal background (maximal signal to noise The amount of antibody required for optimal staining in any ratios). Once the optimal working dilution has been found, larger given test has to be determined by different antibody dilutions. volumes can be prepared according to need and stability. For polyclonal antisera the amount of speciic antibodies is often not measurable, so the optimal staining titer is deter- The extent to which monoclonal antibodies can be diluted is mined by a series of antiserum dilutions. Afinity puriication subject to additional criteria. Because of their restricted molec- of polyclonal antisera produces little beneit for immunohis- ular conformation and well deined pI, monoclonal antibodies tochemical applications, because non-speciic antibodies are more sensitive to the pH and ions of the diluent buffer (1). and soluble aggregates – frequent sources of non-speciic Indeed, it has been demonstrated that almost all monoclonal background become enriched also. For monoclonal antibody antibodies could be diluted higher and stained more intense- preparations, the absolute concentration of speciic antibodies ly at pH 6.0, especially after the use of heat-induced antigen can be readily determined, and frequently forms the basis for retrieval and almost all monoclonal antibodies stained more in- making required dilutions. tensely in the absence of NaCl (1). Of several diluents used in this investigation, phosphate buffered saline (PBS), although An optimal antibody dilution is also governed by the intrinsic still widely used as a diluent for primary antibodies, was found afinity of an antibody. If the titer is held constant, a high-afinity to suppress the reactivity of all monoclonal antibodies tested antibody is likely to react faster with the tissue antigen and give (1). Differences in the net negative electrostatic charges of more intense staining within the same incubation period than the target antigen are likely the explanation for these pH- and an antibody of low afinity. ion-related observations (1, 2). In more practical terms, titers may vary from 1:100 to 1:2000 Dilutions are usually expressed as the ratio of the more con- dilution of polyclonal antisera and from 1:10 to 1:1000 dilution centrated stock solution to the total volume of the desired 209 Appendix B | Basic Immunochemistry dilution. For example, a 1:10 dilution is made by mixing one Appendix B.4 Incubation Temperature part of stock solution with nine parts diluent. Two fold serial dilutions are made by successive 1:2 dilutions of the Because antigen antibody reactions happen more quickly at previous dilution. In order to make a very small volume of a 37 °C compared with room temperature, some workers prefer highly diluted solution, it may be necessary to make it in two to incubate at the higher temperature. However, while increas- steps. For example, to prepare 1.0 mL of a 1:1000 dilution, es in incubation temperature allow for greater dilution of the irst make 100 µL of a 1:10 dilution (10 µL + 90 µL), and then antibody and/or a shortened incubation time, consistency in 1000 µL of a 1:100 dilution using 10 µL of the intermediate incubation time becomes even more critical. It is not known dilution (10 µL + 990 µL). whether an increased temperature promotes the antigen antibody reaction selectively, rather than the various reactions that Antibody Incubation give rise to background. As mentioned above, incubation time, temperature and antibody titers are interdependent. A change in one factor may In