Evaluating State Dependence and Subtype Selectivity of Calcium Channel Modulators in Automated Electrophysiology Assays

Abstract

Voltage-gated Ca²⁺ channels play essential roles in control of neurosecretion and muscle contraction. The pharmacological significance of Ca_v channels stem from their identification as the molecular targets of calcium blockers used in the treatment of cardiovascular diseases, such as hypertension, angina, and arrhythmia, and neurologic diseases, such as pain and seizure. It has been proposed that state-dependent Ca_v inhibitors, that is, those that preferentially bind to channels in open or inactivated states, may improve the therapeutic window over relatively state-independent Ca_v inhibitors. High-throughput fluorescent-based functional assays have been useful in screening chemical libraries to identify Ca_v inhibitors. However, hit confirmation, mechanism of action, and subtype selectivity are better suited to automated patch clamp assays that have sufficient capacity to handle the volume of compounds identified during screening, even of modest sized libraries (≤500,000 compounds), and the flexible voltage control that allows evaluation of state-dependent drug blocks. IonWorks™ Barracuda (IWB), the newest generation of IonWorks instruments, provides the opportunity to accelerate the Ca_v drug discovery studies in an automated patch clamp platform in 384-well format capable of medium throughput screening and profiling studies. We have validated hCa_v1.2, hCa_v2.1, hCa_v2.2, and hCa_v3.2 channels assays on the IWB platform (population patch clamp mode) and demonstrated that the biophysical characteristics of the channels (activation, inactivation, and steady-state inactivation) obtained with the IWB system are consistent with known subtype-specific characteristics. Using standard reference compounds (nifedipine, BAY K8644, verapamil, mibefradil, and pimozide), we demonstrated subtype-selective and state- and use-dependent characteristics of drug–channel interactions. Here we describe the design and validation of novel robust high-throughput Ca_v channel assays on the IWB platform. The assays can be used to screen focused compound libraries for state-dependent Ca_v channel antagonists, to prioritize compounds for potency or to counterscreen for Ca_v subtype selectivity.

Introduction

Voltage-gated Ca²⁺–selective channels play essential roles in the control of neurosecretion and muscle contraction. The channels have been categorized as L-, N-, P/Q-, R-, or T-type based on functional and pharmacological properties. Sequence similarity has identified three subfamilies Ca_v1, Ca_v2, and Ca_v3, where Ca_v1 and Ca_v3 correspond to L- and T-type, respectively. Ca_v2 encompasses N-, R-, and P/Q types. Each subtype contains a pore-forming α₁-subunit accompanied by auxiliary subunits (e.g., α₂δ, β_1–4, and γ) that modify gating, subcellular trafficking, and drug binding characteristics.¹

Development of drugs that act as calcium channel blockers has been successful for L-type channels in cardiovascular indications, for example, hypertension, heart failure, and atrial fibrillation. The success in this area has been due largely to state-dependent and use-dependent characteristics of the drug–channel interaction such that the inhibitory effect is most pronounced at partially depolarized membrane potential, thus preserving normal function.²

Another area of calcium channel drug development has focused on neurologic disorders, including pain and epilepsy. Cav2.2 (N-type) has been identified as a valid pain target that can be inhibited by the selective blocker ziconotide to relieve chronic pain.³ However, recent development has centered on identifying state-dependent Ca_v2.2 inhibitors, which preferentially bind to channels in open or inactivated states that may improve the therapeutic window over relatively state-independent Ca_v2.2 inhibitors, such as ziconotide.⁴ Antiepileptic calcium channel inhibitors are exemplified by ethosuximide, broadly selective blockers of T-type (Cav3.x) channels.⁵ Another example is gabapentin, which inhibits Cav2.x trafficking through interaction with the α₂δ₁ auxiliary subunit and is approved for treatment of both seizures and pain.⁶

State- and use-dependent characteristics, and subtype selectivity of voltage-gated channels can be measured in a straightforward manner using patch clamp methods. Automated, high-throughput platforms are preferred in screening and profiling studies, but until recently, automated patch clamp instruments lacked the required levels of throughput and flexibility. Systems such as IonWorks™ Barracuda (IWB) that operate in a 384-well format can evaluate small chemical libraries in the range of 20,000–100,000 compounds. In the present study, we show that state dependence and subtype selectivity can be readily assessed in IWB assays in recombinant cell lines that stably express L-, N-, P/Q, or T-type calcium channels.

