The voltage-gated sodium channel inhibitor, 4,9-anhydrotetrodotoxin, blocks human Na_v1.1 in addition to Na_v1.6

Abstract

Voltage-gated sodium channels (VGSCs) are responsible for the initiation and propagation of action potentials in neurons. The human genome includes ten human VGSC α-subunit genes, SCN(X)A, encoding Na_v1.1–1.9 plus Na_x. To understand the unique role that each VGSC plays in normal and pathophysiological function in neural networks, compounds with high affinity and selectivity for specific VGSC subtypes are required. Toward that goal, a structural analog of the VGSC pore blocker tetrodotoxin, 4,9-anhydrotetrodotoxin (4,9-ah-TTX), has been reported to be more selective in blocking Na⁺ current mediated by Na_v1.6 than other TTX-sensitive VGSCs, including Na_v1.2, Na_v1.3, Na_v1.4, and Na_v1.7. While SCN1A, encoding Na_v1.1, has been implicated in several neurological diseases, the effects of 4,9-ah-TTX on Na_v1.1-mediated Na⁺ current have not been tested. Here, we compared the binding of 4,9-ah-TTX for human and mouse brain preparations, and the effects of 4,9-ah-TTX on human Na_v1.1-, Na_v1.3- and Na_v1.6-mediated Na⁺ currents using the whole-cell patch clamp technique in heterologous cells. We show that, while 4,9-ah-TTX administration results in significant blockade of Na_v1.6-mediated Na⁺ current in the nanomolar range, it also has significant effects on Na_v1.1-mediated Na⁺ current. Thus, 4,9-ah-TTX is not a useful tool in identifying Na_v1.6-specific effects in human brain networks.

GRAPHICAL ABSTRACT

1. Introduction

Voltage-gated sodium channels (VGSCs) are transmembrane protein complexes that drive the initiation and propagation of action potentials in excitable cells. VGSCs consist of a single, pore-forming α-subunit (220–260 kDa) and two smaller β-subunits (30–40 kDa) that have many roles, including the modulation of channel expression, location, and function. To date, ten major human VGSC α-subunit subtypes (Na_v1.1–1.9 plus Na_x), encoded by the SCN(X)A genes have been identified [1]. Toxins have long served as pharmacological probes, playing an instrumental role in elucidating the structure and function of VGSCs [2]. Among the most widely used is tetrodotoxin (TTX), a potent guanidinium neurotoxin found in many species of aquatic and terrestrial organisms [3]. TTX binds with high affinity to Na_v1.1, Na_v1.2, Na_v1.3, Na_v1.4, Na_v1.6, and Na_v1.7. These channels are deemed “TTX sensitive” due to their susceptibility to half-maximal conduction blockade at nanomolar concentrations of TTX. To further understand the unique role each VGSC plays in normal and pathophysiological function in neural networks, compounds with high affinity for a specific VGSC subtype are required [4]. A structural analog of TTX, 4,9-an-hydro-TTX (4,9-ah-TTX) has been reported to be more selective in blocking Na⁺ current (I_Na) mediated by Na_v1.6 (IC₅₀ = 7.8 nM) than the other TTX sensitive VGSCs, including Na_v1.2 (IC₅₀ = 1260 nM), Na_v1.3 (IC₅₀ = 341 nM), Na_v1.4 (IC₅₀ = 988 nM), and Na_v1.7 (IC₅₀ = 1270 nM) [5]. While SCN1A, encoding Na_v1.1, has been implicated in neurological diseases [6], the effect of 4,9-ah-TTX on Na_v1.1-mediated I_Na has not been tested. Despite this, numerous studies have utilized 4,9-ah-TTX as a Na_v1.6 selective inhibitor following the initial report of its selective properties [7–11]. In this study, we determine the affinity of 4,9-ah-TTX for human and mouse brain VGSCs using radioligand competition binding. We investigate the effects of 4,9-ah-TTX on human Na_v1.1-, Na_v1.3- and Na_v1.6-mediated I_Na using the whole-cell patch clamp technique in HEK cells. Finally, we use mass spectrometry to evaluate the pH and time-dependence of 4,9-ah-TTX conversion to 4-epiTTX and TTX, as previous reports have indicated these factors can influence the pharmacological analysis of 4,9-ah-TTX [12,13]. Our work shows that, while 4,9-ah-TTX administration results in significant blockade of Na_v1.6-mediated I_Na at nanomolar concentrations, it also has significant effects on Na_v1.1-mediated I_Na at similar concentrations. Thus, 4,9-ah-TTX is not useful to resolve Na_v1.6-specific effects in brain networks.

2. MATERIALS & METHODS

2.1. Cell lines

Human Embryonic Kidney (HEK) 293 T cells stably expressing human VGSC α subunits Na_v1.1 (GenBank accession number NP_001159435.), Na_v1.3 (GenBank accession number NP_008853.3) or Na_v1.6 (GenBank accession number NP_055006.1) cDNAs were cultured in Dulbecco’s Modified Eagle Medium containing (4.5 g/L D-glucose, L-glutamine, 110 mg/L sodium pyruvate, 400 μg/mL G418, 100 U/mL penicillin/streptomycin). All cells were maintained in an incubator at 37 °C with 5% CO₂ until the time of recording.