some cases, e.g. a research/experimental setting, a temper- affect the others. ature of 4 °C is may be used in combination with overnight or longer incubations. Slides incubated for extended periods, or at elevated temperature, should be placed in a humidity-con- Appendix B.3 Antibody Dilution trolled environment to prevent evaporation and drying of tissue sections. Similarly, tissue incubated at room temperature in a There is an inverse relationship between incubation time and very dry or drafty environment will be at risk of drying out. Most antibody titer: the higher the antibody titer, the shorter the in- automated staining instruments used in the pathology labora- cubation time required. In practice however, it is important to tories today are designed to take into account the potential consider the alignment of protocol incubation times for the an- issues of drying of tissue sections. tibodies used the laboratory in order to achieve optimal worklow (see Chapter 5). Appendix B.5 Basic Enzymology in IHC Incubation times for the primary antibody may vary within up to 24 hours, with 10-30 minutes probably being the most wide- Immunoenzymatic staining methods utilize enzyme substrate ly used incubation time. For an antibody to react suficiently reactions to convert colorless chromogens into colored end strongly with the tissue antigen in a short period of time, it must products. Of the enzymes used in these applications, only be of high afinity and concentration, as well as have the op- horseradish peroxidase and calf intestine alkaline phosphatase timal reaction milieu (pH and diluent ions). Variables believed will be described briely. Detailed descriptions and information to contribute to increased non-speciic background staining for the preparation of the most commonly used substrate-chro- should be kept to a minimum (see Chapter 15). Low afinity mogen mixtures and their appropriate use and advantages or and/or low titer antibodies must be incubated for longer pe- disadvantages are available in references 4-7. riods than antibodies of high afinity and concentration. But 210 nothing can be gained by prolonging primary antibody incu- Horseradish Peroxidase (HRP) bation beyond the time at which the tissue antigen is saturated The enzyme horseradish peroxidase (HRP) has a molecular with antibody. Consistent timing and temperature (see below) weight 40 kDa and is isolated from the root of the horserad- of the antibody incubation step is important to reduce varia- ish plant (Cochlearia armoracia). HRP has an iron-containing tions in overall staining quality and intensity, which otherwise heme group (hematin) as its active site, and in solution is color- may lead to incorrect interpretation of results. These criteria ed brown. The hematin of HRP irst forms a complex with hy- are particularly essential in efforts that attempt to assess the drogen peroxide (H202), and then causes it to decompose, degree of tumor differentiation. resulting in water and atomic oxygen. HRP oxidizes sever- Basic Immunochemistry | Appendix B al substances, two of which are polyphenols and nitrates. It AP had not been used extensively in IHC until publication of should be noted that similar to many other enzymes, HRP and the unlabeled alkaline phosphatase-antialkaline phosphatase some HRP like activities can be inhibited by excess substrate. (APAAP) procedure (8, 9). The soluble immune complexes The complex formed between HRP and excess hydrogen per- utilized in this procedure have molecular weights of