Review The road map to oral bioavailability: an industrial perspective 2. Solubilit y V Hayden T homas, Shobha Bhattachar, Linda Hitchingham, Philip Zocharski, Maryanne Naath, Narayanan Surendran, Chad L Stoner & Ayman El-Kattan† 3. Permeabilit y 2800 Plymouth Road, Ann Arbor, MI 48105, USA 1. Int roduct ion Pfizer Global Research and Development, Department of Pharmacokinetics, Dynamics and Metabolism, 4. M et abolism 5. Conclusion 6. Expert opinion Opt imisat ion of oral bioavailabilit y is a cont inuing challenge f or t he pharmaceut ical and biot echnology indust ries. The number of pot ent ial drug candidat es requiring in vivo evaluat ion has signif icant ly increased w it h t he advent of combinat orial chemist ry. In addit ion, drug discovery programmes are increasingly f orced int o more lipophilic and low er solubilit y chemical space. To aid in t he use of in vit ro and in silico t ools as w ell as reduce t he number of in vivo st udies required, a t eam-based discussion t ool is proposed t hat provides a ‘road map’ t o guide t he select ion of prof iling assays t hat should be considered w hen opt imising oral bioavailabilit y. This road map divides t he f act ors t hat cont ribut e t o poor oral bioavailabilit y int o t w o int errelat ed cat egories: absorpt ion and met abolism. This road map provides an int erf ace f or cross discipline discussions and a syst emat ic approach t o t he experiment at ion t hat drives t he drug discovery process t ow ards a common goal – accept able oral bioavailabilit y using minimal resources in an accept able t ime f rame. Keywords: absorption, ADME optimisation, bioavailability, distribution, drug discovery, elimination, metabolism, pharmacokinetics, solubility Expert Opin. Drug Metab. Toxicol. (2006) 2(4):591-608 1. For reprint orders, please contact: ben.fisher@informa.com Introduction Discovering novel therapeutic agents is an increasingly time-consuming and costly process. Most estimates indicate that it takes ∼ 10 – 15 years and > $800 million to discover and develop a successful drug product [1]. During the past decade, the pharmaceutical industry used parallel medicinal chemistry and high-throughput screening (HT S) approaches in drug discovery programmes [2]. T his approach enabled the creation of millions of compounds that were screened against hundreds of potential targets with a common goal of discovering chemical matter with high affinity for the intended pharmacological target. One of the major limitations to this approach is that the newly discovered leads tend to have high molecular weight and lipophilicity with low aqueous solubility, resulting in new chemical entities (NCEs) that are usually associated with poor oral bioavailability [3]. It is well established in the literature that poor oral bioavailability is one of the leading causes of compound failure in preclinical and clinical development [4]. Compounds with poor oral bioavailability tend to have low plasma exposure and high interindividual variability, which would limit their therapeutic usefulness. To improve the oral plasma exposure of a leading candidate and increase the probability of a successful development, its physicochemical properties and pharmacological activity should be optimised in parallel. However, the cost and time required to screen all newly synthesised compounds in relevant assays are prohibitive. T herefore, the current approach is to perform selective biopharmaceutical and pharmacokinetic ‘spot checks’ on these leading compounds to funnel information 10.1517/17425255.2.4.591 © 2006 Informa UK Ltd ISSN 1742-5255 591 The road map to oral bioavailability: an industrial perspective Rat F% ≥ 30% Manageable risk < 30% Lipinski RO5 compliant PSA ≤ 140 Å2 GI chemical stability No Physicochemical SAR Yes Absorption or metabolism Absorption Metabolism Permeability Solubility Figure 1. Entry into the road map w ith the initial evaluation of a compound series. Compounds w ith poor oral bioavailability in rat are evaluated against a subset of physicochemical properties prior to entering the main body of the road map. GI: Gastrointestinal; PSA: Polar surface area; RO5: Rule of 5; SAR: Structure–activity relationship. into rapid optimisation loops based on structure–activity relationship (SAR) screens [5]. Assays developed to fit this approach have advanced considerably over the past few years in the areas of kinetic solubility, metabolism screens and plate-based permeability assays [6,7]. T hese assays, in conjunction with improvements in predictive software for pKa, logP and polar surface area (PSA), and other relevant physicochemical parameters, provide a useful semiquantitative guide to discovery scientists during lead optimisation [3,4,8]. T his multiparameter optimisation generates a great deal of data requiring clear interpretation and a firm understanding of the various pharmacokinetic issues involved. T his is often a multifaceted challenge that tends to slow down drug discovery and development timelines due to misdirected attempts at solving incorrectly identified problems. Hence, it is pivotal that the discovery team visualises and interprets the data in a manner that will enable the next round of compound design and synthesis [9]. T he task of resolving challenging pharmacokinetic behaviours with respect to oral bioavailability becomes less troublesome if broken down into manageable modules of experimentation. T his article provides discovery teams with a ‘road map’ of experimentation that determines the most probable causes of poor oral bioavailability, when optimising a chemical series and ultimately selecting lead compounds for preclinical development. T his systematic approach divides the potential factors leading to the compound’s (template’s) low oral bioavailability into two interrelated categories: absorption (product of solubility and permeability) and metabolism. In each category, moderate- to high-throughput assays are used to rapidly evaluate various attributes of a chemical series with 592 two possible outcomes to each assay: a manageable risk where a discovery team proceeds to the next assay, or high risk where SAR and troubleshooting is warranted to overcome the liablity that is contributing to the limited oral bioavailability of the chemical series. Using this approach to identify key properties that contribute to the poor oral bioavailability of a chemical series will decrease discovery cycle time and cost required in the optimisation of a chemical series. In addition, it will play a key role in the selection of drugable candidates for preclinical development. T his road map was developed as a result of an extensive literature review. It should be noted that corporate business practices and technological resources are anticipated to vary depending on project focus and individual business models. As the road map described in this paper is based on fundamental aspects of physicochemical and pharmacokinetic concepts, similar systematic approaches may be developed based on the proposed model to assist discovery teams. Initial evaluation T here are many parallel activities taking place in the mid to late phases of the drug discovery cycle to bring an NCE into full preclinical development. T hese include chemical scalability, biopharmaceutical characterisation and pharmacokinetic/pharmacodynamic profiling [10]. T hese studies are focused on entering preclinical development in the shortest time frame possible, giving little consideration to potential development hurdles. Given the high cost and lengthy time frame for drug development, it is pivotal to achieve a minimally acceptable oral bioavailability standard for selecting compounds prior to moving forward. In discovery projects, 1.1 Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan 10,000 Log solubility (µg/ml) 2100 520 1000 207 52 100 100 21 10 10 5 1 1 0.1 0.1 high 0.1 avg 0.1 low 1.0 high 1.0 avg 1.0 low 10 high 10 avg Ka Ka Ka Ka Ka Ka Ka Ka 10 low Ka Projected dose in mg/kg Figure 2. The predicted solubility required for a compound series to achieve a minimally acceptable oral absorption. Bars show the minimum solubility in µg/ml (absolute values are denoted above the bar) for low, medium and high permeability (Ka – apparent absorption rate) at three projected clinical doses (0.1, 1 and 10 mg/kg, denoted by bar colour). M odified w ith permission from Lipinsky CA. In Pharmaceutical Profiling in Drug Discovery for Lead Selection . Borchardt RT, Kerns EH, Lipinski CA, Thakker DR, Wang B, (Eds), Arlington, VA: American Association of Pharmaceutical Scientists (2004):95, Copyright 2004, American Association of Pharmaceutical Scientists [13,20]. oral bioavailability is usually first assessed in rats as they are readily available. Moreover, pharmacokinetic parameters obtained in rat can be commonly scaled to human, providing a basis for rational compound selection [11]. A minimal acceptable oral bioavailability of 30% has been established for a typical oral programme [12-14]. T his minimum threshold will help the team to achieve an early proof of concept. If this criterion is not achieved during preliminary in vivo rat testing, the impact on drugability must be determined by conducting key studies on compounds in the selected chemical series. Once representative compounds have been evaluated and a chemical series has been shown to exhibit low oral bioavailability, the project team may use this road map to determine the physicochemical and/or pharmacokinetic parameters that contribute to the poor oral bioavailability and optimise them accordingly (Figure 1 ). For candidates with < 30% oral bioavailability in rat, the first assessment is to ensure that it meets desirable physicochemical properties, such as the Lipinski’s rule of 5 (RO5), PSA < 140 Å2 , and acceptable chemical stability in the dosing vehicle and various simulated gastrointestinal conditions (Figure 1 ). Failure of the