Cationic Modulation of Voltage-Gated Sodium Channel (Nav1.5): Neonatal Versus Adult Splice Variants—2. Divalent (Cd²⁺) and Trivalent (Gd³⁺) Ions

Abstract

Background: A “neonatal” splice-form of the voltage-gated sodium channel Nav1.5 is functionally expressed in human cancers and potentiates metastatic cell behaviors. Splicing causes the replacement of 7 amino acids, including a negatively charged aspartate²¹¹ in the “adult” Nav1.5 (aNav1.5) to a positively charged lysine in the “neonatal” (nNav1.5). These changes occur in the region surrounding the DI:S3–S4 extracellular linker. The splice variants respond differently to changes in extracellular H⁺ and this could be of pathophysiological significance. However, how the two differentially charged splice variants would react to cations of higher valency is not known.

Materials and Methods: We used patch-clamp recording to compare the electrophysiological effects of Cd²⁺ and Gd³⁺ on “adult” and “neonatal” Nav1.5 expressed stably in EBNA-293 cells. Several parameters were determined for the two channels and statistically compared.

Results: Both cations inhibited peak I_Na through reducing G_max and induced a positive shift in the voltage range of activation. However, unlike Gd³⁺, Cd²⁺ had only a weak effect on voltage dependence of activation, and no effect on voltage dependence of inactivation, recovery from inactivation, or the kinetics of activation/inactivation.

Conclusions: The electrophysiological effects of Cd²⁺ and Gd³⁺ studied were essentially the same for “neonatal” and “adult” Nav1.5, although these splice variants possess differences in their external charges. In contrast, the effects of H⁺ were shown earlier to be significantly differential. Taken together, these results suggest that limited adjustment of the charged structure of pharmacological agents could enable selective targeting of neonatal Nav1.5 associated with several cancers.

Introduction

Alternative splicing occurs in the genes coding for the α-subunits of voltage-gated sodium channels (VGSCαs).¹ In SCN5A (the gene encoding Nav1.5), there is a major splicing of exon-6 that affects the voltage sensor/“paddle” region of domain I (DI), including segment 3 (DI:S3), most of segment 4 (DI:S4), and the DI:S3–S4 extracellular linker.^2,3 This splicing is developmentally regulated; transcripts possessing the 5′-exon occur abundantly at birth but are replaced within a few days by the 3′-exon.^4,5 Accordingly, the 5′ and the 3′ variants are referred to as “neonatal” and “adult,” respectively. Importantly, the two splice variants differ by 7 amino acids, including a negatively charged aspartate at position 211 in the adult (aNav1.5), which is replaced by a positively charged lysine in the neonatal form (nNav1.5).²

Previously, we demonstrated that aNav1.5 and nNav1.5 expressed stably in EBNA-293 cells differ significantly in their electrophysiological profiles.⁶ Thus, nNav1.5 channels exhibited significantly more depolarized voltage of activation, hyperpolarized inactivation, slower recovery from inactivation, and slower activation and inactivation kinetics compared with aNav1.5. Changing Lys²¹¹ back to Asp²¹¹ in nNav1.5 abolished most of the differences between aNav1.5 and nNav1.5, suggesting that the positive electric field induced by Lys²¹¹ in the voltage-sensing region of DI was primarily responsible for the neonatal phenotype.⁶ Such a change of local electrostatic potential could also affect the interaction of the channel protein with cations and impact differentially upon channel functioning, especially since the charge reversal is near the voltage sensor.

In our previous article, we studied the effect of the monovalent cation H⁺ on nNav1.5 and aNav1.5 in a comparative approach.⁷ We found that H⁺ inhibited peak I_Na through reducing G_max and induced a positive shift in voltage range of activation, which was significantly greater for aNav1.5, compared with nNav1.5.⁷ Whether cations of higher valency would bring about additional effects is not known. Here, we extended the comparison to cations of increasing valency. Cd²⁺ was chosen as a divalent cation since it blocks Nav1.5 at relatively lower concentrations compared with most other VGSCαs.⁸ As a trivalent cation, Gd³⁺ was studied for the first time on Nav1.5 in isolation.

Materials and Methods

Details of the materials and methods are given in the accompanying article.⁷ Below, we outline any additional information.

Solutions

CdCl₂ and GdCl₃ (anhydrous) were freshly dissolved and serially diluted in mammalian physiological saline (MPS) to various working concentrations on the day of the recordings. Although hydrolysis of Gd³⁺ may result in a decreased free Gd³⁺ concentration (e.g., by formation of hydroxides), at the concentrations and the pH value used, this reduction should be stable and <5%.^9,10 The solutions were bath applied at a rate of 1.5 mL/min using a gravity-fed perfusion system. Recordings started 4–5 min after seal rupture to allow time for stabilization.¹¹ In each recording set, cells were exposed to “control” and “test” MPS solutions alternatively for at least 2 min each (perfusion switch time <2 s) and voltage pulses were applied in each condition. The effects of several concentrations of Cd²⁺ could be measured from the same cell since their effects were rapidly reversible. In all experiments, the order of applying the various doses was randomized. Data were accepted only if the pre- and postmeasurements were within 10%. For Gd³⁺, recovery was notably less and only one concentration per cell could be studied.

