Development of a Rapid Throughput Assay for Identification of hNav1.7 Antagonist Using Unique Efficacious Sodium Channel Agonist, Antillatoxin

Voltage-gated sodium channels (VGSCs) are responsible for the generation of the action potential. Among nine classified VGSC subtypes (Nav1.1–Nav1.9), Nav1.7 is primarily expressed in the sensory neurons, contributing to the nociception transmission. Therefore Nav1.7 becomes a promising target for analgesic drug development. In this study, we compared the influence of an array of VGSC agonists including veratridine, BmK NT1, brevetoxin-2, deltamethrin and antillatoxin (ATX) on membrane depolarization which was detected by Fluorescence Imaging Plate Reader (FLIPR) membrane potential (FMP) blue dye. In HEK-293 cells heterologously expressing hNav1.7 α-subunit, ATX produced a robust membrane depolarization with an EC50 value of 7.8 ± 2.9 nM whereas veratridine, BmK NT1, and deltamethrin produced marginal response. Brevetoxin-2 was without effect on membrane potential change. The ATX response was completely inhibited by tetrodotoxin suggesting that the ATX response was solely derived from hNav1.7 activation, which was consistent with the results where ATX produced a negligible response in null HEK-293 cells. Six VGSC antagonists including lidocaine, lamotrigine, phenytoin, carbamazepine, riluzole, and 2-amino-6-trifluoromethylthiobenzothiazole all concentration-dependently inhibited ATX response with IC50 values comparable to that reported from patch-clamp experiments. Considered together, we demonstrate that ATX is a unique efficacious hNav1.7 activator which offers a useful probe to develop a rapid throughput screening assay to identify hNav1.7 antagonists.


Introduction
Voltage-gated sodium channels (VGSCs) are responsible for the rising phase of the action potential in excitable cells such as neurons, cardiac myocytes and skeletal muscle myocytes [1,2]. VGSCs are composed of voltage-sensing and pore-forming elements in one principal α-subunit and one or two auxiliary β-subunits which alter the channel physiological properties and subcellular cells. Among the agonists tested, only ATX produced a potent and efficacious membrane depolarization, providing a good signal/noise ratio. We further demonstrated that six VGSC antagonists including lidocaine, lamotrigine, phenytoin, carbamazepine, riluzole, and 2-amino-6-trifluoromethylthiobenzothiazole (SKA- 19) all concentration-dependently inhibited ATX response with IC 50 values comparable to that reported from patch-clamp experiments. These data suggested that ATX might represent a useful probe for developing an HTS assay to identify Na v 1.7 antagonists.

Influence of VGSC Agonists on Membrane Depolarization in HEK-293 Cells Stably Expressing hNa v 1.7
Previous studies have demonstrated that VGSC agonists such as veratridine, brevetoxin-2 (PbTx-2) had minimal effects on the stimulation of sodium influx or membrane depolarization in HEK-293 cells stably expressing VGSCs [23,25]. To identify an efficacious VGSC agonist which can provide a good signal/noise ratio, we examined the response on membrane depolarization in hNa v 1.7 expressed HEK-293 cells of five VGSC agonists which bound to topologically distinct neurotoxin sites including veratridine (VER, neurotoxin site 2), BmK NT1 (a site 3 α-scorpion toxin) [22], PbTx-2 (neurotoxin site 5) [5], deltamethrin (unrecognized neurotoxin site) [21], and ATX (unrecognized neurotoxin site) [23]. ATX produced a robust membrane depolarization in a concentration-dependent manner ( Figure 1A). Veratridine (up to 20 µM), BmK NT1 (up to 10 µM) and deltamethrin (up to 10 µM) produced a marginal response on the membrane depolarization ( Figure 1B,C,E). The neurotoxin site 5 agonist PbTx-2 was without effect on the membrane potential change in hNa v 1.7-HEK-293 cells ( Figure 1D). The EC 50 value for ATX-stimulated membrane depolarization (area under curve, AUC vs Log (concentration)) was 7.8˘2.9 nM with a maximal response of 11.7-fold of vehicle control ( Figure 2). Compared to the maximal ATX response (efficacy defined as 1), the maximal responses of veratridine, deltamethrin and BmK NT1 were only 0.11, 0.10 and 0.05, respectively ( Figure 2). hNav1.7-HEK-293 cells. Among the agonists tested, only ATX produced a potent and efficacious membrane depolarization, providing a good signal/noise ratio. We further demonstrated that six VGSC antagonists including lidocaine, lamotrigine, phenytoin, carbamazepine, riluzole, and 2amino-6-trifluoromethylthiobenzothiazole (SKA-19) all concentration-dependently inhibited ATX response with IC50 values comparable to that reported from patch-clamp experiments. These data suggested that ATX might represent a useful probe for developing an HTS assay to identify Nav1.7 antagonists.