candidate to meet one of these simple descriptors would trigger the discovery team to execute specific studies to determine their potential impact on the candidate oral absorption. For example, PSA is a parameter, which is defined as the area occupied by nitrogen and oxygen atoms and the hydrogen atoms attached to these heteroatoms [15]. T his descriptor is widely used to assess compound absorption potential. Veber et al. [14] have shown that compounds with a PSA ≤ 140 Å2 will have a high probability of good oral absorption in the rat. T herefore, a compound exhibiting a PSA value > 140 Å2 would require an assessment for its permeability potential, possibly using the Caco-2 cell line to determine if its intestinal permeability is limited by its PSA. Furthermore, if the candidate meets RO5 and standard physicochemical requirements, it would undergo a parallel evaluation in each of the discussion tool’s two interrelated categories: absorption and metabolism. Absorption Absorption is a dynamic process of drug transfer from the site of administration, the gastrointestinal lumen, across the intestinal epithelium and into portal blood. Taking the simplest approach, oral drug absorption can be expressed using Fick’s First Law applied to membranes [16]: 1.2 (1) J wall = P wall ⋅ C int where J wall is the drug flux across a homogeneous intestinal membrane, Pwall is the effective permeability, which is the rate that dissolved drug will cross the intestinal wall to reach the portal blood circulation, and C int is the drug concentration in the luminal fluid. T herefore, drug flux is a product of drug permeability and solubility [17]. A qualitative understanding of permeability, solubility and their impact on drug absorption is essential in achieving acceptable oral bioavailability. A proper balance must be established between these two parameters during early SAR optimisation, while still considering other factors such as dose and drug stability in the intestinal medium, which can significantly impact oral absorption. Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 593 The road map to oral bioavailability: an industrial perspective HTS solubility (kinetic) ≥ 60 µg/ml high solubility at pH 6.5 < 60 µg/ml low solubility at pH 6.5 High order ≥ 250˚C Consider permeability and metabolism tree Solubiity (thermodynamic) ≥ 60 µg/ml Assess crystal lattice energy (melting point) Low to mild order < 250˚C Yes Yes LogP ≤ 3.5 No No Consider crystal packing/H-bonding (computations) LogP SAR optimisation to improve solubility Team discussion – consider other factors affect solubility including dissolution, media, kinetics, pKa, dose, etc. Figure 3. Solubility section of the road map. This section of the road map should be used to as a guide to determine w hich of the main solubility related factors are resulting in low aqueous solubility and thus poor oral bioavailability. HTS: High-throughput screening; SAR: Structure–activity relationship. 2. Solubility As shown with Equation 1, the prerequisite for good oral drug absorption is ensuring a sufficient compound amount is in solution at the primary site of absorption. Various approaches are available to measure compound aqueous solubility. In general, the different types of solubility measurements may be classified into thermodynamic and kinetic solubility methods [3]. T hermodynamic solubility is defined as the equilibrium concentration of a compound that is saturated in a given solvent. T hermodynamic solubility is dependent on many variables such as compound crystal lattice, temperature and pressure [18]. On the other hand, kinetic solubility is a measurement of compound solubility where it is predissolved in an organic solvent (typically dimethyl sulfoxide) then titrated with aqueous medium over a period of time until the compound precipitates [19]. Kinetic solubility determination is the preferred method at the discovery stage as it is more amenable to HTS, requiring virtually no equilibration time. However, the main 594 disadvantage of this approach is the lack of consideration for any crystal lattice effects as shown with thermodynamic solubility measurements [20]. It should be stressed that both thermodynamic and kinetic solubility are generally measured in a phosphate-based buffer (pH range: 6.5 – 7) as this medium is considered to be a surrogate for intestinal pH. Determining the minimum solubility necessary to achieve adequate oral absorption is dependent on the compound projected human dose and permeability [13,20]. Lipinski depicted this relationship graphically, as shown in Figure 2 , which serves as a quick reference guide to determine the minimum level of solubility that is needed by a compound to have reasonable absorption in human. For example, given a compound series with a projected human dose of ∼ 1 mg/kg and average permeability, assessed using the apparent absorption rate (Ka), an aqueous solubility of ≥ 52 µg/ml would be needed to ensure that oral bioavailability is not solubility limited. In actual practice, discovery teams routinely target a kinetic solubility > 60 µg/ml (pH 6.5) when developing a Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan O Kow is the octanol–water partition coefficient of the compound. T his equation clearly demonstrates that decreasing lipophilicity by one logP unit or decreasing melting point by 100°C will increase compound aqueous solubility by ∼ 10-fold. Even though this equation cannot be used to estimate the solubility for all compounds, specifically ionisable molecules, it demonstrates the importance of considering these pivotal factors when optimising compound aqueous solubility. O O O N N HN N O O O O Thalidomide N-methyl-thalidomide Crystal lattice energy and solubility A compound that can assemble into a highly ordered crystal tends to have a higher melting point and lower aqueous solubility [22]. T his highly ordered crystal lattice primarily originates from two types of strong cohesive interactions: i) intermolecular hydrogen bonding and/or ii) short-ranged dispersion forces. T hese cohesive interactions are most noted in compounds that contain one or more of the following properties: high number of hydrogen-bond acceptor (HBA)/donor (HBD) groups, planar or inflexible conformation or a high degree of symmetry [23]. All of these factors promote tight packing within the crystal lattice that is not readily solubilised by aqueous media. In general, a compound with a high melting point (> 250°C) will have poor thermodynamic solubility, requiring a discovery team to develop an SAR to reduce its crystal lattice energy (Figure 3 ). T his approach was taken by Goosen et al. when evaluating the effect of structural modification on the physicochemical properties of thalidomide, which has an aqueous solubility of 52 µg/ml and a melting point of 275°C [24]. T halidomide’s poor aqueous solubility is attributed to an acidic imido hydrogen on the glutarimide ring, which is responsible for the strong hydrogen and dipolar bonding in the crystalline state. When this group was replaced with an alkyl group, the highly ordered crystal lattice was disrupted (Figure 4 ). As a result, N -methyl thalidomide has an aqueous solubility of 276 µg/ml and 159°C melting point [24]. PNQX had desirable in vitro and in vivo activity; however, it suffered from low solubility (Figure 5 ). In an effort to improve its physicochemical properties and potency, Nikam et al. synthesised a novel ring-opened analogue series to reduce packing efficiency in the crystal lattice, which significantly increased aqueous solubility [25]. 2.1 Figure 4. N-alkylation of thalidomide disrupts the highly ordered crystal lattice, improving solubility by more than fourfold. O OH N N O2N H N O N H O PNQX (pH 7.4, solubility: 8.6 µg/ml) O2N H N O N H O Nonplanar analogue PNQX (pH 7.4, solubility: 420 µg/ml) Figure 5. PNQX ring opening and out-of-plane substitution results in a reduced crystal lattice energy and hence higher aqueous solubility. compound series, as this solubility value is expected to lead to reasonable absorption provided that the compound has average permeability and dose (Figure 3 ). In the case where a compound series has a kinetic solubility < 60 µg/ml (pH 6.5) and low oral bioavailability, aqueous solubility could be the rate-limiting factor in achieving acceptable compound oral plasma exposure. T herefore, the discovery team should determine the main factors that are contributing to the compound low aqueous solubility and thus poor oral bioavailability. In general, these factors are demonstrated in the general solubility equation developed by Jain et al. (Equation 2) [21]: (2) solid log S W = 0.5 – 0.01 ( MP – 25 ) – log K ow where SW solid is the molar aqueous solubility, MP is the melting point (widely used to assess crystal lattice energy), and M odulating lipophilicity for improved compound solubility A more commonly encountered reason for poor aqueous solubility than that of a highly ordered crystal lattice is lipophilicity. LogP is a widely accepted measure of lipophilicity, and compounds demonstrating a clogP > 3.5 generally have poor aqueous solubility [19], and should warrant a logP SAR optimisation to improve oral absorption (Figure 3 ). In general, decreasing lipophilicity will improve solvation potential by increasing solvent–solute interactions in aqueous media. A common approach to the reduction of lipophilicity is through the introduction of ionisable or polar groups. T his approach was considered in the development of indinavir (Figure 6 ), 2.2 Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 595 The road map to oral bioavailability: an industrial perspective Ionisable centre N O OH O OH OH H N HN OH N H N N O N H O O L-685,434 Indinavir Figure 6. Solubilisation SAR w here an ionisable basic amine group w as substituted into the backbone of L-685,434 to increase aqueous solubility. SAR: Structure–activity relationship. Table 