An Automated Electrophysiological Assay for Differentiating Ca V 2.2 Inhibitors Based on State Dependence and Kinetics

Abstract

Ca_V2.2 (N-type) calcium channels are key regulators of neurotransmission. Evidence from knockout animals and localization studies suggest that Ca_V2.2 channels play a critical role in nociceptive transmission. Additionally, ziconotide, a selective peptide inhibitor of Ca_V2.2 channels, is clinically used to treat refractory pain. However, the use of ziconotide is limited by its low therapeutic index, which is believed, at least in part, to be a consequence of ziconotide inhibiting Ca_V2.2 channels regardless of the channel state. Subsequent efforts have focused on the discovery of state-dependent inhibitors that preferentially bind to the inactivated state of Ca_V2.2 channels in order to achieve an improved safety profile relative to ziconotide. Much less attention has been paid to understanding the binding kinetics of these state-dependent inhibitors. Here, we describe a novel electrophysiology-based assay on an automated patch platform designed to differentiate Ca_V2.2 inhibitors based on their combined state dependence and kinetics. More specifically, this assay assesses inactivated state block, closed state block, and monitors the kinetics of recovery from block when channels move between states. Additionally, a use-dependent assay is described that uses a train of depolarizing pulses to drive channels to a similar level of inactivation for comparison. This use-dependent protocol also provides information on the kinetics of block development. Data are provided to show how these assays can be utilized to screen for kinetic diversity within and across chemical classes.

Introduction

Functional Ca_V2.2 (N-type) channels are made up of the pore-forming α1B subunit, co-expressed with the auxiliary α2δ and β subunits that modulate channel expression and function.¹ Biological, genetic, and pharmacological evidence support an important role for Ca_V2.2 channels in nociceptive transmission. First, these channels are localized on the presynaptic nerve terminals of nociceptive primary afferents that terminate in the superficial layers of the spinal dorsal horn and are responsible for triggering neurotransmitter release.^2
–4 Additionally, multiple strains of Ca_V2.2 knockout mice have decreased responses to noxious stimulation,⁵ and the Ca_V2.2 selective peptide inhibitor ziconotide (Prialt^®) has been shown to be effective in attenuating pain in both preclinical and clinical studies.^6
–8 However, the clinical application of ziconotide is limited by the requirement for intrathecal delivery as well as its relatively narrow therapeutic index.^7,9 For these reasons, research efforts have focused on discovering an orally available, small-molecule inhibitor of Ca_V2.2 channels with an improved safety window compared with ziconotide.¹⁰

Ca_Vs exist in at least three different conformational states (closed, open, and inactivated). Although ziconotide has a degree of state dependence at very hyperpolarized membrane potentials, it inhibits Ca_V2.2 channels regardless of the channel state when the potentials are within physiological ranges.^11,12 It has been suggested that a more state-dependent Ca_V2.2 inhibitor would have an improved therapeutic index relative to ziconotide.^13,14 Supporting this hypothesis, the state-dependent Ca_V2 inhibitor TROX-1 shows efficacy in preclinical models of pain with a satisfactory preclinical therapeutic window against cardiovascular and CNS side effects.¹⁵ A number of assays have been described for identifying Ca_V2.2 inhibitors in a high throughput format^16

–19 with some specifically designed to identify state-dependent inhibitors.^20,21 Additionally, a more moderate throughput electrophysiological assay has been recently described that can assess state dependence.²¹

Despite this focus on the discovery of state-dependent Ca_V2.2 inhibitors, little attention has been paid to understanding the underlying kinetics of channel block. There are several important considerations for state-dependent compounds that could have a profound impact on the overall efficacy and safety. For example, after neuronal excitation and channel inactivation, a state-dependent Ca_V2.2 inhibitor produces inactivated state block. How quickly does this inhibition diminish as the channel moves to the closed state on repolarization of the neuron? Similarly, how quickly does the inactivated state block develop on the induction of neuronal activity? The limited attention directed toward the kinetics of Ca_V2.2 inhibitors may partly stem from the lack of higher throughput assays for assessing this information. Here, we describe a novel automated electrophysiological assay that can be used to differentiate compounds based on both state dependence and the kinetics of channel recovery in the presence of the compound. We also describe a use-dependent assay, where inactivation is preferentially driven through trains of brief depolarization. This assay can be used in parallel with the state-dependent assay to compare compounds and to provide information on the kinetics of block development. These assays can ensure that a wide diversity of kinetic profiles are explored and are incorporated into structure-activity relationships within chemical classes. Additionally, novel tool Ca_V2.2 inhibitors may be indentified that will facilitate investigation of how kinetic and state-dependent profiles relate to efficacy and safety.

Methods

Ca_V2.2 Cell Line

The Ca_V2.2 cell line stably co-expressing rat recombinant Ca_V2.2 (201719) e[Δ18a, Δ24a, 31a, 37b], Ca_Vβ3, and α2δ1subunits was purchased from Dr. Diane Lipscombe (Brown University). Cells were maintained at 37°C and 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 1% GlutaMAX, 100 IU/mL penicillin, 100 mg/mL streptomycin, 10 μg/mL blasticidin, 50 μg/mL hygromycin B, and 50 μg/mL zeocin. All cell culture reagents were purchased from Invitrogen (Carlsbad, CA).