Data analysis

Values of G_max and peak I_Na at various cation concentrations were normalized relative to control values. Data were plotted and fit with a logistic function of the form:

where A₁ and A₂ are the maximum and the minimum; C is the cation concentration; IC₅₀ is concentration giving half-maximum block; and n is the Hill coefficient. Details of all the statistical analyses are given in the accompanying article.⁷

Results

The data obtained for the electrophysiological effects of Cd²⁺ and Gd³⁺ on nNav1.5 and aNav1.5 are given and compared in Tables 1–3.

Table 1.

Summary of Effects of Cd²⁺ and Gd³⁺ on Nav1.5 Neonatal and Adult Splice Variants

Cation	Parameter	Splice variant	n	IC50	Maximum asymptote	Minimum asymptote	Hill coefficient (nH)	R
Cd2+	Relative Gmax (%)	nNav1.5	≥11	132.89 ± 13.01	100.75 ± 3.14	4.13 ± 4.36	1.30 ± 0.17	0.99
		aNav1.5	≥11	132.54 ± 17.81	98.11 ± 4.23	2.11 ± 6.22	1.10 ± 0.19	0.99
Gd3+	Relative INa at −20 mV	nNav1.5	≥7	48.69 ± 4.93	1.03 ± 0.03	0.00 ± 0.03	1.05 ± 0.10	0.99
		aNav1.5	≥8	58.25 ± 9.10	1.03 ± 0.04	−0.01 ± 0.04	0.92 ± 0.13	0.99

For each parameter (given columns), data are shown for nNav1.5 (above) and aNav1.5 (below). There was no statistically significant difference between the respective pairs (p > 0.05; unpaired Student's t-tests). In addition, the minimum asymptotes were not statistically different from zero.

Table 2.

Effects of Cd²⁺ on Nav1.5 Neonatal and Adult Splice Variants

Parameter	Splice variant	Control	50 μM	100 Mm	250 μM	500 μM	Recovery
Relative Gmax (%)	nNav1.5	100.0 ± 0.0	81.1 ± 3.6	59.2 ± 1.2	35.1 ± 2.0	18.2 ± 0.9	104.4 ± 4.7
	aNav1.5	100.0 ± 0.0	74.8 ± 3.2	55.7 ± 2.2	35.8 ± 1.4	18.9 ± 1.1	98.1 ± 6.4
Vthres (mV)	nNav1.5	−55.4 ± 0.7	−53.6 ± 0.9	−53.8 ± 0.9	−48.8 ± 0.8	−44.8 ± 0.9	−56.3 ± 0.8
	aNav1.5	−60.0 ± 1.0	−57.9 ± 0.9	−59.1 ± 1.4	−54.0 ± 1.2	−49.4 ± 1.4	−61.4 ± 1.3
Vpeak (mV)	nNav1.5	−5.6 ± 1.6	−3.2 ± 2.1	−5.8 ± 1.6	−0.4 ± 1.7	2.9 ± 2.5	−7.0 ± 2.4
	aNav1.5	−18.1 ± 1.3	−18.1 ± 1.3	−16.8 ± 1.4	−12.3 ± 0.9	−5.4 ± 1.3	−21.1 ± 1.8
Activation V1/2 (mV)	nNav1.5	−25.8 ± 1.3	−24.3 ± 1.9	−24.9 ± 1.5	−20.6 ± 1.8	−18.7 ± 2.0	−26.8 ± 1.7
	aNav1.5	−37.3 ± 1.1	−34.9 ± 1.1	−37.4 ± 1.2	−30.9 ± 1.0	−27.2 ± 1.1	−39.8 ± 1.2
Activation k	nNav1.5	7.8 ± 0.3	8.1 ± 0.2	8.1 ± 0.4	8.9 ± 0.4	9.3 ± 0.3	7.5 ± 0.5
	aNav1.5	6.1 ± 0.3	6.8 ± 0.4	6.7 ± 0.2	7.9 ± 0.3	8.6 ± 0.5	6.5 ± 0.4
Inactivation V1/2 (mV)	nNav1.5	−91.5 ± 1.5	—	—	−90.2 ± 1.2	—	−91.8 ± 2.0
	aNav1.5	−90.5 ± 0.8	—	—	−91.2 ± 0.5	—	−90.8 ± 0.8
Inactivation k	nNav1.5	−5.9 ± 0.2	—	—	−5.6 ± 0.2	—	−6.1 ± 0.3
	aNav1.5	−6.3 ± 0.2	—	—	−6.2 ± 0.2	—	−6.2 ± 0.1
V0 (mV)	nNav1.5	21.2 ± 0.8	22.9 ± 3.7	17.0 ± 1.6	23.0 ± 3.0	23.1 ± 5.0	18.1 ± 0.7
	aNav1.5	18.8 ± 1.7	18.5 ± 3.8	19.1 ± 1.9	21.5 ± 4.0	18.0 ± 3.9	19.2 ± 1.7
Tpeak at 0 mV (ms)	nNav1.5	0.63 ± 0.04	0.65 ± 0.04	0.67 ± 0.04	0.69 ± 0.06	0.72 ± 0.06	0.65 ± 0.04
	aNav1.5	0.51 ± 0.02	0.55 ± 0.01	0.48 ± 0.03	0.56 ± 0.03	0.58 ± 0.02	0.49 ± 0.03
τfast at 0 mV (ms)	nNav1.5	0.72 ± 0.05	—	—	0.77 ± 0.02	—	0.76 ± 0.04
	aNav1.5	0.63 ± 0.02	—	—	0.69 ± 0.03	—	0.68 ± 0.02
τslow at 0 mV (ms)	nNav1.5	5.79 ± 0.35	—	—	6.06 ± 0.40	—	5.88 ± 0.27
	aNav1.5	5.21 ± 0.23	—	—	5.79 ± 0.38	—	5.16 ± 0.21
τrecovery (ms)	nNav1.5	21.3 ± 1.2	—	—	23.8 ± 1.3	—	22.8 ± 1.8
	aNav1.5	22.4 ± 0.9	—	—	26.3 ± 1.3	—	24.3 ± 1.1

