Identification and targeting of a unique NaV1.7 domain driving chronic pain

Abstract

Small molecules directly targeting the voltage-gated sodium channel (VGSC) Na_V1.7 have not been clinically successful. We reported that preventing the addition of a small ubiquitin-like modifier onto the Na_V1.7-interacting cytosolic collapsin response mediator protein 2 (CRMP2) blocked Na_V1.7 function and was antinociceptive in rodent models of neuropathic pain. Here, we discovered a CRMP2 regulatory sequence (CRS) unique to Na_V1.7 that is essential for this regulatory coupling. CRMP2 preferentially bound to the Na_V1.7 CRS over other Na_V isoforms. Substitution of the Na_V1.7 CRS with the homologous domains from the other eight VGSC isoforms decreased Na_V1.7 currents. A cell-penetrant decoy peptide corresponding to the Na_V1.7-CRS reduced Na_V1.7 currents and trafficking, decreased presynaptic Na_V1.7 expression, reduced spinal CGRP release, and reversed nerve injury-induced mechanical allodynia. Importantly, the Na_V1.7-CRS peptide did not produce motor impairment, nor did it alter physiological pain sensation, which is essential for survival. As a proof-of-concept for a Na_V1.7 -targeted gene therapy, we packaged a plasmid encoding the Na_V1.7-CRS in an AAV virus. Treatment with this virus reduced Na_V1.7 function in both rodent and rhesus macaque sensory neurons. This gene therapy reversed and prevented mechanical allodynia in a model of nerve injury and reversed mechanical and cold allodynia in a model of chemotherapy-induced peripheral neuropathy. These findings support the conclusion that the CRS domain is a targetable region for the treatment of chronic neuropathic pain.

Keywords: CRMP2; NaV1.7; SUMO; chronic pain; gene therapy.

Fig. 1.

Identification of a unique CRMP2 regulatory domain on Na_V1.7. (A) Cartoon of the domain structure of human Na_V1.7 with intracellular loops labeled. These loops were divided into 384 15-mer peptides with 12 overlapping amino acids and were printed onto a peptide array. (B) Fluorescent intensity of CRMP2 binding to the peptide array from rat, pig, and human DRG and spinal cord lysate (n = 4). (C) Microscale thermophoresis of NTA-labeled His-CRMP2 (200 nM) with different concentrations of the CRMP2-binding peptide from Na_V1.7 or the homologous region of NaV1.X channels, fitted to a one-site binding model (r²= 0.97 for Na_V1.7 and 0.76 for NaV1.1; n = 4; Bottom). Peptides with higher sequence similarity to Na_V1.7 bound more tightly with CRMP2 in this assay. Those with less sequence similarity bound with little to no detectable affinity. (D) Sequence alignment of the CRMP2 binding domain on Na_V1.7 with the CRS analogous domains. (E) Cartoon depiction of the Halo-Na_v1.7 construct that was adapted for transfection into rat DRG neurons. (F) Diagram showing the steady-state availability protocol that was used to isolate TTX-R from TTX-S currents using a prepulse to −40 mV to determine the TTX-R current fraction (shown in pink). (G) An electrical subtraction protocol was used to isolate the currents originating from Halo-NaV1.7 channels. This was achieved by subtracting the current obtained during the test pulse (at 0 mV) preceded by a 1-s pulse at −40 mV (pink trace) from the current obtained at the test pulse (at 0 mV) previously subjected to different prepulses ranging from −120 to +10 mV. This approach allowed for the isolation of Halo-NaV1.7 channel currents, as these channels exhibit TTX resistance and possess distinct NaV1.7 inactivation kinetics. (H) Representative traces of sodium currents recorded from DRGs transfected with Halo-Na_V1.7_(WT) (green) and Halo-Na_V1.7_(1.3) (teal) in the presence or absence of 5 nM ProTx-II. (I) Current density–voltage relationship for Halo-Na_V1.7_(WT), Halo-Na_V1.7_(1.3), and both groups following treatment with 5 nM ProTx-II. Treatment with ProTx-II did not reduce the current density further suggesting maximal reduction was achieved by swapping out the Na_V1.7-CRS domain with the analogous domain from NaV1.3. (J) Peak Halo-Na_V1.7 current density for Halo-Na_V1.7_(WT), Halo-Na_V1.7_(WT) + 5 nM ProTx-II, Halo-Na_V1.7_(1.3), and Halo-Na_V1.7_(1.3) + 5 nM ProTx-II. (K) TTX-R peak current density for the conditions listed above. Each group was compared to its own on-plate Halo-Na_V1.7_(WT) control group and statistical significance was determined using a Mann–Whitney U test. n = 11 to 16 cells; error bars indicate mean ± SEM; P values as indicated (Dataset S1); Kruskal–Wallis test with the Dunn post hoc test. All biophysical parameters are shown in SI Appendix, Table S1. SDH—Spinal Dorsal Horn; DRG—Dorsal Root Ganglia. Panels E–H Green circles—Halo-Na_V1.7_(WT) + 500 nM TTX; Blue circles Halo-Na_V1.7_(1.3) + 500 nM TTX; Green circles with gray center - Halo-Na_V1.7_(WT) + 500 nM TTX + 5 nM ProTx-II; Blue circles with gray center Halo-Na_V1.7_(1.3) + 500 nM TTX + 5 nM ProTx-II.