Materials and Methods

Cell Lines

Stable cell lines expressing (under tetracycline induction) human Ca_v1.2/β₂/α₂δ₁ (Chinese hamster ovary [CHO]; CACNA1C/CACNB2/CACNA2D1), Ca_v2.1/β₄/α₂δ₁ (CHO; CACNA1A/CACNB4/CACNA2D1), Ca_v2.2/β₃/α₂δ₁ (CHO; CACNA1B/CACNB2/CACNA2D1), and Ca_v3.2 (HEK293; CACNA1H) ion channels were constructed as described previously.⁷ CHO cells were maintained in Ham's F-12 CHO media supplemented with 10% fetal bovine serum, 100 U/mL of penicillin G sodium, 100 μg/mL of streptomycin sulfate, and the appropriate selection antibiotics. HEK293 cells were maintained in the Dulbecco's modified Eagle's medium/nutrient mixture F-12 (DMEM/F-12) supplemented with 10% fetal bovine serum, 100 U/mL of penicillin G sodium, 100 μg/mL of streptomycin sulfate, and 500 μg/mL of G418 as the selection antibiotic. For experiments, the cells were passed in a medium lacking selection antibiotics 2–4 days. Expression was induced with tetracycline 16–14 h before recording. Verapamil at 3 μM was included in the induction medium to avoid Ca²⁺ overload toxicity. Cell density was ∼50%–70% confluent at the time of harvest; two 150-mm plates (∼1.2×10⁷ cells) were used per population patch clamp (PPC) experiment.

Cells were harvested by washing twice with 15–20 mL of Hank's Balanced Salt Solution (HBSS) and treatment with 5 mL of Accutase™ (Innovative Cell Technologies, San Diego, CA) solution for 20 (CHO cells) or 60 (HEK293 cells) minutes. Cells were resuspended in a 50-mL conical tube with addition of 10 mL of HBSS and triturated with a serological pipette to resuspend the cells and break up cell clusters. Cells were pelleted at 500 g for 2.5 min, the supernatant was removed, and the cell pellet was resuspended in 10 mL of HBSS. The cell suspension was centrifuged again at 500 g for 2.5 min and the supernatant removed. Finally, the cell pellet was resuspended in 5 mL of HEPES-buffered physiological saline (HB-PS).

Solutions and Electrophysiological Procedures

Chemicals used in a solution preparation were purchased from Sigma-Aldrich (St. Louis, MO) and were of ACS reagent grade purity or higher. Stock solutions of test articles were prepared in dimethyl sulfoxide (DMSO) and stored frozen. Each test article formulation was sonicated (Model 2510/5510; Branson Ultrasonics, Danbury, CT) at ambient room temperature for 20 min to facilitate dissolution. Test article concentrations were prepared fresh daily by diluting stock solutions into extracellular solutions (HB-PS buffer). The solution composition was 137 mM NaCl, 4 mM KCl, 7 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM glucose, pH adjusted to 7.4 with NaOH. All test and control solutions contained 0.3% DMSO and 0.05% F-127. The test article formulations were prepared in 384-well compound plates using an automated liquid handling system (Cyclone, Caliper). The internal HEPES-buffered solution consisted of 90 mM CsF, 50 mM CsCl, 2 mM MgCl₂, 0.5 mM EGTA, and 10 mM HEPES, pH 7.2 adjusted with CsOH. A stock solution of amphotericin B was prepared in DMSO (30 mg/mL) and added to the solution at a final concentration of 100 μg/mL. An extracellular buffer was loaded into the PPC plate wells (11 μL/well) and a cell suspension was added into the wells (9 μL/well). After establishment of a whole-cell configuration (a 10-min perforation), membrane currents were recorded by on-board patch clamp amplifiers in IWB. The data acquisition frequency was 5 kHz. Inward peak current amplitudes were measured. Under these conditions, each assay was completed in 45 min, and 5–10 experiments could be conducted each day.

Data Analysis

Initial data acquisition and analyses were performed using the IWB system operation software (version 2.0.2; Molecular Devices Corporation, Union City, CA). Data were corrected for leak current. The decrease in current amplitude after test article application was used to calculate the percent block relative to control. Results for each test article concentration (n=3–4) were averaged, mean and standard deviation values were calculated and used to generate dose–response curves in XLfit add-in for Excel 2003 (Microsoft, Redmond, WA).

Drug effects were calculated as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} {\rm \%} {\rm Block} = \left[ \frac {(1- {\rm I} _{\rm TA})} {{\rm I_ {Baseline}}} \right] \times 100 {\rm \% } ,\end{align*} \end{document}

where I_Baseline and I_TA were the current amplitudes measured at baseline (before addition of a test article) and in the presence of a test article (or vehicle control), respectively.

If not specified, the data were corrected for run-down using the following equation: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \% {\rm Block}^{\prime} = 100 \% - \left[ (\% {\rm Block} - \% {\rm PC}) \times \frac {100 \%} {\% {\rm VC} - \% {\rm PC}} \right] ,\end{align*} \end{document}

where %VC and %PC were the mean values of the current inhibition with the vehicle and positive controls, respectively; inhibition by saturating concentrations of nifedipine (Ca_v1.2) or mibefradil (Ca_v2.1, Ca_v2.2, Ca_v3.2) was used as %PC value.