approx- oxide is catalytically inactive, and in the absence of an electron imately 560 kDa. The major advantage of the APAAP proce- donor (eg, chromogenic substance), is reversibly inhibited. It is dure compared to the earlier peroxidase techniques was the the excess hydrogen peroxide and the absence of an electron lack of interference posed by endogenous peroxidase activ- donor that brings about quenching of endogenous peroxidase ity. Because of the potential distraction of endogenous per- activities. Cyanide and azide are two other strong (reversible) oxidase activity, the alkaline phosphatase techniques were inhibitors of peroxidase. particularly recommended for use on blood and bone marrow smears. Endogenous alkaline phosphatase activity from HRP can be attached to other proteins either covalently or non- bone, kidney, liver and some white cells can be inhibited by covalently. Covalent binding of HRP is used in the dextran-pol- the addition of 1 mM levamisole to the substrate solution (10), ymer visualization systems, such as EnVision™ FLEX. although 5 mM has been found to be more effective (11). Intestinal alkaline phosphatases are not adequately inhibited HRP Substrate – DAB by levamisole. 3,3' diaminobenzidine tetrahydrochloride (DAB) is a HRP substrate that produces a brown end product which is highly insolu- In the immunoalkaline phosphatase staining method, the en- ble in alcohol and other organic solvents. Oxidation of DAB also zyme hydrolyzes naphthol phosphate esters (substrate) to causes polymerization, resulting in the ability to react with osmium phenolic compounds and phosphates. The phenols couple to tetroxide, and thus increasing its staining intensity and electron colorless diazonium salts (chromogen) to produce insoluble, density. Of the several metals and methods used to intensify the colored azo dyes. Several different combinations of substrates optical density of polymerized DAB, gold chloride in combination and chromogens have been used successfully. with silver sulide appears to be the most successful (4). DAB has been classiied as a potential carcinogen and therefore should be AP Substrate – Naphthol AS-MX Phosphate handled and disposed of with appropriate care. This can be used in its acid form or as the sodium salt. The chromogens Fast Red TR and Fast Blue BB produce a bright red or HRP Substrate – AEC blue end product, respectively. Both are soluble in alcoholic and 3 amino-9-ethylcarbazole (AEC) is a HRP substrate that forms other organic solvents, so aqueous mounting media must be a rose red end product which is alcohol soluble. Therefore, used. Fast Red TR is preferred when staining cell smears. specimens processed with AEC must not be immersed in alcohol or alcoholic solutions (for example Harris' hematoxylin). In- AP Substrate – New Fuchsin stead, an aqueous counterstain and mounting medium should This substrate also gives a red end product. Unlike Fast Red be used. AEC is unfortunately susceptible to further oxidation TR and Fast Blue BB, the color produced by New Fuchsin is and when exposed to excessive light will fade in intensity. Stor- insoluble in alcohol and other organic solvents, allowing for age in the dark is therefore recommended. the specimens to be dehydrated before coverslipping. The staining intensity obtained by use of New Fuchsin is greater Alkaline Phosphatase (AP) than that obtained with Fast Red TR or Fast Blue BB. Addi- The enzyme calf intestine alkaline phosphatase (AP) has a molec- tional substrates include naphthol AS-BI phosphate, naphthol ular weight of 100 kDa and removes (by hydrolysis) and transfers AS-TR phosphate and 5-bromo-4-chloro-3-indoxyl phosphate phosphate groups from organic esters by breaking the P 0 bond; (BCIP). Other possible chromogens include Fast Red LB, an intermediate enzyme-substrate bond is briely formed. The Fast Garnet GBC, Nitro Blue Tetrazolium (NBT) and iodoni- chief metal activators for AP are Mg++, Mn++ and Ca++. trotertrazolium Violet (INT). 