6081 J Physiol 589.24 (2011) pp 6081–6092 Shaping a new Ca2+ conductance to suppress early afterdepolarizations in cardiac myocytes Roshni V. Madhvani1 , Yuanfang Xie2,3 , Antonios Pantazis1 , Alan Garfinkel2,5 , Zhilin Qu2,3 , James N. Weiss2,3,4 and Riccardo Olcese1,3 1 Division of Molecular Medicine – Department of Anesthesiology, 2 Department of Medicine (Cardiology), 3 Cardiovascular Research Laboratory, Departments of Physiology, and 5 Departments of Integrative Biology & Physiology, David Geffen School of Medicine at University of California, Los Angeles, CA 90095-7115, USA The Journal of Physiology 4 Non-technical summary Diseases, genetic defects, or ionic imbalances can alter the normal electrical activity of cardiac myocytes causing an anomalous heart rhythm, which can degenerate to ventricular fibrillation (VF) and sudden cardiac death. Well-recognized triggers for VF are aberrations of the cardiac action potential, known as early afterdepolarizations (EADs). In this study, combining mathematical modelling and experimental electrophysiology in real-time (dynamic clamp), we investigated the dependence of EADs on the biophysical properties of the L-type Ca2+ current (I Ca,L ) and identified modifications of I Ca,L properties which effectively suppress EAD. We found that minimal changes in the voltage dependence of activation or inactivation of I Ca,L can dramatically reduce the occurrence of EADs in cardiac myocytes exposed to different EAD-inducing conditions. This work assigns a critical role to the L-type Ca2+ channel biophysical properties for EADs formation and identifies the L-type Ca2+ channel as a promising therapeutic target to suppress EADs and their arrhythmogenic effects. Abstract Sudden cardiac death (SCD) due to ventricular fibrillation (VF) is a major world-wide health problem. A common trigger of VF involves abnormal repolarization of the cardiac action potential causing early afterdepolarizations (EADs). Here we used a hybrid biological–computational approach to investigate the dependence of EADs on the biophysical properties of the L-type Ca2+ current (I Ca,L ) and to explore how modifications of these properties could be designed to suppress EADs. EADs were induced in isolated rabbit ventricular myocytes by exposure to 600 μmol l−1 H2 O2 (oxidative stress) or lowering the external [K+ ] from 5.4 to 2.0–2.7 mmol l−1 (hypokalaemia). The role of I Ca,L in EAD formation was directly assessed using the dynamic clamp technique: the paced myocyte’s V m was input to a myocyte model with tunable biophysical parameters, which computed a virtual I Ca,L , which was injected into the myocyte in real time. This virtual current replaced the endogenous I Ca,L , which was suppressed with nifedipine. Injecting a current with the biophysical properties of the native I Ca,L restored EAD occurrence in myocytes challenged by H2 O2 or hypokalaemia. A mere 5 mV depolarizing shift in the voltage dependence of activation or a hyperpolarizing shift in the steady-state inactivation curve completely abolished EADs in myocytes while maintaining a normal Cai transient. We propose that modifying the biophysical properties of I Ca,L has potential as a powerful therapeutic strategy for suppressing EADs and EAD-mediated arrhythmias. (Received 26 August 2011; accepted after revision 21 October 2011; first published online 24 October 2011) Corresponding author R. Olcese: Division of Molecular Medicine, BH 570 CHS, Department of Anesthesiology, D. Geffen School of Medicine, University of California Los Angeles, CA, USA. Email: rolcese@ucla.edu Abbreviations AP, action potential; APD, action potential duration; CaMKII, Ca2+ –calmodulin-dependent protein kinase II; EAD, early afterdepolarization; I Ca,L , L-type Ca2+ current; RTXI, real-time experimental interface; SR, sarcoplasmic reticulum; VF, ventricular fibrillation; VT, ventricular tachycardia.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society DOI: 10.1113/jphysiol.2011.219600 Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6082 R. V. Madhvani and others Introduction Early afterdepolarizations (EADs) are transient reversals of normal repolarization (i.e. positive dV /dt), occurring in late phase 2 or phase 3 of the cardiac action potential (AP). These events, observable both at the cellular and tissue level, are recognized as triggers of cardiac arrhythmias (Sato et al. 2009; Weiss et al. 2010b). When EADs propagate through cardiac tissue, they can generate premature heart beats, ventricular tachycardia (VT) and ventricular fibrillation (VF), common causes of sudden cardiac death in a variety of clinical settings. EADs classically occur during bradycardia under conditions of reduced repolarization reserve, due to either abnormally decreased outward currents, or increased inward currents, or both. The predominant inward currents in ventricular myocytes which have been implicated in EAD genesis are the persistent I Na (Maltsev et al. 1998), I NCX (Volders et al. 1997; Szabo et al. 1994) and I Ca,L (January & Riddle, 1989). Previous work has shown that reactivation of the L-type Ca2+ current (I Ca,L ) during phase 2 and 3 of the action potential (AP) plateau plays a major role in EAD formation (January et al. 1988), although other factors are also important (Weiss et al. 2010a). Most EADs are initiated between −40 and 0 mV (Fig. 1), corresponding to the range of membrane potentials where the steady-state activation and inactivation curves of I Ca,L overlap, often referred to as the ‘window current’ region (January & Riddle, 1989; Antoons et al. 2007a). As the AP repolarizes into this voltage ‘window’, a fraction of the L-type Ca2+ channels not inactivated may be available for reactivation to generate the upstroke of the EAD. In this study, we sought to determine the relevance of I Ca,L in EAD formation, and investigated whether EAD suppression can be achieved solely by manipulating I Ca,L without altering other ionic conductances or signalling pathways. We analysed whether EAD formation can be suppressed by minimal perturbation of the biophysical properties of L-type Ca2+ channels affecting the I Ca,L window voltage region. Since it is difficult to control these properties experimentally with any degree of precision, we adopted a hybrid biological–computational approach using the dynamic clamp technique (Dorval et al. 2001; Berecki et al. 2005, 2006, 2007; Wilders, 2006). This allowed us to replace the native I Ca,L of the myocyte with a computed virtual I Ca,L with programmable properties. Our findings show that EAD formation is highly sensitive to I Ca,L properties and properties of the window region (such as the slopes of the activation and inactivation curves (Tran et al. 2009)), such that subtle shifts in the voltage dependence of steady-state I Ca,L activation or inactivation can abolish EADs. Moreover, these changes are predicted to have no significant effects on the amplitude or kinetics of the intracellular Ca2+ transient, suggesting that pharmacological or J Physiol 589.24 genetic manipulation of the voltage-dependent properties of I Ca,L could have therapeutic potential for suppressing EAD-mediated arrhythmias without adversely affecting excitation–contraction coupling. Methods Ethical approval All animal protocols were approved by the UCLA Institutional Animal Care and Use Committee and conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health. Electrophysiology Ventricular myocytes were isolated from New Zealand White rabbits as previously described (Mahajan et al. 2008a). All animals were anaesthetized with 50 mg ml−1 intravenous pentobarbital, before excision of the hearts, which were subsequently submerged in Ca2+ -free Tyrode solution containing (in mmol l−1 ): 136 NaCl, 5.4 KCl, 1 MgCl2 , 0.33 NaH2 PO4 , 10 glucose, and 10 Hepes, adjusted to pH 7.4. The hearts were cannulated and perfused retrogradely on a Langendorff apparatus with Ca2+ -free Tyrode buffer containing 1.65 mg ml−1 collagenase and 0.8 mg ml−1 bovine serum albumin for 30–40 min. After washing out the enzyme solution, the hearts were swirled in a beaker to dissociate cells. The Ca2+ concentration was gradually increased to 1.8 mmol l−1 , and the cells were stored at room temperature. This procedure typically yielded 50% rod-shaped Ca2+ -tolerant myocytes. Only Ca2+ -tolerant, rod-shaped ventricular myocytes with clear striations were randomly selected for electrophysiological studies within 8 h of isolation. All recordings were measured with AxoPatch 200B (Axon Instruments). Whole-cell patch-clamp recordings (in current or voltage clamp mode) were performed at 34–36◦ C using pipettes of 1–2 M. The intracellular solution contained (in mmol l−1 ): 110 potassium aspartate, 30 KCl, 5 NaCl, 10 Hepes, 0.1 EGTA, 5 MgATP, 5 creatine phosphate, 0.05 cAMP adjusted to pH 7.2. For I Ca,L recordings the solutions were as described above, except that K+ was replaced with Cs+ to abolish K+ conductance and 10 μmol l−1 TTX was added to the extracellular solution to block Na+ conductance. The L-type-specific Ca2+ current was determined by subtracting the current after 20 μmol l−1 nifedipine from the total current. Data were acquired and analysed using custom-made software (G-Patch, Analysis). Currents were filtered at 1/5 of the sampling frequency. The steady-state activation curves were constructed by dividing the peak I–V curve by the driving force to calculate conductance (G) and dividing G by G max . The inactivation curves were constructed by plotting the normalized peak  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 J Physiol 589.24 ICa,L window current properties set EAD susceptibility current during a test pulse at 10 mV following a 300 ms inactivating pulse at various voltages. The slope and half-activation/inactivation potential of these curves were estimated by fitting to a Boltzmann distribution, and these parameters were used to constrain the dynamic clamp model for both control and H2 O2 conditions. These two versions of the model were then implemented for Dynamic Clamp in RTXI (www.rtxi.org) (Lin et al. 2010). Sampling/computation frequency was 10 kHz. Rudy, 1994), where P o is formulated as: Po = d f q where d is the voltage-dependent activation gate, f is the voltage-dependent inactivation gate and q is the Ca2+ -dependent inactivation gate. The steady states of these gating variables as functions of voltage are formulated in the following manner: 1.0 1.0 + exp(−(v − dhalf )/dslope ) (3) 1.0 − p dest + p dest 1.0 + exp((v − dhalf )/dslope ) (4) 1.0  C s γ 1.0 + cst (5) d∞ (1 − exp(−v − (dhalf + a1 )) a2 (dslope + a2 )(v − (dhalf + a1 )) (6) d∞ = Intracellular calcium measurements Myocytes were loaded with 10 μmol l−1 Fluo4-AM (Molecular probes) for 30 min at room temperature. Cells were then washed twice in regular Tyrode solution and placed in a heated chamber at 34–36◦ C for fluorescence and electrophysiology measurements. Intracellular calcium (Cai ) transients were recorded using a CMOS camera (Silicon Imaging SI-1300, Niskayuna, NY, USA). Dynamic clamp f∞= q∞ = τd = τf = b1 exp(−b2 (v − (f half + b3 ))2 ) + b4 The ionic conductances included in the model are I Ca,L , the fast sodium current I Na , the Na+ –K+ pump current I NaK , the Na+ –Ca2+ exchange current I NCX , and the Ca2+ -dependent slow component of the delayed