Effects of 50–500 μM Cd²⁺ are shown. For each parameter (given rows), data are given for nNav1.5 (above) and aNav1.5 (below). Data are presented as mean ± SEM (n ≥ 11). There was no statistically significant difference between the respective pairs (p > 0.05; unpaired Student's t-tests). Recovery data are shown for the values obtained from the experiment with 250 μM Cd²⁺.

G_max, maximal Na⁺ conductance; k, slope factor; SEM, standard error of the mean; T_peak, time to peak; V₀, coefficient describing the voltage dependence of time to peak; V_1/2, half-maximal voltage; V_peak, voltage at which Na⁺ current is maximal; V_thres, threshold voltage for activation; τ_fast, fast inactivation decay constant; τ_recovery, recovery from inactivation time constant. τ_slow, slow inactivation decay constant.

Table 3.

Effects of Gd³⁺ on Nav1.5 Neonatal and Adult Splice Variants

Parameter	Splice variant	Control	50 μM	100 μM	Recovery
Relative Gmax (%)	nNav1.5	100.0 ± 0.0	64.2 ± 4.0	57.3 ± 3.2	92.4 ± 10.0
	aNav1.5	100.0 ± 0.0	67.0 ± 5.3	57.7 ± 7.7	93.4 ± 9.8
Vthres (mV)	nNav1.5	−53.3 ± 0.7	−42.2 ± 1.2	−36.6 ± 1.7	−51.1 ± 1.3
	aNav1.5	−59.4 ± 1.1	−48.1 ± 1.4	−42.8 ± 1.1	−58.4 ± 1.4
Vpeak (mV)	nNav1.5	−4.4 ± 2.1	9.4 ± 1.9	10.6 ± 2.2	−0.7 ± 2.8
	aNav1.5	−20.0 ± 2.1	−3.7 ± 1.6	3.8 ± 2.5	−8.8 ± 4.1
Activation V1/2 (mV)	nNav1.5	−24.7 ± 1.6	−11.4 ± 1.5	−8.1 ± 2.1	−21.4 ± 1.9
	aNav1.5	−35.9 ± 1.8	−23.3 ± 1.6	−16.7 ± 1.7	−31.4 ± 2.7
Activation k	nNav1.5	7.7 ± 0.5	9.6 ± 0.5	9.7 ± 0.5	8.7 ± 0.6
	aNav1.5	5.8 ± 0.5	7.9 ± 0.4	8.8 ± 0.5	7.6 ± 0.7
Inactivation V1/2 (mV)	nNav1.5	−87.0 ± 1.1	—	−77.4 ± 2.0	−83.0 ± 1.8
	aNav1.5	−87.4 ± 1.8	—	−78.0 ± 2.3	−83.7 ± 0.8
Inactivation k	nNav1.5	−5.5 ± 0.1	—	−6.4 ± 0.6	−6.1 ± 0.2
	aNav1.5	−5.5 ± 0.2	—	−5.4 ± 0.3	−5.9 ± 0.2
V0 (mV)	nNav1.5	18.4 ± 1.8	19.7 ± 2.2	21.8 ± 4.9	20.0 ± 3.1
	aNav1.5	17.6 ± 2.3	19.9 ± 2.0	19.6 ± 1.8	19.7 ± 2.0
Tpeak at 0 mV (ms)	nNav1.5	0.86 ± 0.06	1.42 ± 0.15	1.65 ± 0.23	1.01 ± 0.13
	aNav1.5	0.64 ± 0.03	0.91 ± 0.04	1.06 ± 0.08	0.70 ± 0.04
τfast at 0 mV (ms)	nNav1.5	1.17 ± 0.15	1.83 ± 0.28	1.78 ± 0.25	1.03 ± 0.27
	aNav1.5	0.75 ± 0.07	1.19 ± 0.12	1.38 ± 0.26	0.81 ± 0.12
τslow at 0 mV (ms)	nNav1.5	7.44 ± 0.70	9.99 ± 1.18	11.58 ± 1.52	7.79 ± 0.66
	aNav1.5	4.91 ± 0.27	7.65 ± 0.51	7.11 ± 1.71	5.23 ± 0.31
τrecovery (ms)	nNav1.5	17.8 ± 1.3	—	12.0 ± 2.2	17.5 ± 1.7
	aNav1.5	16.5 ± 1.8	—	11.5 ± 1.3	12.5 ± 2.1