1. Example of functional group addition. Dimethylaminoethyl group added to 6-amino-seco-cyclopropylindole compound to improve solubility 5 6 Cl R 7 N H N O NH2 R Solubility (µg/ ml) 5,6,7-triOCH3 13 5-OCH3 8 5-O(CH2)2N(CH3)2 290 5-OM e, 6-O(CH2)2N(CH3)2 > 500 where an ionisable basic amine (and a pyridine) group were incorporated into the hydroxyl ethylene backbone of L-685,434 (logP 5.36), reducing the logP by 2.5 units and notably increasing aqueous solubility [26]. In another example, Milbank et al. decreased lipophilicity in a series of 6-amino-seco-cyclopropylindole compounds, by incorporating a dimethylaminoethyl group at the 5- or 6-position on the ring, which significantly increased its aqueous solubility (Table 1 ) [27]. When oral absorption is solubility limited in a series that has a clogP < 3.5 and a low to mid order crystal lattice, little can be gained by reducing lipophilicity without adversely affecting intestinal permeability. As a general rule of thumb, a logP value of 2 – 3 provides a good balance with respect to solubility and permeability [5]. In this case, the discovery team must carefully consider the contributions of other solubility 596 associated factors, including compound polarity, ionisation, and so on, to improve solubility (Figure 3 ). Solubility associated factors limiting oral absorption A compound series can have sufficient solubility (thermodynamic, ≥ 60 µg/ml) and still have limited oral absorption due to solubility related physicochemical factors, such as dissolution kinetics in biorelevant media, pKa, and so on [28-31]. Therefore, the discovery team should consider these factors when optimising a chemical series (Figure 3 ). For example, a slight shift in pKa, decreasing acidic and increasing basic, can improve a compound series’ solubility in physiologically relevant conditions. As seen with Jiang et al. in optimising a series of phosphodiesterase type 5 inhibitors, they demonstrated the effect of shifting a basic pKa from 4.4 to 5, and thereby increasing rat oral bioavailability from 12 to 38% [32]. Other factors such as dissolution rate can limit oral absorption of a compound series, with moderate solubility. The most common approach to confirm dissolution-limited oral absorption is through particle size reduction [33]. Discovery teams can use a mathematical model, microscopic mass balance approach, developed by Oh et al. to determine what particle size would be necessary to achieve an acceptable oral absorption [34]. Particle size reduction was shown to significantly increase the dissolution rate of phenytoin and improve its oral absorption in humans [35]. Another example of particle size reduction to increase dissolution rate was in the dosing of a danazol nanoparticulate dispersion, which improved the oral bioavailability in dogs [36]. 2.3 3. Permeability Various routes exist by which administered drugs can cross the intestinal membrane. In general, the main routes are by passive diffusion or active transport (Figure 7 ). Passive diffusion is the most common mechanism of absorption across the intestinal membrane and is divided into two pathways: the paracellular Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan Passive diffusion Apical (gut) Transcellular Paracellular Active transport Influx Efflux Enterocytes Basolateral (blood) Figure 7. Primary mechanisms of intestine permeability. Passive diffusion can be paracellular or transcellular and active transport can be influx or efflux. pathway, in which drug diffuses through the aqueous pores at the tight junctions between the intestinal enterocytes; and the transcellular (lipophilic) pathway, which requires drug diffusion across the lipid cell membrane of the intestinal enterocyte. T he active transport pathway is mediated by transporters and is divided into active drug influx and efflux. T he relevance of each route is determined by the compound’s physicochemical properties and its potential affinity for transport proteins, as will be discussed in the following sections. T herefore, it is pivotal to investigate and determine the major absorption pathways when optimising oral drug absorption. In paracellular diffusion, drug molecules can cross the intestinal enterocytes through the water-filled pores between these cells [37]. In general, drugs that are absorbed through this pathway are quite small molecules (e.g., molecular weight [MW] < 250 Da) and hydrophilic in nature (logP < 0). Because the junctional complex has a net negative charge, positively charged molecules pass more readily, whereas negatively charged molecules are repelled [38]. It is interesting to note that the paracellular pathway offers a limited window for absorption as the tight junctions between cells become tighter travelling from the jejunum towards the colon [39]. T he paracellular pathway of absorption is a minor pathway due to the tight junctions and small surface area, which accounts for ∼ 0.01% of the total surface area of intestinal membrane [40]. T herefore, compounds with paracellular absorption have dose- and regional-dependent absorption, as seen with cimetidine [41]. It is, therefore, preferred to design compounds without a significant paracellular component. On the other hand, the transcellular pathway is the major route of absorption of compounds absorbed. The passive transcellular transport starts with the penetration of apical membrane, followed by diffusion through the cytoplasm. Finally, the drug molecule exits through the basolateral membrane into the portal blood [42]. In general, the rate of passive transcellular permeability is mainly determined by the rate of transport across the apical cell membrane, which is controlled by the physicochemical properties of the absorbed compound. According to solubility diffusion model, pH partition theory and RO5, compounds that are absorbed through the transcellular pathway are unionised, with reasonable lipophilicity and molecular weight (logP > 0 and MW > 300 Da, respectively). In addition, the hydrogen-bonding capacity determined by the number of HBAs and HBDs is < 10 and 5, respectively [3,8]. A large number of influx transporters are expressed by the small intestinal mucosa and play a major role in the absorption of nutrients and vitamins. In addition, these influx transporters mediate the absorption of some drugs and xenobiotics. T here are several reviews that discussed the involvement of these transporters in intestinal drug absorption [42,43]. Examples of these transporters are di-/tripeptides (PEPT 1), large neutral amino acids (system L), bile acids, nucleosides and monocarboxylic acid transporters. In general, compounds that are substrates for these influx transporters exhibit intestinal absorption higher than expected from their diffusion across intestinal cell membranes. For example, PEPT 1, which is expressed predominantly, but not exclusively, in the small intestine, is a well-characterised influx transporter with many substrates, such as angiotensin-converting enzyme inhibitors, β-lactam antibiotics (both cephalosporins and penicillins), PD-158473 and renin inhibitors [43,44]. Discovery teams commonly use PEPT 1 as a promising strategy for oral drug delivery due to its broad substrate specificity. Unlike absorption influx transporters, efflux transporters function as an absorptive barrier that limits oral bioavailability of many drugs and xenobiotics. T hese transporters belong to the AT P-binding cassette superfamily of transporters and are expressed at the apical surface of the small intestine enterocytes. T hey include P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), multi-drug resistance-associated protein (MRP) and organic ion transporters [42,45,46]. P-gp, the best-characterised member of the apical efflux transporters, is a product of the multi-drug resistance (MDR1) gene. Different research groups have shown P-gp to limit the intestinal absorption of a large number of drugs such as digoxin, talinolol, UK-343,664 and ciclosporin, to name a few [47-49]. Several research groups are beginning to establish a better understanding in P-gp substrate recognition [50-54]. In an interesting study, Didziapetris et al. established a rule of four to predict P-gp substrate interaction. T his rule is roughly determined by the following factors: i) compound’s size expressed through molecular weight; ii) number of HBA; and iii) extent of ionisation determined by the acid/base pKa values. In general, compounds with HBA > 8, MW > 400 Da, and acid pKa > 4 are likely to be P-gp substrates. On the other hand, compounds with HBA < 4, MW < 400 Da and base pKa < 8 are not likely to be P-gp substrates [51]. 3.1 Using Caco-2 cell-based assay to classify compound permeability A variety of in silico, in situ, in vivo and in vitro models are available for the assessment of intestinal drug permeability [55]. However, in vitro models using Caco-2 cells are the most commonly used techniques, as it is amenable to HTS [56]. For passively absorbed compounds, several reports demonstrated a Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 597 The road map to oral bioavailability: an industrial perspective Permeability Yes Caco-2 ≥ 1 x 10-6 cm/s Acceptable permeability and no evidence of efflux or paracellular pathway Consider solubility and metabolism tree No No Yes Evidence of high efflux or paracellular pathway Have discussion with team regarding the impact of low Caco-2 results. Consider Caco-2 SAR No Spot check MDCK/MDR1 BA/AB results ≥ 2.5 Not a P-gp substrate Have discussion with team to consider other efflux transporters Yes P-gp substrate P-gp SAR optimisation Consider solubility and metabolism tree Figure 8. Permeability section of the road map. Once a compound series is determined to have Caco-2 permeability < 1 x 10 -6 cm/s, steps should be taken to determine if paracellular or active efflux is playing a significant role and seek to eliminate this potential liability through structural modifications. AB: Apical to basal; BA: Basal to apical; M DCK: M adin–Darby Canine Kidney; M DR: M ulti-drug resistance; P-gp: P-glycoprotein; SAR: Structure–activity relationship. good correlation between Caco-2 permeability and the oral fraction absorbed in humans [57-60]. Caco-2 cells also have active transport mechanisms similar to those present in the human intestine [61]. T here are a few factors that limit the use of the Caco-2 cell line, such as a long culture time (21 days) to reach confluence and the reported underexpression of metabolic enzymes as compared with levels observed in human small intestine [56,62]. Caco-2 is also not a good model for the assessment of paracellular absorption due to the tight junctions between the cells as shown by the high transepithelial resistance (∼ 400 Ω cm2) compared with human intestine (∼ 60 – 120 Ω cm2) [63]. Even considering these relevant points, Caco-2 has been proven valuable as a model to project human absorption and to answer mechanistic questions regarding intestinal permeability. It should be emphasised that other institutions may use other tools to assess passive permeability such as Parallel Artificial Membrane Permeability Assay or Mardin–Darby Canine Kidney cell line [52]. 