The concentration–response data for inhibitors were fit to an equation of the following form: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} {\rm \%} {\rm Block} = \frac {100 {\rm \%}} {1 + ([{\rm Test} ] / {\rm IC} _ {50}) ^N} ,\end{align*} \end{document}

where [Test] was the concentration of test article, IC₅₀ was the concentration of the test article producing half-maximal inhibition, N was the Hill coefficient, and %Block was the mean value of percentage of the current inhibited at each concentration of a test article. Nonlinear least square fits were solved with the XLfit add-in for Excel 2003 (Microsoft).

Acceptance Criteria

Individual well data were filtered according to the criteria listed in Table 1 and the experiments were accepted based on the criteria listed in Table 2.

Table 1.

Acceptance Criteria for Wells

Parameter	Acceptance criterion
R_SEAL (baseline)^a	>300 MΩ
Current amplitude (baseline)	>0.2 nA
R_SEAL stability	<50% decrease
Current stability	<50% decrease in mean of vehicle control
Voltage clamp quality	Visual inspection

^aTypical R_SEAL values ranged between 500 and 1,000 MΩ.

Table 2.

Acceptance Criteria for 384-Well Population Patch Clamp Plate

Parameter	Acceptance criterion
Z′ factor	≥0.4
SW	≥2
Success rate	>80% accepted wells per PPC plate
IC₅₀ for reference inhibitors	≤0.5 log from historical mean

SW, signal window; PPC, population patch clamp.

The Z′ factor and signal window (SW) in each experiment were calculated as follows: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} &\qquad\qquad\qquad\qquad Z^{\prime} = 1 - \frac{3 \times {\rm SDVC} + 3 \times {\rm SDPC}} {{\rm ABS} ({\rm Mean VC - Mean PC})} , \\& {\rm SW} = \frac {({\rm MeanPC} - 3 \times{\rm SDPC}) - ({\rm MeanVC} + 3 \times {\rm SDVC})} {{\rm SDVC}}, \end{align*} \end{document}

where MeanVC and SDVC are the mean and standard deviation values for a vehicle control, and MeanPC and SDPC are the mean and standard deviation values for a positive control.

Results

We developed cell lines that express recombinant voltage-gated Ca²⁺ channels belonging to the L-, N-, P/Q-, and T-type classes: hCa_v1.2 (α₁C/β₂/α₂δ₁), hCa_v2.1 (α₁A/β₄/α₂δ₁), hCa_v2.2 (α₁B/β₃/α₂δ₁), and hCa_v3.2 (α₁H), respectively; where α1 is the pore-forming subunit, and β_1–4 and α₂δ₁ are auxiliary subunits. Based on the relatively depolarized voltage range required to activate the channels, Ca_v1.2, Ca_v2.1, and Ca_v2.2 are high-voltage-activated channels, whereas Ca_v3.2 is activated at more negative membrane potentials and classified as a low-voltage-activated channel. Figure 1A illustrates representative families of current traces for each channel recorded in IWB. Currents were evoked by voltage pulses that span the range from −80 to +60 mV in 10 mV increments. The current–voltage relationships (Fig. 1B) show the expected separation of voltage ranges between the high- and low-voltage-activated channels. Thus, hCa_v1.2, hCa_v2.2, and hCa_v2.1 show peak responses at positive membrane potentials, whereas hCa_v3.2 peaks at about −30 mV. All four channel subtypes display voltage-dependent inactivation during prolonged depolarization (Fig. 1A), as indicated by the return to baseline current. The voltage dependence of inactivation is plotted in Figure 1C. Steady-state inactivation curves fitted using the Boltzmann equation with half-inactivation (h_0.5) values were −65.8, −31.6, −27.4, and −26.7 mV for Ca_v3.2, Ca_v2.2, Ca_v2.1, and Ca_v1.2, respectively. Thus, as compared with the high-voltage-activated subtypes, inactivation of Ca_v3.2 was shifted nearly 40 mV toward more negative potentials. These results confirm that patch clamp recordings in the IWB platform can recapitulate the known voltage-dependent gating characteristics of these Ca²⁺ channel subtypes. Based on these results, we selected conditioning pulses that allow evaluation of state-dependent inhibition and test pulse potentials that would be expected to minimize the effect of uncompensated series resistance error.

Fig. 1.