Cell Preparation

On the day of the experiment, cells that were plated in T175 cell culture flasks and grown to 60%–80% confluency were rinsed once in Dulbecco's phosphate buffered saline and dissociated with 7 mL Detachin (Genlantis, San Diego, CA) at 37°C for 5 min. Dissociated cells were then centrifuged in a Beckman GS-15 centrifuge (Brea, CA) at 1000 rpm for 2 min. at room temperature, the supernatant was removed, and the pellet was re-suspended in serum-free media (CHO-SFM-II; Invitrogen) supplemented with 25 mM Hepes (Sigma, St. Louis, MO) at a final cell density of 3×10⁶ cells/mL. The cells were added to the Sophion QPatch (Sophion Bioscience, Ballerup, Denmark) automated cell preparation station where the cells were stirred and maintained in a single cell suspension. After 10 min of cell recovery, experiments were run unattended for several hours.

QPatch-HT Electrophysiology

Whole-cell patch-clamp experiments were carried out on the QPatch-HT automated platform that performs 48 parallel and independent recordings on a disposable QPlate. Cells were automatically prepared by the QPatch and re-suspended in extracellular solution using an on-board mini centrifuge (Qfuge). Positioning of the cells on the QPlate measuring site was done by application of negative pressure. Gigaseal and whole-cell configurations were established using QPatch software (Sophion Biosciences). To evaluate inactivated state inhibition and recovery kinetics, an alternating voltage protocol was used to stabilize the current at the inactivated state. The holding membrane potential was alternated between −110 mV and −75 mV with a 50 ms step to 0 mV at a frequency of 0.067 Hz. The addition of a brief step to 0 mV after 3 s of recovery at the hyperpolarized membrane potential was added to further monitor recovery kinetics (Fig. 1A). Using this protocol, ∼50%–70% of the N-type channels are inactivated at the depolarized holding potential compared with −110 mV holding potential. The actual depolarized holding potential ranged between −80 and −75 mV so that the inactivation levels would be similar across passages and assay runs. To evaluate the closed state inhibition, the same protocol was used as just mentioned; however, the baseline holding potential remained at −110 mV throughout the assay. The extracellular solution contained (in mM) 87.5 CsCl, 40 (TEA)-Cl, 10 HEPES, 10 glucose, 5 BaCl₂, and 1 MgCl₂, pH 7.2, 310 mOsm. The intracellular solution contained (in mM) 112 CsCl, 27 CsF, 8.2 EGTA, 10 HEPES, and 2 NaCl. On the day of the experiment, 4 mM MgATP was added, and the pH was adjusted to 7.2 with CsOH, 290 mOsm. The addition of antagonists was set up using the QPatch Assay Software and delivered with an eight-pipette liquid handling robot. All compounds were prepared first as serial dilutions in dimethyl sulfoxide (DMSO) and then brought up in extracellular solution for final testing concentrations in the presence of 0.3% DMSO. All the compound test solutions were kept in glass inserts held in 96-well plates (BioTech Solutions, Vineland, NJ). DMSO (0.3%) controls were run in all experiments. DMSO control currents measured from the inactivating period exhibited from 1% run-up to 7% run-down; from the hyperpolarized period, from 1% to 6% run-up; and from the 2–3 s recovery period, from 2% run-up to 2% rundown. Due to this small differential change at the different measuring periods, data outside of the screening data from Figure 4 were corrected based on the run-up or run-down observed for DMSO controls, for the respective measuring period within the same run.

Fig. 1.

The two-voltage recovery protocol can be used to differentiate Ca_V2.2 inhibitors. (A) For the two-voltage recovery protocol, cells were alternated between a depolarized and a hyperpolarized membrane potential every 15 s. Cells were stepped to 0 mV for 50 ms to open Ca_V2.2 channels and elicit current from available channels. Current measurements were made following the inactivating depolarization (circle) to assess inactivated state block. Current measurements were also made 3 s (triangle) and 15 s (square) after returning to the hyperpolarized potential to monitor recovery from inactivated state block. This cycle was repeated continuously during an initial control period and into the drug addition period. (B) Example of current versus time plots showing peak currents at the three measurement time points in each cycle during the initial control period and during the addition of the compound. Examples illustrate compounds exhibiting fast, intermediate, or slow recovery kinetics. For the slowly recovering compound, multiple additions of the same compound concentration are shown in the current vs. time (IT) plot, as block development was also slow and continued to accumulate. For this slower compound, the raw traces and percent block calculations were taken from the end of the second compound application period where the block was getting closer to s.s. levels. For the fast and intermediate compounds, raw traces and percent block calculations were taken from the end of the first compound period, as it appeared that the s.s. block had been reached. Ca_V, voltage-gated calcium channel; s.s., steady state.