598 As shown in Figure 8 , if a chemical series has a Caco-2 permeability rate > 1 x 10 -6 cm/s, then a moderate to high absorption (Fa = 30 – 100%) in human is predicted [60]. In this case, other factors in the road map, solubility and/or metabolism, should be evaluated for their role in limiting oral bioavailability of the series. If a compound has a Caco-2 permeability < 1 x 10 -6 cm/s, human absorption is predicted to be permeability limited (Fa = 0 – 30%). Factors affecting intestinal permeability should be considered, including paracellular absorption and significant efflux. To increase passive transcellular permeability and reduce paracellular permeability contribution to the total absorption of a compound, three common approaches are taken: i) increase compound molecular weight and lipophilicity, if below the optimum range (MW ∼ 300 – 400 Da and logP ∼ 2 – 3); ii) reduce hydrogen-bond functionality by decreasing the number of HBDs or HBAs; and/or iii) reduce net ionisation in the physiological pH range 5.5 – 7.0. Examples Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan highlighting these approaches are common in the literature and here only a few are noted. A popular way to effectively increase lipophilicity, while maintaining target potency, is through bioisosteric replacement of hydrophilic core groups. Lin et al. demonstrated this approach in optimising a series of very late antigen-4 antagonists, by replacing the anilide core with benzoxazole. T his resulted in an ∼ 30-fold net increase in rat oral bioavailability without compromising target potency [64]. In another interesting study, Walker et al. demonstrated the importance of hydrogen-bond functionality with a series of endothelin antagonists. Earlier compounds in this series had poor intestinal absorption due in part to a high number of HBDs and HBAs, and by effectively reducing the number of these groups permeability significantly improved, which was seen in Caco-2 cells [65]. Another approach to reducing hydrogen-bond functionality is through the promotion of intramolecular hydrogen bonding. T his technique reduces a compound’s polar group interaction with water, as hydrogen bonding is internally satisfied. For example, sildenafil, which does not meet RO5 criteria, still has a good oral absorption profile, largely due to significant intramolecular hydrogen bonding [3]. In addition, altering or reducing a compound net charge in the pH range 5.5 – 7.0 is usually associated with significant improvements in intestinal permeability. Depending on the series, this can be done by increasing pKa of acids, decreasing pKa of bases or avoiding zwitterions [66]. Marsilje et al. used this approach to optimise the oral exposure of a set of melanocortin 4 receptor antagonists. T he group demonstrated that reducing the compound basicity by substituting an amidine with an imidazole significantly improved its oral exposure [67]. In the above examples, physicochemical changes mainly influenced compound intestinal permeability. However, discovery teams should also consider the impact of these changes on other relevant parameters such as solubility, metabolism and pharmacological potency. T herefore, discovery teams are advised to establish a balance between these interrelated parameters. Efflux transporters and permeability Efflux transporters play a major role in limiting the intestinal permeability of various xenobiotics. T his is usually demonstrated by using Caco-2 cell system by comparing permeability in two directions, apical to basal (A→B) and basal to apical (B →A). When a compound series has a BA/AB permeability ratio > 2.5, with low passive permeability, efflux transporters are playing a significant role in reducing the effective intestinal permeability [68]. It is interesting to observe that Caco-2 cell line expresses not only P-gp transporter, but also other relevant efflux transporters such as BCRP, MRP and organic ion transporters [69]. In order to determine if efflux is primarily due to P-gp, BA/AB permeability ratio is measured in the Madin–Darby Canine Kidney/MDR1 cell line (Figure 8 ). T his cell line has the advantage of short culture times and a capacity to stably overexpress specific transporters, in this case MDR1 (P-gp) [46]. If the ratio in these cells is 3.2 similar or greater than that obtained in Caco-2 cells, efflux is likely due to P-gp. If lower, other transporters present in Caco-2 cells may be responsible for the high efflux. As a result of efflux involvement, it is quite likely that the effective permeability of the compound series will be concentration dependent and the rate of permeation should be evaluated at concentrations relevant to the projected dose. Walker et al. provides a detailed study in defining the role of efflux transporters and the superproportional dose–exposure relationship seen with UK-427,857 [70]. Many successful compounds are substrates of P-gp, such as, digoxin [71], erythromycin [72] and atorvastatin [73], to name a few. P-gp substrates can also cause competitive inhibition of coadministered medications, resulting in altered absorption profiles, as is the case when atorvastatin is administered with digoxin [74,75]. It is key to understand the implications of having an efflux transporter on the oral bioavailability of compounds evaluated and if possible eliminate this potential liability through structural modifications via a P-gp SAR optimisation (Figure 8 ). It should be emphasised that excipients can inhibit P-gp activity and/or CYP-mediated gut metabolism, resulting in a significant increase in the plasma exposure of the administered drug [76]. As a general approach, reduction in compound lipophilicity and HBA groups is a good starting point [77]. 4. M etabolism In 2002, hepatic metabolism was the major route of elimination for around three-quarters of the top 200 prescribed drugs in the US [201]. Hepatic metabolism is an enzymatic process in which a drug is chemically modified into a polar metabolite that can be more easily excreted through urine or bile [78]. In general, discovery teams strive to identify molecules that have adequate hepatic stability to allow for once-a-day dosing. To that end, selected chemical series are inherently metabolically stable and/or can be modified to reduce metabolic liability. Metabolism is divided into phase I and II processes [79]. In phase I, the drug undergoes oxidative attack that introduces or exposes a polar functional group on the drug molecule. For example, a hydroxyl group is introduced onto the phenyl group on mephenytoin by aromatic hydroxylation to form hydroxy(S)-mephenytoin [80]. Dextromethorphan is demethylated to form dextrophan [81]. A list of common reactions and enzymes involved in these processes is shown in Tables 2 and 3 , respectively [82,83]. From this list, the CYP superfamily is the major family of enzymes that is responsible for the metabolism of most marketed drugs and xenobiotics [84-87]. In phase II, a polar moiety is usually added into either the parent molecule or its phase I metabolites [88]. T he resulting polar metabolites are then excreted from the body through urine or bile, or in some cases sweat or exhalation. Phase II processes involve conjugation reagents that are normally derived from biochemical compounds involved in carbohydrate, fat and Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 599 The road map to oral bioavailability: an industrial perspective Table 2. Phase I reactions. Phase I reactions F = ( 1 – E h ) ⋅ Fa ⋅ Fg Examples Oxidation Hydroxylation Phenytoin, acetaminophen Dealkylation Diazepam, phenacetin Deamination Amfetamine Sulfoxidation Chlorpromazine Reduction Sulfasalazine, chloramphenicol Hydrolysis Aspirin, phenacetin (3) where F is the oral drug bioavailability, Fa is the fraction of the dose that is absorbed after oral administration and Fg is the fraction of the dose that escapes intestinal metabolism. Eh is the hepatic extraction ratio, which is a measure of the liver’s ability to extract drug from the systemic circulation and is used to assess the impact of hepatic extraction on oral bioavailability. Eh is calculated using Equation 4 [89]: (4) Table 3. Phase I enzymes. Phase I reactions Enzymes Oxidation P450 monooxygenase CL h E h = ---------Q Xanthine oxidase Peroxidases Amine oxidase Dioxygenase Semicarbizide-sensitive amine oxidase M onoamine oxidase Reduction where CLh is the hepatic drug blood clearance and Q is the hepatic blood flow. Because hepatic metabolism is the primary clearance mechanism for most marketed drugs, it is reasonable to assume that the total blood clearance (CL) is equal to CLh; shown in Equation 5: P450 