ATX-Induced Membrane Depolarization Was Dependent on the Activation of hNav1.7
Given the efficacious response on the membrane potential change in hNav1.7-HEK-293 cells, we examined whether this membrane depolarization was from the activation of hNav1.7. Pre-treatment of TTX, a pore blocker of VGSC, concentration-dependently suppressed the ATX (10 nM)-induced membrane depolarization with an IC50 value of 49.8 nM (31.0-80.4 nM, 95% CI) ( Figure 3A,B). In null HEK-293 cells, ATX produced minimal response on the membrane depolarization which, compared to the ATX-induced response in hNav1.7-HEK-293 cells, was marginal ( Figure 3C-E).

ATX-Induced Membrane Depolarization Was Dependent on the Activation of hNav1.7
Given the efficacious response on the membrane potential change in hNav1.7-HEK-293 cells, we examined whether this membrane depolarization was from the activation of hNav1.7. Pre-treatment of TTX, a pore blocker of VGSC, concentration-dependently suppressed the ATX (10 nM)-induced membrane depolarization with an IC50 value of 49.8 nM (31.0-80.4 nM, 95% CI) ( Figure 3A,B). In null HEK-293 cells, ATX produced minimal response on the membrane depolarization which, compared to the ATX-induced response in hNav1.7-HEK-293 cells, was marginal ( Figure 3C-E).

Z 1 Factor Determination
Given the efficacious ATX response on the membrane depolarization which was solely dependent on the hNa v 1.7 activation, we determined Z 1 factor to test the suitability to use ATX and membrane potential dye for HTS assay. As shown in Figure 4, 30 nM ATX produced a robust, yet consistent response on the membrane depolarization. The calculated Z 1 value was 0.7598.

Z′ Factor Determination
Given the efficacious ATX response on the membrane depolarization which was solely dependent on the hNav1.7 activation, we determined Z′ factor to test the suitability to use ATX and membrane potential dye for HTS assay. As shown in Figure 4, 30 nM ATX produced a robust, yet consistent response on the membrane depolarization. The calculated Z′ value was 0.7598. The cells in Columns 2 to 11 were exposed to 30 nM of ATX. The Columns 1 and 12 were negative controls (0.1% DMSO). ATX produced a robust, yet consistent membrane depolarization. The Z′ factor was calculated to be 0.7589. This experiment was performed in two independent cultures.

Influence of an Array of VGSC Antagonists on ATX-Induced Membrane Depolarization
We next tested the influence of six VGSC antagonists on ATX (10 nM)-induced membrane depolarization. All the VGSC antagonists tested including lidocaine, lamotrigine, phenytoin, carbamazepine, riluzole, and SKA-19 produced concentration-dependent inhibition of ATX (10 nM)induced membrane depolarization (  Table 1). The IC50 values generated here are consistent to that generated from patch clamp (Table 1). It should be noted that riluzole, SKA-19, carbamazepine, and lamotrigine all produced nearly complete inhibition on ATX-induced membrane depolarization. However, the maximal inhibition of lidocaine and phenytoin on ATX-induced depolarization was somewhat smaller representing a maximal suppressing of 80.2% ± 5.8% and 78.8% ± 5.5%, respectively.  The cells in Columns 2 to 11 were exposed to 30 nM of ATX. The Columns 1 and 12 were negative controls (0.1% DMSO). ATX produced a robust, yet consistent membrane depolarization. The Z 1 factor was calculated to be 0.7589. This experiment was performed in two independent cultures.