Fig. 2.

Myr-TAT-Na_V1.7-CRS causes CRMP2 dependent reduction in Na_V1.7 currents in DRG neurons. (A) The Myr-TAT-Na_V1.7-CRS peptide competes for binding to Na_V1.7 regulatory proteins. (B) Representative immunoblots and summary (C) of CRMP2 immunoprecipitation (IP) to detect Na_V1.7 from CAD cells treated with the indicated peptides (n = 5). (D) Representative current traces recorded from small-sized DRG neurons in the presence of 5 µM Myr-TAT-SCR peptide (blue circles, n = 15) and Myr-TAT-Na_V1.7-CRS peptide (green circles, n = 16). (E) Summary of Boltzmann fits for current density–voltage curves, (F) peak current densities, and (G) electrically isolated TTX-R currents. (H) Representative current traces recorded from small-sized DRGs incubated with Myr-TAT-SCR and Myr-TAT-Na_V1.7-CRS in the presence and absence of the Na_V1.7-specific inhibitor ProTx-II. (I) Boltzmann fits of the current density–voltage curves for each treatment group. (J) Summary of peak current densities. N = 8 to 12 cells; (K) representative current traces recorded from small-diameter DRGs transfected with either siRNA-Control or siRNA-CRMP2 and treated with Myr-TAT-Na_V1.7-CRS or Myr-TAT-SCr as indicated. (L) Boltzmann fits for current density–voltage curves and (M) peak current densities. n = 12 to 13 cells; (N) representative immunoblots and (O) summary (Bottom) of Na_V1.7 expression in CAD cells treated with the indicated peptides. βIII-Tubulin is used as a loading control (n = 4). (P) Representative immunoblots of streptavidin-enriched surface fractions probed for Na_V1.7 and Na+/K+ ATPase as a control. (Q) Bar graph with scatter plot of mean surface localized Na_V1.7 in CAD cells treated with the indicated peptides. (R) Representative current traces recorded from small-diameter DRG neurons in the presence of Myr-TAT-SCR or Myr-TAT-Na_V1.7-CRS ± 20 µM Pitstop2 as indicated; (S) Boltzmann fits for current density-voltage curves and (T) summary of peak current densities (pA/pF) showing pitsopt2 blocked the current reduction imposed by Myr-TAT-Na_V1.7-CRS. n = 7 to 14 cells; error bars indicate mean ± SEM; P values as indicated; Kruskal–Wallis test with Dunnett’s post hoc comparisons (Dataset S1). All biophysical parameters are shown in SI Appendix, Table S2 and statistical comparisons shown in Dataset S1. CRS—CRMP2 regulatory sequence; SCR—Scrambled control peptide (Sequence: KYHPWACFRQWRSPK).