monooxygenase CL F = ⎛ 1 – -------⎞ ⋅ Fa ⋅ Fg ⎝ Q⎠ Ketoreducatase Glutathione peroxidases Hydrolysis Epoxide hydrolase Table 4. Phase II reactions. Phase II reactions Examples Glucoronidation Phenytoin, chloramphenicol M ethylation Noradrenaline Acetylation Proconamide, isoniazid protein metabolism [89]. It is interesting to note that these reactions include a high-energy form of the conjugating agent, such as uridine diphosphate-glucuronic acid, acetyl CoA, 3 ′-phosphoadenosine-5 ′-phosphosulfate or S-adenosylmethionine, which in the presence of appropriate transferase enzyme, combines with the drug or its phase I metabolite to form the phase II conjugate [88]. T he major reactions and enzymes involved in these processes are shown in Tables 4 and 5 , respectively [90]. Oral bioavailability can be defined as the product of absorption and metabolism, as shown in Equation 3 [55,89]: 600 (5) As shown in Figure 9 , if a compound series has a rat blood CL value < 50 ml/min/kg and considering that the rat Q is equal to 70 ml/min/kg [91], then its E h is < 0.7 and its poor oral bioavailability can be assumed to be predominately a result of poor absorption and/or significant intestinal metabolism rather than high hepatic metabolism. T herefore, it is pivotal to consider factors in the absorption section of the road map that may influence compound oral bioavailability. As for the intestinal metabolism, it should be emphasised that our current understanding of intestinal metabolism has increased significantly as a result of major strides in the fields of molecular sciences and biochemistry. In the literature, it has been clearly demonstrated that the small intestine plays a pivotal role in first-pass metabolism, especially with compounds with poor aqueous solubility and low oral dose (dose < 100 mg/day) [55]. However, it is clear that both the protein level and catalytic activity of drug-metabolising enzymes in the small intestine are generally lower than those in the liver and that this is Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan Table 5. Phase II enzymes. Phase II enzymes Reactions Glucuronosyltransferase M ethylation Enzymes Sulfotransferase O-methyltransferase Glutathione S-transferase N-methyltransferase Glucosyltransferase S-methyltransferase Amide synthesis (transcylase) Acetylation N-acetyltransferase Acetyltransferase Thiosulfate Sulfotransferase particularly true for CYP enzymes. Considering the limited drug-metabolising capacity in the small intestine, the contribution of this organ to the overall metabolism of a drug is less likely to be quantitatively as important as that of the liver, unless a very small oral dose is given [78]. It should be noted that each species has a different value for Q (blood flow) and most likely different values for blood CLh. T herefore, comparing E h values simplifies comparisons between rat, dog, man and other evaluated species [78]. If a compound series has a rat CL > 50 ml/min/kg, then the discovery team should determine if the compound series has species differences in its hepatic stability. Importance of monitoring species differences It is important to note, as seen with the variable type and expression of transporters in oral absorption, species differences also exist in drug-metabolising enzymes. Several investigators attributed these findings to differences in physiological factors such as hepatic blood flow, or enzyme type and expression [78,89]. For example, Nelson et al. reported that so far 14 CYP gene families have been identified in mammals with significant variations in the primary sequence of amino acids across species. However, these members of the superfamily had highly conserved regions of amino acid sequence [92]. Similar findings were also reported with uridine diphosphoglucose transferases and carboxylesterases [93,94]. Overall, these small differences in the amino acid sequences can lead to significant differences in substrate affinity and specificity which translates into differences in the metabolism rate and metabolism profiles. As a general rule, compounds with good passive absorption, high rat hepatic extraction ratio, and poor oral bioavailability tend to have better oral bioavailability in higher species such as dogs, monkeys and humans. T here are many cited examples that are consistent with this trend. For example, the pharmacokinetics of remoxipride was studied in rodents (mice, rats, hamsters), dogs and humans [95]. Remoxipride was rapidly and completely absorbed through the intestinal wall in all species evaluated. However, the bioavailability was low in the rodents (< 10% in mice and hamsters and < 1% in rats) due to extensive first-pass elimination in the hepato–portal system. Blood clearance estimated after the same intravenous doses was high in rodents and similar to or exceeding normal liver blood flow. On the other hand, in dogs and humans, clearance values were low and the bioavailability high (> 90%) [95]. Other examples of marketed drugs that have poor oral bioavailability in rodents yet good bioavailability in higher species include reboxetine, diazepam and indinavir [26,95-98]. T hese successful medications would not be on the market if the discovery team solely depended on rat oral bioavailability to evaluate their metabolism in humans. 4.1 4.2 Liver microsomes as an indicative model for reducing metabolism-limited absorption Liver microsomes are widely used as a tool to assess the contribution of phase I enzymes such as CYP enzymes, flavin-containing monooxygenases and other metabolising enzymes (Table 3 ) [99]. Microsomes are convenient as they can be stored frozen, for long periods of time (years) without compromising their enzymatic activity [100]. The most common approach to using human microsomes is in determining hepatic stability, by measuring the intrinsic hepatic clearance (CLint) detailed in Equation 6 summarised by Obach et al. in a previous report [100]: (6) gm liver wt 1 CL int = 0.693 × -------- × ----------------------------t1 ⁄ 2 kg body wt ml incubation 45 mg microsomal prot × ----------------------------------------------------- × -----------------------------------------------------------mg microsomal prot gm liver wt A human CLint value > 40 µl/min/mg protein is considered high and warrants a metabolism SAR optimisation to improve the metabolic profile (Figure 9 ). One of the most widely used approaches to improve the metabolic stability of various compounds is to reduce drug lipophilicity, as higher Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 601 The road map to oral bioavailability: an industrial perspective Metabolism Rat CL > 50 ml/min/kg Yes Yes No Human microsomes CLint > 40 µl/min/mg Discuss with team absorption and/or metabolism-limiting factors. Consider factors in the absorption tree Metabolism SAR optimisation Discuss with team soft spot analysis (MetaSite) No No Human hepatocytes Eh > 0.7 Yes Metabolism SAR optimisation No Discuss with team species for toxicology studies Yes Dog hepatocytes Eh > 0.7 Metabolism SAR optimisation Figure 9. M etabolism section of the road map. When a compound series has a rat CL > 50 ml/min/kg, then the discovery team should determine if the compound series has species differences in its hepatic stability through the use of microsomes. These studies should then be follow ed by a thorough hepatic stability evaluation in human hepatocytes. CLint : Intrinsic hepatic clearance; Eh: Hepatic extraction ratio; SAR: Structure–activity relationship. lipophilicity increases the binding affinity of various drugs and xenobiotics to metabolising enzymes [8,101,102]. T here are two generally accepted approaches to reduce the lipophilic nature of a molecule. T he first is by replacing a bulky lipophilic group, which has minimal relevance to the pharmacological activity of the compounds, with groups known to lower lipophilicity. For example, Dragovich et al. replaced a benzyl group in the rhinovirus 3C protease inhibitor lead compound with ethyl and propyl groups. T hese changes were associated with significant improvement in the monkey oral exposure [103]. T he second approach is to introduce polar functional groups (e.g., pyridine) or isosteric atoms such as nitrogen or oxygen to improve the metabolic stability of compounds tested. Tagat et al. replaced the benzamide group 602 in the CCR5 antagonist lead compound with nicotinamide, which was associated with a significant improvement in its metabolic stability in rat, dog and monkey [104]. Another approach that is gaining interest from medicinal chemists is the use of in silico models of biotransformation such as MetaSite. T his technique is a new in silico method that provides the metabolism site for any human CYP-mediated reaction [105]. T he methodology can be applied automatically to all major CYP enzymes and used by medicinal chemists to detect positions that should be protected in order to avoid metabolic degradation or to check the suitability of a new scaffold or prodrug. Although this and other similar approaches do not identify rates of hepatic clearance, they provide information on molecular ‘soft spots’ and hence Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan provide for a handle to medicinal chemists to design analogues with superior properties [106,107]. 