Effects of 50 and 100 μM Gd³⁺ are shown. For each parameter (given rows), data are given for nNav1.5 (above) and aNav1.5 (below). Data are presented as mean ± SEM (n ≥ 7). There was no statistically significant difference between the respective pairs (p > 0.05; unpaired Student's t-tests). Recovery data are shown for the values obtained from the experiment with 100 μM Cd²⁺.

Effects of Cd²⁺

G_max and voltage dependence of activation

Effects of Cd²⁺ were studied in the concentration range from 50 to 500 μM. At all concentrations tested, Cd²⁺ blocked I_Na. This effect reached steady state within <20 s and recovery from block upon washout was complete within <40 s even at 500 μM (Fig. 1A, B). I-V and G-V relationships for both Nav1.5 variants revealed that I_Na and G_max were dose dependently blocked by Cd²⁺ over the concentration range used (Fig. 1C, D). Cd²⁺ was equally potent in blocking G_max of nNav1.5 and aNav1.5 with statistically similar half-inhibitory concentrations (IC₅₀s) (Table 1). Cd²⁺ block of I_Na was only weakly voltage dependent (not shown), consistent with the Cd²⁺-induced depolarization of activation being limited. For example, Cd²⁺ (500 μM), which reduced G_max by ∼80%, depolarized activation-V_1/2 similarly and by only 7.7 ± 1.3 and 10.3 ± 0.6 mV for nNav1.5 and aNav1.5, respectively (p < 0.01 vs. control for both; p = 0.09 between the variants). Positive shifts in V_thres and V_peak induced by ≥250 μM Cd²⁺ were statistically significant for the two variants (p < 0.05 vs. control for both) (Table 2). Finally, activation k was increased dose dependently by Cd²⁺; this effect was statistically significant for [Cd²⁺] ≥ 250 μM for both Nav1.5 variants (p < 0.05 vs. controls; Table 2). For all activation parameters described (V_1/2, k, V_thres, and V_peak), Cd²⁺ application produced statistically similar effects for nNav1.5 versus aNav1.5 (p > 0.05 for all comparisons; Table 2).

FIG. 1.

Effects of Cd²⁺: Time course of current block, effects on peak conductance, and voltage dependence of activation. (A, B) Typical whole-cell currents for nNav1.5 (A) and aNav1.5 (B), before and after 100 and 500 μM Cd²⁺ application, elicited by 12 ms depolarizing pulses to −20 mV at 0.1 Hz; holding potential was −100 mV. Traces shown were recorded at times (a–e) indicated correspondingly in parts (C, D). Cd²⁺ block was fully reversible upon washout. Dashed horizontal lines indicate zero current. (C, D) Time course of onset and offset of Cd²⁺ block of currents for nNav1.5 (C) and aNav1.5 (D) (n = 6 and 5, respectively). a, control; b, 100 μM Cd²⁺; c, washout (for 100 μM Cd²⁺); d, 500 μM Cd²⁺; e, washout (for 500 μM Cd²⁺). (E, F) Normalized I-V relationships for nNav1.5 (E) and aNav1.5 (F); data are shown for control, 50–500 μM Cd²⁺, and recovery (symbols defined at bottom) (n ≥ 11). Data were normalized relative to the maximal control current (I/Imax_Control). Insets, corresponding normalized conductance transforms (G/Gmax_Control). Recovery was within 10% of controls. (G, H) Normalized G-V relationships (G/Gmax) for control, 50–500 μM Cd²⁺, and recovery (symbols defined at bottom) (n ≥ 11). Data (mean ± SEM) were fitted with Boltzmann functions for nNav1.5 (G) and aNav1.5 (H). Some error bars are smaller than symbols. SEM, standard error of the mean.

