Differential regulation of cardiac sodium channels by intracellular fibroblast growth factors

Abstract

Voltage-gated sodium (NaV) channels are responsible for the initiation and propagation of action potentials. In the heart, the predominant NaV1.5 α subunit is composed of four homologous repeats (I-IV) and forms a macromolecular complex with multiple accessory proteins, including intracellular fibroblast growth factors (iFGF). In spite of high homology, each of the iFGFs, iFGF11-iFGF14, as well as the individual iFGF splice variants, differentially regulates NaV channel gating, and the mechanisms underlying these differential effects remain elusive. Much of the work exploring iFGF regulation of NaV1.5 has been performed in mouse and rat ventricular myocytes in which iFGF13VY is the predominant iFGF expressed, whereas investigation into NaV1.5 regulation by the human heart-dominant iFGF12B is lacking. In this study, we used a mouse model with cardiac-specific Fgf13 deletion to study the consequences of iFGF13VY and iFGF12B expression. We observed distinct effects on the voltage-dependences of activation and inactivation of the sodium currents (INa), as well as on the kinetics of peak INa decay. Results in native myocytes were recapitulated with human NaV1.5 heterologously expressed in Xenopus oocytes, and additional experiments using voltage-clamp fluorometry (VCF) revealed iFGF-specific effects on the activation of the NaV1.5 voltage sensor domain in repeat IV (VSD-IV). iFGF chimeras further unveiled roles for all three iFGF domains (i.e., the N-terminus, core, and C-terminus) on the regulation of VSD-IV, and a slower time domain of inactivation. We present here a novel mechanism of iFGF regulation that is specific to individual iFGF isoforms and that leads to distinct functional effects on NaV channel/current kinetics.

Figure S1.

Targeted disruption of the Fgf13 locus and generation/validation of cardiac-specific Fgf13 targeted deletion mice. (A) Schematics of the endogenous Fgf13 locus (top), located on the X chromosome, and the floxed Fgf13 locus (middle). Following sequencing of the first and second generation of offspring, heterozygous Fgf13 floxed (Fgf13^fl/+) female mice and hemizygous Fgf13 floxed (Fgf13^fl/y) male mice were crossed to also generate homozygous Fgf13 floxed (Fgf13^fl/fl) females. The Fgf13^fl/y and Fgf13^fl/fl animals were then crossed with transgenic mice expressing Cre recombinase driven by the cardiac-specific α-MHC promoter. Crossing the Cre recombinase positive, hemizygous floxed male (Fgf13^fl/y) offspring with Fgf13^fl/fl females (or Cre recombinase heterozygous Fgf13^fl/+ females with Fgf13^fl/y males) provided cardiac-specific Fgf13 targeted deletion hemizygous male and homozygous female animals, referred to collectively here as cardiac-specific Fgf13 knockouts, cFgf13KO. Three sets of primers were employed to detect the 5′ LoxP site, the 3′ LoxP site, and the α-MHC-driven Cre-recombinase (see Materials and methods). (B) Representative results from PCR analyses of homozygous Fgf13^fl/fl, heterozygous Fgf13^fl/+, and WT (^+/+) female mice using primers targeting the 5′ LoxP site (left) and the 3′ LoxP site (middle) are shown. Further screening of these animals with primers to the α-MHC–Cre-recombinase transgene identified Cre recombinase positive animals (right). (C) Western blot analyses confirmed loss of iFGF13 protein expression. Protein lysates (1 µg), prepared from WT, cFgf13KO, and Fgf12 targeted deletion (Fgf12KO) ventricles, were fractionated, transferred, and membranes were probed with an anti-iFGF13 antibody, as described in Materials and methods. A prominent band at ∼32 kD and a faint band at ∼24 kD were detected with the anti-iFGF13 antibody in the WT (left and right panels) and Fgf12KO (right panel) LV protein samples, whereas these bands were not detected in the cFgf13KO LV samples (left panel), consistent with the elimination of iFGF13 proteins. Source data are available for this figure: SourceData FS1.

Figure S2.