4.3 Hepatocytes: the more definitive guide to metabolism In the case where human microsomal data of a chemical series suggest low to moderate intrinsic clearance (CLint < 40 µl/min/mg protein), it is recommended to continue the hepatic stability evaluation in human hepatocytes ( Figure 9 ). Microsomes metabolise compounds primarily through phase I (e.g., CYP-mediated metabolism), whereas hepatocytes are known to metabolise compounds through both phase I and II reactions [108]. Examples have been cited where compounds were extensively metabolised by phase II, but not phase I and had significantly lower oral bioavailability. For example, raloxifene is extensively glucuronidated in humans, effectively eliminating its oral bioavailability (F% = 0.02) [109,110]. Other examples on drugs with extensive conjugation and poor oral bioavailability (F% < 30) are morphine, nalbuphine and propofol, to name a few [111-113]. T herefore, evaluating hepatic stability in hepatocytes is critical for developing a compound series with good oral bioavailability in humans. In general, compound stability in hepatocytes is assessed by determining E h using Equations 7 and 8 [100,114]: (7) gm liver wt 1 CL int = 0.693 × -------- × ----------------------------t1 ⁄ 2 kg body wt ml incubation cells × ------------------------------------------ × ----------------------------cells incubation gm liver wt For compounds to be developed in the clinic, an appropriate plasma exposure should be achieved in nonrodent species to study the drug toxicological profile. Dogs are usually the species of choice because of availability and resemblance to human physiology [116,117]. T he authors recommend evaluating the metabolism profiles of compounds in dog hepatocytes. If the compound continues to show high clearance in dog hepatocytes (Eh > 0.7), a metabolism SAR optimisation is necessary, especially when compounds demonstrate poor plasma exposure following oral dosing in dogs. T his will limit the ability of discovery teams to evaluate the compound toxicological profile in dogs. If the compound demonstrates low to moderate clearance in dog hepatocytes (E h < 0.7) then the team should consider other factors in the absorption section of the road map that may influence the compound oral bioavailability such as permeability or/and solubility. 5. T he road map concept provides a common language to focus discussions around the potential causes of poor oral bioavailability when evaluating options in optimising a chemical series. It is based on dividing the factors that influence the oral bioavailability in two interrelated categories: absorption (product of solubility and permeability) and metabolism. Each category has its unique role, but with a great degree of interdependence in many cases. Such a tool should allow discovery teams to streamline the process of addressing poor oral bioavailability by using key assays for evaluation of each factor that may limit oral exposure with significant emphasis on in vitro and in silico tools to reduce costs and length of time for optimising the critical physicochemical and pharmacokinetic properties of a chemical series. 6. Q h ⋅ fu ⋅ CL int CL h = -----------------------------------Q h + fu ⋅ CL int (8) where fu is the drug unbound fraction. An E h value > 0.7 suggests the potential for high clearance that would limit oral drug bioavailability, as shown in Equation 3. Metabolism SAR optimisation should be initiated to address the projected high human clearance using approaches shown in Section 3.2. In comparison, an E h value < 0.7 may suggest a low to moderate risk to metabolism-limiting drug disposition in humans. It should be noted that sometimes in vitro systems of metabolism can greatly overpredict the biotransformation of compounds that have high plasma protein binding and, therefore, it is critical to put the E h values in context [95,111,115]. Conclusion Expert opinion It is often difficult and costly to generate a complete set of physicochemical and pharmacokinetic data in early discovery settings due to limited compound supply and aggressive discovery timelines. As a result, it is pivotal to quickly identify major barriers contributing to poor bioavailability with as little compound as possible. T his tool provides a road map of physicochemical property assessments and molecular modelling to address relationships between molecular structure and various pharmacokinetic properties. T he Biopharmaceutics Classification System [202] and the Biopharmaceutics Drug Disposition Classification System [118] are examples of tools that move beyond molecular structure and consider both physicochemical and biological characteristics that impact drug absorption and disposition. T hese approaches to drug classification are useful for a particular molecular entity; however, as one progresses through the timelines of a discovery programme, several molecular templates are typically considered in parallel, greatly increasing the number of species to be evaluated. A discovery team must be able to triage molecules Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 603 The road map to oral bioavailability: an industrial perspective from a particular chemical template for potential bioavailability limitations using available high-throughput in vitro and in silico assays to minimise raw material demands. T he concept of developing a road map provides a foundation from which a team of scientists may begin to address the limitations to oral plasma exposure. T he map may be used either to pursue causal factors leading to poor in vivo exposure or to evaluate potential absorption/metabolism barriers prior to in vivo evaluation as an effort to conserve compound supply and animal resources. In either case, the map describes a systematic approach towards evaluating in vivo and in vitro data related to absorption based on Bibliography established pharmaceutical principles. It does not, however, predict the extent of oral absorption or provide a ranking of drug-likeness based on one’s progression through the various sections of the road map. Conclusions that may be made are ones of possible risk for certain molecules or chemical templates based on early in vitro, in vivo and in silico data. Acknow ledgements T he authors wish to thank M Mollan and J Brodfuehrer for their helpful review of this manuscript. AVDEEF A, T ESTA B: Physicochemical 8. 16. FLYNN GL, YALKOWSKY SH, Papers of special note have been highlighted as profiling in drug research: a brief survey of ROSEMAN T J: Mass transport phenomena either of interest (•) or of considerable interest the state-of-the-art of experimental and models: theoretical concepts. (••) to readers. techniques. Cell. Mol. Life Sci. (2002) 1. 59 (10):1681-1689. DIMASI JA, HANSEN RW, GRABOWSKI HG: T he price of 9. innovation: new estimates of drug development costs. J. Health Econ. (2003) 22 (2):151-185. • 2. 10. GRIBBON P, SEWING A: High-throughput drug discovery: what can 11. and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv. Drug Deliv. Rev. (2001) 46 (1-3):3-26. •• 4. T horough overview describing several 12. GRANT DJW, HIGUCHI T: Solubility MASUCCI JA, HAGEMAN W, YAN Z: behavior of organic compounds. T he new pre-preclinical paradigm: WH Saunders Jr (Ed.), John Wiley and Sons, New York, USA (1990):12-88. 19. DEHRING KA, WORKMAN HL, MILLER KD, MANDAGERE A, PARROT T N, PAQUEREAU N, POOLE SK: Automated robotic liquid COASSOLO P, LAVE T: An evaluation of handling/laser-based nephelometry system the utility of physiologically based models for high throughput measurement of kinetic of pharmacokinetics in early drug discovery. aqueous solubility. J. Pharm. Biomed. Anal. J. Pharm. Sci. (2005) 94 (10):2327-2343. (2004) 36 (3):447-456. HELLRIEGEL ET, BJORNSSON T D, 20. LIPINSKI CA: Solubility in water and DMSO: Issues and potential solutions. In: bioavailability is related to the extent of Pharmaceutical Profiling in Drug Discovery VAN DE WAT ERBEEMD H, absorption: implications for bioavailability for Lead Selection. RT Borchardt, EH Kerns, and bioequivalence studies. Clin. CA Lipinski et al. (Eds), AAPS Press, Arlington, VA, USA (2004):93-126. Pharmacol. Ther. (1996) 60 (6):601-607. 13. LIPINSKI CA: Drug-like properties and the 21. JAIN N, YALKOWSKY SH: Estimation of pharmacokinetics. J. Med. Chem. (2001) causes of poor solubility and poor the aqueous solubility I: application to 44 (9):1313-1333. permeability. J. Pharmacol. Toxicol. Methods organic nonelectrolytes. J. Pharm. Sci. DI L, KERNS EH: Profiling drug-like (2000) 44 (1):235-249. properties in discovery research. Curr. Opin. •• Chem. Biol. (2003) 7 (3):402-408. BALANI SK, MIWA GT, GAN L-S, WU J-T, LEE FW: Strategy of utilizing 14. in vitro and in vivo ADME tools for lead KERNS EH: High throughput physicochemical profiling for drug discovery. J. Pharm. Sci. (2001) 90 (11):1838-1858. YALKOWSKY SH: Solubility and miscibility. In: Solubility and solubilization rational drug design approaches. in aqueous media . Oxford University Press, VEBER DF, JOHNSON SR, CHENG HY New York, USA (1999):49-80. J. Med. Chem. (2002) 45 (12):2615-2623. 15. (2001) 90 (2):234-252. 22. output obtained from both H TS and the oral bioavailability of drug candidates. Curr. Top. Med. Chem. (2005) 5 :1033-1038. Provides a prospective to chemical matter et al.: Molecular properties that influence optimization and drug candidate selection. 604 Pharm. Res. (1995) 12 (3):413-420. 18. solubility and permeability. optimization of drug absorption and 7. CALDWELL GW, RITCHIE DM, HAUCK WW: Interpatient variability in WALKER DK: Property-based design: 6. dissolution and in vivo bioavailability. discovery approaches for assessing SMIT H DA, BEAUMONT K, 5. the correlation of in vitro drug product selection tool in drug discovery. Chem. (2001) 1 (5):353-366. LIPINSKI CA, LOMBARDO F, DOMINY BW, FEENEY PJ: Experimental for a biopharmaceutic drug classification: projection using in vitro screening data as a phase drug discovery. Curr. Top. Med. (2005) 10 (1):17-22. 3. SHAH VP, CRISON JR: A theoretical basis et al.: Integrated oral bioavailability compound optimization in early and late we expect from HTS? Drug Discov. Today AMIDON GL, LENNERNAS H, STONER CL, CLETON A, JOHNSON K Int. J. Pharm. (2004) 269 (1):241-249. A good overview for the financial impact of poor decision making. J. Pharm. Sci. (1974) 63 (4):479-510. 17. 