Gating kinetics, availability, and recovery from inactivation

Cd²⁺ had no effect on T_peak for either Nav1.5 variant (p > 0.05 vs. controls; Table 2). Inactivation τ_fast and τ_slow were also unaffected by Cd²⁺ (250 μM) treatment (p > 0.05 vs. controls; Table 2). With regard to steady-state inactivation, V_1/2 and k values remained unchanged by application of Cd²⁺ (250 μM) for either Nav1.5 variant (p > 0.05 vs. controls; Table 2). The time constant of recovery from inactivation (τ_recovery) was significantly but similarly slowed by Cd²⁺ (250 μM) for nNav1.5 and aNav1.5 (p < 0.05 vs. controls, p = 0.36 for the variants; Table 2).

In conclusion, none of the parameters studied showed any difference between nNav1.5 and aNav1.5 for the effects of Cd²⁺.

Effects of Gd³⁺

G_max and voltage dependence of activation

Effects of Gd³⁺ were studied within the concentration range from 1 to 5000 μM. Gd³⁺ demonstrated a dose-dependent block of I_Na (Fig. 2A). The block stabilized rapidly, but recovery was slower and incomplete, reaching only ∼80% after 4 min of washout even at a midrange concentration (Fig. 2B). Logistic fits to dose/response relationships for I_Na block at −20 mV yielded statistically similar IC₅₀ values for nNav1.5 and aNav1.5 (Fig. 2C and Table 1). The incomplete recovery from Gd³⁺ block, which has been reported previously for VGSCs,¹² prevented the study of the effects of multiple Gd³⁺ concentrations on given cells. Thus, the remainder of the characterization was performed at only 50 and 100 μM (Figs. 3 and 4). Families of I_Na for control and Gd³⁺ are shown in Figure 3A. Gd³⁺ block of I_Na was voltage dependent with significantly larger suppression at negative potentials (Fig. 3B). This effect was the same for both nNav1.5 and aNav1.5 and could be due to accompanying positive shifts in activation voltage (Fig. 3C, D).

FIG. 2.

Effects of Gd³⁺: Time course and dose dependence of current block. (A, B) Typical whole-cell currents for nNav1.5 (A) and aNav1.5 (B), before and after 20–500 μM Gd³⁺. Currents were elicited by 12 ms depolarizing pulses to −20 mV at 0.1 Hz; holding potential was −100 mV. Dashed horizontal lines indicate zero current. Effects were only partially reversible. (C) Time course of onset and offset of current block by 50 μM Gd³⁺ for nNav1.5 (open circles) and aNav1.5 (black circles) (n = 7 and 8, respectively). Again, recovery after washing was incomplete (∼80%) for both Nav1.5 variants. (D) Dose dependence of Gd³⁺ block (1–5000 μM) of peak Na⁺ current at −20 mV (n ≥ 7). Current blockage was similar for nNav1.5 and aNav1.5 at each Gd³⁺ concentration tested (p > 0.05; unpaired t-tests). Dashed (nNav1.5) and solid (aNav1.5) lines represent logistic fits to the data points (mean ± SEM; some error bars are smaller than symbols).

FIG. 3.

Effects of Gd³⁺: Peak conductance and voltage dependence of activation. (A) Typical effects of 50 and 100 μM Gd³⁺ on whole-cell currents for nNav1.5 (i) and aNav1.5 (ii). Currents were elicited by 60 ms depolarizing pulses to between −80 and +45 mV from a holding potential of −100 mV; interpulse duration was 2 s. (B) Bar diagram illustrating voltage dependence of current block by 100 μM Gd³⁺; blockage (%) of peak current is plotted as a function of voltage for nNav1.5 (gray bars) and aNav1.5 (black bars) (n = 8 for both). (C, D) Normalized I-V relationships for nNav1.5 (C) and aNav1.5 (D) obtained under control (pretreatment and recovery) conditions and during application of 50 and 100 μM Gd³⁺. Data were normalized relative to the maximal control current (I/Imax_Control). Insets, corresponding normalized conductance transforms (G/Gmax_Control). (E, F) Normalized G-V relationships (G/Gmax) obtained under control (pretreatment and recovery) conditions and during application of 50 and 100 μM Gd³⁺. Data were fitted with Boltzmann functions for nNav1.5 (E) and aNav1.5 (F). All data are shown as mean ± SEM (n ≥ 7). Some error bars are smaller than symbols.

FIG. 4.

Effects of Gd³⁺: Gating kinetics. (A, B) Time to peak (T_peak), plotted as a function of test potential and fitted to single exponential functions for control (pretreatment and recovery) conditions and during application of 50 and 100 μM Gd³⁺ for nNav1.5 (A) and aNav1.5 (B). (C, D) Inactivation τ_fast and τ_slow, determined from double exponential fits to normalized current decays for control and 50 μM Gd³⁺, plotted as a function of test potential for nNav1.5 (C) and aNav1.5 (D). Data points are connected with straight lines for clarity. All data are shown as mean ± SEM (n ≥ 7); some error bars are smaller than symbols. Significance (paired t-tests vs. control): *p < 0.05, **p < 0.01.