Loss of Fgf13 or Fgf12 does not measurably affect the expression levels of other Fgf transcripts in adult mouse LV. The expression levels of the Fgf11, Fgf12A, Fgf12B, Fgf13S(A), Fgf13VY, Fgf13U(B), Fgf14A, and Fgf14B transcripts in WT, cFgf13KO, and Fgf12KO LV samples were measured and normalized to Hprt expression in the same sample, as described in Materials and methods. These analyses revealed that the Fgf13 transcripts are undetectable in cFgf13KO LV. In addition, the mean ± SEM relative expression levels of the other Fgf transcripts in cFgf13KO and WT LV are indistinguishable.

Figure 1.

Cardiac-specific deletion of Fgf13 (cFgf13KO) alters inactivation, but not activation, of I_Na in LV myocytes. (A) Representative I_Na waveforms recorded from LV myocytes isolated from WT (black) and cardiac-specific Fgf13 deletion (cFgf13KO; orange) mice; the voltage-clamp protocol is illustrated below the current records. (B) Mean ± SEM normalized conductances (G) of I_Na activation and inactivation were plotted as a function of voltage and fitted to single Boltzmann functions (solid lines). There is a marked hyperpolarizing shift in the voltage-dependence of I_Na inactivation in cFgf13KO (orange) compared with WT (black) LV myocytes (P value <0.001; Table 1), whereas the voltage-dependences of I_Na activation are indistinguishable in cFgf13KO and WT LV myocytes (P value = 0.12; Table 1). Control experiments revealed that the voltage-dependences of I_Na activation and inactivation in Fgf13 floxed (grey) LV myocytes are indistinguishable from WT LV myocytes (P values are 0.11 and 0.97, respectively; Table 1). It was not possible, however, to conduct additional control experiments to independently determine if there are any of the previously reported cardiotoxic effects (Pugach et al., 2015) of the prolonged expression of Cre-recombinase in our cFgf13KO mice. (C) Analysis of I_Na recovery from inactivation, examined using a three-step protocol as described in Materials and methods, reveals that recovery follows a biexponential time course (see Table 1). The decay phases of I_Na in all WT adult mouse LV myocytes were best fitted with the sum of two exponentials as described in Materials and methods. (D) The mean ± SEM (n = 19) fast (Ƭ_f) and slow (Ƭ_s) time constants of I_Na decay determined in WT cells are plotted as a function of the test potentials in D. As is evident, neither time constant displays any appreciable voltage dependence (P value = 0.26; Table 1). Inset: Overlaying representative I_Na waveforms recorded from cFgf13KO and WT LV myocytes reveal that current decay is accelerated in cFgf13KO compared with WT cells. Indeed, in 13 (of 26) cFgf13KO LV myocytes, the decay phases of I_Na were best described by single exponentials characterized (see text near Fig. 1 D citation) by Ƭ values indistinguishable from Ƭ_f determined in WT cells, i.e., the slow component of I_Na decay was undetectable in these 13 cells. For the remaining 13 (of 26) cFgf13KO LV myocytes, I_Na decay was well described by two exponentials, with Ƭ_f and Ƭ_s values similar to those determined in WT cells at all test potentials (see values in Table S2). (E) The mean ± SEM fractional amplitudes of the slow component of I_Na decay (%A_slow) in these (13 of 26) cFgf13KO LV myocytes, however, were significantly lower than those determined in WT cells at all test potentials (see values in Table S2).

Figure S3.

The amplitudes/densities of I_Na in mouse LV myocytes are not affected by the loss of iFGF13, and the properties of I_Na in mouse LV myocytes are not affected by the loss of iFGF12. (A) The cardiac-specific deletion of Fgf13 does not affect peak I_Na densities in LV myocytes: peak I_Na densities at all voltages in WT, Fgf13 floxed, and cFgf13KO mouse LV myocytes are not significantly different (P values >0.15 at all voltages). (B and C) Consistent with the negligible expression of Fgf12 transcripts in adult mouse LV (see Fig. S2), peak I_Na densities at all voltages (B) and the voltage-dependences of I_Na activation and inactivation (C) determined in adult Fgf12KO LV myocytes are indistinguishable (P values >0.13) from I_Na in WT LV myocytes at all voltages.