23. ANDO HY, RADEBAUGH GW: Property-based drug design and preformulation. In: Remington: The Science PALM K, ST ENBERG P, LUT HMAN K, and Practice of Pharmacy. Philadelphia ART URSSON P: Polar molecular surface UotSi (Ed.) Lippincott, Williams and properties predict the intestinal absorption Wilkins, Philidelphia, PA, USA of drugs in humans. Pharm. Res. (1997) (2005):720-744. 14 (5):568-571. Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan 24. GOOSEN C, LAING T J, DU PJ, human bioavailability of phenytoin. Chem. gastrointestinal secretion of glutathione GOOSEN TC, FLYNN GL: Pharm. Bull. (1986) 34 (10):4400-4402. conjugates in rats. J. Pharmacol. Exp. Ther. LIVERSIDGE GG, CUNDY KC: Particle (2000) 292 (1):433-439. Physicochemical characterization and solubility analysis of thalidomide and its 36. size reduction for improvement of oral N -alkyl analogs. Pharm. Res. (2002) 19 (1):13-19. 25. UK-343,664 in man. Xenobiotica (2001) Int. J. Pharm. (1995) 125 (1):91-97. 31 (8-9):665-676. 37. 42 (12):2266-2271. VACCA JP, DORSEY BD, SCHLEIF WA et al.: L-735,524: an orally bioavailable 38. 39. (6-amino-seco-CI) DNA minor groove alkylating agents and structure–activity relationships for their cytotoxicity. J. Med. Chem. (1999) 42 (4):649-658. CHOWHAN ZT: pH-solubility profiles or 40. 30. 41. 43. structure requirements for pluronic block tract of the rate. J. Pharm. Sci. (1998) copolymers in modifying P-glycoprotein 87 :360-366. drug efflux transporter activity in bovine LENNERNAS H: Does fluid flow across brain microvessel endothelial cells. Classification analysis of P-glycoprotein FRIEND DR: Drug delivery to the small 18 (1):1-15. Good overview of intestinal transporters. 44. OH DM, HAN HK, WILLIAMSON RM et al.: Improved intestinal transport of PD-158473, an N -methyl-D -aspartate JOHNSON KC, SWINDELL AC: (NMDA) antagonist, by involvement of Guidance in the setting of drug particle size multiple transporters. J. Pharm. Sci. (2002) specifications to minimize variability in 91 (12):2579-2587. DOHERT Y MM, CHARMAN WN: T he • relevant as an organ of drug metabolism? Estimating the fraction dose absorbed from Clin. Pharmacokinet. (2002) suspensions of poorly soluble compounds in 41 (4):235-253. 46. GOTOH Y, SUZUKI H, KINOSHITA S et al.: Involvement of an organic anion YAKOU S, YAMAZAKI S, SONOBE T transporter (canalicular multispecific et al.: Particle size distribution affects the organic anion transporter/multidrug resistance-associated protein 2) in Expert Opin. Drug M etab. Toxicol. (2006) 2(4) Detailed analysis into P-gp substrate recognition. 52. GOMBAR VK, POLLI JW, HUMPHREYS JE, WRING SA, SERABJIT-SINGH CS: Predicting P-glycoprotein substrates by a quantitative structure–activity relationship model. J. Pharm. Sci. (2004) 93 (4):957-968. 53. SEELIG A, LANDWOJTOWICZ E: Structure-activity relationship of P-glycoprotein substrates and modifiers. Eur. J. Pharm. Sci. (2000) 12 (1):31-40. 54. T EODORI E, DEI S, SCAPECCHI S, GUALT IERI F: T he medicinal chemistry of multidrug resistance (MDR) reversing drugs. Farmaco (2002) 57 (5):385-415. 55. LIN J, SAHAKIAN DC, DE MORAIS SM et al.: T he role of absorption, distribution, mucosa of the small intestine: how clinically OH DM, CURL RL, AMIDON GL: (1993) 10 (2):264-270. substrate specificity. J. Drug Target. (2003) 11 (7):391-406. KAT SURA T, INUI K: Intestinal Drug Metab. Pharmacokinet. (2003) 45. DIDZIAPET RIS R, JAPERTAS P, AVDEEF A, PET RAUSKAS A: with improved pharmaceutical profiles for humans: a mathematical model. Pharm. Res. 51. PIYAPOLRUNGROJ N, ZHOU YS, LI C • 13 :1795-1798. 304 (2):845-854. 12 (11):1573-1582. et al.: Pyrroloquinolone PDE5 inhibitors absorption. Pharm. Res. (1996) J. Pharmacol. Exp. Ther. (2003) transporters: mechanisms and regulation. J. Med. Chem. (2005) 48 (6):2126-2133. 35. MILLER DW, KABANOV AV: Optimal absorption of drugs mediated by drug JIANG W, GUAN J, MACIELAG MJ BAT RAKOVA EV, LI S, ALAKHOV VY, drugs in different regions of the intestinal 6 (5):371-376. Pharm. Res. (2003) 20 (3):406-408. clinical studies on erectile dysfunction. 50. LENNERNAS H: Membrane transport of intestine. Curr. Gastroenterol. Rep. (2004) depends on the dose/solubility ratio. 34. disposition in rat. Pharmacology (1998) 56 (6):308-313. Drug Metab. Dispos. (2000) 28 (1):65-72. MACHERAS P: T he mean dissolution time ketoconazole on digoxin absorption and UNGELL A-L, NYLANDER S, elimination in rat small intestine. 42. SALPHAT I L, BENET LZ: Effects of Eur. J. Pharm. Sci. (1999) 9 (1):47-56. SERAJUDDIN AT: IV-IVC considerations RINAKI E, DOKOUMET ZIDIS A, Pharm. Pharmacol. (2004) 56 (8):967-975. 49. et al.: Cimetidine absorption and 94 (7):1396-1417. 33. pharmacokinetics of selected substrates. J. LI S, HE H, PART HIBAN LJ, YIN H, oral dosage form. J. Pharm. Sci. (2005) 32. KARLSSON J, UNGELL A, GRASJO J, reevaluation? Pharm. Res. (1995) KRAMER SF, FLYNN GL: Solubility of in the development of immediate-release 31. psychoactive compounds-implications for drug absorption? Is it time for a J. Pharm. Sci. (1978) 67 (9):1257-1260. (1972) 61 (12):1896-1904. substrates and inhibitors among (1990) 16 (2-3):101-108. the intestinal mucosa affect quantitative oral organic carboxylic acids and their salts. organic hydrochlorides. J. Pharm. Sci. et al.: Identification of P-glycoprotein BERGST RAND S, SJOBERG A, 3-(chloromethyl)-6-aminoindoline EL ELA AA, HART T ER S, SCHMIT T U cellular pathways. Miner. Electrolyte Metab. influence of charge and induced water flux. USA (1994) 91 (9):4096-4100. et al.: Synthesis of 1-substituted 48. absorption. Interplay of paracellular and transport across intestinal epithelia: protease inhibitor. Proc. Natl. Acad. Sci. MILBANK JB, T ERCEL M, AT WELL GJ NELLANS HN: Intestinal calcium ART URSSON P: Paracellular drug human immunodeficiency virus type 1 29. the non-proportional pharmacokinetics of ORT WINE DF et al.: Design and synthesis acid derivatives. J. Med. Chem. (1999) 28. et al.: Potential role for P-glycoprotein in Absolute oral bioavailability of nanocrystalline danazol in beagle dogs. AMPA/GlyN receptor antagonists: amino 27. ABEL S, BEAUMONT KC, CRESPI CL NIKAM SS, CORDON JJ, of novel quinoxaline-2,3-dione 26. 47. bioavailability of hydrophobic drugs: I. metabolism, excretion and toxicity in drug discovery. Curr. Top. Med. Chem. (2003) 3 (10):1125-1154. 56. LI Y, SHIN YG, YU C et al.: Increasing the throughput and productivity of Caco-2 cell permeability assays using liquid chromatography-mass spectrometry: application to resveratrol absorption and 605 The road map to oral bioavailability: an industrial perspective 57. 58. metabolism. Comb. Chem. High Throughput molecules. Chem. Pharm. Bull. (2004) Screen. (2003) 6 (8):757-767. 52 (5):561-565. PORT ER CJH: An in vitro examination of MARSILJE T H, ROSES JB, the impact of polyethylene glycol 400, LENNERNAS H: Human intestinal 67. permeability. J. Pharm. Sci. (1998) CALDERWOOD EF et al.: Synthesis and 87 (4):403-410. biological evaluation of imidazole-based POLLI JE, GINSKI MJ: Human drug small molecule antagonists of the 68. Pharm. Res. (1998) 15 (10):1520-1524. of analytical solutions for k(a) and F(a). 69. SUGIYAMA Y: Impact of drug transporter YEE S: In vitro permeability across Caco-2 studies on drug discovery and development. cells (colonic) can predict in vivo (small 70. WAT KINS PB et al.: First-pass midazolam metabolism catalyzed by 1 α,25-dihydroxy 71. 73. in microsomal tolbutamide, S-mephenytoin, and omeprazole hydroxylations. Arch. Biochem. Biophys. (1998) 353 (1):16-28. 81. iv [14 C]erythromycin LIN LS, LANZA T J, JR., potent and orally bioavailable antagonists of VLA-4. Bioorg. Med. Chem. Lett. (2004) 74. liver microsomes as a prototype reaction to monitor cytochrome P450 db1 activity. Clin. Pharmacol. Ther. (1989) 45 (1):34-40. breath and urine 82. DICKINSON RP et al.: Absorption, distribution, metabolism, and excretion PELKONEN O, RAUNIO H, RAUT IO A, LANG M: the proton-monocarboxylic acid Xenobiotic-metabolizing enzymes and co-transporter. Pharm. Res. (2000) cancer risk: correspondence between 17 (2):209-215. genotype and phenotype. IARC Sci. Publ. BOYD RA, ST ERN RH, ST EWART BH (1999) 148 :77-88. 84. LIU L, PANG KS: T he roles of transporters increase digoxin concentrations by and enzymes in hepatic drug processing. inhibition of intestinal Drug Metab. Dispos. (2005) 33 (1):1-9. J. Clin. Pharmacol. (2000) 40 (1):91-98. indole-containing endothelin antagonist. IARC Sci. Publ. (1999) 148 :13-22. 83. model: contributions of P-glycoprotein and P-glycoprotein-mediated secretion. considerations in selection of orally active LANG M, PELKONEN O: Metabolism of xenobiotics and chemical carcinogenesis. WU X, WHIT FIELD LR, ST EWART BH: et al.: Atorvastatin coadministration may WALKER DK, DACK KN, DAYER P, LEEMANN T, ST RIBERNI R: Dextromethorphan O-demethylation in LEMAHIEU WP, MAES BD, GHOOS Y Atorvastatin transport in the Caco-2 cell replacement of anilide with benzoxazole: 85. MEIBOHM B, BEIERLE I, DERENDORF H: How important are ST EWART BH, KUGLER AR, gender differences in pharmacokinetics? 29 (11):1424-1431. T HOMPSON PR, BOCKBRADER HN: Clin. Pharmacokinet. (2002) RUELL JA, TSINMAN O, AVDEEF A: A saturable transport mechanism in the Acid-base cosolvent method for determining aqueous permeability of amiodarone, itraconazole, tamoxifen, terfenadine and other very insoluble 606 CAVET ME, WEST M, SIMMIONS NL: Physiol. (2003) 285 (3):G470-G482. CASTONGUAY LA et al.: Bioisosteric 66. CYP2C9 from human liver: respective roles + segments. Pharm. Res. (1993) Drug Metab. Dispos. (2001) Characterization of CYP2C19 and the impact of species difference. test. Am. J. Physiol. Gastrointest. Liver intestinal epithelial cells and rat intestinal 65. ARAMSOMBAT DEE E, RAUCY JL: CYP3A4 and PGP activity by combined po absorption: cultured monolayers of human LASKER JM, WEST ER MR, A very well-designed study to determine et al.: Measurement of hepatic and intestinal permeability in two models of intestinal 14 (9):2331-2334. Curr. Drug Metab. (2000) 1 (2):107-132. 80. Dispos. (2005) 33 (4):587-595. Br. J. Pharmacol. (1996) 118 :1389-1396. 