The IC₅₀ for Gd³⁺ block of I_Na was determined at −20 mV to be ∼50 μM (Table 1). This is likely to be an overestimate for more negative potentials and an underestimate for more positive potentials. For instance, on the basis of the blockage profile shown in Figure 3B and respective Hill coefficients of 1.05 and 0.92 obtained at −20 mV for nNav1.5 and aNav1.5 (Table 1), the IC₅₀s in the range −45 to −25 mV would be re-estimated as <25 μM for both Nav1.5 variants. Conversely, the potency of Gd³⁺ was much lower at potentials >10 mV, the estimated IC₅₀ values being >100 μM. Li and Baumgarten have described a similar voltage dependency of IC₅₀ estimations for Gd³⁺ block of cardiac sodium currents.¹²

Gd³⁺ had two effects on I_Na: G_max was suppressed and voltage dependence of activation was significantly depolarized (Fig. 3C, D). Gd³⁺ (50 μM) blocked G_max of nNav1.5 and aNav1.5 equally (Fig. 3C, D, and Table 3). The same Gd³⁺ concentration depolarized V_thres, V_peak, and activation-V_1/2 by 10–15 mV (p < 0.05 vs. controls; Table 3). However, there was no difference between the two splice variants (Table 3). In addition, activation k became significantly less steep with Gd³⁺ and this effect was also similar for both variants (p < 0.01 vs. controls; Table 3).

Kinetics of activation and inactivation

Activation kinetics was greatly slowed by both Gd³⁺ concentrations tested (Fig. 4A, B). In the presence of Gd³⁺, T_peak at 0 mV nearly doubled (Table 3). These effects were highly significant but not different between nNav1.5 and aNav1.5 (p < 0.01 vs. controls; Table 3). Voltage dependence of T_peak (V₀) was not altered by Gd³⁺ in both cases (Table 3). With regard to inactivation kinetics, Gd³⁺ (50 μM) slowed τ_fast and τ_slow significantly in the voltage range −15 to 10 mV; there was no difference between the two variants (Fig. 4C, D, and Table 3).

Availability and recovery from inactivation

Gd³⁺ (100 μM) depolarized inactivation—V_1/2, significantly in both Nav1.5 variants, while the slope factor was not affected (p < 0.01 vs. controls; Fig. 5A, C, E, and Table 3). Interestingly, the Gd³⁺-induced shifts in inactivation-V_1/2 were smaller compared with activation (Table 3). Gd³⁺ also significantly quickened recovery from inactivation overall (Fig. 5B, D, F); the associated time constant (τ_recovery) was similar for both Nav1.5 variants (p < 0.05 vs. controls; Table 3).

FIG. 5.

Effects of Gd³⁺: Steady-state inactivation and recovery from inactivation. (A) Typical whole-cell nNav1.5 currents used to assess voltage dependence of steady-state inactivation, in control solution and 100 μM Gd³⁺. (B) Superimposed whole-cell nNav1.5 currents used to assess kinetics of recovery from inactivation in (i) control solution and (ii) 100 μM Gd³⁺. Arrows indicate Gd³⁺-induced acceleration of current recovery. (C, E) Availability/voltage relationships in control solution and 100 μM Gd³⁺ fitted with Boltzmann functions for nNav1.5 (C) and aNav1.5 (E). (D, F) Current recovery (I_t/I_c) plotted as a function of time interval and fitted with single exponential functions for nNav1.5 (D) and aNav1.5 (F). Data are mean ± SEM (n ≥ 6). Significance (paired t-tests vs. control): *p < 0.05, **p < 0.01.

In conclusion, the effects of Gd³⁺ on channel block and kinetics were similar for nNav1.5 and aNav1.5.

Discussion

The present study aimed to extend our previous work on effects of H⁺ on “adult” versus “neonatal” splice variants of Nav1.5 to cations of higher valency (Cd²⁺ and Gd³⁺). The main findings were as follows. (1) Both the divalent and trivalent cations inhibited peak I_Na through reducing G_max and inducing a positive shift in voltage dependence of activation. (2) There was no splice variant selective effect of Cd²⁺ or Gd³⁺ on any of the parameters tested. (3) Effects of Gd³⁺ on channel gating were qualitatively similar, but distinct from Cd²⁺. In particular, Cd²⁺ was less potent in depolarizing the voltage dependence of activation and, unlike Gd³⁺, had no effect on voltage dependence of inactivation, time course of recovery from inactivation, and kinetics of activation and inactivation. These results are discussed in relation to possible mode(s) of action of the cations and the charge difference between nNav1.5 versus aNav1.5.