Compound plates were set up for 32 two-point titrations (control, concentration 1, and concentration 2). Each condition was run in duplicate (16 total conditions) with a request for a total n of 8 in the QPatch assay software. This typically resulted in a run of 4–5 QPlates, at ∼1 plate/h, with n's of 3–5 of useable data for each condition. As an example, in the screening mode, 16 compounds could be assessed at two concentrations over 4–5 h with n values of 3–5 for each concentration (∼30 data points/h or ∼32 compounds/day at two concentrations). Although low by high throughput screening (HTS) standards, this is an extremely high-content electrophysiological assay with greatly improved throughput over manual electrophysiology (mEP).

Data Analysis and Statistical Comparisons

Peak current amplitudes were measured using the Sophion QPatch software and exported to Microsoft Excel, where percent inhibition values were calculated. GraphPad Prism (GraphPad Software, San Diego, CA) was used for data plotting, data fitting, and statistical comparisons. Statistical comparisons and data fitting were performed as noted in the figure legends.

Results

Modified Two-Voltage Protocol for Identifying Compounds with Different Kinetic Profiles

In an effort to develop a protocol that provides informative kinetic information, we modified the two-voltage protocol for Ca_V2.2 channels first described in Vortherms et al.²¹ The Vortherms protocol alternated between a depolarized and hyperpolarized holding potential every 15 s. During the 15 s depolarization, a portion of Ca_V2.2 channels are inactivated, and compound inhibition can be measured under inactivating conditions. When the membrane potential is hyperpolarized, channels recover from inactivation and inactivated-state block. To better monitor this recovery from the block, we added a brief step to 0 mV after 3 s of recovery at the hyperpolarized membrane potential (Fig. 1A). By comparing the degree of Ca_V2.2 block under inactivating conditions (measured at circle), with that of the block after 3 s (triangle) and 15 s (square) of recovery at a hyperpolarized membrane potential, compounds can start being differentiated based on their recovery kinetics. Figure 1B shows examples of different compounds displaying fast, intermediate, and slow recovery kinetics. For the fast-recovering compound, there is 86% block under inactivated conditions, but much of this block is relieved within 3 s at the hyperpolarized potential (36% inhibition) without much additional recovery at 15 s (31% inhibition). The compound in the middle panel displays more intermediate recovery kinetics with 77% inhibition under inactivated conditions, partial recovery to 41% inhibition after 3 s, and further recovery to 21% block after 15 s. In sharp contrast to these first two examples, the compound in the lower panel of Figure 1B shows little, if any, recovery from the block during the 15 s at the hyperpolarized membrane potential (80% under inactivated conditions, 80% after 3 s, and 77% after 15 s).

Evaluation of Closed State Block Provides a More Complete Assessment of the Compound Inhibition Profile

For compounds with slower recovery kinetics, full recovery to closed state block levels may not be complete within the 15 s recovery period. To further explore this, we utilized a protocol that measures the Ca_V2.2 closed state block more directly (Fig. 2B) The value of this protocol is illustrated by the comparison of two compounds within the same chemical series that displayed sharply contrasting kinetic profiles (Fig. 2C). Compound A1 displayed very fast kinetics with a substantial recovery by 3 s (solid vs. open bar) using the two-voltage recovery protocol. The lack of any additional recovery at 15 s (hashed bar) suggests that recovery was complete by 3 s (i.e., fully relaxed to closed state, hyperpolarized levels). This was confirmed by assessing the block at the same concentration utilizing the closed state protocol (reverse hashed bar). In contrast, Compound A2 showed a similar block at all three measurement points in the two-voltage recovery protocol, suggesting very slow recovery kinetics with essentially no recovery from inactivated state block occurring up through 15 s. This profile would also be consistent with the compound having no actual state dependence. The closed state protocol can be used to distinguish between these two possibilities. In the closed state protocol, Compound A2 exhibited reduced blocking, indicating that it is indeed a state-dependent compound. Therefore, although these two compounds display a similar degree of state dependence (solid bars compared with reverse hashed bars for each), they differ dramatically in their recovery kinetics.

Fig. 2.

Follow up with the closed state protocol is required to assess state dependence for slow kinetic compounds. (A, B) The closed state protocol (B) was designed for a direct comparison with the two-voltage recovery protocol (A). The 3 s recovery step was retained, so the protocols only differed in the absence or presence of the 15 s inactivating period. (C) Examples of two compounds within the same chemical series that showed marked differences in kinetics but similar degrees of state dependence (solid bars compared with reverse hashed bars) as assessed by follow up with the closed state protocol. Solid bars, inactivated state block; open bars, 3 s recovery block; hashed bars, 15 s recovery block; reverse hashed bars, closed state block. Data for inactivated, 3 s recovery, and 15 s recovery were compared using a one-way repeated measures ANOVA with a post-hoc Tukey comparison for significance. Closed state data, which had been obtained from a different population of cells, were compared with other data using a one-way ANOVA with a Dunnet's comparison for significance. *P<0.05; **P<0.01. For the two-voltage recovery protocol, n=4 for both compounds; for the closed state protocol, n=8 for Compound A1, and n=5 for Compound A2. Compound A1 was assessed at 10 μM, and Compound A2 was assessed at 1 μM to achieve similar levels of inactivated state block for the direct comparison. ANOVA, analysis of variance.