211 Appendix B | Basic Immunochemistry References 1. Boenisch T. Diluent buffer ions and pH: their inluence on the per formance of monoclonal antibodies in immunohistochemistry. Appl Immunohistochem Mol Morph 1999;7:300-6. 2. Boenisch T. Heat-induced antigen retrieval restores electrostatic forces: prolonging the antibody incubation as an alternativeAppl Immunohistochem Mol Morph 2002;10:363-7. 3. Sternberger LA and Sternberger NH. The unlabeled antibody method: comparison of peroxidase-antiperoxidase with avidin-biotin complex by a new method of quantiication. J Histochem Cytochem 1986;34:599 605. 4. Newman GR, Jasani B, Williams ED. The visualization of trace amounts of diaminobenzidine (DAB) polymer by a novel gold-sulphide-silver method. J Microscopy 1983; 132; RP1-2. 5. Clark CA, Downs E, Primus FJ . An unlabeled antibody method using glucose oxidase-antiglucose oxidase complexes. J Histochem Cytochem 1982;30:27-34. 6. Gay et al. J Histochem Cytochem 1984;32:447-51. 7. Bondi A Chieregatti G, Eusebi V, Fulcheri E, Bussolati G. The use of beta-galactosidase as a tracer in immunocytochemistry. Histochemistry 1982;76:153-8. 8. Mason DY, Stein H, Naiem M, Abdulaziz Z: Immunohistological analysis of human lymphoid tissue by double immunoenzymatic labeling. J Cancer Res Clin Oncol 1981;101:13 22. 9. Cordell JL, Falini B, Erber WN, Ghosh AK, Abdulaziz Z, MacDonald S, Pulford KA, Stein H, Mason DY. Immunoenzymatic labeling of monoclonal antibodies using immune complexes of alkaline phosphatase and monoclonal anti-alkaline phosphatase (APAAP complexes). J Histochem Cytochem 1984;32:219 29. 10. Ponder BA and Wilkinson MM. Inhibition of endogenous tissue alkaline phosphatase with the use of alkaline phosphatase conjugates in immunohistochemistry. J Histochem Cytochem 1981;29:981-4. 11. Gown AM In DeLellis RA (ed). Advances in Immunohistochemistry. New York: Raven Press, 1988, pp 31 45. 212 213 Index | Immunohistochemical Staining Methods Index ABC 79, 80, 81, 106, 173 Accessioning 11, 12, 21, 22, 123-125, 130 Adjuvant 135, 174, 202 AEC 69, 84, 86, 187, 188, 211 Affinity 13, 48-50, 53, 56, 57, 79, 84, 85, 106, 107, 175, 182, 202-205 Algorithm Chap 7 and 12, 86, 87, 91 Alkaline phosphatase 13, 84, 86, 105, 153, 172, 173, 182, 186, 188, 189, 191, 210, 211 Animal Research Kit 85, 205 Antibodies adsorption 206 aggregates 176, 209 cocktail 49, 85 dilution 12, 65, 66, 106-109, 162, 175, 181 incubation temp. 104, 181, 183, 209, 210 incubation time Chap 5 polymerization 176, 206, 211 secondary 11-14, 79-87, 106, 107, 164, 176, 181-191 Antigen(s) (see also Epitope(s)) Chap 2, Chap 3, Chap 4, Chap 5, App A and 80-84, 87, 90, 99, 103, 104, 106, 107, 111, 113, 114, 115, 116, 118, 123, 145, 146, 160-167, 171, 174-177, 181-189, 191, 194, 209, 210 Antigen retrieval Chap 1, Chap 2, Chap 3 and 47, 49, 52, 54, 56, 57, 61, 64, 65, 67, 69, 70, 74, 103, 104, 111, 113-116, 118, 123, 174,176, 182, 183, 185, 188, 189, 202, 205, 209 APAAP 13, 205, 211 Artifact(s) 21, 26, 103, 115, 149, 164, 196 ASCO or CAP guidelines 21-23, 32, 61, 99, 171, 193 Avidin-Biotin 