72. LOFROT H JE: Selective paracellular 10 (8):1123-1129. enzymes: mechanisms and functions. intestinal epithelial (Caco-2) cells. 289 (2):1134-1142. 64. WALKER DK, ABEL S, COMBY P et al.: cardiac glycoside, digoxin, by human monolayers. J. Pharmacol. Exp. Ther. (1999) ART URSSON P, UNGELL AL, discovery and development. Pharmacol. Rev. (1997) 49 (4):403-449. Transport and epithelial secreation of the vitamin D3-modified Caco-2 cell 63. pharmacokinetics and metabolism in drug A good overview. potential treatment for HIV. Drug Metab. •• LIN JH, LU AYH: Role of SHEWEITA SA: Drug-metabolizing CCR5 antagonist, UK-427,857, a new absorption. Expert Opin. Drug Metab. 78. • VAN BREEMEN RB, LI Y: Caco-2 cell FISHER JM, WRIGHTON SA, Mol. Pharmacol. (2002) 61 (5):974-981. 79. Species differences in the disposition of the permeability assays to measure drug 62. of P-glycoprotein inhibitors and substrates. Pharmacol. Rev. (2003) 55 (3):425-461. myth. Pharm. Res. (1997) 14 (6):763-766. Toxicol. (2005) 2005 (1):175-185. quantitative structure–activity relationships MIZUNO N, NIWA T, YOT SUMOTO Y, 314 (1):391-399. EKINS S, KIM RB, LEAKE BF et al.: Application of three-dimensional MDR1-MDCK and Caco-2 monolayers. development and evaluation and derivation 61. 4 (4):1-13. 77. cysteine protease inhibitor, across absorption-disposition model: model intestinal) absorption in man – fact or intestine. AAPS PharmaSci (2002) KO2, a novel vinylsulfone peptidomimetic Caco-2 permeability using an 60. enterocyte-based metabolism in excised rat of P-glycoprotein mediated transport of human drug oral absorption kinetics from J. Pharmacol. Exp. Ther. (2005) ZHANG Z, BENET LZ: Characterization polyethylene glycol 1000 succinate on P-glycoprotein efflux and 14 (14):3721-3725. Pharm. Res. (1998) 15 (1):47-52. USANSKY HH, SINKO PJ: Estimating D - α-tocopheryl Bioorg. Med. Chem. Lett. (2004) Caco-2 monolayer permeabilities. JOHNSON BM, CHARMAN WN, pluronic P85, and vitamin E melanocortin 4 receptor (MC4-R). absorption kinetics and comparison to 59. 76. 75. intestinal absorption of gabapentin is the underlying cause of the lack of proportionality between increasing dose and drug levels in plasma. Pharm. Res. (1993) 10 (2):276-281. Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 41 (5):329-342. 86. SCHMUCKER DL: Liver function and phase I drug metabolism in the elderly: a paradox. Drugs Aging (2001) 18 (11):837-851. Thomas, Bhattachar, Hitchingham, Zocharski, Naath, Surendran, Stoner & El-Kattan 87. 88. LIN JH, CHIBA M, BALANI SK et al.: angiotensin II receptor antagonists. VON MOLT KE LL, GREENBLAT T DJ: Species differences in the pharmacokinetics J. Med. Chem. (2005) 48 (13):4389-4399. Human drug metabolism and the and metabolism of indinavir, a potent cytochromes P450: application and human immunodeficiency virus protease relevance of in vitro models. J. Clin. inhibitor. Drug Metab. Dispos. (1996) Pharmacol. (2001) 41 (11):1149-1179. 24 (10):1111-1120. VENKATAKRISHNAN K, XU C, LI CY, KONG AN: Induction of 98. 99. phase I, II and III drug clearance from hepatic microsomal metabolism/transport by xenobiotics. metabolism data. Curr. Opin. Drug Discov. Devel. (2001) 4 (1):36-44. Arch. Pharm. Res. (2005) 28 (3):249-268. 89. 90. 100. OBACH RS, BAXT ER JG, LISTON T E (1999):353-398. et al.: T he prediction of human SHEWEITA SA, T ILMISANY AK: Cancer pharmacokinetic parameters from DAVIES B, MORRIS T: Physiological NELSON DR, KAMATAKI T, superfamily: update on new sequences, gene mapping, accession numbers, early trivial preclinical and in vitro metabolism data. J. Pharmacol. Exp. Ther. (1997) 283 (1):46-58. 101. LEWIS DF, DICKINS M: Substrate SARs in human P450s. Drug Discov. Today (2002) 7 (17):918-925. 102. MANNHOLD R: T he impact of clearance. Drug Metab. Dispos. (2002) 30 (6):694-700. 110. MORELLO KC, WURZ GT, DEGREGORIO MW: Pharmacokinetics of selective estrogen receptor modulators. Clin. Pharmacokinet. (2003) 42 (4):361-372. 111. HIRAOKA H, YAMAMOTO K, OKANO N et al.: Changes in drug plasma on β-blockers. Mini. Rev. Med. Chem. concentrations of an extensively bound and (2005) 5 (2):197-205. to altered plasma binding. Clin. Pharmacol. lipophilicity in drug research: a case report 103. DRAGOVICH PS, PRINS T J, ZHOU R highly extracted drug, propofol, in response Ther. (2004) 75 (4):324-330. names of enzymes, and nomenclature. et al.: Structure-based design, synthesis, and DNA Cell Biol. (1993) 12 (1):1-51. biological evaluation of irreversible human et al.: Pharmacokinetics of nalbuphine in BURCHELL B, EBNER T, BAIRD S et al.: rhinovirus 3C protease inhibitors. Part 7: infants, young healthy volunteers, and structure–activity studies of bicyclic elderly patients. Clin. Pharmacol. Ther. 2-pyridone-containing peptidomimetics. (1989) 46 (2):226-233. Use of cloned and expressed human liver UDP-glucuronosyltransferases for analysis of drug glucuronide formation and assessment of drug toxicity. Environ. Health Perspect. (1994) 102 (Suppl. 9):19-23. HOSOKAWA M, MAKI T, SATOH T: Characterization of molecular species of liver microsomal carboxylesterases of several animal species and humans. Arch. Biochem. Biophys. (1990) 277 (2):219-227. WIDMAN M, NILSSON LB, BRYSKE B, LUNDST ROM J: Disposition of remoxipride in different species. Species differences in metabolism. Arzneim. Forsch. (1993) 43 (3):287-297. BERGMANN JF, LANEURY JP, DUCHENE P et al.: Pharmacokinetics of reboxetine in healthy, elderly volunteers. Eur. J. Drug Metab. Pharmacokinet. (2000) 25 (3-4):195-198. 97. Characterization of raloxifene Appleton & Lange, Stamford, CT, USA WAXMAN DJ et al.: T he P450 96. Dispos. (2005) 33 (12):1852-1858. 109. KEMP DC, FAN PW, ST EVENS JC: glucuronidation in vitro: contribution of 10 (7):1093-1095. 95. for five benzodiazepines. Drug Metab. intestinal metabolism to presystemic humans. Pharm. Res. (1993) 94. human hepatocytes: kinetic characteristics microsomes in a discovery setting. parameters in laboratory animals and 93. of metabolic clearance using cryopreserved Biopharmaceutics and Pharmacokinetics. Curr. Drug Metab. (2003) 4 (1):45-58. 92. •• HAKOOZ N, HOUSTON JB: Prediction Provides an overview to the use of SHARGEL L, YU A: Applied and phase II drug-metabolizing enzymes. 91. OBACH RS: T he prediction of human 108. HALLIFAX D, RAWDEN HC, COCCHIARA G, BAT TAGLIA R, PEVARELLO P, ST ROLIN BENEDET T I M: Comparison of the disposition and of the metabolic pattern of Reboxetine, a new antidepressant, in the rat, dog, monkey and man. Eur. J. Drug Metab. Pharmacokinet. (1991) 16 (3):231-239. Bioorg. Med. Chem. Lett. (2002) 12 (5):733-738. 104. TAGAT JR, MCCOMBIE SW, 112. JAILLON P, GARDIN ME, LECOCQ B 113. WEST ERLING D, PERSSON C, HOGLUND P: Plasma concentrations of morphine, morphine-3-glucuronide, and NAZARENO D et al.: Piperazine-based morphine-6-glucuronide after intravenous CCR5 antagonists as HIV-1 inhibitors. IV. and oral administration to healthy Discovery of volunteers: relationship to nonanalgesic 1-[(4,6-dimethyl-5-pyrimidinyl)carbonyl]- actions. Ther. Drug Monit. (1995) 4-[4-[2-methoxy-1(R)-4-(trifluoromethyl) phenyl]ethyl-3(S)-methyl-1-piperaz inyl]4-methylpiperidine (Sch-417690/Sch-D), a potent, highly selective, and orally bioavailable CCR5 antagonist. J. Med. Chem. (2004) 47 (10):2405-2408. 105. ZAMORA I, AFZELIUS L, CRUCIANI G: Predicting drug metabolism: a site of metabolism prediction tool applied to the cytochrome P450 2C9. J. Med. Chem. (2003) 46 (12):2313-2324. 106. CRUCIANI G, CAROSAT I E, DE BOECK B et al.: MetaSite: understanding metabolism in human cytochromes from the perspective of the chemist. J. Med. Chem. (2005) 48 (22):6970-6979. 107. BERELLINI G, CRUCIANI G, MANNHOLD R: Pharmacophore, drug metabolism, and pharmacokinetics models 17 (3):287-301. 114. SHARGEL L, YU A: Applied Biopharmaceutics and Pharmacokinetics. Appleton & Lange, Stamford, CT USA (1999). 115. MASIMIREMBWA CM, BREDBERG U, ANDERSSON T B: Metabolic stability for drug discovery and development: pharmacokinetic and biochemical challenges. Clin. Pharmacokinet. (2003) 42 (6):515-528. 116. CALABRESE EJ: Suitability of animal models for predictive toxicology: theoretical and practical considerations. Drug Metab. Rev. (1984) 15 (3):505-523. 117. SCHIAVO DM: T he use of laboratory animals in toxicology: an ophthalmoscopic assessment. Toxicol. Pathol. (1990) 18 (1 Pt 2):222-223. on non-peptide AT 1, AT 2, and AT 1/AT 2 Expert Opin. Drug M etab. Toxicol. (2006) 2(4) 607 The road map to oral bioavailability: an industrial perspective 118. WU CY, BENET LZ: Predicting drug disposition via application of BCS: transport/absorption/elimination interplay and development of a biopharmaceutics drug disposition classification system. Pharm. Res. (2005) 22 (1):11-23. Websites 201. http://www.rxlist.com/02top.htm T he top 200 prescriptions for 2002 by number of US prescriptions dispensed (2002). 202. http://www.fda.gov/cder/OPS/BCS_ guidance.htm T he Biopharmaceutics Classification System (BCS) guidance (2001). Affiliation V Hayden T homas1 , Shobha Bhattachar2 , Linda Hitchingham3 , Philip Zocharski4 , Maryanne Naath3 , Narayanan Surendran5 , Chad L Stoner2 & Ayman El-Kattan†1 † Author for correspondence 1 Senior Principal Scientist, Pfizer Global Research and Development, Department of Pharmaceutical Sciences, 2800 Plymouth Road, Ann Arbor, MI 48105, USA 2 Principal Scientist, Pfizer Global Research and Development, Department of Pharmaceutical Sciences, 2800 Plymouth Road, Ann Arbor, MI 48105, USA 3 Senior Associate Scientist, Pfizer Global Research and Development, Department of Pharmaceutical Sciences, 2800 Plymouth Road, Ann Arbor, MI 48105, USA 4 Senior Scientist, Pfizer Global Research and Development, Department of Pharmaceutical Sciences, 2800 Plymouth Road, Ann Arbor, MI 48105, USA 5 Director, Pfizer Global Research and Development, Department of Pharmaceutical Sciences, 2800 Plymouth Road, Ann Arbor, MI 48105, USA 608 View publication stats Expert Opin. Drug M etab. Toxicol. (2006) 2(4)