2.2. Mouse models

Mouse studies were performed in compliance with protocols approved by the University of Michigan IACUC and were in accordance with the policies of the NIH Guide for the Care and Use of Laboratory Animals. Scn1a^tm1Kea mice were a generous gift from the laboratory of Dr. Jennifer Kearney at Northwestern University [14]. The Scn1a^tm1Kea colony was maintained by breeding heterozygous (129S6.Scn1a^+/−) animals to 129S6/SvEvTac mice (Taconic #129SVE). For all experiments, 129S6.Scn1a^+/− mice were crossed to generate −/−, +/−, and +/+ offspring. Genotyping of F1 pups was performed by PCR amplification of mouse genomic DNA. Primers used in genotyping were: 5′-AGTCTGTACCAGGCAGAACTTG-3′; 5′-CTGTTTGCTCCATCTTGTC ATC-3′ and 5′-GCTTTTGAAGCGTGCAGAATGC-3′. The PCR products are 1 kb for the Scn1a WT allele and 650 bp for the transgenic allele. All mice were maintained on a 12:12 h light:dark cycle and had ad libitum access to food and water throughout the experiments. Male and female mice were used in all experiments.

2.3. Human brain samples

De-identified, frozen, human cortical white matter samples from neurologically normal patients were obtained from the Human Brain and Spinal Fluid Resource Center in Los Angeles, CA under IRB approval (HUM00012142).

2.4. [³H]-Saxitoxin competition binding

To measure the competitive binding of 4,9-ah-TTX to VGSCs, whole brain membranes were prepared from postnatal day (P) 10–17 mice or from frozen human brain samples as described previously [15]. Equilibrium [³H]-saxitoxin (STX) binding in the presence of 4,9-ah-TTX was measured at 4 °C following an incubation period of one hour using a vacuum filtration assay with a saturating concentration (5 nM) of C-11 labelled [³H]-STX (20 Ci/mmol, American Radiolabeled Chemicals Inc.). In a subset of samples, 10 μM unlabeled TTX (Alomone Labs) was added to assess nonspecific binding. To quantify [³H]-STX binding, counts per minute (CPM) values obtained from liquid scintillation counting (Packard Tri-carb 1900TR) were corrected for specific binding by subtraction of nonspecific (10 μM TTX) values and then converted to decays per minute (DPM) before normalization to total protein concentration using the bicinchoninic acid (BCA) protein assay (Thermo-Fisher). A stock of 4,9-ah-TTX was diluted in 1XP binding buffer (50 nM HEPES pH 7.4, 130 mM choline chloride, 5.4 mM KCl, 80 μM MgSO4, 0.1 % dextrose) to give a final concentration of 10000, 1000, 100, 10, 1, 0.1 or 0.01 nM. A 7-point concentration-response curve was generated for 4,9-ah-TTX in competition with 5 nM [³H]-STX. Four independent experiments were conducted, and in each experiment, each sample was tested in duplicate. The average DPM reading of duplicate values was used for analysis. Competitive binding data were analyzed using Microsoft Excel and Graph Pad Prism v8.2 (San Diego, CA). The specific binding curve in [³H]-STX (fmol)/protein (mg) was initially generated using the following one-site IC₅₀ equation:

$$\text{Y}\text{=}\frac{\text{Bottom}\text{+}\left(\text{Top}\text{-}\text{Bottom}\right)}{1+{10}^{\text{X-log}\left({\text{IC}}_{50}\right)}}$$

where Top and Bottom are plateaus in the units of the Y-axis. IC₅₀ represents the concentration in nanomolar of unlabeled competitor that results in binding half-way between Top and Bottom values. Data were then normalized for each concentration point to the Top value obtained in the assay for each unlabeled toxin to obtain fraction of total binding. Each Y value thus represents the fraction of maximal [³H]-STX displacement at the corresponding concentration (X value). To obtain K_i values from the IC₅₀, the following equation was used:

$${\text{K}}_{\text{i}}=\frac{{\text{IC}}_{50}}{1+\frac{\left[\mathit{\text{radioligand}}\right]}{{\text{K}}_{\mathrm{d}}}}$$

where [³H-STX] = 5 nM in each assay at equilibrium. K_d of [³H]-STX = 2.5 nM. K_i = equilibrium dissociation constant of the unlabeled competitor in nM.

2.5. Whole-cell patchclamp electrophysiology

I_Na was measured at room temperature using the whole-cell patch clamp technique using previously described electrophysiological methods [17]. Cells were plated on 35 × 10 mm treated polystyrene culture dishes (Thermo-Fisher) and used for electrophysiological recordings within 1–3 days after plating. Cells were identified using an Eclipse TE300 upright microscope (Nikon). Micropipettes were obtained from 1.5 mm outer diameter capillary glass tubing (Warner Instruments) using a P-97 horizontal puller (Sutter Instrument Co.). Micropipettes were then polished using a MF-830 micro forge (Narishige) to obtain a resistance between 2.0–6.0 MΩ. The intracellular solution contained the following (in mM): 115 CsCl, 5 EGTA, 0.4 GTP, 2 MgATP, 0.5 CaCl₂, 10 HEPES, 5 Na⁺ phosphocreatine, 20 tetraethylammonium chloride, pH 7.2 with CsOH. Extracellular solution contained the following (in mM): 120 NaCl, 1 BaCl₂, 2 MgCl₂, 0.2 CdCl₂, 1 CaCl₂, 20 sucrose, 10 glucose, 10 HEPES, 20 tetraethylammonium chloride, pH 7.35 with NaOH. Signals were amplified using a Multiclamp 700B amplifier (Molecular Devices). Data were acquired with a Digidata 1440A interface (Molecular Devices) and analyzed using pClamp10 offline. Pipette and whole-cell capacitance were compensated at 70 %. Signals were low pass-filtered at 5 kHz, and data were sampled at 40 kHz. For the data generated in Figs. 2–4 and Tables 1 and 2 4,9-ah-TTX (Cayman Chemical) was evenly dissolved into the static extracellular bath solution at a concentration of 100 nM prior to recording. Cells were bathed in 4,9-ah-TTX for at least 10 min prior to recording. Peak I_Na was normalized to cell capacitance to obtain current density, used to plot I–V curves and calculate conductance with the following equation:

$$g=\frac{I}{V-V_{rev}}$$

where g is conductance, I is current, V is the test potential, and V_rev is the measured reversal potential. Peak currents were normalized to the maximum peak I_Na amplitude. The V_1/2 of activation represents the voltage of the membrane at which half-maximal conductance occurred. To determine the I_Na amplitude and the voltage dependence of activation, Na⁺ currents were evoked by 250 ms depolarizing test pulses (from −100 to +30 mV at 5 and 10 mV intervals) from a holding potential of −120 mV. Voltage-dependence of inactivation was determined by applying a 50 ms test pulse to 0 mV after 250 ms prepulses to the same voltages as described for the voltage dependence of activation. Normalized activation and inactivation curves were fit with a Boltzmann equation:

$$\frac{1}{1+e^{\frac{V-V_{1/2}}{k}}}$$

where V_1⁄2 is the membrane potential in the midpoint of the curve, and k is the slope factor. Persistent sodium current was analyzed as the average current in a 2 ms time period exactly 50 ms after the peak current was observed at a depolarizing step from −120 to 0 mV. Persistent current values were normalized to cell capacitance to obtain current density. The mean of all values obtained in the absence of 4,9-ah-TTX (control) were compared to the mean of all values obtained in the presence of 100 nM 4,9-ah-TTX. For data generated in Figs. 5 and 6 and Table 3, cells were plated on coverslips (Thermo-Fisher) and extracellular solution containing 4,9-ah-TTX was perfused into the RC-26 recording chamber (Warner Instruments) by a gravity-driven perfusion system with a flow rate of 2–3 mL/min. Peak I_Na was evoked via a 250 ms depolarizing pulse to 0 mV from a holding potential of −120 mV. Baseline recordings were established, followed by a 3 min perfusion of 4,9-anhydro-TTX at the indicated concentrations. Extracellular recording solution was then perfused on for ≥ 3 min and a washout recording obtained. The effect of 4,9-ah-TTX was evaluated as a decrease in peak I_Na compared to baseline control of that cell. Series resistance was monitored throughout perfusion recordings. If series resistance changed ≥ 20 % at any time, the recording was not included in the analysis. Signals were low pass-filtered at 10 kHz, and data were sampled at 20 kHz. Capacitive transients and leak conductance were subtracted using a P/4 protocol.

Fig. 2.

Na_v1.1, Na_v1.3 and Na_v1.6 I-V curves and I_Na traces in the presence and absence of 100 nM 4,9-ah-TTX. (A) Current-voltage (I-V) relationship obtained from recordings in the presence and absence of 100 nM 4,9-ah-TTX via whole-cell patch clamp of HEK cells stably expressing human Na_v1.1, Na_v1.3 or Na_v1.6. I_Na was evoked by stepping to different 250 ms test pulses in 5–10 mV intervals. Peak I_Na was normalized to cell capacitance to obtain current density. Values represent the mean ± S.E. (B) Representative peak I_Na traces recorded in the presence of 100 nM 4,9-ah-TTX and absence (control) via whole-cell patch clamp from HEK cells stably expressing Na_v1.1, Na_v1.3 or Na_v1.6. Cells were voltage-clamped at −120 mV and peak I_Na was elicited via stepping to a test pulse of 0 mV for 250 ms.

Fig. 4.

Normalized conductance to voltage and voltage dependence of availability relationships. Curves represent the voltage-dependence of steady-state activation (A) and voltage dependence of availability (B) in the presence (blue) and absence (black) of 100 nM 4,9-ah-TTX via whole-cell patch clamp from HEK cells stably expressing human Na_v1.1, Na_v1.3 or Na_v1.6. (B) Data points represent the mean ± S.E. Data obtained from recordings depicted in Fig. 2.

Table 1.

Average peak and persistent I_Na density and recorded in the presence of 100 nM 4,9-ah-TTX and absence (control) via whole-cell patch clamp from HEK cells stably expressing human (h) Na_v1.1, Na_v1.3 or Na_v1.6. Values represent the mean ± S.E.

Peak INa Density (pA/pF)				Persistent INa Density (pA/pF)
	Control	4,9-anhydro-TTX (100 nM)			Control	4,9-anhydro-TTX (100 nM)
hNavl.l	−131.6 ± 9.47	−47.24 ± 4.04		−5.25 ± 0.39	−2.58 ± 0.94
hNav1.3	−108.7 ± 8.54	−66.78 ± 11.61		−2.77 ± 0.82	−3.71 ± 0.26
hNav1.6	−136.9 ± 12.72	−23.63 ± 4.45		−1.52 ± 0.13	−1.84 ± 0.54

Table 2.

Voltage-dependence of activation and inactivation parameters recorded in the presence of 100 nM 4,9-ah-TTX and absence (control) via whole-cell patch clamp from HEK cells stably expressing human (h) Na_v1.1, Na_v1.3 or Na_v1.6. Values represent the mean ± S.E.

	V1/2 of Activation (mV)					k (mV−1)
	control	N	4,9-anhydro-TTX (100nM)	N		control	N	4,9-anhydro-TTX (100nM)	N
hNav1.1	− 17.36 ± 1.05	9	−15.31 ± 0.83	7		5.501 ± 0.27	9	5.705 ± 0.14	7
hNav1.3	− 16.88 ± 1.58	7	−14.47 ± 1.25	7		5.261 ± 0.24	7	5.332 ± 0.28	7
hNav1.6	− 17.97 ± 0.80	7	−17.28 ± 0.99	7		4.857 ± 0.28	7	6.265 ± 0.41*	7
	V1/2 of Activation (mV)					k (mV−1)
	control	N	4,9-anhydro-TTX (100nM)	N		control	N	4,9-anhydro-TTX (100nM)	N
hNav1.1	− 50.41 ± 2.37	9	−45.14 ± 1.55	7		− 5.04 ± 0.22	9	− 4.83 ± 0.55	7
hNav1.3	− 53.41 ± 2.22	7	−55.97 ± 1.95	6		− 5.85 ± 0.59	7	− 9.16 ± 1.51	6
hNav1.6	− 50.73 ± 1.54	7	−55.83 ± 2.79	5		− 4.63 ± 0.41	7	− 8.61 ± 2.17	5

^*p = 0.0158.