Figure S4.

The time-dependence of AAV9-mediated expression of iFGF12B and eGFP in adult mouse LV and the validation of the anti-iFGF12 antibody. (A and B) Western blots of fractionated protein lysates prepared from cFgf13KO ventricles 2, 3, or 4 wk following retro-orbital injections of a 1:1 mixture of the hFGF12B-expressing and eGFP-expressing AAV9 viruses, probed with an anti-EGFP (A) or anti-iFGF12 (B) antibody, as described in Materials and methods. (C) To validate the anti-iFGF12 antibody used in the Western blot in B, protein lysates prepared from WT and Fgf12KO adult mouse left and right atria were fractionated, transferred, and the membranes were probed with an anti-iFGF12 antibody, as described in Materials and methods. A prominent band at ∼20 kD was detected with the anti-iFGF12 antibody in the WT, but not in the Fgf12KO, atrial protein samples, consistent with the elimination of iFGF12 proteins and validating the anti-iFGF12 antibody for Western blot analyses of iFGF12 protein expression. Source data are available for this figure: SourceData FS4.

Figure 2.

Expression of iFGF12B in cFgf13KO LV myocytes alters the voltage-dependence of I_Na activation and inactivation. (A) Representative I_Na waveforms recorded from cFgf13KO LV myocytes without (orange) and with (blue) iFGF12B expression; the voltage-clamp protocol is illustrated below the current records. (B) The expression of iFGF12B in cFgf13KO LV myocytes resulted in depolarizing shifts in I_Na activation and inactivation (blue) compared with cFgf13KO LV myocytes (orange; P values are <0.001 for both G-V activation and inactivation; Table 1). (C) Expression of iFGF12B in cFgf13KO LV myocytes also accelerated I_Na recovery from inactivation compared with the currents in both WT and cFgf13KO LV myocytes (P value = 0.04; Table 1). (D) The decay phases of the currents in cFgf13KO cells expressing iFGF12B (blue), however, are indistinguishable from cFgf13KO (orange) LV myocytes (inset). In addition, in 13 (of 19) cFgf13KO + iFGF12B-expressing LV myocytes, the decay phases of I_Na were best described by single exponentials with Ƭ values indistinguishable from Ƭ_f determined in WT cells, i.e., the slow component of I_Na decay was undetectable in these 13 cells. In the remaining (6 of 19) cFgf13KO + iFGF12B-expressing LV myocytes, I_Na decay was well described by the sum of two exponentials (D) with Ƭ_f and Ƭ_s values similar to those in cFgf13KO and WT (Fig. 1 D) cells (see values in Table S2). (E) In addition, the fractional amplitudes of the slow component of I_Na decay (%A_slow) in cFgf13KO + iFGF12B-expressing LV myocytes are indistinguishable at all test potentials from those determined in cFgf13KO cells (see values in Table S2).

Figure 3.

Coexpression of iFGF12B modulates the time- and voltage-dependent properties of Na_V1.5-encoded currents in Xenopus oocytes. (A) Representative human Na_V1.5-encoded (I_Na) waveforms recorded from Xenopus oocytes expressing Na_V1.5 alone (black) or together with iFGF12B (blue); the voltage-clamp protocol is illustrated below the current records. (B) The mean ± SEM normalized conductance-versus-voltage plots reveal marked depolarizing shifts in the voltage-dependences of activation and inactivation of the currents with iFGF12B coexpression (blue), relative to Na_V1.5 expressed alone (black; P values are 0.012 and <0.001, respectively; Table 2). (C) The rate of I_Na recovery from inactivation is also accelerated with iFGF12B coexpression (P value = 0.04; Table 2). (D and E) Inset: Overlaying representative I_Na waveforms, recorded on membrane depolarizations to −20 mV, illustrates the difference in the decay phases of the currents recorded from Xenopus oocytes expressing Na_V1.5 alone (black) or Na_V1.5 with iFGF12B (blue). Analyses of the decay phases of the currents revealed that the time constants of the fast and slow components of I_Na decay (Ƭ_f and Ƭ_s, respectively; D) are indistinguishable in the absence and the presence of iFGF12B, whereas the fractional amplitude of the slow component of current decay (%A_slow; E) was reduced markedly with iFGF12B coexpression at all test potentials (see values in Table S3).