79, 80, 81, 84, 106, 173, 174 Avidity 48, 106, 203, 204 Autoluorescence 35, 87, 89, 90 Automation, IHC Chap 9 and 8, 9, 12-14, 18, 28, 33, 34, 36, 41, 42, 47, 49, 57, 58, 61, 64, 71, 91, 95, 98, 100, 127, 177, 178, 181, 183185, 192-196 B5 ixative 176 Background Chap 15 and 9, 26, 36, 63-66, 73, 81, 82, 87, 91, 103, 106, 109, 111, 115, 142, 164, 166 181, 183-91, 195, 196, 209, 210 Bacterial 90, 177, 182, 185, 196 B cell(s) 24, 52, 55, 175, 198 Basal cell(s) 55, 83 Batch staining 114, 116, 117, 123 Biopsy 9, 11-13, 17, 26, 83, 145, 164, 183, 184, 191 Biotin (See also avidin-biotin) 79, 80-82, 84, 85, 106, 153, 163, 164 214 Biotinylated 80, 85, 177, 190 Breast cancer 14, 15, 23, 27, 32, 51, 55, 57, 83, 99, 133-135, 140, 146-148, 153-157 BSA 106, 176, 201, 206 Buffer(s) Chap 3 and 12, 47, 65, 69, 103, 106109, 111, 113, 116, 119, 153, 171, 174, 176, 177, 181-184, 186-189, 195, 196, 204, 209 CAP 18, 22, 23, 28, 61, 99, 123, 130, 162164, 168, 171, 178, 193 Capillary Gap 13, 111, 112 Cell line(s) 14, 17, 22, 27, 29, 53, 54, 90, 137, 146, 148, 163, 167, 168, 202, 203 Cell smear(s) 35, 84, 184, 211 Chromogen(s) 12, 61, 62, 64, 69, 82, 84-87, 91, 105, 109, 153, 155, 164, 171, 172, 181183, 185-187, 191, 196, 210, 211 CISH Chap 13 and 15, 119, 120, 135, 153 Class II/Class III 21, 24, 28, 57, 99, 141 Clone(s) Chap 4 and 12, 16, 28, 33, 34, 65, 72, 75, 108, 161, 165, 166, 175, 192, 200, 202, 203 Companion diagnostics Chap 11 and 9, 13, 14, 51, 91, 157, 167 Conjugation 79-82, 85, 87, 88, 106, 134, 153, 171, 176, 177, 182, 205, 206 Connective tissue 22, 108, 176, 183-185 Connectivity Chap 10 Controls, Reagent Chap 14 and 16, 171, 184-190, 193 Controls, On-slide 28, 163, 165, 166, 169 Controls, Tissue Chap 14, Chap 16 and 12, 27, 28, 47, 49, 52, 54, 57, 58, 64, 65, 70-76, 148, 149 Cores Chap 12 and 26, 34, 50, 163, 165, 166, Counterstain 12, 67, 69, 76, 86, 111, 118, 153, 155, 177, 181-183, 187-189, 211 Cross-links 23, 31, 32, 103, 104, 176, 183-185 Cross-reaction 31, 35, 48, 52-55, 84-86, 138, 175, 183-185, 201-205 Cryostat 184, 191 CSA 79, 81, 82 Cut-off 12, 47, 52, 125, 126, 136-139, 151 Cytology 153 Cytoplasmic 49, 52, 55, 58, 66, 83, 104, 173, 174, 184, 186, 193, 205 Degradation 17, 23, 62, 64, 76, 164, 165 Dextran 11, 13, 14, 80-82, 106, 205, 211 DAB 11, 63, 69, 74, 75, 79, 80, 84-86, 105, 109, 164, 171, 182, 187, 188, 190, 211 Dako Omnis 37, 41, 61, 105, 113, 114, 168 Digital imaging Chap 7 and 14, 69, 119, 169 Diluent 103, 106, 107, 164, 176, 177, 181, 182, 190, 192, 209, 210 Disulide bridges 199, 200 DNA 32, 35, 36 , 39, 83, 88, 90, 99, 119, 135, 145, 152, 157, 201 Double staining 13, 79, 84-86, 173 Drying 12, 16, 21, 26, 47, 109, 115, 123, 177, 185, 193, 196, 210 Dynamic gap 112, 113 EGFR 135, 140-142 ELISA 15, 18, 36, 53, 161, 196 Embedding 11, 12, 17, 21, 26, 76, 80, 90, 105, 111, 123, 124, 126, 129, 130, 149, 177, 182, 185, 193 EnVision™ 67, 68, 75, 81, 85, 86, 192, 211 Enzyme(s) 12-14, 21, 22. 27, 34, 35, 39, 62, 69, 80-82. 84-87, 89, 91, 103-106, 120, 153, 164, 173, 181-183, 191, 210, 211 EQA 18, 61, 70, 104, 168, 169 Epitope(s) 21, 23, 24, 28, 48, 49, 53, 63, 70, 76, 81, 84, 104-107, 169, 171, 175, 182, 194, 200-205 Epitope retrieval (see Antigen retrieval) ER/PR 15, 17, 28, 48, 99, 133, 147, 166 Ethanol 35, 104 Expression Chap 5 and 17, 27, 28, 51, 53, 54, 56, 58, 107, 108, 133, 161, 165, 166 Expiry 109, 161, 182, 194, 195, 205, 206 Fab/F(ab’)2 fragment 82, 85, 176, 177, 186, 199, 200, Fade 151, 154, 211 False, negative or positive 15, 24, 33, 55, 61, 63, 64, 67, 71, 73, 75, 