Fig. 5.

Percent reduction in Na_v1.1 and Na_v1.6 peak I_Na by 4,9-ah-TTX. Percent reduction from control in peak I_Na mediated by Na_v1.1 and Na_v1.6 following perfusion of 10 nM 4,9-ah-TTX (light blue), 100 nM 4,9-ah-TTX (cyan), and 500 nM (dark blue). Data points represent the mean ± S.E.

Fig. 6.

Na_v1.1 and Na_v1.6 peak I_Na traces before, during and after perfusion of 10, 100 and 500 nM 4,9-ah-TTX. Representative I_Na traces for the control (black), perfusion of 10 nM, 100 nM, and 500 nM 4,9-ah-TTX (blue) and after washout (red) via whole-cell patch clamp from HEK cells stably expressing human Na_v1.1 (A-C) or human Na_v1.6 (D-F). Cell were voltage-clamped at −120 mV and peak I_Na was elicited via stepping to a test pulse of 0 mV for 250 ms.

Table 3.

Percent reduction in Na_v1.1 and Na_v1.6 peak I_Na from control following the perfusion of 10, 100 or 500 nM 4,9-ah-TTX. Values represent the mean ± S.E.

Reduction in Peak INa Following Perfusion of 4,9-ah-TTX
[4,9-ah-TTX]	(% of Control)
	Nav 1.1	N	Nav 1.6	N
10nM	11 % (± 4)	4	18 % ( ±2)	4
100nM	46 % ( ± 5)	5	39 % ( ±2)	4
500 nM	64 % ( ± 4)	5	76 % ( ±2)	3

2.6. Mass spectrometry

A 500 μM stock solution of 4,9-ah-TTX (Cayman Chemical) was prepared in 1XP binding buffer. Duplicate 10 μL samples containing 100 μM 4,9-ah-TTX were prepared in plastic 1.7 mL centrifuge tubes according to the conditions outlined in Fig. 9 varying buffer solution (1XP or external recording solution), pH (range 6.4–8.4), incubation time, and storage temperature. Samples were diluted with 30 μL acetonitrile, vortexed to mix, and centrifuged at 12,000 x g for 20 min. 10 μL each sample was further diluted with 80 μL acetonitrile and 10 μL 0.1 % formic acid in MilliQ water with 2.5 μg/mL [¹⁵N]-arginine as an internal standard (Cambridge Isotopes). Samples 2 and 6, indicated as being incubated for 60 min, were incubated at room temperature for 1 h prior to the first acetonitrile dilution. For assessment of time dependence, triplicate samples of 10 μL volume were prepared with 1XP buffer at pH 7.4 and incubated at ambient temperature (22 °C) for 0, 1, and 5 weeks. Samples were quenched by the addition of 30 μL acetonitrile and stored at −20 °C until prep as described previously and analysis of all samples after the 5-week timepoint.

Fig. 9.

4,9-ah-TTX, 4-epiTTX and TTX in solution over time. Plot of mass spectrometry data assessing the relative abundances of 4,9-ah-TTX, 4-epi-TTX, and TTX over time comparing ratios of the analyte peak area to the peak area of the internal standard.

Samples were injected in 0.5 μL volumes on a hydrophobic liquid interaction chromatography (HILIC) BEH Amide 1.7 μM, 2.1 × 100 mm UPLC column (Waters) and analyzed using an Agilent G6545A quadrupole-time of flight (Q-TOF) mass spectrometer equipped with a dual AJS ESI source and an Agilent 1290 Infinity series diode array detector, autosampler, and binary pump. Solvent A was water with 0.1 % formic acid and solvent B was 95 % acetonitrile, 5% water, and 0.1 % formic acid. Separations were performed using a 20 % solvent A isocratic method for 5 min followed by a 1 min gradient from 20 % to 40 % A and an 8 min re-equilibration at 20 % A. Data was analyzed using MassHunter software (Agilent). The peak areas of 4,9-ah-TTX, [¹⁵N]-arginine, epi-TTX, and TTX were obtained by integration.

2.7. Statistical analysis

All data analysis of the recorded I_Na was performed using the software packages Clampfit v10.4 (Molecular Devices), Microsoft Excel, or Graph Pad Prism v8.2 (San Diego, CA). The statistical significance of differences between mean values was evaluated using Student’s unpaired t-test, or one-way ANOVA with p < 0.05 considered significant. Results are presented as mean ± S.E unless indicated otherwise.