Figure 4.

iFGF12B regulates the activation of the voltage sensor in repeat IV of Na_V1.5. VCF recordings were obtained as described in Materials and methods. Representative fluorescence signals recorded from Xenopus oocytes expressing Na_V1.5 alone (black) or with iFGF12B (blue) in response to 50-ms depolarizing voltage steps to −140, −80, −20, and 40 mV, are presented; only four fluorescence traces are shown for clarity. The mean ± SEM normalized steady-state fluorescence signals were plotted and fitted with a single Boltzmann function (solid lines). Coexpression of iFGF12B induced a hyperpolarizing shift in the VSD-IV F-V curve (P value <0.001, Table 3), but caused no change in the F-V curves of the other VSDs (see Table 3).

Figure 5.

Coexpression of iFGF13VY affects the voltage dependence of steady-state inactivation and the inactivation kinetics of Na_V1.5-encoded currents in Xenopus oocytes. (A) Representative I_Na waveforms recorded from Xenopus oocytes expressing Na_V1.5 alone (black) or with iFGF13VY (orange); the voltage-clamp protocol is illustrated below the current records. (B) The mean ± SEM conductance versus voltage plots reveal marked depolarizing shifts in the voltage dependence of inactivation with iFGF13VY coexpression compared with Na_V1.5 expressed alone (P value <0.001; Table 2). The voltage dependence of I_Na activation, however, is not affected by iFGF13VY (P value = 0.3; Table 2). (C and D) Coexpression of iFGF13VY accelerated I_Na recovery from inactivation (C; P value <0.001; Table 2) and overlay of representative I_Na waveforms (D, inset), evoked at −20 mV, illustrates the difference in the decay phases of I_Na in recordings from Xenopus oocytes expressing Na_V1.5 without (black) and with (orange) iFGF13VY. (E) Analyses of the decay phases of the currents revealed that the time constants, Ƭ_f and Ƭ_s, of the fast and slow components of I_Na decay, respectively, (D) are indistinguishable for the currents in the absence and presence of iFGF13VY, whereas the fractional amplitude of the slow component of current decay (%A_slow; E) was reduced markedly with iFGF13VY coexpression at all test potentials (see values in Table S3).

Figure 6.

Distinct modulation of Na_V1.5 voltage sensor conformations with iFGF13VY coexpression. The F-V curves for all four VSDs of Na_V1.5 recorded from Xenopus oocytes expressing Na_V1.5 alone (black) or with iFGF13VY (orange) show that iFGF13VY affects the voltage-dependences of activation of VSD-I and VSD-IV. For VSD-I, iFGF13VY coexpression shifts the F-V curve toward depolarizing potentials (P value = 0.002; Table 3). For VSD-IV, the F-V curves show that iFGF13VY results in a steeper slope and a small shift toward more hyperpolarized membrane potentials relative to Na_V1.5 expressed alone (P values are 0.022 and <0.001, respectively; Table 3). The effect on the VSD-IV F-V curve with iFGF13VY coexpression is distinct from that observed with iFGF12B coexpression (compare with Fig. 4).

Figure 7.

Analysis of iFGF12B, iFGF13VY, and iFGF12B/13VY chimeras reveals a strong correlation between the slopes of the VSD-IV F-V curves and the slow component of I_Na inactivation. (A) Alignment of the iFGF12B and iFGF13VY sequences highlights the prominent difference in N-termini. (B–D) Several iFGF12B/13VY chimeras were constructed by swapping the N-termini, C-termini, and both C- and N-termini domains. Each chimera was coexpressed with Na_V1.5, and effects on the time-and voltage-dependent properties of I_Na and VSD-IV activation were determined. Correlation and linear regression analyses were performed on these results obtained from iFGF chimeras, iFGF12B, and iFGF13VY. (B and C) The analyses revealed that the slow time constants of I_Na decay (Ƭ_s; B) and the fractional amplitudes of the slow component (%A_slow; C), determined for currents evoked at 0 mV, are both strongly (negatively [B] or positively [C]) correlated with the VSD-IV F-V k values. (D) The k and V_1/2 values, determined from the VSD-IV F-V curves, can be used to predict %A_slow.