rectifier potassium channel I Ks . The intracellular Ca2+ concentration was divided into four compartments, i.e. submembrane Ca2+ (C s ), cytosolic Ca2+ (C i ) and network SR Ca2+ and junctional SR Ca2+ . The average submembrane Ca2+ concentration, C s , is used to compute Ca2+ -dependent inactivation in the I Ca,L formulation. The Ca2+ flux into the cell due to I Ca,L is given by (Mahajan et al. 2008b): J Ca = g Ca P o i Ca , 4P Ca VF 2 C s e 2a − 0.341[Ca2+ ]o RT e 2a − 1 (7) where τd and τf are the time constants of d gate and f gate respectively, d half is the voltage at half maximum of activation/inactivation (f half ), d slope is the slope factor of the activation curve, pdest is the non-inactivating pedestal of the inactivation gate, C s is the submembrane Ca concentration in units of mM, Cst is the affinity for Ca of the inactivation gate, and a1, a2, b1, b2, b3, and b4 are additional factors used for fitting. All the parameters listed in the equations are fitted to our experimental data (the fitting parameters are shown in online supplemental materials). The Ca2+ cycling is the same as that in the rabbit ventricular myocyte model (Mahajan et al. 2008b). (1) where J Ca is the current flux into the cell via L-type Ca2+ channels, g Ca is determined by fitting to the nifedipine-sensitive current traces measured under voltage clamp, P o is the probability of finding an L-type Ca2+ channel in the open state, and iCa is the driving force for Ca2+ given by: i Ca = 6083 (2) where C s is the submembrane concentration in units of mmol l−1 , P Ca (0.00054 cm/s) is the Ca channel permeability and V is the membrane voltage, F is the Faraday constant, and T is temperature. In order to facilitate shifts of the activation and inactiation curves, we replaced the Markovian calcium channel open probability P o (Mahajan et al. 2008b) with the Hodgkin–Huxley formulation from the Luo–Rudy dynamic model (Luo & Results Oxidative stress-induced and hypokalaemia-induced EADs in isolated rabbit ventricular myocytes To determine the relevance of I Ca,L biophysical properties to EAD formation and/or suppression, we used two methods to induce EAD in whole-cell patch-clamped isolated rabbit ventricular myoyctes: (i) oxidative stress, by exposure to 600 μmol l−1 H2 O2 (Xie et al. 2009), or (ii) hypokalaemia, by lowering extracellular [K+ ] (from 5.4 to 2.0–2.7 mmol l−1 ). Myocytes were superfused in the recording chamber at 34–36◦ C, and stimulated at a pacing cycle length (PCL) of 5 s (Fig. 1). Under current clamp, both H2 O2 (n = 14) and hypokalaemia (n = 8) prolonged the APD significantly and consistently produced EADs in ∼70% of paced APs (Fig. 1A and B) within 5–10 min. In our analysis, we individually inspected each action  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6084 R. V. Madhvani and others potential for EAD occurrence, defined as well-resolved deflections of the voltage with a positive dV /dt during late phase 2 or phase 3 of the AP. EAD morphology varied among experiments and also within the same action potential as in Fig. 1B (under hypokalaemia). Typically, EADs exhibited a positive voltage deflection of ≥5 mV with durations (from inflection point to peak) between 30 and 150 ms). Among other effects, oxidative stress with H2 O2 is known to promote EADs by activating Ca2+ –calmodulin-dependent protein kinase II (CaMKII) (Xie et al. 2009), which alters I Ca,L (Dzhura et al. 2000) and induces late I Na (Wagner et al. 2006; Wagner et al. 2011). To characterize how H2 O2 altered I Ca,L properties under our experimental conditions, we recorded the nifedipine-sensitive Ca2+ current in voltage-clamped isolated myocytes before and after exposure to 600 μmol l−1 H2 O2 (n = 5) (Fig. 2): in addition to an overall slowing J Physiol 589.24 of inactivation (Fig. 2A and B), H2 O2 shifted the peak of the current–voltage (I–V ) relationship by −5 mV as shown in Fig. 2C. By constructing steady-state activation and inactivation curves, we found that H2 O2 produced an ∼−5 mV shift of the steady state half-activation potential (V 1/2,act ), accompanied by an ∼+5 mV shift in the steady state half-inactivation potential (V 1/2,inact ) (Fig. 2D). Effectively, H2 O2 increased the height of the I Ca,L window current region, changes which are expected to promote EAD formation (Antoons et al. 2007b; Qi et al. 2009; Tran et al. 2009). Role of oxidative stress-mediated I Ca,L alterations in EAD generation To understand how the H2 O2 -induced changes in I Ca,L described above contributed to EAD formation, we next used the dynamic clamp technique (Tan & Joyner, 1990) to Figure 1. Oxidative stress and hypokalaemia generate EADs in dissociated rabbit cardiomyocytes A, representative AP recordings in current-clamped ventricular myocytes stimulated at a pacing cycle length of 5 s. EADs, absent in control conditions, were consistently induced by addition of 600 µmol l−1 H2 O2 to the bath solution or by reducing extracellular [K+ ] from 5.4 mmol l−1 to 2.7 mmol l−1 (hypokalaemia). The histogram to the left of each trace shows the distribution of action potential duration measured at 90% repolarization (APD90 ) for each condition. B, action potentials in the dotted box in A are shown on an expanded scale (arrows point to EADs). C, percent of action potentials displaying EADs in Control, H2 O2 and hypokalaemia conditions.