Enhanced slow inactivation contributes to dysfunction of a recurrent SCN2A mutation associated with developmental and epileptic encephalopathy

Abstract

Key points: The recurrent SCN2A mutation R853Q is associated with developmental and epileptic encephalopathy with typical onset after the first months of life. Heterologously expressed R853Q channels exhibit an overall loss-of-function as a result of multiple defects in time- and voltage-dependent channel properties. A previously unrecognized enhancement of slow inactivation is conferred by the R853Q mutation and is a major driver of loss-of-function. Enhanced slow inactivation is potentiated in the canonical splice isoform of the channel and this may explain the later onset of symptoms associated with R853Q.

Abstract: Mutations in voltage gated sodium (Na_V ) channel genes, including SCN2A (encoding Na_V 1.2), are associated with diverse neurodevelopmental disorders with or without epilepsy that present clinically with varying severity, age-of-onset and pharmacoresponsiveness. We examined the functional properties of the most recurrent SCN2A mutation (R853Q) to determine whether developmentally-regulated alternative splicing impacts dysfunction severity and to investigate effects of the mutation on slow inactivation. We engineered the R853Q mutation into neonatal and adult Na_V 1.2 splice isoforms. Channel constructs were heterologously co-expressed in HEK293T cells with human β1 and β2 subunits. Whole-cell patch clamp recording was used to compare time- and voltage-dependent properties of mutant and wild-type channels. The R853Q mutation exhibits an overall loss-of-function attributed to multiple functional defects including a previously undiscovered enhancement of slow inactivation. The mutation exhibited altered voltage dependence of activation and inactivation, slower recovery from inactivation and decreased channel availability during high-frequency depolarizations. More notable were effects on slow inactivation, including a 10-fold slower rate of recovery from slow inactivation exhibited by mutant channels. The impairments in fast inactivation properties were more severe in the neonatal splice isoform, whereas slow inactivation was more pronounced in the splice isoform of the channel expressed predominantly in later childhood. Enhanced later-onset slow inactivation may be a primary driver of the later onset of neurological features associated with this mutation.

Keywords: channelopathy; developmental and epileptic encephalopathy; epilepsy; sodium channel.

Figure 1. SCN2A R853Q channel topology and representative current traces of neonatal and adult splice variant of WT and R853Q.

(A) Transmembrane topology of Na_V1.2 showing the portion of the protein encoded by exon 6 (red cylinders). Also shown are the location of the amino acid change (filled purple circle) that distinguishes the adult splice variant (Asp-209) from the neonatal splice variant (Asn-209), as well as the location of the R853Q mutation (filled yellow circle). (B) Representative whole-cell current density traces from tsA201 cells transfected with either neonatal (6N) or adult (6A) splice variants of WT or R853Q. Peak current densities of representative traces shown: 6A-WT = −652.6 pA/pF, 6A-R853Q = −404.2 pA/pF, 6N-WT = −685.5 pA/pF, 6N-R853Q = −294.9 pA/pF (currents were normalized to cell capacitance as a proxy for cell size to calculate current density).

Figure 2. Functional properties of WT and R853Q.

(A) Current-voltage relationships in which peak currents were normalized to cell capacitance (proxy for cell size) to approximate current density. (B) Normalized conductance-voltage relationships illustrating the voltage-dependence of activation for WT and R853Q. Solid lines represent average Boltzmann fits to the data. In panels A and B, currents were elicited by depolarizing to a range of potentials to trigger Na_v1.2 channel activation. Statistical analysis method: One-Way ANOVA with Holm-Sidak post-hoc test. (C) Voltage-dependence of fast inactivation was measured during a 4 ms 0 mV voltage step after triggering fast inactivation at different potentials. Solid lines represent average Boltzmann fits to the data. Statistical analysis method: One-Way ANOVA with either Holm-Sidak (V_1/2,SSI) or Dunn’s (k_SSI) post-hoc test. (D) Recovery from fast inactivation comparing R853Q to WT in 6A and 6N. Channel recovery was examined by first delivering a 100 ms 0 mV depolarizing pulse to enter the channel into fast inactivated state, recovering for range of recovery times, and then delivering a 0 mV test pulse to measure recovered current. The plotted current was calculated by normalizing current elicited by the test pulse to the current measured during the inactivating voltage step. Statistical analysis method: One-Way ANOVA with Holm-Sidak post-hoc test. (E) Voltage protocols used for panels A-D. Each experiment had an inter-sweep interval of 5 seconds at the holding potential of −120mV. Quantitative values and cell counts (n) are presented in Table 1. These data were recorded from tsA201 cells from 7 separate transient transfections of each group.

Figure 3. Sodium channel behavior in response to repetitive depolarizing pulses.