Electrophysiological characteristics of Ca_v channels in IonWorks™ Barracuda. (A, B) Voltage-dependent activation of channels. Calcium current families, where 7 mM Ca²⁺ is the charge carrier: Ca_v1.2 (i), Ca_v2.1 (ii), Ca_v2.2 (iii), and Ca_v3.2 (iv). Currents were elicited by applying test pulses from −80 to +60 mV in 10 mV increments; the holding potential of −80 mV for Ca_v1.2 and −90 mV for other channels. In case of Ca_v3.2, a 500-ms conditioning voltage prepulse to −120 mV preceded each test pulse. The averaged current–voltage relationship is shown in (B) (mean±SE; n=16). Superimposition of steady-state inactivation curves of Ca_v channels is illustrated in (C). The currents were elicited by 400 ms test pulses, after 1-s, the conditioning prepulses (CPs) ranged from −120 to 50 mV (10 mV increments); the test pulse potentials were 10, 10, 30, and −30 mV for Ca_v1.2, Ca_v2.1, Ca_v2.2, and Ca_v3.2 channels, respectively.

The channel subtypes also are distinguished on the basis of their pharmacological properties. The pore-forming α₁ subunits confer selective sensitivity to most of the known activators and inhibitors (with the exception of gabapentin and pregabalin that bind the auxiliary α₂δ subunit). In particular, Ca_v1.2 and other L-type channels are sensitive to dihydropyridine activators (e.g., BAY K8644) and inhibitors (e.g., nifedipine). Other high-voltage-activated and all low-voltage-activated Ca²⁺ channels have low sensitivity to dihydropyridines.⁸ Figure 2 illustrates Ca_v1.2 versus Ca_v2.2 differences in dihydropyridine sensitivity. Current traces recorded pre- (vehicle; baseline) and postcompound addition show that Ca_v1.2 currents (Fig. 2A-i) are completely blocked by 1 μM nifedipine and potentiated by 0.3 μM (S)-(−)-BAY K8644. By contrast, peak Ca_v2.2 (Fig. 2A-ii) current amplitude were unaffected by 1 μM nifedipine (compared with vehicle control records). Similarly, Ca_v2.1 and Ca_v3.2 channels were insensitive to nifedipine at concentrations up to 3 μM (not shown). Conversely, BAY K8644, a dihydropyridine compound that potentiates the L-type channel current, increased the Ca_v1.2 channel at low nanomolar concentrations (EC₅₀=6 nM), but produced nonspecific inhibition of all four channels at concentrations ≥10 μM. Nifedipine block of Ca_v1.2 and BAY K8644 block of Ca_v2.2 were concentration dependent (Fig. 2B) with apparent IC₅₀ values of 23 nM and 13.6 μM, respectively.

Fig. 2.

Selectivity of pharmacological responses of Ca_v subtypes to dihydropyridine compounds. (A) Representative current traces illustrating effects of nifedipine and BAY K8644 on Ca_v1.2 (i) and Ca_v2.2 (ii) channels. The currents were elicited with test pulses to 10 mV (Ca_v1.2) and 30 mV (Ca_v2.2); the holding potential was −90 mV with 60-s CP to −50 mV. (B) Dose–response curves of the compounds effect on Ca_v1.2 (i) and Ca_v2.2 (ii) channels. Data presented as mean±SD; n=4. Data were fit to the Hill equation with coefficient in the range 0.6–2.0. Nifedipine inhibited Ca_v1.2 channels (IC₅₀=0.016 μM) and had no effects on Ca_v2.2 channels. BAY K8644 produced twofold potentiation of Ca_v1.2 channels (EC₅₀=0.006 μM), but inhibited Ca_v2.2 channel (IC₅₀=14.3 μM).

To evaluate the pharmacological responses of Ca_v channels in the IWB platform, we tested a set of well-characterized reference inhibitors (diltiazem, nifedipine, verapamil, mibefradil, and pimozide) for use- and/or voltage-dependent inhibition. The assays were optimized by testing several different voltage protocols to obtain the best balance between sensitivity and success rate. Examples of results obtained with two different voltage protocols are illustrated in Figure 3. A simple two-test pulse protocol (Fig. 3A) demonstrated the use dependence of Ca_v1.2 inhibition with verapamil; the inhibitory potency detected by the second test pulse (TP2) was 3.5-times greater than that produced during the first test pulse (TP1). IC₅₀ values were 35.3 and 9.1 μM, respectively, for TP1 and TP2. Similar results were obtained for diltiazem block of Ca_v1.2 and mibefradil block of Ca_v2.2 and Ca_v3.2 channels (not illustrated). In contrast, inhibition of Ca_v1.2 by nifedipine showed no use dependence in the two-test pulse protocol; the IC₅₀ values were 0.59 and 0.57 μM, respectively, for the TP1 and TP2.

Fig. 3.

Voltage and use dependence of Ca_v1.2 block by verapamil. (A) Examples of voltage protocols used for detection of use-dependent (two-pulse protocol; i) and voltage-dependent (preconditioning protocol; ii) effects of verapamil. (B) Representative current traces for the two-pulse (i) and preconditioning (ii) voltage protocols. (C) Dose–response curves for Ca_v1.2 inhibition by verapamil in the two-pulse (i) and preconditioning (ii) protocols.