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 J Physiol 589.24 ICa,L window current properties set EAD susceptibility replace the myocyte’s native I Ca,L with a virtual I Ca,L with programmable properties. The principle of the dynamic clamp operation is illustrated in Fig. 3A. The recorded membrane potential of a current-clamped myocyte (V m ) is fed into a Linux-based computer with the Real-time Experimental Interface (RTXI) software (Dorval et al. 2001) running the UCLA ventricular myocyte AP model (Mahajan et al. 2008b). The model computes, in real time, the macroscopic L-type Ca2+ channel current, which is injected continuously into the myocyte, in turn changing its membrane potential. This cycle (V m sampling → I Ca,L computation → I Ca,L injection → V m sampling . . .) is iterated at 10 kHz, creating a dynamic, bidirectional relationship between the model and the cell. The virtual I Ca,L computed in the AP model was tuned to fit experimental recordings of native I Ca,L 6085 recorded from rabbit ventricular myocytes (Fig. 3B). In addition to computing I Ca,L , the model also computes all other elements of Cai cycling in the UCLA rabbit ventricular myocyte model (Mahajan et al. 2008b), which is critical for simulating the effects of intracellular Ca2+ on Ca2+ -dependent inactivation of the virtual I Ca,L being injected into the myocyte. Thus, this model also generates a virtual intracellular Ca2+ transient, which showed good agreement with the intracellular Ca2+ transient recorded directly from a myocyte under control conditions (Fig. S1). The model computes Ca2+ concentrations in four discrete subcellular compartments (Fig. 3A), with the average submembrane Ca2+ concentration (C s ) used for the calculation of Ca2+ -dependent inactivation of I Ca,L , as described previously (Mahajan et al. 2008b). Although we initially used a seven-state Markovian description of I Ca,L , Figure 2. H2 O2 alters both the time-dependent and steady-state properties of ICa,L in rabbit ventricular myocytes A, voltage clamp recordings of nifedipine (20 µmol l−1 ) sensitive ICa,L in response to depolarizing pulses from −40 mV (HP = −80 mV), under control conditions (black) or after 2 min of exposure to 600 µmol l−1 H2 O2 (red). B, voltage clamp recordings of nifedipine-sensitive ICa,L in response to a two-pulse protocol from a holding potential of −90mV, under control conditions (black) or after 2 min of exposure to 600 µmol l−1 H2 O2 (red). The inactivating pulse was 300 ms to approximate the duration of a normal action potential. C, the I–V relationship for ICa,L (normalized to the peak current) under control conditions (black circles, n = 4) and H2 O2 (red circles, n = 5). Note the negative shift of the I–V curve in H2 O2 . D, mean steady-state activation and inactivation curves for ICa,L in control (black circles, n = 4) and H2 O2 conditions (red diamonds, n = 5). Note that the H2 O2 -induced shift of steady-state curves results in a higher ICa,L window area. External solution (mmol l−1 ): 136 NaCl, 5.4 CsCl, 1 MgCl2 , 0.33 NaH2 PO4 , 1.8 CaCl2 , 10 glucose, 10 Hepes, and 10 µmol l−1 TTX, pH 7.4. Pipette solution (in mmol l−1 ): 100 caesium aspartate, 30 CsCl, 5 NaCl, 10 Hepes, 0.1 EGTA, 5 MgATP, 5 creatine phosphate, 0.05 cAMP, pH 7.2. Temperature = 35–37◦ C.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6086 R. V. Madhvani and others we achieved equivalent accuracy of fitting the experimental I Ca,L data with a Hodgkin–Huxley type I Ca,L formulation, with the advantage that the biophysical parameters relevant to this work could be directly modified, such as V 1/2,act and V 1/2,inact , which determine the steady-state activation and inactivation properties of the injected virtual I Ca,L . In our experimental protocol utilizing the dynamic clamp technique, we first paced a rabbit ventricular myocyte to record the AP in the current clamp mode (Fig. 4A, trace 1). The recorded AP waveform was used to compute a virtual Cai transient, displayed below the AP trace (Fig. 4B, trace 1). The myocyte was then exposed to 600 μmol l−1 H2 O2 until EADs appeared consistently (Fig. 4A, trace 2). Nifedipine (20 μmol l−1 ) was then added to block the native I Ca,L , which markedly shortened APD and abolished EADs (Fig. 4, trace 3). Current injection by the dynamic clamp was then turned on, adding the virtual I Ca,L computed from the UCLA rabbit ventricular AP cell model (Mahajan et al. 2008b), whose properties had been adjusted to simulate the previously analysed effects of H2 O2 on I Ca,L gating properties shown in Fig. 2. Injection J Physiol 589.24 of this ‘H2 O2 -modified’ virtual I Ca,L effectively restored the electrical properties of the myocyte membrane, resulting in AP prolongation and the reappearance of EADs (Fig. 4, trace 4). The associated computed virtual Cai transients are shown in Fig. 4B. To investigate the sensitivity of EAD formation to the properties of the virtual I Ca,L , we then shifted the voltage dependence of I Ca,L activation (V 1/2,act ) by +5 mV, to resemble the normal steady activation curve prior to H2 O2 exposure. This small shift, which effectively suppressed the I Ca,L ‘window’ current by reducing the overlap between the steady state activation and inactivation curves to (Fig. 4C), abolished EADs and restored the AP duration towards the control value (Fig. 4, trace 5). Moreover, the virtual Cai transient (Fig. 4B, trace 5) had similar amplitude and kinetics to that computed for the normal control AP (Fig. 4B, trace 1). Similar results were also achieved by shifting the voltage dependence of I Ca,L inactivation (V 1/2,inact ) in the negative direction to reduce the window current region (not shown). The overall results, summarized in Fig. 5, demonstrate that the small shifts in the V 1/2 of steady-state activation Figure 3. The dynamic patch clamp implements a virtual Ca2+ current computed with detailed Ca2+ cycling dynamics A, diagram of the dynamic clamp system. A ventricular