Figure S5.

Switching N-terminal domains between iFGF12B and iFGF13VY is not sufficient to replicate the iFGF-specific modulation of Na_V1.5 gating and function. (A) Schematics of iFGF12B, iFGF13VY, and the iFGF chimeras generated with the iFGF12B and iFGF13VY N-termini exchanged (iFGF12B/13 and iFGF13VY/12). (B) Normalized inactivation, G-V, and VSD-IV F-V curves for Na_V1.5 expressed alone (grey dashed line), with iFGF12B (blue dashed line), iFGF13VY (orange dashed line), or one of the iFGF chimeras, FGF12B/13 (blue triangle) or FGF13VY/12 (orange diamond). The effects of the two iFGF chimeras on the G-V and F-V curves are distinct from those observed with either iFGF12B or iFGF13VY, implying that the N-termini of iFGF12B and iFGF13VY alone do not confer iFGF-specific regulation of Na_V1.5.

Figure S6.

iFGF chimeras differentially modulate the voltage dependence of VSD-IV activation (VSD-IV F-V). Three pairs of iFGF chimeras were constructed by switching the various iFGF domains of iFGF12B and iFGF13VY: the N-termini (left), the C-termini (middle), and both the N- and C-termini (right). The VSD-IV F-V curves of the various iFGF chimeras (black circle) are plotted together with those of either iFGF12B (blue dashed line) or iFGF13VY (orange dashed line). The top row presents data for iFGF chimeras with the iFGF13VY core domain and the iFGF12B N- (left), C- (middle), or both N- and C-termini (right) exchanged, and the bottom row shows the reverse configurations, i.e., the iFGF12B core domain with the iFGF13VY N- (left), C- (middle), or both N- and C-termini (right) exchanged. In both cases, the exchanged domain(s) is (are) underlined. As is evident, none of these chimeras confer functional effects on the activation of VSD-IV identical to those seen with native iFGF12B or iFGF13VY.

Figure S7.

Correlation and linear regression analyses reveal no correlations between the VSD-IV F-V and the G-V curves of activation and inactivation determined for Na_V1.5 coexpressed with iFGF12B, iFGF13VY, or the iFGF12B/13VY chimeras. The data derived from the voltage-clamp and VCF recordings of Na_V1.5 coexpressed with iFGF12B, iFGF13VY, and the iFGF12B/13VY chimeras were plotted to determine if there were any correlations between the parameters (V_1/2 and k) derived from the Boltzmann fits to the VSD-IV F-V and the G-V curves. (A and B) As is evident in the plots shown, there are no correlations between V_1/2 determined from the VSD-IV F-V curves and the G-V curves for I_Na inactivation (A) or activation (B). (C and D) Comparison of the slope factors, k, of VSD-IV F-V and the G-V curves reveals a modest correlation for I_Na inactivation (C), but no correlation for I_Na activation (D). (E and F) In addition, the VSD-IV F-V V_1/2 values are not well correlated with either the slow inactivation time constants (E) or the fractional amplitudes of the slow component of I_Na decay (F). (G) The k and V_1/2 values, determined from the VSD-IV F-V curves, cannot be used to predict the inactivation G-V V_1/2 values.

Figure 8.

The mechanism of iFGF regulation of the Na_V channel gating. At the resting membrane potential, the closed-state Na_V channel has all voltage sensors (VSD-I–VSD-IV) down in their resting conformations and the activation gate is closed. This schematic portrays two interactions between the resting VSD-IV and the C-terminal domain, and the C-terminal domain and the III–IV linker in the Na_V channel, as proposed by Clairfeuille et al. (2019). The activation of VSD-I to VSD-III leads to the transition of Na_V1.5 from a closed to an open state, allowing for Na⁺ conductance. The resulting activation of VSD-IV facilitates Na_V channel fast inactivation. The coexpression of iFGF facilitates the activation of VSD-IV and thus facilitates the Na_V1.5 channel inactivation.