Z′ Factor Determination
Given the efficacious ATX response on the membrane depolarization which was solely dependent on the hNav1.7 activation, we determined Z′ factor to test the suitability to use ATX and membrane potential dye for HTS assay. As shown in Figure 4, 30 nM ATX produced a robust, yet consistent response on the membrane depolarization. The calculated Z′ value was 0.7598. The cells in Columns 2 to 11 were exposed to 30 nM of ATX. The Columns 1 and 12 were negative controls (0.1% DMSO). ATX produced a robust, yet consistent membrane depolarization. The Z′ factor was calculated to be 0.7589. This experiment was performed in two independent cultures.

Discussion
In primary cultured neuronal preparation, the VGSC agonists ATX, veratridine, BmK NT1, PbTx-2 and deltamethrin which bound to topologically distinct neurotoxin sites all produced robust and significant sodium influx with distinct efficacies [5,22,23]. However, several studies have pointed out that in heterologous expression systems, these VGSC agonists produced a minimal response on both sodium influx and membrane potential [5,25]. In this study, we examined the ability of five VGSC agonists which bound to topologically distinctive neurotoxin sites to stimulate the membrane depolarization in hNa v 1.7-HEK-293 cells. Consistent with the previous studies [23,25], veratridine (neurotoxin site 2) only produced minimal response on the membrane depolarization whereas PbTx-2 (neurotoxin site 5) was without effect in hNa v 1.7-HEK-293 cells. We further demonstrated that deltamethin, which bound to an undefined neurotoxin site delaying the inactivation of the VGSCs [32], produced minimal response on the membrane potential changes. In addition, a scorpion toxin, BmK NT1 which likely bound to neurotoxin site 3 and prolonged the inactivation of the VGSCs in neurons [22] only produced marginal response to stimulate membrane depolarization. However, ATX produced an efficacious response on the membrane depolarization. Although the sodium channel expression density may partially account for the response discrepancy between neuronal and heterologously expression system [33], the hNa v 1.7-HEK-293 cells had little β-subunits co-expression. Sodium channel β-subunits regulate α-subunit function at multiple levels including mRNA expression, channel stabilization/trafficking and direct channel modulation [34]. In addition, the resting membrane potential for HEK-293 cells is relatively depolarized (´35˘5 mV) [35]. At this depolarized resting membrane potential, most hNa v 1.7 channels are in inactivated state [35]. Veratridine, BmK NT1, PbTx-2 and deltamethrin were demonstrated to primarily delay the VGSCs inactivation kinetics, but not the activation kinetics [22,32,36], an alternative explanation for these four VGSC agonists only producing marginal response in hNa v 1.7-HEK-293 cells. Although the detailed electrophysiological characterization of ATX on the VGSC remained to be established, the efficacious response on the membrane depolarization in hNa v 1.7-HEK-293 cells highly suggested that ATX response was not dependent on the β-subunits. The unique efficacious response of ATX also suggested that ATX may interact preferentially with the inactivated state of VGSC α-subunits.
The ATX response in hNa v 1.7-HEK-293 cells was from activation of hNa v 1.7 inasmuch as TTX completely suppressed the ATX-induced membrane depolarization. Furthermore, in null HEK-293 cells, ATX produced marginal membrane depolarization. The marginal membrane depolarization possibly was derived from the endogenously expressed VGSCs in null HEK-293 cells [37]. A rapid throughput assay to identify Na v 1.7 antagonists has been developed by co-application of veratridine and a scorpion toxin SVqq to achieve the robust fluorescence signals through activating VGSCs. However, this co-application of two agonists resulted in a low Z 1 value (0.15-0.45) [25], which was not suitable for the HTS assay. We demonstrated here that ATX produced a robust as well as consistent fluorescence change in a whole 96-well plate with a Z 1 value of 0.7589. The Z 1 value greater than 0.5 was thought to be suitable for the HTS assay [25].