Because continuous (5–10 min) clamping of Ca_v1.2, Ca_v2.1, and Ca_v2.2 cells at holding potentials of −50 or −40 mV, as commonly done in native cells under voltage clamp,⁹ resulted in extensive run down of the currents, we used conditioning depolarizing prepulses to evaluate voltage-dependent inhibition in the IWB platform (Fig. 3B). In this voltage protocol, Ca_v1.2 cells were held at −80 mV before stimulation. Test pulses were preceded by 50-s conditioning steps to −40 mV. The preconditioning enhanced the potency of Ca_v1.2 inhibition compared with the two-test pulse protocol described above. IC₅₀ values for the preconditioning protocol were 1.89, 0.013, and 2.43 μM, respectively, for verapamil, nifedipine, and diltiazem.

As an extension of the two-pulse voltage protocol, we evaluated use-dependent block augmentation by using stimulus patterns composed of repetitive test pulses. Figure 4 illustrates an experiment in which Ca_v2.2 channels were stimulated by repetitive 400-ms test pulses to +20 mV from a holding potential −90 mV at frequency 0.1 Hz (30 test pulse train); mibefradil was added 5 min before stimulation by the TP1. Dose–response curves for the first (TP1) and last (TP30) test pulses were generated. Inhibition with mibefradil obtained at TP30 was about seven times more potent compared with the inhibition registered at TP1; IC₅₀ values were 4.8 and 32.7 μM, respectively.

Fig. 4.

Use-dependent block of Ca_v2.2 by mibefradil. (A) Time course of Ca_v2.2 inhibition with mibefradil during repetitive stimulation. The currents were elicited with test pulses to +20 mV from a holding potential −90 mV with 0.1 Hz frequency (30 stimulations in total); data were run down, corrected using time-matching vehicle control and presented as mean±SD (n=4 per concentration). (B) Dose–response curves for the first (TP1) and last (TP30) stimulation test pulses (mean±SD). Mibefradil tonic block IC₅₀=32.7 μM, use-dependent block IC₅₀=4.8 μM.

To utilize the high-throughput capabilities and increased stimulus flexibility in the IWB system, and to obtain more accurate prediction of both use- and voltage-dependent effects, we validated a multiple-mode voltage protocol that could be used with Ca_v1.2, Ca_v2.1, Ca_v2.2, and Ca_v3.2 channels. The voltage waveform of the protocol is shown in Figure 5 and parameters specific for each Ca_v channel are provided in Table 3. IC₅₀ values for reference compounds are presented in Table 4. The stimulus pattern consists of a combination of the two-pulse and use-dependent repetitive pulse patterns. As shown in the Ca_v2.2 example (Fig. 5), the first test pulse (TP1) consists of a brief depolarizing stimulus (+30 mV amplitude, 400 ms duration) preceded by a long, depolarizing conditioning pulse (−50 mV amplitude, 1 s duration) that would be expected to cause partial inactivation. TP1 is applied to both pre (baseline)- and post-compound application to assess the inactivation-sensitive component of block. The second test pulse (TP2) consists of the same depolarizing stimulus as TP1, preceded by a hyperpolarizing conditioning potential (−100 mV amplitude, 2-s duration). At baseline, TP2 evaluates maximum activation of channels from their resting state. After test compound application, TP2 can be used as a measure of tonic block of resting channels. An additional use-dependent block arising from repetitive activation of the channels is evaluated by TP3 to TP32, consisting of a train of brief depolarizing pulses (+30 mV amplitude, 400 ms duration, from −90 mV holding potential, delivered repetitively at 10-s intervals). It is noteworthy that TP3, the initial pulse in the train, is preceded by a 30-s conditioning interval at the holding potential (−90 mV) and provides an alternative index of tonic block of the resting state. Block augmentation (postcompound addition) beyond the level attained in the resting state (i.e., use-dependent block) is measured by comparing peak current amplitudes, TP3 versus TP32, normalized to the effects of repetitive stimulation in channels exposed to vehicle alone. Ca_v2.2 current traces elicited by the multiple mode assay (Fig. 5B) illustrate the blocking characteristics of mibefradil and verapamil. Mibefradil shows use-dependent, but not inactivation-dependent block augmentation. By contrast, verapamil shows both inactivation- and use-dependent block amounting to an approximately threefold increase in block compared with tonic levels at −90 mV preconditioning, as illustrated in Figure 6 that plots the concentration–response curves for each Ca_v channel. Table 4 presents IC₅₀ values for voltage- and use-dependent block in the multiple mode assay. In these experiments, calculation of the Z′ statistic for TP1, using reference antagonists (Table 5) gave values exceeding 0.5 in each case, indicative of a robust assay.¹⁰

Fig. 5.