3. Results

3.1. 4,9-ah-TTX binding affinity for human and mouse VGSCs

The binding curve of unlabeled 4,9-ah-TTX in competition with 5 nM [³H]-STX for the VGSCs present in Scn1a^+/+ mouse whole brain membranes is shown in Fig. 1 A. 4,9-ah-TTX showed normal steepness (Hill slope of −1.04) and was determined to have an inhibitory dissociation constant (K_i) of 63.02 nM with 95 % CI (47.42–84.59 nM). At 10 μM, 4,9-ah-TTX displaced > 95 % of the [³H]-STX present. Fig. 1B depicts the binding curve of 4,9-ah-TTX to Scn1a^−/− mouse whole brain membranes. The K_i was determined to be 87.67 nM 95 % CI (68.8–113.5), IC₅₀ of 248.6 nM with 95 % CI (186.6–333.4), and a Hill slope of −1.08. Fig. 1C shows the results of 4,9-ah-TTX binding to human cortical white matter membrane preparations. The human brain binding data resulted in a K_i of 62.27 nM with 95 % CI (29.2–136.3), IC₅₀ of 172.4 nM 95 % CI (85.8–463.4) and a Hill slope of −1.21. No significant differences in the K_i, IC₅₀ or Hill slope of 4,9-ah-TTX were detected between the Scn1a^+/+, Scn1a^−/− mouse, or human brain membranes (one-way ANOVA p = 0.05).

Fig. 1.

[³H]-Saxitoxin competition binding. Concentration response curves showing binding of 4,9-ah-TTX to human and mouse brain membrane samples in competition with 5 nM [³H]-saxitoxin. Unlabeled 4,9-ah-TTX was tested at 10 μM, 1000 nM, 100 nM, 10 nM, 1 nM, 0.1 nM and 0.01 nM. (A) Binding of 4,9-ah-TTX to Scn1a^+/+ mouse whole brain membranes (N = 4). (B) Binding of 4,9-ah-TTX to Scn1a^−/− mouse whole brain membranes (N = 4). (C) Binding of 4,9-ah-TTX to human cortical white matter membranes (N = 2). Each data point represents the mean fraction of total specific binding ± S.E.

3.2. Effects of 100 nM 4,9-ah-TTX on I_Na density in human HEK-Na_v1.1, -Na_v1.3 or -Na_v1.6 cells

Fig. 2. shows the current-voltage (I–V) relationships for HEK-Na_v1.1, -Na_v1.3 or -Na_v1.6 cells in the presence and absence of 100 nM 4,9-ah-TTX. Differences in mean I_Na density evoked at 0 mV following the bathing of cells in 100 nM 4,9-ah-TTX compared to control for each cell line are shown in Table 1. Mean I_Na density in the absence of 4,9-ah-TTX (control) for HEK-Na_v1.1 was −131.6 ± 9.47, HEK-Na_v1.3–108.7 ± 8.54 and HEK-Na_v1.6–136.9 ± 12.72. Exposure of cells to 100 nM extracellular 4,9-ah-TTX reduced the mean peak I_Na density in HEK-Na_v1.1 cells by 64 ± 3% (84.31 pA/pF ± 11.39 S.E.), in HEK-Na_v1.3 cells by 39 ± 11 % (41.87 pA/pF ± 14.41 S.E.), and in HEK-Na_v1.6 cells by 83 ± 3% (113.3 ± 13.79 S.E.). The conductance-voltage (G–V) relationships for HEK-Na_v1.1, HEK-Na_v1.3 or HEK Na_v1.6 cells in the presence and absence of 4,9-ah-TTX are shown in Fig. 4. The V_1/2 of activation was not significantly different in any of the cell lines tested in the presence of 100 nM extracellular 4,9-ah-TTX compared to control (Table 2). The slope factor (k) was not significantly different when comparing control to 100 nM 4,9-ah-TTX conditions for HEK-Na_v1.1 or HEK-Na_v1.3. In contrast, the slope factor for HEK Na_v1.6 was significantly increased when comparing control (k = 4.857 ± 0.280) to the 100 nM 4,9-ah-TTX condition (k = 6.265 ± 0.415) (p = 0.015). The voltage dependence of availability relationship for HEK-Na_v1.1, HEK-Na_v1.3 or HEK-Na_v1.6 cells in the presence and absence of 100 nM 4,9-ah-TTX is also shown in Fig. 4. The V_1/2 of inactivation and slope factor were not significantly different between the control and 4,9-ah-TTX conditions for HEK-Na_v1.1, HEK-Na_v1.3 or HEK-Na_v1.6 (Table 2). An analysis of persistent current at a test pulse to 0 mV in the presence and absence of 100 nM 4,9-ah-TTX revealed a significant reduction (49.2 %) in HEK-Na_v1.1 persistent I_Na density (Fig. 3). There was no significant difference in the persistent I_Na density observed in the control versus 100 nM 4,9-ah-TTX condition for HEK-Na_v1.3 or HEK-Na_v1.6.

Fig. 3.

Average peak and persistent I_Na density recorded in the presence and absence of 100 nM 4,9-ah-TTX in HEK cells stably expressing human Na_v1.1, Na_v1.3 or Na_v1.6. (A) Cells were voltage-clamped at −120 mV and peak I_Na was elicited via stepping to a test pulse of 0 mV for 250 ms. (B) Persistent I_Na density present in a 2 ms time period 50 ms after peak I_Na was observed at a test pulse to 0 mV. Bar graphs represent mean ± S.E. Significant differences between control and toxin treated conditions were tested using a Student’s unpaired t-test. Data obtained from recordings depicted in Fig. 2.

3.3. Reduction in Na_v1.1 and Na_v1.6 mediated peak I_Na following perfusion of 4,9-ah-TTX

Perfusion of 4,9-ah-TTX onto HEK-Na_v1.1 and HEK-Na_v1.6 cells resulted in a concentration dependent reduction in peak I_Na (Figs. 5 and 6). Na_v1.1 peak I_Na was reduced by (11 % ± 4) following the perfusion of 10 nM 4,9-ah-TTX. At 100 nM, the mean reduction in Na_v1.1 peak I_Na was (46 % ± 5). The perfusion of 500 nM 4,9-ah-TTX reduced Na_v1.1 peak I_Na by (64 % ± 4) (Table 3). Washout of 4,9-ah-TTX restored peak I_Na on average to 94.5 % (range 84–106 %) of baseline control in all recordings. Na_v1.6 peak I_Na was reduced by (18 % ± 2) following the perfusion of 10 nM 4,9-ah-TTX. At 100 nM, the mean reduction in Na_v1.6 peak I_Na was (39 % ± 2). The perfusion of 500 nM 4,9-ah-TTX reduced Na_v1.6 peak I_Na by (76 % ± 2) (Table 3).