myocyte is whole-cell patch-clamped with an amplifier in current-clamp mode. The myocyte’s V m signal (red arrow) is digitized and input to a computer running a cardiac action potential model (Mahajan et al. 2008) containing macroscopic Hodgin–Huxley (HH)-type formulations for five conductances (ICa,L ; the fast Na+ channel INa ; the Na+ /K+ pump INaK ; the Na+ /Ca2+ exchanger INCX ; and the Ca2+ -dependent K+ rectifier IKs ) and [Ca2+ ] in four discrete cellular compartments (C s , C i , C jp and C j ). Ionic conductances and concentrations can be fine-tuned by altering model parameters online. The calculated ionic conductances can be output in any combination to produce ICommand . In this case, ICommand = ICa,L (blue arrow). ICommand is combined with a pacing AP stimulus and converted to analog before being input to the amplifier to be injected into the clamped myocyte. This alters the myocyte V m , which is in turn sampled for the next computation. Thus, there is a dynamic, bidirectional relationship between V m and model conductance output, at a sampling/computation frequency of 10 kHz. B, the predicted ICa,L conductance from the model in panel A (red) accurately predicts a current similar in amplitude and kinetics to experimentally recorded ICa,L from rabbit cardiomyocytes (black) in response to depolarizing pulses from a holding potential of −40 mV.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 J Physiol 589.24 ICa,L window current properties set EAD susceptibility and inactivation of I Ca,L strongly suppress EAD formation. A +4 mV shift of the V 1/2,act from −5 mV to −1 mV (Fig. 5A) shortened APD90 from 1197 ± 112 (n = 7) to 373 ± 15 ms (n = 4) (Fig. 5B) and suppressed EAD occurrence from 91 ± 4% to 24 ± 20% of paced APs (Fig. 5C). An additional +1 mV shift to 0 mV (Fig. 5A) shortened APD90 to 331 ± 15 ms (n = 6) (Fig. 5B) and further suppressed EAD occurrence to ∼2% (Fig. 5C). Similar results were achieved with small hyperpolarizing shifts in the V 1/2,inact . A −3 mV shift from −17 mV to −20 mV (Fig. 5D) reduced APD90 to 696 ± 42 ms (n = 3) (Fig. 5E) and EAD occurrence to 52% (Fig. 5F). Further shifting V 1/2,inact by an additional −2 mV to −22 mV (Fig. 5D) reduced APD90 to 488 ± 21 ms (n = 4) (Fig. 5E) and EAD occurrence to 10%. 6087 Reducing I Ca,L window current also suppresses hypokalaemia-induced EADs To demonstrate that the results described above are not unique to H2 O2 -induced EADs, we also assessed the effectiveness of shifting V 1/2,act or V 1/2,inact at suppressing hypokalaemia-induced EADs. Lowering extracellular K+ from 5.4 to 2.0 mmol l−1 hyperpolarized the resting membrane potential to nearly −100 mV and induced EADs within 5 min (Fig. 1 and 6). As with H2 O2 -induced EADs, blocking I Ca,L with 20 μmol l−1 nifedipine abolished EADs and shortened the action potential (not shown). Adding a virtual I Ca,L with dynamic clamp prolonged APD90 to 2170 ± 190 ms (n = 6) (Fig. 6C and E, hatched bars) and reconstituted EADs in 83% of APs Figure 4. Reconstitution of EADs by a virtual I Ca,L and their suppression by shifting I Ca,L V 1/2 activation A, APs from rabbit ventricular myocytes recorded at a pacing cycle length of 5 s at 36–37◦ C in control conditions (1). Addition of 600 µmol l−1 H2 O2 to the bath solution induced EADs within 5–10 min (2). Addition of 20 µmol l−1 nifedipine blocked the native ICa,L , shortened APD and abolished EADs (3). The dynamic clamp was then turned on to inject a virtual ‘ICa,L ’ computed in real time (red trace), which restored EADs (4). Varying a single ICa,L parameter (V 1/2 activation) by +5 mV to effectively reduce the ‘ICa,L window region’ abolished EADs, despite the presence of H2 O2 (5). B, representative action potentials from each condition (numbered) in A and their corresponding virtual Cai transients. Note that both action potential duration and the shape and amplitude of the computed Cai transient in conditions (1) and (5) are similar. C, graphical representation of the change in the overlap of the activation and steady-state inactivation curves upon a +5 mV shift in the V 1/2 of activation.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6088 R. V. Madhvani and others (Fig. 6D and F, hatched bars). A +5 mV depolarizing shift in the steady state activation curve from −1 to +4 mV shortened APD90 to 476 ± 39 ms (n = 3) (Fig. 6C, open bar) and reduced EAD occurrence to 0% (Fig. 6D, open bar). Similarly, a hyperpolarizing shift in the steady state inactivation curve from −18 to −23 mV shortened APD90 to 869 ± 167 ms (n = 3) and reduced EAD occurrence to 10% of APs. Discussion It is widely accepted that EADs are important cellular triggers for VT and VF when repolarization reserve is reduced, as can occur clinically in acquired and congenital long QT syndromes and heart failure (Huffaker et al. 2004; Maruyama et al. 2011). In the present study, we used two different conditions, oxidative stress J Physiol 589.24 (H2 O2 ) and hypokalaemia (Fig. 1), which reliably induce bradycardia-dependent EADs (positive dV /dt during phase 2 or 3) by different mechanisms in >90% of rabbit ventricular myocytes within 5–10 min of exposure. After blocking the native I Ca,L with nifedipine, we used the dynamic clamp technique to inject a virtual I Ca,L , which successfully reconstituted EADs. Moreover, the biophysical properties of the virtual I Ca,L could be reprogrammed to explore its role in EAD formation. Our major findings are: (i) EAD formation is very sensitive to small shifts in the voltage sensitivity of steady state activation and inactivation of I Ca,L , such that a +5 mV depolarizing shift in V 1/2,act , or a −5 mV hyperpolarizing shift