Kinetic Profile of the State-Dependent Ca_V2 Inhibitor TROX-1

TROX-1 was recently reported as being a small-molecule state-dependent inhibitor of Ca_V2 channels with a preclinical therapeutic window for analgesic efficacy against cardiovascular and central nervous system side effects.¹⁵ Therefore, we assessed the kinetic and state-dependent profile of TROX-1 on Ca_V2.2 using our two-voltage recovery protocol (Fig. 3). At concentrations of both 3 μM and 10 μM, TROX-1 displayed intermediate to slow kinetics with little recovery by 3 s (open bar) but significant recovery by 15 s (hashed bar). Recovery at 15 s, however, was still not complete, as follow-up assessment in the closed state protocol revealed a further reduction in the block under closed state conditions (reverse hashed bar). As reported in Abbadie et al.,¹⁵ TROX-1 demonstrated a good degree of state dependence as illustrated from the difference between the inactivated state (solid bar) and closed state (reverse hashed bar) block at both concentrations. Although TROX-1 exhibits relatively slow kinetics with strong state dependence, it is unclear what overall profile provides the optimal safety window in vivo. For this reason, it is important to be able to screen for diversity within and across series to maximize the chances of success.

Fig. 3.

TROX-1 exhibits relatively strong state dependence with intermediate recovery kinetics for Ca_V2.2. Similar profiles are obtained at both 3 μM and 10 μM TROX-1. Little recovery was observed within the first 3 s. More significant recovery was observed by 15 s although still not complete for the closed state blocking levels. Solid bars, inactivated state block; open bars, 3 s recovery block; hashed bars, 15 s recovery block; reverse hashed bars, closed state block. The statistical comparison was the same as described in Figure 2. *Statistically different from inactivated state block; ^†statistically different from the closed state block. One symbol, P<0.05; two symbols, P<0.01; three symbols P<0.001. The 3 s and 15 s recovery measurements at both 3 μM (P<0.001) and 10 μM (P<0.05) were also significantly different but not indicated on the graph for purposes of clarity. For 3 and 10 μM TROX-1 in the two-voltage recovery protocol, n=7 and n=4, respectively. For 3 and 10 μM TROX-1 in the closed state protocol, n=7 and n=5, respectively. TROX-1, (3R)-5-(3-chloro-4-fluorophenyl)-3-methyl-3-(pyrimldin-5-ylmethyl)-1-(1H-1,2,4-triazol-3-yl)-1,3-dihydro-2H-indol-2-one.

Screening for Diversity: Triazine and Pyrimidine Ca_V2.2 Inhibitors

The triazine Compound T1 (Fig. 4A) was initially identified as a Ca_V2.2 inhibitor in a high throughput screen of an Abbott chemical library. Subsequent synthetic chemistry efforts resulted in a number of structurally related triazine and pyrimidine compounds with activity against Ca_V2.2 channels. The two-voltage recovery protocol was used to look at a subset of these compounds to identify members with differentiated state-dependent and/or kinetic properties Figure 4 gives examples of some triazines and pyrimidines found to display a range of blocking profiles. For the triazines, Compounds T1 and T2 displayed relatively fast recovery kinetics with a substantial recovery of the block at 3 s and little to no additional recovery at 15 s. Compound T3 also showed relatively fast kinetics with a substantial recovery from the block at 3 s and a further recovery after 15 s to essentially no block. Although structurally similar in many aspects, the pyrimidines exhibited different profiles. Compound P1 demonstrated more intermediate kinetics with little recovery at 3 s but a significant recovery at 15 s, while Compound P2 showed very little recovery from block even up to 15 s, –a profile consistent with either very slow recovery kinetics (>15 s) or very little state dependence.

Fig. 4.

Diversity of profiles observed for the structurally similar triazine and pyrimidine series using the two-voltage recovery protocol. (A) Compounds ranged from having fast recovery kinetics with varying degrees of state dependence (Compound T1–T3) to having more intermediate kinetics (Compound P1) to having very slow kinetics and/or very weak state dependence (Compound P2). Data were compared using a one-way repeated measures ANOVA with a post-hoc Tukey test. Concentrations that yielded inactivated state block in the 50%–75% range were chosen for comparison. About 20 μM Compound T1, n=6; 5 μM Compound T2, n=4; 3 μM Compound T3, n=5; 5 μM Compound P1, n=4; 10 μM Compound P2, n=4. Asterisks indicate statistical difference (*P<0.01, **P<0.001) from the inactivated state block. ⁺Statistically different (P<0.01) from the 3 s recovery block.