3.4. Time and pH dependent effects of 4,9-ah-TTX conversion to 4-epiTTX and TTX

Relative abundances of 4,9-ah-TTX, 4-epiTTX, and TTX were monitored at different pH conditions and incubation times at ambient temperature. The results shown in Fig. 8 indicate that differences in pH over the range 6.4–8.4 do not significantly impact the distribution of 4,9-ah-TTX, 4-epiTTX, and TTX, suggesting 4,9-ah-TTX was not significantly converted to 4-epiTTX or TTX under our analysis conditions. Assessing the relative abundances of these molecules over time similarly did not result in increased amounts of 4-epiTTX or TTX relative to 4,9-ah-TTX; however, a significant decrease in all three molecules was observed over time, with a minute amount remaining after 5 weeks of incubation at ambient temperature, as shown in Fig. 9.

Fig. 8.

Impact of pH on 4,9-ah-TTX, 4-epiTTX and TTX. (Top) Plot of mass spectrometry results comparing peak area/internal standard ratio based on peak integration corresponding to 4,9-ah-TTX, 4-epiTTX, and TTX. (Inset) Zoomed-in graph of represented data. (Below) Reaction conditions used in analysis. ER: external recording solution used in electrophysiology experiments. 1XP: binding buffer used in competition binding experiments.

4. Discussion

TTX binds with high specificity to the VGSC pore in a 1:1 stoichiometry to occlude the conduction of Na⁺ ions by forming a web of electrostatic interactions with the channel. A large body of evidence from numerous studies strongly supports that TTX interacts with a narrow region of residues within the channel involved in permeation and the selectivity of Na⁺ ions [19]. Major coordinating residues are located on the membrane re-entrant P-loops (pore loops) of segments S5-S6 on all 4 domains. This region of the channel heavily influences channel conductance and ion selectivity, forming an inner and outer ring of charged amino acid residues (six carboxylates and one lysine) known as the selectivity filter ([20]; Favre et al. [21]). The asymmetric charge distribution of each ring coordinates solvated Na⁺ ions as they pass through the pore [16]. Mutagenesis and electrophysiological studies have revealed four key residues (D: aspartate, E: glutamate, K: lysine, and A: alanine), each present in one of the domains, that form the inner ring of the mammalian VGSC selectivity filter. The DEKA motif found in human VGSCs is conserved across all subtypes. Mutating K and A of the DEKA motif to glutamates, generating DEEE, abolishes Na⁺ selectivity and permits Ca²⁺ to permeate the channel [22]. Three or four residues downstream of the DEKA selectivity filter in each domain lie the outer ring of the human VGSC selectivity filter composed of an EE(M/D)D motif (see Fig. 10). The outer EE(M/D)D ring is comprised of four negatively charged carboxylates shown to attract Na⁺ ions into the outer vestibule of the pore (Ding et al. [23]). Neutralizing mutations in the inner or outer ring of the selectivity filter, and any residues linking the two regions, can dramatically reduce TTX binding [19,24–26].

Fig. 10.

VGSC schematic with TTX binding site and selectivity filter residue alignment. (Top) Schematic of the mammalian voltage-gated sodium channel α and β subunit complex. Segments are labelled numerically (1–6) and domains are labelled in roman numerals (I-IV). Voltage-sensing S4 segments in each domain are labelled in light red. The pore loop regions formed by segments S5 and S6 are labelled in green. The amino acid residues representing the Na⁺ ion selectivity filter and canonical TTX binding site are depicted in dark red circles. (Below) Amino acid residues of the selectivity filter in human Na_v1.1, Na_v1.3, Na_v1.6 or mouse Na_v1.6. The selectivity filter is −2 through +4 residues from the DEKA locus. The inner DEKA ring residues of each domain are shown in red. The outer EE(M/D)D ring residues of each domain are shown in blue. The following accession numbers were used to generate the basic local alignment (hNav1.1: NP_001159435.1; hNav1.3: NP_008853.3; hNav1.6: NP_055006.1; mNav1.6 NP_001070967.1).

Another key residue in TTX binding is Y401 (Na_v1.4 numbering). TTX-sensitive channels have an aromatic Y/F (tyrosine or phenylala-nine) at this position that is crucial for coordinating TTX in the pore [27]. Na_v1.5, Na_v1.8, and Na_v1.9 achieve “TTX-resistance” through replacement of the Y/F residue with a nonaromatic cystine (C) or serine (S) [28]. Intriguingly, all the TTX-resistant VGSC subtypes are located on human chromosome 3. Adaptive resistance strategies in TTX-sensitive channels have been observed in many TTX-producing predators and their prey (Soong and Venkatesh [29]). Various species of puffer-fish also possess nonaromatic substitutions for the Y/F residue present in domain I, dramatically reducing their sensitivity to TTX. Many species of newts and snakes have evolved TTX resistant Na_v1.4 channels via substitutions in the pore loops of domain III and IV [30]. Amino acid substitutions of the inner and outer selectivity filter ring of domain III and IV that alter TTX sensitivity have been identified [31,32]. A threonine (T) substitution of the methionine (M) in the domain III outer selectivity filter ring on Na_v1.4 is observed in all TTX-harboring newts and the Thamnophis couchii garter snake [18].