in V 1/2,inact powerfully suppressed EAD formation (Figs 4–6); (ii) these small shifts were predicted to have minimal effects on peak I Ca,L during the AP, such that the Cai transient, hence excitation–contraction coupling Figure 5. EADs induced by oxidative stress are steeply dependent on V 1/2 of activation and inactivation A, steady-state activation and inactivation of the virtual ICaL under control conditions (black) and with the V 1/2 activation shifted by −1 (blue) or −5 mV (red) to partially simulate the effects of oxidative stress increasing the ICaL window region (cross-hatched). B and C, APD90 (B) and the percentage of AP’s exhibiting EADs (C) corresponding to the virtual ICa,L activation/inactivation properties in A, when the respective currents were injected by the dynamic clamp into ventricular myocytes in the presence of 0.6 mmol l−1 H2 O2 and 20 µmol l−1 nifedipine. The virtual ICa,L simulating oxidative stress caused marked AP prolongation and EADs, which were suppressed by returning the V 1/2 activation towards normal (blue and black). D, steady-state activation and inactivation of the virtual ICa,L under control conditions (black) and with the V 1/2 inactivation shifted by +3 (blue) or +5 mV (red) to partially simulate the effects of oxidative stress increasing the ICa,L window region (cross-hatched). E and F, APD90 (E) and the percentage of APs exhibiting EADs (F) corresponding to the virtual ICa,L activation/inactivation properties in A, when the respective currents were injected by the dynamic clamp into ventricular myocytes in the presence of 0.6 mmol l−1 H2 O2 and 20 µmol l−1 nifedipine. The virtual ICa,L corresponding to the red curve simulating oxidative stress caused marked AP prolongation and EADs, which were suppressed by returning the V 1/2 inactivation towards normal (blue and black).  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 J Physiol 589.24 ICa,L window current properties set EAD susceptibility and contractility, remain normal (Fig. 4B); (iii) the same strategy was effective at suppressing EADs generated by two different mechanisms (Figs 4 and 6): oxidative stress in which CaMKII activation potentiating inward Na+ and Ca2+ currents plays a critical role (Dzhura et al. 2000; Anderson, 2004; Wagner et al. 2006; Xie et al. 2009; Koval et al. 2010; Wagner et al. 2011), and hypokalaemia, in which decreased outward K+ currents play the major role (Sato et al. 2010). Moreover, the degree of hypokalaemia used in this study is clinically relevant to values associated with EAD-mediated arrhythmias in patients (Osadchii, 2010). These findings have therapeutic implications, namely that novel agents which alter the voltage dependence and/or kinetics of I Ca,L could be developed to suppress EAD-mediated arrhythmias without adversely depressing excitation–contraction coupling. Sensitivity of EADs to I Ca,L properties In the present study, we analysed the effects of small shifts in the V 1/2,act and V 1/2,inact on EAD formation, 6089 as a convenient way to suppress I Ca,L reactivation in the window voltage region. However, I Ca,L reactivation causing EADs during AP repolarization is not solely dependent on V 1/2,act and V 1/2,inact , but instead is a complex function of many I Ca,L parameters juxtaposed with other factors, such as K+ currents influencing the speed of repolarization (Tran et al. 2009). With respect to I Ca,L , other parameters in addition to V 1/2,act and V 1/2,inact influence I Ca,L reactivation during repolarization, including the maximal L-type Ca2+ current conductance; the steepness (slope factor) of voltage-dependent activation and inactivation; the time constants of activation, inactivation (both Ca2+ dependent and voltage dependent), and recovery from inactivation; and the amplitude of the non-inactivating pedestal current. These factors can be combined into a single mathematical expression giving a necessary condition for EADs (Tran et al. 2009). In the present study, our modifications to the virtual I Ca,L assumed that all of these other factors remained constant, allowing us to examine the isolated effects of V 1/2,act and V 1/2,inact . Thus, a major advantage of the dynamic clamp approach is that the Figure 6. EADs induced by hypokalemia are steeply dependent on V1/2 of activation and inactivation. A, Under hypokalemic conditions (2.0 mM external K+ ), EADs (blue) were reconstituted in the presence of 20 mM nifedipine by injecting via the dynamic clamp a virtual ICa,L , with normal activation (blue) and inactivation (black) properties. Shifting the V1/2 of activation by +5 mV (black) to reduce the ICa,L window region reversed AP prolongation and abolished EADs. B, Similar effects were observed when V1/2 of inactivation was shifted by −5 mV to reduce the ICa,L window region. C, The +5 mV shift in the activation curve shortened and D, abolished EADs. E, The −5 mV shift in the inactivation curve also shortened APD90 and F, dramatically reduced EAD occurrence.