As demonstrated from the data in Figure 4, the two-voltage recovery protocol can be used as a first-pass screen to identify compounds with different state-dependent and/or kinetic profiles. Following the first-pass screen, compounds of interest can be further evaluated with follow-up concentration-response curves for the inactivated state, the two recovery periods, and the closed state, and, thus, provide a more thorough characterization of the state dependence and kinetics. Since the triazines displayed very fast recovery kinetics, in some cases even appearing to be fully recovered within 3 s (Fig. 4B, Compounds T1–T3), the first recovery measurement was shortened to 2 s in the follow-up concentration-response curves. Figure 5 shows the full concentration-response curves for three compounds with contrasting profiles. Compound T4, a triazine closely related to the Figure 4 triazines, displayed similarly fast recovery kinetics, and almost complete recovery to closed state block values within 2 s. Compound T4 also displayed a strong state dependence with an IC₅₀=5.9 μM for the inactivated state and an IC₅₀>30 μM for the closed state. The pyrimidine Compound P1 had similar state dependence (IC₅₀=6.3 μM for the inactivated state and IC₅₀>30 μM for the closed state) compared with Compound T4; however, it had a more intermediate kinetic profile with little recovery at 2 s and a more complete recovery by 15 s. Diversity was found within the pyrimidine chemical series, as Compound P2 showed only minimal, although statistically significant, state dependence (IC₅₀=4.6 μM for the inactivated state and 8.9 μM for the closed state). There was not enough separation between the inactivated and closed state inhibition curves to resolve the kinetics of Compound P2.

Fig. 5.

Follow-up full titrations for select compounds showing various kinetic/state-dependent profiles on Ca_V2.2. Compound T4 displayed strong state dependence (compare the inactivated state curve with the closed state curve) and very fast recovery kinetics with the block largely recovered within 2 s. Compound P1 exhibited more intermediate recovery kinetics with a similar degree of state dependence. In contrast, Compound P2 displayed very little if any state dependence. For clarity, statistics are only included in the figure legend. Inactivated, 2 s recovery, and 15 s recovery data at each concentration were compared using a one-way repeated measures ANOVA with a post-hoc Tukey comparison for significance. Closed state data, which has been obtained from a different population of cells, were compared with other data using a one-way ANOVA with Dunnet's comparison for significance. For Compound T4, the inactivated state block was different from the 2 s recovery block (1 μM*, 3 μM**, 10 μM**, 30 μM***), the 15 s recovery block (1 μM*, 3 μM***, 10 μM***, 30 μM***), and the closed state block (1 μM**, 3 μM**, 10 μM***, 30 μM***). The 2 s recovery block was different from the 15 s recovery block at 30 μM** and the closed state block at 30 μM**. For Compound P1, the inactivated state block was different from the 15 s recovery block (1 μM*, 3 μM***, 10 μM***, 30 μM***) and the closed state block (3 μM*, 10 μM***, 30 μM***). The 2 s recovery block was different from the 15 s recovery block (3 μM***, 10 μM***, 30 μM***) and the closed state block (10 μM***, 30 μM***). The 15 s recovery block was different from the closed state block at 30 μM**. For Compound P2, the closed state block was different from the inactivated state block (1 μM**, 3 μM*, and 10 μM*), the 2 s recovery block (10 μM**), and the 15 s recovery block (10 μM*). For Compound T4, P1, and P2 n=5–8 cells per data point for the two-voltage recovery protocol data and n=3–9 cells per data point for the closed state protocol. The data were fit using a four-parameter logistic equation with the minimum fixed at zero and the maximum fixed to 100% of the control. A shared hill slope was used within each data set except for Compound P2 for which the closed state data was fit independently, as the shared slope did not fit the data well. The IC₅₀ values for the inactivated state, 3 s recovery, 15 s recovery, and closed state were 5.9, 26, and >30 μM (49 μM extrapolated) for Compound T4; 6.3, 7.6, 22, and >30 μM (41 μM extrapolated) for Compound P1; and 4.6, 5.2, 6.4, and 8.9 μM for Compound P2. *P<0.05, **P<0.01, ***P<0.001.

Further Differentiation of Compounds Utilizing a Matched Use-Dependent Protocol

Compounds were further differentiated by employing a use-dependent protocol, which relies on trains of brief, channel-opening depolarizations to drive inactivation (Fig. 6A). The frequency in the use-dependent protocol was chosen in order to attain similar levels of Ca_V2.2 inactivation as had been achieved in the two-voltage recovery protocol, and, thus, allow for a more direct comparison of results between protocols. Although the degree of block attained between the two protocols was very similar for some compounds (Fig. 6B, TROX-1 and Compound P1, solid vs. open bars), other compounds showed striking differences. Both Compounds T2 and T4 showed substantially less block in the use-dependent protocol with Compound T4 showing essentially no inhibition.

Fig. 6.

Use-dependent protocol for further differentiation of compounds. (A) The use-dependent protocol was modeled after the two-voltage protocol but with a train of brief depolarizations to drive inactivation. (B) Comparison of Ca_V2.2 inactivated state inhibition by TROX-1 and triazine/pyrimidine compounds in the two-voltage (solid bars) and use-dependent (open bars) protocols. *P<0.05; ***P<0.001. (C, D) Plotting percent inhibition versus pulse number provides additional information on the development of inhibition during the 3 Hz train. For clarity, inhibition data are only shown for pulses 1 through 5 and then every 5th pulse thereafter. Compound concentrations in (B) were as labeled for TROX-1, 5 μM for Compound P1, and 10 μM for Compounds T2 and T4. Compound concentrations in (C) were 10 μM for TROX-1 and 5 μM for Compound P1. Inactivation levels were determined by comparing the current amplitude at the end of the inactivating step/train with the amplitude after the 15 s at the hyperpolarized potential during the last control sweep before drug addition. Inactivation levels were slightly higher for TROX-1 in the two-voltage protocol relative to the use-dependent protocol; 71%±4% versus 51%±3% (P<0.01) at 3 μM and 71%±6% versus 53%±4% (P=0.04) at 10 μM. Inactivation levels between protocols for the other compounds were statistically indistinguishable; 56%±5% versus 52%±4% for Compound P1, 54%±3% versus 57%±4% for Compound T2, and 69%±4% versus 62%±7% for Compound T4.