Effects of Cd²⁺

The effects of Cd²⁺ on Nav1.5 were generally inhibitory, irrespective of the splicing, and included the following: (1) reduction of G_max, (2) weak voltage dependence, (3) slowing of activation and but not inactivation kinetics, and (4) slowing of time course of recovery from inactivation. These were significantly different to the effects of Gd³⁺ (see section of Effects of Gd³⁺) and H⁺.⁷ Thus, although Cd²⁺ reduced G_max, it was significantly less potent in modifying the parameters associated with channel gating. This would suggest that a binding site for Cd²⁺ is away from the primary voltage sensors in S4. Indeed, such a site (a Cys residue) has been identified within domain I/segments 5–6.¹³ Overall, our results agree with the notion that tetrodotoxin-resistant VGSCs are more sensitive to blockage by Cd²⁺ (and Zn²⁺) than tetrodotoxin (TTX)-sensitive channels and these cations may share the same binding site as TTX itself.^8,14 In contrast, H⁺ and Gd³⁺, which induced substantial shifts in voltage dependence of activation, may bind near the S4 voltage sensors.^12,15

In agreement with our findings here, Cd²⁺ caused depolarizing shifts in the conductance/voltage relationship in cardiac Purkinje cells but had no effect on steady-state inactivation.¹⁵ In addition, the block was voltage independent,¹⁶ as has also been observed for the effect of Cd²⁺ in ventricular myocytes.¹⁷ However, some differences in the effects of Cd²⁺ have also been observed. For example, in our study, Cd²⁺ had no effect on time to peak or steady-state inactivation. On the contrary, in cardiac Purkinje cells, the former was slowed, while the latter was shifted to depolarized potentials.¹⁵ Cd²⁺ block was also found to be voltage dependent for both cardiac and skeletal muscle VGSCs.^13,14 The effects of other divalent cations (including Ba²⁺, Ca²⁺, Co²⁺, Mg²⁺, Mn²⁺, Ni²⁺, and Zn²⁺), studied in cardiac Purkinje cells and/or ventricular myocytes, were similar to Cd²⁺ in causing depolarizing shifts in the conductance/voltage relationship. Some cations, however, were found also to shift steady-state inactivation to depolarized potentials, unlike for Cd²⁺.^15,17,18 In addition, while channel block was voltage dependent for Ba²⁺, Co²⁺, Mg²⁺, and Mn²⁺, it was voltage independent for Zn²⁺ and Hg²⁺.^16,17 In contrast, Zn²⁺ blockage of brain VGSCs was voltage dependent.¹⁹

In conclusion, Cd²⁺ blocked Nav1.5 but did not differentiate between the two splice variants. The main effects included a reduction of G_max, slowing of activation, but not inactivation kinetics, and a slowing of time course of recovery from inactivation. These effects were weakly voltage dependent. Overall, the effects of divalent cations on VGSCs appear to depend on the cation itself as well as the subtype of channel and the cell type under investigation.

Effects of Gd³⁺

The effects of Gd³⁺ on Nav1.5 included the following: (1) reduction of G_max, (2) depolarization of activation voltage and channel availability, (3) slowing of activation and inactivation kinetics, and (4) speeding of time course of recovery from inactivation. Thus, Gd³⁺ proved to be a potent inhibitor of Nav1.5 activity irrespective of the splice variant. Previous studies of effects of Gd³⁺ were conducted in native cardiomyoytes^12,20 and VGSCs in myelinated axons of Xenopus laevis.⁹ On the whole, as reported in the accompanying article for H⁺,⁷ Gd³⁺ effects on VGSCs can be dissected into the following two mechanisms.^15,21

Inhibition of Na⁺ permeation

The blocking efficacy of Gd³⁺ was similar in cardiomyocytes and the Nav1.5 variants. For instance, 50 μM Gd³⁺ reduced G_max of rabbit ventricular cardiomyocytes by 27%,¹² which is comparable with an ∼33% reduction for the Nav1.5 variants found in the present study. The IC₅₀ values for I_Na block were also comparable: 48 μM in cardiomyocytes¹² and ∼55 μM for the Nav1.5 variants (Table 1). Thus, taken together, the available data suggest that the Gd³⁺ sensitivity of Na⁺ permeation is similar in different VGSCs.

Modification of channel gating

Modification of various aspects of a/nNav1.5 gating by Gd³⁺ was different from past studies on cardiomyocytes. For instance, in cardiomyocytes, 50 μM Gd³⁺ depolarized V_1/2 by 7.9 mV,¹² which was considerably larger for nNav1.5 and aNav1.5 in the present study (13.3 and 12.6 mV, respectively). In addition, we detected a significant increase in activation slope factor by 50 μM Gd³⁺ application, but no such effect was reported in cardiomyocytes. Furthermore, 100 μM Gd³⁺ was found to markedly accelerate the rate of recovery from inactivation in Nav1.5 variants, while no such effect of 50 μM Gd³⁺ on this parameter has been reported in rabbit cardiomyocytes.¹² Species, recording techniques, and/or solutions could account for the quantitative differences in the effects of Gd³⁺. On the contrary, the qualitative differences reported (e.g., effect vs. no-effect on recovery from inactivation) would suggest strongly that such effects of Gd³⁺ may be dependent upon the individual VGSCs. Here, the effects of Gd³⁺ (and H⁺, as described in the accompanying article⁷) on the parameters associated with channel gating were found to be greater than for Cd²⁺.