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6090 R. V. Madhvani and others virtual I Ca,L can be programmed to simulate all of the effects that a candidate agent has on these I Ca,L properties, once they have been characterized experimentally. The appropriately reshaped virtual I Ca,L can then be evaluated to determine whether or not it suppresses EADs in real myocytes when introduced by the dynamic clamp technique. Limitations J Physiol 589.24 2009) such as α2 δ or β subunits. These accessory subunits are highly specific for the channel and are known to modulate its biophysical properties (Birnbaumer et al. 1998; Olcese et al. 1994; Platano et al. 2000). Thus, genetic interventions to tune the biophysical properties of I Ca,L by altering its subunit composition could be a promising approach to suppress EAD formation in a highly specific manner. References Although the dynamic clamp provides a technique to systematically dissect the role of key ionic currents such as I Ca,L in EAD formation, an important limitation is that a virtual I Ca,L does not trigger SR Ca2+ release, so it is not possible to directly test the effects of a modified virtual I Ca,L on the Cai transient or cell shortening. This limitation was partially offset by using an AP model incorporating detailed Ca2+ cycling dynamics which allowed us to make predictions about the Cai transient, validated by experimental recordings of the Cai transient in rabbit ventricular myocytes using a Ca2+ -sensitive fluorescent dye (Fig. S1). In fact, the lack of Ca2+ entering the cell during the virtual I Ca,L injection also provides an advantage, allowing us to distinguish the purely electrical effects of ionic currents from their biochemical consequences (Ca2+ -mediated signalling). Thus, these results demonstrate that EADs can occur in the absence of Ca2+ influx or SR Ca2+ release and can be a purely electrical phenomenon. For example, we computed the virtual Na+ –Ca2+ exchange current accompanying the virtual Cai transient in our model, but chose not to inject this current into the myocyte via the dynamic clamp, since our initial goal was to investigate whether I Ca,L alone was sufficient to reconstitute EADs without other Ca2+ -sensitive currents. Furthermore, through the dynamic clamp, the contributions of other currents to EAD formation can be analysed systematically, assuming they have been accurately formulated in the model. Therapeutic implications Our study demonstrates ‘proof-of-concept’ for a hybrid biological–computational approach designed to predict how modifications of ion channel properties affect cellular arrhythmogenic phenomena such as EADs. Thus, in this study, the dynamic clamp has highlighted potential therapeutic targets (i.e. I Ca,L properties such as V 1/2,act , V 1/2,inact ) for new drugs or genetic interventions to suppress EAD-mediated arrhythmias. Interstingly, whereas pharmacological agents may have off-target effects in both cardiac and non-cardiac tissue, a specific way to target I Ca,L could be to take advantage of the modulatory accessory subunits of the L-type Ca2+ channel (Birnbaumer et al. 1998; Davies et al. 2007; Dolphin, Anderson ME (2004). Calmodulin kinase and L-type calcium channels; a recipe for arrhythmias? Trends Cardiovasc Med 14, 152–161. Antoons G, Volders PG, Stankovicova T, Bito V, Stengl M, Vos MA & Sipido KR (2007b). Window Ca2+ current and its modulation by Ca2+ release in hypertrophied cardiac myocytes from dogs with chronic atrioventricular block. J Physiol 579, 147–160. Antoons G, Volders PG, Stankovicova T, Bito V, Stengl M, Vos MA & Sipido KR (2007a). Window Ca2+ current and its modulation by Ca2+ release in hypertrophied cardiac myocytes from dogs with chronic atrioventricular block. J Physiol 579, 147–160. Berecki G, Zegers JG, Bhuiyan ZA, Verkerk AO, Wilders R & Van Ginneken AC (2006). Long-QT syndrome-related sodium channel mutations probed by the dynamic action potential clamp technique. J Physiol 570, 237–250. Berecki G, Zegers JG, Verkerk AO, Bhuiyan ZA, de Jonge B, Veldkamp MW, Wilders R & Van Ginneken AC (2005). HERG channel (dys)function revealed by dynamic action potential clamp technique. Biophys J 88, 566–578. Berecki G, Zegers JG, Wilders R & Van Ginneken AC (2007). Cardiac channelopathies studied with the dynamic action potential-clamp technique. Methods Mol Biol 403, 233–250. Birnbaumer L, Qin N, Olcese R, Tareilus E, Platano D, Costantin J & Stefani E (1998). Structures and functions of calcium channel beta subunits. J Bioenerg Biomembr 30, 357–375. Davies A, Hendrich J, Van Minh AT, Wratten J, Douglas L & Dolphin AC (2007). Functional biology of the α2 δ subunits of voltage-gated calcium channels. Trends Pharmacol Sci 28, 220–228. Dolphin AC (2009). Calcium channel diversity: multiple roles of calcium channel subunits. Curr Opin Neurobiol 19, 237–244. Dorval AD, Christini DJ & White JA (2001). Real-time linux dynamic clamp: a fast and flexible way to construct virtual ion channels in living cells. Ann Biomed Eng 29, 897–907. Dzhura I, Wu Y, Colbran RJ, Balser JR & Anderson ME (2000). Calmodulin kinase determines calcium-dependent facilitation of L-type calcium channels. Nat Cell Biol 2, 173–177. Gaborit N, Le BS, Szuts V, Varro A, Escande D, Nattel S & Demolombe S (2007). Regional and tissue specific transcript signatures of ion channel genes in the non-diseased human heart. J Physiol 582, 675–693.  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 J Physiol 589.24 ICa,L window current properties set EAD susceptibility Huffaker R, Lamp ST, Weiss JN & Kogan B (2004). Intracellular calcium cycling, early afterdepolarizations, and reentry in simulated long QT syndrome. Heart Rhythm 1, 441–448. January CT & Riddle JM (1989). Early afterdepolarizations: mechanism of induction and block. A role for L-type Ca2+ current. Circ Res 64, 977–990. January CT, Riddle JM & Salata JJ (1988). A model for early afterdepolarizations: induction with the Ca2+ channel agonist Bay K 8644. Circ Res 62, 563–571. Koval OM, Guan X, Wu Y, Joiner ML, Gao Z, Chen B, Grumbach IM, Luczak ED, Colbran RJ, Song LS, Hund TJ, Mohler PJ & Anderson ME (2010). CaV 1.2 β-subunit coordinates CaMKII-triggered cardiomyocyte death and afterdepolarizations. Proc Natl Acad Sci U S A 107, 4996–5000. Lin RJ, Bettencourt J, Wha IJ, Christini DJ & Butera RJ (2010). Real-time