This use-dependent protocol also provides information on the development of the block following compound application, and can be monitored throughout the inactivating train. Examples of this are illustrated for TROX-1 (Fig. 6C) and Compound P1 (Fig. 6D). TROX-1 takes 30–40 pulses (∼10–13 s) to reach steady state (s.s.) block, while Compound P1 reaches s.s. block by ∼15 pulses (∼5 s). These data are derived from the first train in the presence of the compound, and the s.s. block achieved at the end of the compound application period is shown for comparison. This s.s. measurement was typically obtained after four to six trains. Trains were repeated once every 15 s in order to be comparable to the inactivating depolarizations in the two-voltage recovery assay.

Discussion

There has been a significant effort devoted toward the discovery and development of small-molecule Ca_V2.2 inhibitors with an improved therapeutic index compared with ziconotide (Prialt).¹⁰ These efforts have largely focused on finding state-dependent inhibitors with little consideration of kinetics. This is likely due, in part, to the lack of higher throughput assays for obtaining useful kinetic information. Here, we describe an automated electrophysiological assay for acquiring kinetic information in parallel with state dependence. This assay can be used to identify compounds with a range of state-dependent and kinetic profiles, thereby providing tools and potential clinical candidates for investigating the relevance of these properties to an improved therapeutic window. The utility of this assay was demonstrated by differentiating triazine/pyrimidine compounds based on their overall state-dependent and kinetic profiles. We also describe a follow-up use-dependent protocol that can be used to identify compounds which differentially block the channel when inactivation is driven through trains of brief strong depolarizations. This use-dependent protocol has the potential to identify compounds that are more reliant on open channel blocking mechanisms, and facilitates monitoring of the development of the block to provide additional information on kinetics.

A number of high-throughput nonelectrophysiological assays for identifying Ca_V2.2 inhibitors have been described,^16

–19 including ones that also have the ability to measure state dependence.^20,21 However, these assays do not have the ability to provide relevant kinetic information. Automated electrophysiological assays for Ca_V2.2 have also been described to look at either state dependence²¹ or use dependence.¹⁷ Although some kinetic information can potentially be gleaned from use-dependent assays by looking at the development of the block during the train, the overall potency measurements that result can be difficult to interpret. In use-dependent protocols, channels are continuously cycling through the different states, and the measured potency of a compound can be influenced by the overall stimulation frequency relative to the compound kinetics. In some cases, compounds could even be missed if a use-dependent protocol were the only assay being utilized (e.g., Compound T4 from Fig. 6B). These limitations make it difficult to apply findings from use-dependent assays more broadly.

The limitations of these existing assays highlight some of the advantages of the assay described in this article. First, in addition to state-dependent information, the two-voltage recovery protocol provides a measure of the time it takes for the channels, after hyperpolarization, to recover from inactivated state block levels and return to closed state block levels. This could be a particularly relevant consideration for Ca_V2.2 state-dependent inhibitors, as these channels cycle between closed, open, and inactivated states under in vivo conditions. Second, the kinetic measurements described here are made in the continuous presence of the compound. This protocol is more physiologically relevant than the kinetic measurements derived from compound washout rates. The kinetics derived from washout rates can appear artificially slow if the time it takes the compound to wash out from the cell and membrane is rate limiting or if a fast and direct continuous perfusion system is not utilized. Lastly, the use-dependent protocol described here is designed to mirror the two-voltage state-dependent protocol but with trains of depolarizing steps driving channel inactivation. This protocol provides a way to more directly compare the development of the block across compounds and, when used in parallel with the two-voltage recovery protocol, can be used to identify potentially interesting compounds such as those illustrated in Figure 6B. It is unclear exactly why there is such a dramatic decrease in the inhibition observed for Compounds T2 and T4 using the use-dependent protocol. It could be related to the faster recovery kinetics of these compounds preventing the accumulation of inhibition during trains. This seems unlikely, however, as there is a similar degree of cumulative channel inactivation in both protocols. An alternative explanation is that the use-dependent assay drives the channels into a different inactivated state that is less sensitive to the block by the triazines. Regardless of the root cause for the decrease in inhibition, the parallel use of the two-voltage and use-dependent protocols would enable these compounds to be singled out for potential follow up with mEP or even behavioral experiments.