Multiple mode voltage protocol for evaluation of voltage- and use-dependent block in Ca_v channels. (A) Optimized voltage protocol to detect use and voltage dependence of test compound interactions in Ca_v channels. (B) Representative Ca_v2.2 current traces elicited with the voltage protocols before (baseline) and after addition of vehicle, 10 μM mibefradil, and 30 μM verapamil.

Fig. 6.

Voltage- and use-dependent block of Ca_v channels by reference compounds recorded in the multiple mode protocol. Dose–response curves for verapamil, mibefradil, and pimozide, multiple mode assay protocol. The voltage protocol is shown in Figure 5; specific voltage parameters for each Ca_v channel are presented in Table 3. For each test compound and Ca_v channel target, three curves are superimposed: TP1 (circles)—the curve generated at the test pulse after the CP, TP3 (triangles)—the curve generated at the first test pulse from the holding potential, and TP32 (inverted triangles)—the curve generated at the last test pulse from the holding potential. IC₅₀ values for the curves are presented in Table 4.

Table 3.

Multiple Mode Voltage Protocol Parameters

Ion channel	Holding potential (mV)	CP potential (mV)	CP duration (s)	TP potential (TP1–TP32, mV)	TP duration (TP1–TP32, ms)	TP frequency (TP3–TP32, Hz)
Cav1.2	−90	−50	60	10	250	0.1
Cav2.1	−90	−50	60	10	400	0.1
Cav2.2	−90	−50	60	30	400	0.1
Cav3.2	−90	−75	60	−30	200	1

CP, conditioning prepulse; TP, test pulse.

Table 4.

Pharmacology Summary: Reference Compound Potencies

	Ca_v1.2			Ca_v2.1			Ca_v2.2			Ca_v3.2
Reference compound	Pre-pulse −50 mV	Pre-pulse −90 mV	0.1 Hz	Pre-pulse −50 mV	Pre-pulse −90 mV	0.1 Hz	Pre-pulse −50 mV	Pre-pulse −90 mV	0.1 Hz	Pre-pulse −75 mV	Pre-pulse −90 mV	1 Hz
Verapamil	6.5±1.5	25.3±5.2	5.3±1.1	5.0±1.1	22.1±4.7	12.4±3.6	13.1±2.7	46.0±8.7	17.8±3.1	14.3±3.3	21.2±3.8	12.7±1.8
Nifedipine	0.023±0.007	0.085±0.026	0.10±0.013	>3.0	>3.0	>3.0	>3.0	>3.0	>3.0	>3.0	>3.0	>3.0
BAY K8644	(0.006; 0.005)^a	ND	ND	5.4±0.9	24.8±7.3	10.0±2.6	(13.6, 14.3)	(54.7, 37.2)	(17.3, 16.1)	10.8±1.9	11.5±0.6	7.2±0.7
Pimozide	>30.0	27.8±9.8	1.9±0.6	>30.0	>30.0	>30.0	>30.0	>30.0	>30.0	8.9±3.6	7.2±1.2	5.5±1.3
Mibefradil	10.8±3.10	16.1±4.6	8.0±1.8	6.5±1.2	4.0±1.1	2.2±0.6	10.8±1.3	8.4±2.1	3.0±0.2	3.30±0.9	3.7±1.0	2.0±0.9

Data are mean±SD IC₅₀s (in μM) from three to four experiments; where only two experiments were run, data from each experiment are listed.

^aEC₅₀, BAY K8644 potentiated Ca_v1.2.

ND, not determined.

Table 5.

Assay Quality Parameters

	Ion channel
	Ca_v1.2			Ca_v2.1			Ca_v2.2			Ca_v3.2
Parameter	Pre-pulse−50 mV	Pre-pulse −90 mV	0.1 Hz (−90 mV)	Pre-pulse−50 mV	Pre-pulse −90 mV	0.1 Hz (−90 mV)	Pre-pulse −50 mV	Pre-pulse −90 mV	0.1 Hz (−90 mV)	Pre-pulse−75 mV	Pre-pulse −90 mV	1 Hz (−90 mV)
RD, %^a	43.9±1.4	N/A	N/A	26.1±7.6	N/A	N/A	36.4±4.5	N/A	N/A	28.8±12.6	N/A	N/A
Z′	0.55±0.09	0.56±0.05	0.40±0.06	0.57±0.04	0.61±0.06	0.46±0.05	0.69±0.03	0.64±0.09	0.42±0.09	0.65±0.11	0.72±0.03	0.64±0.06
SW	4.0±1.4	4.5±0.6	3.3±1.9	5.2±1.2	5.9±1.4	3.1±0.8	6.4±3.5	4.3±1.0	3.0±0.9	6.1±2.2	10.9±2.5	9.1±2.9

Reference compounds for calculating Z′ and SW were verapamil (Ca_v1.2, Ca_v2.1, Ca_v2.2) or mibefradil (Ca_v3.2).