When all the available data are taken together, effects of Gd³⁺ on Na⁺ conductance appear to be similar across different VGSCα isoforms, whereas the effects on channel gating may be isoform dependent. This is consistent with Gd³⁺ interacting with multiple sites on VGSCs, and distinct “blocking” and “modulatory” sites mediate its particular actions, that is, block of conductance versus depolarization of activation.⁹ Similar to other cations, shielding of negative surface charges by Gd³⁺ can account for only some of the complex effects on channel gating, and unidentified binding sites must exist. With regard to block of Na⁺ conductance, the molecular identities of Gd³⁺-interacting sites within the pore are also not known. Uniform block levels observed in different cell types would strongly suggest that Gd³⁺ interacts with residues that are highly conserved across VGSCα subtypes.^9,12 Moreover, weak voltage dependence of open-channel block by Gd³⁺ (assuming that shifts in activation gating are largely responsible for the apparent voltage dependence of Gd³⁺ block of I_Na) indicates that the receptor site for Gd³⁺ is possibly located close to the external surface of the channel, for example, outer ring carboxylates and nearby residues forming the outer vestibule. Interestingly, a negative shift in Na⁺ reversal potential upon Gd³⁺ application has been reported in frog myelinated axons, suggesting that selectivity filter residues may also be part of the “receptor” site for Gd³⁺.⁹ However, we and others did not find any change in the Na⁺ reversal potential in Nav1.5 variants or cardiomyocytes, and determination of whether Gd³⁺ alters Na⁺ selectivity in VGSCs requires further investigation.¹²

Finally, it is of interest to compare our findings using Gd³⁺ with another trivalent cation, La³⁺. In rat dorsal root ganglia neurons, La³⁺ blocked VGSC peak current and slowed activation and inactivation kinetics in agreement, similar to our findings.¹⁰ Also, similar to our results, in hippocampal CA1 neurons, steady-state activation was shifted to more positive potentials in agreement with our findings.²² On the contrary, dissimilar to our findings, La³⁺ had no effect on steady-state inactivation and block was not voltage dependent in the range studied in dorsal root ganglia,¹⁰ hippocampus peak current was increased in a voltage-independent manner, and recovery from inactivation was shifted to more negative potentials.²²

Thus, Gd³⁺ blocked Nav1.5 in a voltage-dependent manner but did not differentiate between “adult” and “neonatal” Nav1.5. The main effects included a reduction of G_max, depolarization of activation voltage and channel availability, slowing of activation and inactivation kinetics, and a speeding up of time course of recovery from inactivation. It is also apparent that the effects of Gd³⁺ on channel gating may be both isoform and cell-type dependent, but these may depend on the type of trivalent cation.

Conclusion

In overall conclusion, the VGSC blockage by cations as a whole is complex. It would appear that these effects are both cation specific and, in part, VGSC-subtype specific. Importantly, charge alone cannot solely explain the observed cationic effects. On the whole, voltage dependence of channel block depends not just on valency but also on the specific cation, subtype of VGSC, and the cell type involved.^{10,16,17,22–24} In fact, it is likely that the effect of a given cation will also depend on two additional factors: (1) binding site(s) relative to the S4 voltage sensors, that is, effective Debye length¹² and (2) membrane partition and possible access to the channel's “interior” as well as its fenestrated structure.²⁵ Further work is needed to elucidate these aspects.

In terms of cancer, the results of the present study and our earlier article, taken together, could provide clues to cancer-specific drug development. Here, we show that the electrophysiological effects of Cd²⁺ and Gd³⁺ studied were essentially the same for “neonatal” and “adult” Nav1.5 (although these splice variants possess marked differences in their external charges). In contrast, previously, we showed that the effects of H⁺ on the two variants were significantly differential.⁷ Since electrostatic interaction can play a fundamental role in drug-receptor interaction, our results taken together would suggest that limited adjustment of the charged structure of pharmacological agents could enable selective targeting of nNav1.5, which is expressed in several cancers and promotes metastasis.

Authors' Contribution

M.B.A.D. and R.O. conceived the experiments. S.P.F. and R.O. performed the experiments and the analyses. All authors contributed to writing the article. All coauthors have reviewed and approved of the article before submission.

Author Disclosure Statement

M.B.A.D. is involved in a spinout company (Celex Oncology Ltd) focused on ion channels and cancer. S.P.F. and R.O. declare no competing financial interests.

Funding Information

We thank the Pro Cancer Research Fund (PCRF) for continuous support (M.B.A.D.; S.P.F.). R.O. was funded by a PhD studentship from the British Heart Foundation (BHF).