The 3 s recovery pulse in the two-voltage recovery protocol facilitates the identification of compounds with fast recovery rates and allows their state dependence to be determined (i.e., Compounds A1 and T4). For these compounds, the degree of inhibition was similar at both recovery time points, implying that the block had relaxed back to closed state levels. This was confirmed using the follow-up closed state protocol. In contrast, very slow recovery kinetics can mask true state dependence in the two-voltage recovery assay. An example of this is Compound A2, where the inactivated, 3 s, and 15 s recovery measurements all showed a similar block. In cases such as this, the follow-up closed state protocol is required to distinguish slow kinetic compounds from ones with no state dependence. For Compound A2, the closed state protocol resulted in reduced block, thereby identifying this compound as a state-dependent inhibitor with very slow kinetics.

When using only one or two concentrations of compounds during compound screening with the two-voltage recovery protocol, it should be recognized that the related kinetics or state-dependence data can be influenced by where the selected concentrations fall in the concentration response curve. Ideally, direct comparisons should only be made between compounds showing similar levels of inactivated state block. In Figures 2 and 4, for example, comparisons were only made across compounds using concentrations that produced inactivated block ranging from 50% to 75%. This range was also selected, because it provided a more sensitive window for visualizing recovery kinetics. To help guide initial concentration selections, results from an IMR-32 FLIPR assay (described in Vortherms et al.²¹) were used. Additionally, we found that the degree of inactivation at a given membrane potential could vary slightly depending on cell passage. An advantage of the current assay is that the degree of inactivation can be monitored for each individual cell and across runs by comparing the current elicited from the hyperpolarizing membrane potential with the current elicited from the inactivating step. The depolarizing voltage can then be adjusted so that similar inactivation levels are used over time. Overall, the difference in the inactivation over passages was relatively small, and the depolarizing voltage was only varied between −75 and −80 mV to maintain the targeted inactivation levels. The use-dependent assay can also be designed to produce a similar degree of inactivation as the two-voltage protocol in order to make comparisons more relevant.

As with any assay, it is also important to consider potential limitations. The two-voltage recovery protocol inactivates the Ca_V2.2 channels using a 15 s depolarization followed by 15 s of recovery at a hyperpolarized membrane voltage. The potential exists to underestimate potencies if the block has not reached s.s. after the 15 s of inactivation. For TROX-1 and Compound P1, the data from the use-dependent protocol suggest that the equilibrium block is largely reached within the 15 s time frame (45 pulses). However, for some compounds, the equilibrium block was not achieved, as the block continued to accumulate during successive cycles. Additionally, right shifts in potency sometimes occur on automated platforms, which are believed to be the result of nonspecific binding to compound plates, pipettes, and/or recording wells.^22,23 These factors could contribute to the current data set, as there are right shifts relative to mEP. TROX-1 showed an approximate eightfold right shift relative to published data (48% block at 3 μM on the QPatch vs. a 0.4 μM IC₅₀ by mEP¹⁵). The triazines Compounds T3 and T4 were assessed by mEP and showed shifts of approximately fivefold (60% block at 3 μM on the QPatch vs. a 0.39 μM IC₅₀ by mEP for Compound T3 and 5.9 μM on the QPatch vs. a 1.1 μM IC₅₀ value by mEP for Compound T4). It should also be noted that these mEP protocols utilized a constant depolarizing potential for inducing inactivation. Thus, these mEP values may be related to the block involving slower inactivated states of the channel as has been suggested for varied sodium and calcium channel inhibitors.^24
–26 Regardless, the utility of the current assay is that it can differentiate compounds based on both kinetics and state dependence. Furthermore, the inactivated, recovery and closed state block data are all generated on the same instrument using the same or closely related voltage protocols. These data can then be readily compared, and compounds of interest with differential profiles can be prioritized for follow-up characterization in additional in vivo and in vitro assays.

There is a relatively poor understanding of the exact neuronal activities that drive pain and underlie potential Ca_V2 mechanism-based cardiovascular and CNS side effects. For this reason, it is difficult to predict what combination of state dependence and kinetics will provide the optimal therapeutic utility for a Ca_V2 inhibitor. This uncertainty highlights the need for exploring compounds with a range of kinetic profiles and of identifying tool compounds. From a drug discovery perspective, the assays presented here can be utilized at the multiple stages of a program. In the early stages, when initial leads are being selected, these assays can help assure mechanistic diversity in the compound classes being explored. This strategy should help maximize the chances of success when the ideal blocking mechanism/profile is unknown. During lead optimization, these assays can facilitate a more informed structure-activity relationship (SAR) within and across chemical classes and can continue to be used to identify compounds with differing profiles. Finally, these assays can be used to identify tool compounds for improving our understanding of how compound kinetics might influence in vivo profiles. As the field progresses and additional compounds with favourable safety margins are identified, it will be interesting to see whether they appear to be associated with certain kinetic and/or state-dependent profiles.

Footnotes

Acknowledgments

The authors would like to acknowledge the Abbott Electrophysiology group for helpful discussions concerning state dependence and kinetics and thank their chemistry colleagues for providing the compounds used in these studies.

Disclosure Statement

No competing financial interests exist.

Abbreviations

References

Catterall

. Structure and regulation of voltage-gated Ca2+ channels. Annu Rev Cell Dev Biol, 2000; 16:521–555.

Evans

, Nicol

, Vasko

. Differential regulation of evoked peptide release by voltage-sensitive calcium channels in rat sensory neurons. Brain Res, 1996; 712:265–273.