105, 107, 116, 162, 164, 165, 194-196 Fast Red/Blue 84, 86, 188, 189, 211 Fc fragment/receptor 79, 106, 175, 186, 199, 200 FDA 14, 16, 61, 73, 99, 133-136, 138, 141, 142, 156, 157, 167, 171, 192 Fixatives Chap 2 and 12, 16, 18, 50, 61, 64, 76, 104, 161, 174, 176, 177, 182-185, 191, 192 FITC 13, 82, 87-90, 153, 155 FISH Chap 13 and 15, 26, 27, 86, 90, 119, 133-135, 140, 142 FLEX, RTU Chap 5 and 22, 24, 166 Flow cytometry 35, 36, 85-87 Fluorescent dye 79, 85-89 Fluorescent probe Chap 13 and 90 Fluorescyl-tyramide ampliication 82 Formalin ixation 12-14, 16, 17, 22-27, 31, 35, 40, 42, 49-51, 53, 56, 64, 76, 79, 90, 103, 104, 169, 176, 183-185, 202 Frozen sections 15, 26, 27, 31, 35, 36, 80, 84, 87, 90, 95, 96, 98, 145, 172, 173, 175, 177, 182, 184, 192 FFPE tissue 13- 17, 31-36, 42, 53-56, 65, 76, 79, 80, 84, 85, 87, 95, 103, 109, 136, 137, 154, 155, 172, Grossing 11, 12, 21-23, 26, 123-126, 129, 130, 183, 184, 191 Immunohistochemical Staining Methods | Index HER2 Chap 11, Chap 13 and 13-17, 21-23, 25, 27-29, 32, 57, 75, 98, 99, 146, 147, 162, 168 HercepTest™ 14, 27, 28, 57, 99, 133-135, 142, 167, 168 HIER Chap 3 and 12, 13, 17, 47, 49, 56, 57, 103, 104, 174, 183 192, 209 HRP 14, 67, 68, 74, 75, 79, 81, 84-86, 105, 106, 153, 155, 172, 177, 182, 186-188, 190, 210, 211 Hybridoma 13, 48, 202-204 Hydrophobicity 106, 107, 171, 176, 177, 184, 185 iCSA 79, 82 Image analysis Chap 7 and 83, 87, 91, 142, 150 Immune complex 12, 14, 62, 79-81, 106, 173, 175, 176, 200, 204, 205, 211 Immunization 13, 48, 174, 175, 202, 205 Immunoluorescence 26, 35, 86-91, 108 Immunogen (see also antigen) 49, 54, 202 Ischemic time 21, 22, 63, 64, 76, 115, 164, 167, 169 LDT 57 Liquid Overlay 112 LIS 12, 14, 111, 120, 123, 128, 129, 194 LSAB 79-81, 106 Lymphocyte(s) 24, 25, 46, 58, 186, 202 Lymphoid tissue 24, 104, 105, 172, 173, 186, 187, 189 Macrophages 175, 186 Membrane/membranous 22, 49, 52, 58, 104, 168, 175, 178, 186, 193 Microscope/microscopic Chap 7 and 10, 12, 13, 21, 26, 31, 35, 82, 85-87, 90, 91, 109, 112, 119, 123, 153, 154, 156, 160, 167, 169 Monoclonal antibodies App A and 13, 16, 24, 48-50, 54-56, 58, 85, 109, 133, 164, 175-177, 181, 182, 185, 188-190, 209 Morphology 15, 17, 21, 22, 25, 26, 31, 32, 35, 69, 79, 87, 97, 98, 100, 103, 104, 119, 153-155, 194 Mounting 17, 25, 42, 185, 191, 193 Mounting medium 155, 182, 192, 211 Multi-Staining 34, 83-85, 87, 91, 175 Naphthol 211 NBF 22-25, 27, 28 Next generation sequencing 15, 111, 119, 142, 145, 153 New Fuchsin 188, 211 NordiQC 16, 18, 28, 48-50, 55, 56, 61, 69, 70, 73, 104, 165, 168, 178 Nuclear 33, 49, 52, 69, 83, 104, 153, 174, 184, 205 Overexpression 14, 27, 133, 141, 142 PAP 13, 14, 79, 205 Parallel processing 114, 118 PCR 36, 135, 153 Peroxidase, endogenous 49, 57, 58, 105, 153, 155, 171-173, 186, 187, 211 Pepsin 40, 41, 104, 155, 199, 200 Photobleaching 89 pH value 12, 23, 24, 32-34, 36, 40, 41, 47, 64, 65, 69, 74, 80, 103, 104, 106-108, 174, 176, 177, 181, 192, 194, 204, 205, 209 Polyclonal antibodies App B and 13, 16, 48, 49, 84, 164, 174, 176, 177, 182, 185, 188190, 209 Polymer-based detection systems 13, 14, 25, 67, 79, 80, 81, 106, 108, 153, 163, 171, 174176, 193, 205, 206, 211 Post-analytic Chap 7 and 12, 16, 17, 21, 47, 70 Pre-analytic Chap 2 and 9, 12, 16, 18, 47, 48, 53, 54, 61-63, 70, 71, 76, 115, 156, 163, 167, 169, 203 Precipitation 53, 82, 86, 153, 202, 204 Predictive 17, 21, 28, 29, 47, 51, 79, 91, 132, 133, 135, 136, 139, 142, 147, 156, 163 Pressure Cooker/Autoclave Chap 3 and 183 Pre-treatment Chap 3 and 47, 56, 84, 98, 103, 114, 