In this study, we further demonstrated that an array of VGSC antagonists, including SKA-19, riluzole, phenytoin, lamotrigine, carbamazepine, and lidocaine all concentration-dependently suppressed ATX-induced membrane depolarization. It has been demonstrated that the IC 50 values generated from a fluorescence-based assay is typically five-fold less potent than that generated from patch-clamp [24,30]. However, we demonstrated that the IC 50 values for these VGSC inhibitors generated in this study were consistent with that from patch-clamp experiments [24,[28][29][30][31]38,39] (Table 1). Riluzole has been reported to suppress TTX-sensitive VGSC current in inactivated state with IC 50 values of 2 µM in dorsal root ganglion neurons, in which Na v 1.7 was the major TTX-sensitive VGSC subtype. At resting state, riluzole displayed a much lower affinity with an IC 50 value of 90 µM, suggesting riluzole preferred to bind to the inactivated state of VGSCs [29]. We demonstrated that riluzole inhibited the ATX-induced membrane potential changes with an IC 50 value of 3.58 µM which was comparable to its affinity in the inactivated state [29]. SKA-19, a thioanalog of riluzole, is a use-and state-dependent VGSC antagonist with IC 50 value of 5.8˘2.6 µM. We demonstrated here that SKA-19 suppressed the ATX (10 nM)-induced membrane depolarization in hNa v 1.7-HEK-293 cells with an IC 50 value of 2.02 µM, a value similar to that from voltage-patch clamp experiment [28]. The anticonvulsants phenytoin, lamotrigine, and carbamazepine were also preferred to bind to inactivated state of VGSCs [40][41][42][43][44][45][46][47]. The affinities for these anticonvulsants on the inactivated state are much higher than that on the closed and open states [48]. We demonstrated that the IC 50 values for lamotrigine, carbamazepine and phenytoin suppression of ATX-induced membrane depolarization in hNa v 1.7-HEK-293 cells were 66.3, 77.7 and 18.7 µM, respectively, which were more consistent with their affinities on the inactivated state. The local anesthetics, lidocaine has been reported to affect the steady-state fast inactivation of Na v 1.7 channels with an IC 50 value of 110˘20 µM [31] which was also consistent with current finding (150.6 µM, 92.9-244.0 µM, 95% CI). Considered together, it appeared that the IC 50 values generated from current fluorescence-based HTS assay were more consistent with their respective affinity in the inactivated state.
It has been reported that riluzole, lidocaine, phenytoin bound to distinct sites of the sodium channels [49][50][51][52]. For example, riluzole interacts with amino acids residues TYR 1787, LEU 1843 and GLN 1799 located in the transmembrane segment S6 of domain IV of the α-subunit [53]. Lidocaine binds to the local anesthetics site located in the channel pore [39,54]. Phenytoin binds to the S6 segments of domains III and IV of the Na + channel α-subunit [40,47,55,56]. In addition, TTX binds to another neurotoxin site distinct to those use-and steady-state blockers. The comparable IC 50 values between current study and reported previously suggesting that the fluorescence-based assay developed here was capable of identifying the hNa v 1.7 inhibitors bound to distinct neurotoxin sites on the α-subunits of VGSCs.
It has been reported that fluorescence-based assays are often subject to high false positive hits [57,58]. Further study was required to screen a chemical library to determine the liability of the assay. Nevertheless, the HTS assay developed here may represent a useful alternative for the primary screen to identify hNa v 1.7 antagonists with novel pharmacophores.