Recent high-resolution structural evidence has confirmed that TTX restricts Na⁺ conduction via binding to the vestibule of the VGSC selectivity filter [33,34]. Cryogenic electron microscopy studies showing TTX bound to the American cockroach sodium channel Na_vPaS reveal hydrogen bonds and salt bridges formed between the polar groups on the TTX molecule and acidic residues in the pore loop P2 helix segment of domains I, II and IV [33]. The DE of the inner DEKA selectivity filter also directly binds to TTX in the Na_vPaS channel [33]. In a subsequent study, the authors obtained high-resolution structural data of TTX bound to hNa_v1.7 [34]. Inner and outer selectivity filter residues identified by previous studies were implicated in the coordination of TTX within the pore. The coordination of TTX in hNa_v1.7 was nearly identical to that observed in Na_vPaS, differing only in the P2 segment of domain III [34]. Na_v1.7 contains an isoleucine in place of an aspartate that is present in the other 8 human VGSC subtypes. This outer ring aspartate likely contributes to TTX binding in hNa_v1.1. Aside from that difference, similar TTX coordination strategies observed by Shen et al. [34] would be anticipated in hNa_v1.1, hNa_v1.3 and hNa_v1.6.

It is not known which amino acid residues of VGSCs form the binding site for 4,9-ah-TTX, although it is postulated that 4,9-ah-TTX interacts with the same residues as TTX [12]. Xu et al. explored the differences in binding of TTX and 4,9-ah-TTX to human Na_v1.2 using computational modeling and molecular dynamics. The authors proposed that the H-bond interactions made between the C4 and C9 hydroxyl of TTX and the outer ring carboxylates of the human Na_v1.2 selectivity filter are key factors in the greater inhibitory activity of TTX relative to 4,9-ah-TTX (Xu and Li [35]). It is important to distinguish the species used in an experimental sodium channel preparation when comparing results across studies. The original study reporting 4,9-ah-TTX to be selective for Na_v1.6, utilized the expression of mouse cDNA in Xenopus oocytes [5]. Amino acid sequence alignments shown in Fig. 10 reveal no differences when comparing the selectivity filter regions of human Na_v1.1, Na_v1.3, and Na_v1.6 to mouse Na_v1.6. Despite the high homology seen in human and rodent VGSCs, it is possible that alternative 4,9-ah-TTX binding conformations are achieved in mouse, rat, and human VGSCs, leading to different results observed across studies. To detect possible species-selective differences in 4,9-ah-TTX for mouse and human brain VGSCs, we conducted [³H]-STX binding with cortical white matter membranes from neurologically normal humans (Fig. 1C). The lack of significant differences found in affinity and potency suggest there is no differential sensitivity of 4,9-ah-TTX for human or mouse VGSCs. Further structural studies analyzing the precise amino acid residues of each human TTX-sensitive VGSCs responsible for the coordination of 4,9-ah-TTX are required to understand potential subtype specific binding contributions.

Previous studies conducted on 4,9-ah-TTX, performed in markedly different experimental systems, have yielded variable results. Recording I_Na using voltage-clamp in the squid giant axon, Kao and Yasumoto reported 4,9-ah-TTX to have an IC₅₀ of 298 nM [36]. These experiments were performed at 7 °C with an external solution pH of 7.8. The authors noted the importance of this external pH, as the C-10 hydroxyl of 4,9-ah-TTX has a reported pK_a of 7.95 (Goto et al. 1965). Using [³H]-STX binding in rat brain synaptosomes, Yotsu-Yamashita et al. reported an equilibrium dissociation constant for 4,9-ah-TTX of 180 ± 11 nM [37]. Using the two-electrode voltage clamp technique in Xenopus oocytes, Rosker et al. reported that 4,9-ah-TTX selectively inhibits mouseNa_v1.6 with an IC₅₀ of 7.8 ± 2.3 nM [5]. Teramoto et al. reported that resurgent-like I_Na present in isolated mouse vas deferens myocytes is abolished in the presence of 3 μM 4,9-ah-TTX [11]. Further investigation showed that native I_Na in mouse vas deferens myocytes was sensitive to 4,9-ah-TTX, with a K_i of 512 nM [10]. The authors then compared this to the effects of 4,9-ah-TTX in HEK cells transiently transfected with mouse Na_v1.6 cDNA. In this system, the authors reported 4,9-ah-TTX to have a K_i of 112 nM for Na_v1.6-mediated I_Na and a K_i of 92 nM when Na_v1.6 was co-expressed with β1 subunits. Measuring I_Na using an automated patch clamp system in CHO and HEK cells, Rogers et al. reported an IC₅₀ for 4,9-ah-TTX of 32.9 nM for Na_v1.6- and 1.6 μM for Na_v1.7-mediated I_Na. [9].

4,9-ah-TTX was determined to have an inhibitory dissociation constant (K_i) of 62.03 nM for Scn1a^+/+ mouse brain membrane preparations in our [³H]-STX binding assay, consistent with its reduced affinity for brain VGSCs compared to TTX. In Scn1a^−/− mice, global expression of Na_v1.1 is deleted [14]. While not statistically significant, 4,9-ah-TTX shows a reduced affinity for Scn1a^−/− mouse brain compared to Scn1a^+/+ (K_i of 87.67 nM vs. 63.02 nM), suggesting Na_v1.1 contributes to the binding activity of 4,9-ah-TTX. A heterogenous mixture of Na_v1.1, Na_v1.2, Na_v1.5, Na_v1.6, and Na_v1.7 are known to be present in the adult mouse whole-brain membrane preparations utilized to test 4,9-ah-TTX (Wu et al. [38], [39–41], Kanellopoulos et al. [42]).