experiment interface for biological control applications. Conf Proc IEEE Eng Med Biol Soc 2010, 4160–4163. Lin S, Wang Z & Fedida D (2001). Influence of permeating ions on Kv1.5 channel block by nifedipine. Am J Physiol Heart Circ Physiol 280, H1160–H1172. Luo CH & Rudy Y (1994). A dynamic model of the cardiac ventricular action potential. I. Simulations of ionic currents and concentration changes. Circ Res 74, 1071–1096. Mahajan A, Sato D, Shiferaw Y, Baher A, Xie LH, Peralta R, Olcese R, Garfinkel A, Qu Z & Weiss JN (2008a). Modifying L-type calcium current kinetics: consequences for cardiac excitation and arrhythmia dynamics. Biophys J 94, 411–423. Mahajan A, Shiferaw Y, Sato D, Baher A, Olcese R, Xie LH, Yang MJ, Chen PS, Restrepo JG, Karma A, Garfinkel A, Qu Z & Weiss JN (2008b). A rabbit ventricular action potential model replicating cardiac dynamics at rapid heart rates. Biophys J 94, 392–410. Maltsev VA, Sabbah HN, Higgins RS, Silverman N, Lesch M & Undrovinas AI (1998). Novel, ultraslow inactivating sodium current in human ventricular cardiomyocytes. Circulation 98, 2545–2552. Maruyama M, Lin SF, Xie Y, Chua SK, Joung B, Han S, Shinohara T, Shen MJ, Qu Z, Weiss JN & Chen PS (2011). Genesis of phase 3 early afterdepolarizations and triggered activity in acquired long-QT syndrome. Circ Arrhythm Electrophysiol 4, 103–111. Olcese R, Qin N, Schneider T, Neely A, Wei X, Stefani E & Birnbaumer L (1994). The amino terminus of a calcium channel β subunit sets rates of channel inactivation independently of the subunit’s effect on activation. Neuron 13, 1433–1438. Osadchii OE (2010). Mechanisms of hypokalemia-induced ventricular arrhythmogenicity. Fundam Clin Pharmacol 24, 547–559. Platano D, Qin N, Noceti F, Birnbaumer L, Stefani E & Olcese R (2000). Expression of the α2 δ subunit interferes with prepulse facilitation in cardiac L-type calcium channels. Biophys J 78, 2959–2972. Qi X, Yeh YH, Chartier D, Xiao L, Tsuji Y, Brundel BJ, Kodama I & Nattel S (2009). The calcium/calmodulin/kinase system and arrhythmogenic afterdepolarizations in bradycardia-related acquired long-QT syndrome. Circ Arrhythm Electrophysiol 2, 295–304. 6091 Sato D, Xie LH, Nguyen TP, Weiss JN & Qu Z (2010). Irregularly appearing early afterdepolarizations in cardiac myocytes: random fluctuations or dynamical chaos? Biophys J 99, 765–773. Sato D, Xie LH, Sovari AA, Tran DX, Morita N, Xie F, Karagueuzian H, Garfinkel A, Weiss JN & Qu Z (2009). Synchronization of chaotic early afterdepolarizations in the genesis of cardiac arrhythmias. Proc Natl Acad Sci U S A 106, 2983–2988. Szabo B, Sweidan R, Rajagopalan CV & Lazzara R (1994). Role of Na+ :Ca2+ exchange current in Cs+ -induced early afterdepolarizations in Purkinje fibers. J Cardiovasc Electrophysiol 5, 933–944. Tan RC & Joyner RW (1990). Electrotonic influences on action potentials from isolated ventricular cells. Circ Res 67, 1071–1081. Tran DX, Sato D, Yochelis A, Weiss JN, Garfinkel A & Qu Z (2009). Bifurcation and chaos in a model of cardiac early afterdepolarizations. Phys Rev Lett 102, 258103. Volders PG, Kulcsar A, Vos MA, Sipido KR, Wellens HJ, Lazzara R & Szabo B (1997). Similarities between early and delayed afterdepolarizations induced by isoproterenol in canine ventricular myocytes. Cardiovasc Res 34, 348–359. Wagner S, Dybkova N, Rasenack EC, Jacobshagen C, Fabritz L, Kirchhof P, Maier SK, Zhang T, Hasenfuss G, Brown JH, Bers DM & Maier LS (2006). Ca2+ /calmodulin-dependent protein kinase II regulates cardiac Na+ channels. J Clin Invest 116, 3127–3138. Wagner S, Ruff HM, Weber SL, Bellmann S, Sowa T, Schulte T, Grandi E, Bers DM, Backs J, Belardinelli L & Maier LS (2011). Reactive oxygen species-activated Ca/calmodulin kinase IIδ is required for late I Na augmentation leading to cellular Na and Ca overload. Circ Res 108, 555–565. Weiss JN, Garfinkel A, Karagueuzian HS, Chen PS & Qu Z (2010a). Early afterdepolarizations and cardiac arrhythmias. Heart Rhythm 7, 1891–1899. Weiss JN, Garfinkel A, Karagueuzian HS, Chen PS & Qu Z (2010b). Early afterdepolarizations and cardiac arrhythmias. Heart Rhythm 7, 1891–1899. Wilders R (2006). Dynamic clamp: a powerful tool in cardiac electrophysiology. J Physiol 576, 349–359. Xie LH, Chen F, Karagueuzian HS & Weiss JN (2009). Oxidative-stress-induced afterdepolarizations and calmodulin kinase II signaling. Circ Res 104, 79–86. Author contributions R.V.M. performed research. Y.X. and A.P. provided analytical tools. R.V.M. and R.O. analyzed data. R.V.M, A.G., Z.Q., J.N.W, and R.O. designed research. R.V.M, A.P., A.G., Z.Q., J.N.W and R.O. wrote the paper. All authors have approved the final version of the paper for publication. The authors have no disclosures. Acknowledgements We are grateful to David Cristini and Jonathan Bettencourt for providing expert support with RTXI installation and Maurizio  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 6092 R. V. Madhvani and others Carnesecchi for contributing analytical software. We thank the members of the Olcese, Weiss, Qu, Garfinkel, Karaguezian and Chen laboratories for constructive discussions during the development of the project. This work was supported by research grants NIH/NHLBI P01HL078931 and R01 HL103662 (to J Physiol 589.24 J.W.), NIH/NIGMS R01GM082289 (to R.O.), American Heart Association Predoctoral Fellowship (W.S.A.) 10PRE3290025 to R.M. and American Heart Association Postdoctoral Fellowship (W.S.A.) 11POST7140046 to A.P., and the Laubisch and Kawata endowments (to J.W.).  C 2011 The Authors. Journal compilation  C 2011 The Physiological Society Downloaded from J Physiol (jp.physoc.org) at California Digital Library on March 12, 2014 View publication stats