Materials
FMP blue dye was obtained from Molecular Devices (Sunnyvale, CA, USA). ATX was synthesized as described previously and was characterized to be above 95% purity [59]. G-418, penicillin, streptomycin, heat inactivated fetal bovine serum, poly-D-lysine (molecular weight >300,000), riluzole, veratridine and carbamazepine, deltamethrin, PbTx-2, TTX, and lamotrigine were obtained from Sigma-Aldrich (St. louis, MO, USA). SKA-19 was provided by Prof. Wulff at the University of California, Davis as described previously [28] and was characterized to be greater than 95% purity. Lidocaine was purchased from Abcam (Cambridge, MA, USA). The HEK-293 stably expressed Na v 1.7 was a generous gift from Dr. Lossin (University of California, Davis) and was the same line used as described previously [28].

Membrane Potential Change Detection
Membrane potential changes in HEK-293 or HEK-293 stably expressed hNa v 1.7 were determined using the FMP blue dye (Molecular Devices, Sunnyvale, CA, USA). The cells were plated onto poly-D-lysine (10 µg/mL) coated, 96-well, black-walled, clear-bottom plates at an initial density of 20,000 cells/well and cultured for 6 h. Cells were removed of their medium and 150 µL of 1ˆdye solution (1 bottle dissolved in 50 mL Locke's buffer, in mM: 8.6 HEPES, 5.6 KCl, 154 NaCl, 5.6 Glucose, 1.0 MgCl 2 , 2.3 CaCl 2 , 0.0001 glycine, pH 7.4) was added to each well. Cells were then incubated at room temperature for 30 min. The plate was then transferred to a FLIPR ®TETRA (Molecular Devices, Sunnyvale, CA, USA) chamber. Cells were excited at 510-545 nm and emission at 565-625 nm was recorded at 1 s intervals. After recording the basal fluorescence for 120 s, 50 µL of sodium channel agonists at different concentrations (prepared in 1ˆdye at 4ˆfinal drug concentrations) or vehicle (0.4% DMSO, 4ˆ) were added to different wells by an automated, programmable pipetting system. The fluorescence was recorded for additional 5-6 min at a sampling rate of 1 s. To examine the VGSC antagonist response on ATX-stimulated membrane depolarization, after recording the basal fluorescence for 120 s, different concentrations of VGSC antagonists were added to corresponding wells and the fluorescence was recorded for additional 5 min followed by an addition of 40 nM (4ˆ, final concentration, 10 nM) ATX. The fluorescence signals were presented as F/F 0 , where F was defined as the fluorescence at different time points; F 0 was the basal fluorescence averaged from initial 5 data points.

Data Analysis
Time-response and concentration-response relationships curves were generated using GraphPad Prism 5 software (GraphPad Software, Inc., San Diego, CA, USA). The EC 50 value for VGSC agonists-induced membrane depolarization was determined by non-linear regression analysis using a logistic equation. The IC 50 values of VGSC antagonists against ATX-induced membrane depolarization was determined by non-linear regression analysis using a logistic equation. Z 1 factor was calculated using the following equation as described previously [60]: Z 1 = 1´[3 SD of sample + 3 SD of control]/[mean of sample´mean of control]. Each experiment was repeated at least twice in independent cultures performed at least in triplicate.

Conclusions
The current study investigated an array of VGSC agonists to stimulate the membrane depolarization in hNa v 1.7-HEK-293 cells. We demonstrated that ATX but not other VGSC agonists tested produced efficacious response on the membrane depolarization. We further demonstrated that ATX can serve as a probe to develop an HTS assay for identifying hNa v 1.7 antagonist with distinct binding sites.
Author Contributions: Fang Zhao and Xichun Li performed the experiments, analyzed the data and wrote the manuscript; Fan Zhang, Liang Jin and Boyang Yu analyzed the data and wrote the manuscript; Masayuki Inoue synthesized the antillatoxin; Zhengyu Cao designed the experiments, performed the data analysis and wrote the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.