128, 135, 155, 181, 183, 185, 192, 209 Primary antibody Chap 1, Chap 4, Chap 5, Chap 6, App A and 22, 24, 28, 33, 34, 104-107, 153, 161, 162, 164, 172, 176, 178, 181-190, 209, 210 Prognostic 17, 21, 28, 29, 47, 51, 79, 91, 135, 145, 147, 151 Probes 25, 89, 90, 153-155 Proteolytic (enzymatic) digestion Chap 3 and 17, 24, 25, 49, 56, 103, 104, 155, 183-185, 199, 200 Protocol Chap 4, Chap 5, Chap 8 and 8, 9, 12, 15-18, 22, 25-28, 32-34, 36-42, 83, 84, 98, 111, 113-118, 120, 123-125, 128, 136, 137, 139, 150, 155, 162, 164-168, 173, 178, 181185, 187, 189, 190, 193, 194, 209, 210 Quantitative/quantiication 8, 12, 15, 18, 21, 34-36, 47, 91, 98-100, 120, 142, 154, 167, 168 Quality control 13, 107, 146-149, 161, 162, 168, 169, 178 Red 84-86, 105, 188, 189, 211 Retrieval solution Chap 3 and 24, 74, 103, 104, 174, 192, 194 Robustness 47- 49, 54, 56, 62, 71, 73, 76, 137, 138, 153, 155 RNA/mRNA/miRNA 32, 35, 36, 83, 119, 153 RTU Chap 5 and 22, 24, 28, 33, 39, 49, 85, 96, 113, 116, 166, 178, 182, 206, 209 Saturation 63, 64, 67, 76 Semi-quantitative 14, 15, 17, 27, 61, 69, 79, 99, 100 Sensitivity 13, 14, 17, 47-49, 51-54, 56, 57, 73, 75, 79-82, 91, 107, 115, 117, 137-140, 146, 162, 163, 177, 181, 182, 184, 190 Serum/Serum proteins 13, 85, 106, 107, 164, 176, 183-186, 188, 189, 199-202, 204 Signal-to-noise 48, 53, 62-64, 73-75, 82, 209 Speciicity 12, 13, 17, 47-49, 51-55, 57, 73, 80, 81, 84, 90, 103, 105, 106, 108, 115, 117, 135, 138-140, 146, 161-163, 166, 167, 175, 182, 200, 201, 204, 205 Stability 14, 87, 106, 107, 138, 165, 168, 176, 193, 199, 200, 205, 209 Standardization Chap 1 and 8, 9, 21, 32, 34, 36, 39, 48, 49, 51, 54, 57, 58, 61, 69, 70, 76, 98, 111, 113, 148-150, 182, 183 Staining intensity Chap 4, Chap 5 and 15-18, 22, 33, 35, 87, 89, 90, 97, 99, 100, 104, 106, 108, 165-168, 182, 194, 195, 204, 209-211 TAT Chap 10 and 25, 69, 83, 116-119, 125, 127 T cells 53, 55, 56, 73, 75, 175 Telepathology 95, 96, 98 Test battery 34, 36, 42, 50 Theranostics (see companion diagnostics) Thickness, section 12, 47, 64, 69, 96, 191 Tissue blocks 12, 31, 35, 100, 145, 146, 163, 183 TMA Chap 12 and 27, 28, 34, 50, 51, 53, 70, 165, 166 Tissue processing 11, 12, 24, 25, 53, 124, 126, 130, 161, 174 Total test 13, 14, 16, 18, 48, 169, 191 Trastuzumab 14, 133-135, 140-142, 156, 157 Troubleshooting Chap 16 and 22, 178 UK NEQAS 16, 18, 70, 104, 169, 178 Virtual microscopy (see digital imaging) Chap 7 Validation 15, 16, 25, 27, 28, 29, 34, 41, 48, 51-54, 57, 61, 70, 71, 73, 76, 99, 107, 116, 136-140, 145, 150, 161-163, 169, 192 Water bath Chap 3 and 109, 130, 177, 185, 191 Veriication 22, 29, 49, 61, 73, 76, 127, 161, 162 Wash buffers 37, 41, 42, 108, 109, 119, 153, 171, 177, 182, 184, 186, 188, 189, 195, 196 Wavelength 87-90 Wax/Dewax 21, 25, 26, 182-184 Western blotting 53, 161 Worklow Chap 10 and 64, 66, 68-70, 100, 114, 116-120, 210 215 Australia +61 3 9357 0892 Canada +1 905 335 3256 France +33 1 64 53 61 44 Japan +81 3 5802 7211 Poland +48 58 661 1879 United Kingdom +44 (0)1 353 66 99 11 Austria +43 1 408 43 34 0 China +86 21 3612 7091 Germany +49 40 69 69 470 Korea +82 2 402 6775 Spain +34 93 499 05 06 United States of America +1 805 566 6655 www.dako.com Belgium +32 (0) 16 38 72 20 Denmark +45 44 85 97 56 Ireland +353 1 479 0568 The Netherlands +31 20 42 11 100 Sweden +46 8 556 20 600 Represented in more than 100 countries Brazil +55 11 50708300 Finland +358 9 348 73 950 Italy +39 02 58 078 1 Norway +47 23 14 05 40 Switzerland +41 41 760 11 66 08002 07NOV13 Corporate Headquarters Denmark +45 44 85 95 00