^aRD was measured at TP1 in time-matched vehicle control wells.

RD, run down; TP1, test pulse 1.

Discussion

Voltage-gated calcium (Ca_v) channels activate as a function of membrane potential such that the probability of opening increases with membrane depolarization, but with prolonged depolarization the channels transit to a closed, inactivated state that can be both voltage and calcium dependent. Thus, changes in membrane potential populate different conformational states (closed, open, or inactivated) of the channel. It is well known that a drug binding to Ca_v channels often shows state dependence such that the affinity of antagonist binding changes depending upon the channel conformation.¹¹

The goal of this study was to develop patch clamp assays of recombinant calcium channels that take full advantage of the flexibility and throughput capabilities of the recently released IWB automated patch clamp system. We were able to optimize assays for state- and use-dependent blocks in four different human Ca_v subtypes using a standardized pulse protocol. It should be recognized that the standard protocol developed here provides a useful tool for evaluating Ca_v subtype selectivity and for identifying dynamic characteristics of compounds, but may not be ideal for screening every class of compounds. For instance, compounds that require long exposure time or show a tendency to adsorb to the apparatus may require lengthened preincubation times and multiple applications at the expense of decreasing the number of data points that can be acquired per day. Therefore, in screening programs that require a high volume of output and rapid turnaround, pilot studies that allow further optimization for the classes of compounds to be tested and for the desired characteristics of the actives may be necessary. Another major consideration in designing an appropriate protocol would be data reduction and analysis. Our standard protocol yields multiple endpoints that enable evaluation of state- and use-dependent block, but complicate the analysis and may not be essential for high-volume screening.

Our validation experiments yielded state-and use-dependent IC₅₀ values of reference compounds in all four Ca_v channels, as presented in Table 4. These results highlight the differences between the channel subtypes. For instance, the dihydropyridine compounds nifedipine and antagonist, and BAY K8644, a potentiator, were effective in these roles only in Ca_v1.2. Other compounds showed less subtype selectivity. Thus, mibefradil, a use-dependent blocker that was originally described as a selective T-type channel antagonist and later shown to be a mixed T/L-blocker,¹² was observed in our experiments to be equipotent in the use-dependent evaluation of Ca_v2.1, Ca_v2.2, and Ca_v3.2. Nonetheless, mibefradil's use-dependent block was fourfold more potent in Ca_v3.2 versus Ca_v1.2, in agreement with published reports of Ca_v3.2 selectivity over Ca_v1.2 reported by Martin et al. in manual patch clamp experiments.¹³ However, our Ca_v3.2 IC₅₀ value obtained with −90 mV prepulse (Table 4) was approximately sixfold higher than that reported by Martin et al. ¹³ The use dependence of mibefradil may account for this discrepancy. As is typical in manual patch clamp, the onset of block was evaluated by repetitive stimulation at 0.1 Hz until a steady-state effect was observed,¹³ allowing cumulative use-dependent block, whereas in the IWB assay, stimulation was performed only once after exposure to the test compound for a fixed interval. Nonetheless, in IWB, by stimulating with a brief high-frequency train, we were able to achieve an IC₅₀ value of 2.0 (Table 4, Ca_v3.2, 1 Hz stimulus frequency) similar to that observed in manual patch clamp.¹³ This comparison of manual and automated patch clamp highlights the importance of recognizing the limitations of automated platforms and developing procedures that accommodate differences in the blocking kinetics and state-dependent characteristics of test compounds, as noted by others previously in Ca_v2.2 assays.¹⁴

Pimozide also has been described as a T-type channel inhibitor¹⁵ and our evaluation of tonic block is consistent with this characterization. However, we found that pimozide's selectivity for Ca_v3.2 over Ca_v1.2 disappeared in the use-dependent assay, which gave IC₅₀ values of 4.6 and 1.6 μM, respectively, in the two channel subtypes. As expected from the literature,¹⁶ we found that verapamil was relatively nonselective. It showed a twofold to fivefold increase in potency for state-dependent block in the high-voltage-activated subtypes and state-independent block of the low-voltage-activated channel (Ca_v3.2).

Several studies of Ca_v channels in automated patch clamp have been published recently. Most notably, IonWorks Quattro (IWQ, the predecessor to IWB) has been used in Ca_v1.2 recording with results comparable to those described here. For instance, Morton and Main¹⁷ reported a verapamil IC₅₀ value of 3.3 μM (at a holding potential of −60 mV), similar to the 4.3 μM IC₅₀ value reported here (Table 4, prepulse −50 mV, holding potential, −90 mV). Interestingly, Cao et al. ¹⁸ also reported a verapamil IC₅₀ value of 4.0 μM, but at a holding potential of −90 mV, whereas we observed a marked decrease in potency (IC₅₀=24.2 μM, Table 4, prepulse and holding potential, −90 mV) under these conditions. This discrepancy may reflect the discontinuous voltage clamp characteristic of IWQ, which allows cells to be depolarized between voltage clamp recordings. Therefore, the continuous voltage clamp feature of IWB may be particularly important for characterizing the state and voltage dependence of Ca_v inhibitors.