(A) Representative current traces elicited by repetitive depolarizing pulses (0 mV for 5 ms from −120 mV holding potential) at frequencies of 10Hz and 80Hz. Shown are representative currents measured during pulse 1, 50, 100, 150, 200, 250 and 300 (normalized to respective currents measured during pulse 1). Pulse 1 is purple and pulse 300 is orange. (B,C) Average currents elicited by repetitive depolarizing pulses as described in panel (A). 10 Hz (filled symbols) and 80 Hz (open symbols) (black: 6A-WT; red: 6A-R853Q; grey: 6N-WT; blue: 6N-R853Q). Inter-sweep interval was 120 s. (D) Steady state values are calculated as the averaged current measured during the last 10 pulses normalized to the maximum current measured during the respective stimulation frequency (10, 30, 50, 80, 110, and 135 Hz. 20Hz was not examined in 6A). n = 15 (6A-WT), 10 (6A-R853Q), 5 (6N-WT), and 5 (6N-R853Q). (E) Differences in average normalized steady state current between R853Q and WT for neonatal (filled diamonds) and adult (open diamonds) splice isoforms, respectively. These data were recorded from tsA201 cells from 7 (6A) and 4 (6N) separate transient transfections.

Figure 4. R853Q channels exhibit smaller window current.

(A) Visualization of window current due to overlap between voltage-dependence of activation and fast inactivation curves (normalized Boltzmann fits as shown in Fig. 1B and Fig. 1C) for WT and R853Q expressed in adult (6A) and neonatal (6N) Na_V1.2 splice isoforms. (B) Box plots illustrate window current determined as the integral of the overlapping area between activation and inactivation curves (black: 6A-WT; red: 6A-R853Q; grey: 6N-WT; blue: 6N-R853Q of Panel A). Solid horizontal lines within each box represents the median, the top and bottom of each box are the 75th and 25th percentiles (the interquartile range), and the whiskers represent end of the interquartile range to the furthest observation within the whisker length of 1 times the interquartile range. (*: p < 0.05). Statistical analysis method: One-Way ANOVA with Holm-Sidak post-hoc test. Quantitative values and cell counts (n) are presented in Table 1. These data were recorded from tsA201 cells from 4 separate transient transfections of each group.

Figure 5. R853Q enhances slow inactivation.

(A) Onset of inactivation (OI) measured using depolarizing voltage steps to 20 mV with duration 1 ms to 10 s, a 20 ms step to −120 mV to recover channels from fast inactivation then a final test pulse to 0 mV to assess channel availability. Solid lines represent fits to double exponential decay functions to the averaged normalized data. (B) Voltage dependence of slow inactivation (VDSI) measured using 10 s depolarizing pulses ranging from −120 mV to 30 mV, followed by a 20 ms step to −120 mV to recover channels from fast inactivation and a final test pulse to 0 mV to assess channel availability. Solid lines represent Boltzmann fits to the averaged normalized data (black: 6A-WT; red: 6A-R853Q; grey: 6N-WT; blue: 6N-R853Q). (C,D) Recovery from slow inactivation (RSI) comparing R853Q to WT in both adult (C) and neonatal (D) splice isoforms. Recovery from slow inactivation was examined in an identical manner as recovery from fast inactivation (see Fig. 2D), but the inactivating pulse was 10 s to promote entry into the slow inactivated state. (A-D) Statistical analysis method: One-Way ANOVA with Holm-Sidak post-hoc test. Quantitative values and cell counts (n) are presented in Table 1. (E) Voltage protocols used for panels A-D. Each experiment had an inter-sweep interval of 120 seconds at the holding potential of −120mV. These data were recorded from tsA201 cells from 6 (OI) or 7 (VDSI, RSI) separate transient transfections of each group.

Figure 6. Integrated summary of functional properties of WT and R853Q channels in neonatal and adult splice variants.

The relative differences in 18 functional parameters between WT and R853Q channels is summarized by a radar plot. Individual parameters are labeled as described to the right of radar plots and in Table 1. Slope factors of activation and inactivation were not included because fluctuations in slope factor do not correspond directly to gain or loss of overall channel conductance. The purple shaded quadrants represent voltage-dependent properties and the remaining unshaded quadrants represent time-dependent or kinetic properties. (A) For each parameter, the effect size was calculated to compare R853Q to WT for each respective splice isoform (dotted black line indicates effect size = 0). Separate plots for 6A-R853Q (red) and 6N-R853Q (blue) are shown. Parameter values were scaled such that a relative loss-of-function is represented by points that fall within the dotted circle whereas gain-of-function is indicated by points that fall outside of the dotted circle. The calculated effect sizes for Tau_s,RSI (yellow box) were multiplied by 0.5 to better match the proportional scaling of other parameters. (B) For each parameter, effect sizes were calculated to compare 6N-WT to 6A-WT. Parameter values for 6N-WT were scaled such that a relative loss-of-function is represented by points that fall within the black 6A-WT line whereas gain-of-function is indicated by points that fall outside of the black 6A-WT line.