We showed that bathing cells in 100 nM 4,9-ah-TTX significantly reduced I_Na density in HEK-Na_v1.1 (64 ± 3%), HEK-Na_v1.3 (39 ± 11 %), and HEK-Na_v1.6 cells (83 ± 3%) compared to control. No significant changes in the V_1/2 of activation were observed when comparing control to the presence of 100 nM 4,9-ah-TTX, consistent with 4,9-ah-TTX exerting its inhibitory effects through pore-blocking action, like TTX. To confirm the observed effect was both acute and reversible, as is the case with TTX, we recorded peak I_Na mediated by Na_v1.1 and Na_v1.6 before, during perfusion, and after washout of 4,9-ah-TTX (Fig. 5). Consistent with the results we observed from the bath exposure to 4,9-ah-TTX, the perfusion of 500 nM 4,9-ah-TTX reduced Na_v1.1-mediated peak I_Na by 64 % from baseline, 100 nM 4,9-ah-TTX induced a 46 % reduction from baseline, and 10 nM reduced Na_v1.1-mediated peak I_Na by 11 % from baseline (Table 3). 4,9-ah-TTX reduced Na_v1.6 peak I_Na by (76 %), (39 %), and (18 %) at 500 nM, 100 nM and 10 nM concentrations respectively (Table 3). Together, these data suggest 4,9-ah-TTX has comparable inhibitory activity for human Na_v1.1 and Na_v1.6 at nanomolar concentrations.

Tetrodotoxin has long been considered a voltage and state-independent blocker, inhibiting sodium current without altering the voltage-dependence of activation or inactivation, regardless of the original state of the channel [43]. Reports of TTX causing a modest hyperpolarizing shift in the voltage-dependence of inactivation do exist [44]. 4,9-ah-TTX was reported to induce a hyperpolarizing shift in the voltage-dependence of inactivation of Na_v1.6 [5]. We did not detect any shifts from baseline in the voltage-dependence of inactivation for Na_v1.1, Na_v1.3 or Na_v1.6 in the presence of 100 nM 4,9-ah-TTX (Table 2). Possible state-dependent effects of 4,9-ah-TTX on Na_v1.1 were not directly examined. Neurons expressing Na_v1.1 in vivo would rest at a more depolarized resting membrane potential (~ −65 mV) than the holding potential used in our experiments (−120 mV). Further studies examining the inhibitory activity of 4,9-ah-TTX on hNa_v1.1 at holding potentials from −55 to −70 mV would provide a more physiologically relevant interpretation of our findings.

An important consideration when testing the activity of 4,9-ah-TTX is the spontaneous equilibration that can occur between 4,9-ah-TTX, 4-epiTTX and TTX (Fig. 7: [45]), which can dramatically influence the results of pharmacological testing. Tsukamoto et al. tested 4,9-ah-TTX on human Na_v1.6 mediated I_Na in HEK cells and reported an IC₅₀ value of 294 ± 25 nM from a freshly prepared working solution. However, using a solution prepared from a frozen stock solution that had been placed in storage for 6 months, the authors found a much greater reduction in I_Na following the application of 100 nM 4,9-ah-TTX, presumably due to the conversion of 4,9-ah-TTX to TTX [13]. We performed analytical authentication of 4,9-ah-TTX prior to using it in our experiments to determine the level of contamination present and to validate the presence of 4,9-ah-TTX over a pH range from 6.4–8.4, including the pH of our electrophysiological recording solution (pH = 7.35). The relative amounts of 4,9-ah-TTX compared to 4-epiTTX and TTX were assessed in the electrophysiologicalrecording and binding buffer solutions using mass spectrometry. In contrast to previous findings, in our hands we found little to no evidence of 4,9-ah-TTX conversion to 4-epiTTX or TTX under the different conditions. We also assessed the relative abundances of these molecules over time at pH 7.4 and did not observe interconversion. We did find that incubating 4,9-ah-TTX at ambient temperature over the course of 5 weeks resulted in a significant decrease in the total amount of 4,9-ah-TTX without corresponding increases in 4-epiTTX or TTX, suggesting that 4,9-ah-TTX is not stable under these conditions for extended periods of time. However, the 4,9-ah-TTX used in our experiments had negligible levels of contamination from 4-epiTTX and TTX under multiple conditions.

Fig. 7.

Molecular structures and equilibrium of tetrodotoxin, 4,9-anhydrotetrodotoxin, and 4-epi-tetrodotoxin.

5. Conclusion

Identifying a Na_v1.6 selective inhibitor is of great interest to VGSC biology. Na_v1.6 is widely expressed in the mammalian peripheral and central nervous systems as well as in the heart, although at lower levels [46,47]. In the adult brain, Na_v1.6 is the most abundantly expressed VGSC subtype [1]. Within neurons, Na_v1.6 is highly concentrated at the distal axon initial segment of excitatory and inhibitory neurons and nodes of Ranvier in myelinated axons [48,49]. Variants in SCN8A, which encodes human Na_v1.6, have been implicated in epilepsy, movement disorders, and cognitive impairment [50]. A compound with the selective properties previously reported for 4,9-ah-TTX holds great value. However, caution should be exercised when using this reagent. Our findings demonstrate that, in addition to blocking Na_v1.6-mediated I_Na, 4,9-ah-TTX has significant inhibitory effects on Na_v1.1-mediated I_Na in the nanomolar range and thus cannot be used as a selective Na_v1.6 inhibitor in human cells or tissues where Na_v1.1 is also expressed.

Abbreviations:

4,9-ah-TTX

4,9-anhydro-tetrodotoxin

HEK

human embryonic kidney

IC₅₀

half-maximal inhibitory concentration

I_Na

Na⁺ current

STX

Saxitoxin

TTX

tetrodotoxin

VGSC

voltage-gated sodium channel