Our Ca_v1.2 results (verapamil IC₅₀=4.2 μM and nifedipine IC₅₀=0.016 μM, Table 4) also can be compared with data obtained in giga-ohm seal automated instruments such as PatchXpress^® (Molecular Devices Corporation), where IC₅₀ values of 15 and 0.016 μM, respectively, were obtained for verapamil and nifedipine.¹⁹

In conclusion, we have demonstrated that the advanced capabilities of a high-throughput automated patch clamp, including flexible voltage stimulation patterns, continuous voltage clamp, and rapid throughput, provide a suitable platform to support quantitative drug evaluation of potency, state and use dependence, and subtype selectivity in the therapeutically important voltage-gated Ca²⁺ channel family.

Footnotes

Acknowledgments

We would like to thank Zhiqi Liu and Hung Lee (ChanTest Corporation) for their assistance in cell line development and IWB validation.

Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Catterall

, Perez-Reyes

, Snutch

, Striessnig

: International Union of Pharmacology. XLVIII. Nomenclature and structure-function relationships of voltage-gated calcium channels. Pharmacol Rev, 2005; 57:411–425.

Triggle

: Drug targets in the voltage-gated calcium channel family: why some are and some are not. Assay Drug Dev Technol, 2003; 1:719–733.

Snutch

: Targeting chronic and neuropathic pain: the N-type calcium channel comes of age. NeuroRx, 2005; 2:662–670.

McGivern

, McDonough

: Voltage-gated calcium channels as targets for the treatment of chronic pain. Curr Drug Targets CNS Neurol Disord, 2004; 3:457–478.

Gomora

, Daud

, Weiergräber

, Perez-Reyes

: Block of cloned human T-type calcium channels by succinimide antiepileptic drugs. Mol Pharmacol, 2001; 60:1121–1132.

Field

, Hughes

, Singh

: Further evidence for the role of the α₂δ subunit of voltage dependent calcium channels in models of neuropathic pain. Br J Pharmacol, 2000; 131:282–286.

Wible

, Kuryshev

, Smith

, Liu

, Brown

: An ion channel library for drug discovery and safety screening on automated platforms. Assay Drug Dev Technol, 2008; 6:765–780.

Doering

, Zamponi

: Molecular pharmacology of high voltage-activated calcium channels. J Bioenerg Biomembr, 2003; 35:491–505.

Sanguinetti

, Kass

: Voltage-dependent block of calcium channel current in the calf cardiac Purkinje fiber by dihydropyridine calcium channel antagonists. Circ Res, 1984; 55:336–348.

10.

Zhang

, Chung

, Oldenburg

: A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J Biomol Screen, 1999; 4:67–73.

11.

Hondeghem

, Katzung

: Antiarrhythmic agents: the modulated receptor mechanism of action of sodium and calcium channel-blocking drugs. Annu Rev Pharmacol Toxicol, 1984; 24:387–423.

12.

Mehrke

, Zong

, Flockerzi

, Hofmann

: The Ca²⁺-channel blocker Ro 40–5967 blocks differently T-type and L-type Ca²⁺ channels. J Pharmacol Exp Ther, 1994; 271:1483–1488.

13.

Martin

, Lee

, Cribbs

, Perez-Reyes

, Hanck

: Mibefradil block of cloned T-type calcium channels. J Pharmacol Exp Ther, 2000; 295:302–308.

14.

Swenson

, Niforatos

, Vortherms

: An automated electrophysiological assay for differentiating Cav2.2 inhibitors based on state dependence and kinetics. Assay Drug Dev Technol, 2012; 10:542–550.

15.

Santi

, Cayabyab

, Sutton

, et al.: Differential inhibition of T-type calcium channels by neuroleptics. J Neurosci, 2002; 22:396–403.

16.

Lacinová

: Voltage-dependent calcium channels. Gen Physiol Biophys, 2005; 24 Suppl 1:1–78.

17.

Morton

, Main

: Use of escin as a perforating agent on the IonWorks Quattro automated electrophysiology platform. J Biomol Screen, 2013; 18:128–134.

18.

Cao

, Lee

, Holmqvist

, Lin

, et al.: Cardiac ion channel safety profiling on the IonWorks Quattro automated patch clamp system. Assay Drug Dev Technol, 2010; 8:766–780.

19.

Balasubramanian

, Imredy

, Kim

, et al.: Optimization of Ca_v1.2 screening with an automated planar patch clamp platform. J Pharmacol Toxicol Methods, 2009; 59:62–72.