Gohil

, Bell

, Ramachandran

, Miljanich

. Neuroanatomical distribution of receptors for a novel voltage-sensitive calcium-channel antagonist, SNX-230 (omega-conopeptide MVIIC) Brain Res, 1994; 653:258–266.

Westenbroek

, Hoskins

, Catterall

. Localization of Ca2+ channel subtypes on rat spinal motor neurons, interneurons, and nerve terminals. J Neurosci, 1998; 18:6319–6330.

Saegusa

, Matsuda

, Tanabe

. Effects of ablation of N- and R-type Ca(2+) channels on pain transmission. Neurosci Res, 2002; 43:1–7.

Bowersox

, Gadbois

, Singh

, Pettus

, Wang

, Luther

. Selective N-type neuronal voltage-sensitive calcium channel blocker, SNX-111, produces spinal antinociception in rat models of acute, persistent and neuropathic pain. J Pharmacol Exp Ther, 1996; 279:1243–1249.

Miljanich

. Ziconotide: neuronal calcium channel blocker for treating severe chronic pain. Curr Med Chem, 2004; 11:3029–3040.

Wang

, Pettus

, Gao

, Phillips

, Scott Bowersox

. Effects of intrathecal administration of ziconotide, a selective neuronal N-type calcium channel blocker, on mechanical allodynia and heat hyperalgesia in a rat model of postoperative pain. Pain, 2000; 84:151–158.

Staats

, Yearwood

, Charapata

et al. Intrathecal ziconotide in the treatment of refractory pain in patients with cancer or AIDS: a randomized controlled trial. JAMA, 2004; 291:63–70.

10.

Yamamoto

, Takahara

. Recent updates of N-type calcium channel blockers with therapeutic potential for neuropathic pain and stroke. Curr Top Med Chem, 2009; 9:377–395.

11.

Feng

, Doering

, Winkfein

, Beedle

, Spafford

, Zamponi

. Determinants of inhibition of transiently expressed voltage-gated calcium channels by omega-conotoxins GVIA and MVIIA. J Biol Chem, 2003; 278:20171–20178.

12.

Stocker

, Nadasdi

, Aldrich

, Tsien

. Preferential interaction of omega-conotoxins with inactivated N-type Ca2+ channels. J Neurosci, 1997; 17:3002–3013.

13.

Snutch

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

14.

Winquist

, Pan

, Gribkoff

. Use-dependent blockade of Cav2.2 voltage-gated calcium channels for neuropathic pain. Biochem Pharmacol, 2005; 70:489–499.

15.

Abbadie

, McManus

, Sun

et al. Analgesic effects of a substituted N-triazole oxindole (TROX-1), a state-dependent, voltage-gated calcium channel 2 blocker. J Pharmacol Exp Ther, 2010; 334:545–555.

16.

Benjamin

, Pruthi

, Olanrewaju

et al. Pharmacological characterization of recombinant N-type calcium channel (Cav2.2) mediated calcium mobilization using FLIPR. Biochem Pharmacol, 2006; 72:770–782.

17.

Finley

, Lubin

, Neeper

et al. An integrated multiassay approach to the discovery of small-molecule N-type voltage-gated calcium channel antagonists. Assay Drug Dev Technol, 2010; 8:685–694.

18.

Lubin

, Reitz

, Todd

, Flores

, Qin

, Xin

. A nonadherent cell-based HTS assay for N-type calcium channel using calcium 3 dye. Assay Drug Dev Technol, 2006; 4:689–694.

19.

Zhang

, Kauffman

, Yagel

, Codd

. High-throughput screening for N-type calcium channel blockers using a scintillation proximity assay. J Biomol Screen, 2006; 11:672–677.

20.

Dai

, Haedo

, Warren

et al. A high-throughput assay for evaluating state dependence and subtype selectivity of Cav2 calcium channel inhibitors. Assay Drug Dev Technol, 2008; 6:195–212.

21.

Vortherms

, Swensen

, Niforatos

et al. Comparative analysis of inactivated-state block of N-type (Ca(v)2.2) calcium channels. Inflamm Res, 2011; 60:683–693.

22.

Dubin

, Nasser

, Rohrbacher

et al. Identifying modulators of hERG channel activity using the PatchXpress planar patch clamp. J Biomol Screen, 2005; 10:168–181.

23.

Sorota

, Zhang

, Margulis

, Tucker

, Priestley

. Characterization of a hERG screen using the IonWorks HT: comparison to a hERG rubidium efflux screen. Assay Drug Dev Technol, 2005; 3:47–57.

24.

Errington

, Stohr

, Heers

, Lees

. The investigational anticonvulsant lacosamide selectively enhances slow inactivation of voltage-gated sodium channels. Mol Pharmacol, 2008; 73:157–169.

25.

Hildebrand

, Smith

, Bladen

et al. A novel slow-inactivation-specific ion channel modulator attenuates neuropathic pain. Pain, 2011; 152:833–843.

26.

, Bean

. Inhibition of neuronal voltage-gated sodium channels by brilliant blue G. Mol Pharmacol, 2011; 80:247–257.