Fig. 3.

Disruption of the NaV1.7-CRMP2 interaction decreases presynaptic Na_V1.7, sensory neuron excitability, and spinal cord neurotransmitter release. (A) Representative action potential recordings from DRGs in response to current injections after adding Myr-TAT-SCR (blue) and Myr-TAT-NaV1.7-CRS peptides (green). (B) Quantification of current-evoked action potentials in response to 0 to 120 pA of injected current. (C) Resting membrane potential of cells recorded in A. (D) Representative traces and (E) quantification indicating an increased rheobase with Myr-TAT-NaV1.7-CRS treatment. n = 26 cells; (F) Summary data for membrane input resistance. (G) Immunoblots of NaV1.7, CRMP2, and CaV2.2 expression in the presynaptic fraction of spinal dorsal horn, 1 h after peptide injection [i.t. injection of 20 µg/5 µL of Myr-TAT-SCR (n = 6) or Myr-TAT-NaV1.7-CRS (n = 5)]. (H) Quantification showing NaV1.7 spinal presynaptic localization of data in G. (I) KCl depolarization-evoked CGRP release measured from isolated spinal cord following incubation with control, Myr-TAT-SCR, or Myr-TAT-NaV1.7-CRS peptides. Histograms show normalized CGRP levels (n = 4 animals); error bars indicate mean ± SEM; data analyzed by Mann–Whitney U test or two-way ANOVA (details in Dataset S1), P values indicated.

Fig. 4.

Disruption of the Na_V1.7–CRMP2 interaction alleviates mechanical allodynia without affecting physiological pain. (A) CRMP2-binding intensity to Na_V1.7-derived peptide #141 from contralateral (Contra) and ipsilateral (Ipsi) spinal cords of male rats taken two weeks following SNI (n = 4). (B) Time course for male and female rats following administration of Myr-TAT-SCR, or Myr-TAT-Na_V1.7-CRS (20 µg in 5 µL, i.t.). (C) Area under the curve for paw withdrawal thresholds showing that Myr-TAT-Na_V1.7-CRS reversed mechanical allodynia, n = 6 rats; (D) Cartoon and bar graph with scatter plot of paw withdrawal latency in the hot plate test (52 °C) (n = 10 mice). (E) Cartoon and bar graph with scatter plot of the tail flick (52 °C) test showing no effect of the Myr-TAT-Na_V1.7-CRS (n = 10 mice). (F) Cartoon of the rotarod apparatus to assess motor coordination in rodents. Bar graph with scatter plot showing the latency to fall off a rotating rod was not different between the treatments. n = 7 rats; error bars indicate mean ± SEM; Mann–Whitney U test, Kruskal–Wallis, or two-way ANOVA (details are in Dataset S1), P values as indicated. The experiments were conducted by investigators blinded to treatments.

Fig. 5.

AAV plasmid encoding Na_V1.7-CRS reduces Na_V1.7 currents through an endocytic mechanism and decreases spontaneous activity in the spinal cord. (A) Current traces recorded from DRGs transfected with pAAV-SCR (blue, n = 12) or pAAV-NaV1.7-CRS (green, n = 12). (B) Summary plots of current density–voltage relationship fitted with Boltzmann curve, (C) peak current density, and (D) electrically isolated TTX-R currents. (E) Current traces from transfected DRG neurons treated with vehicle or 5 nM ProTx-II. (F) Summary plots of current density–voltage relationship fitted with Boltzmann curve and (G) peak current density. (H) Current traces from groups transfected with pAAV-SCR or pAAV-Na_V1.7-CRS and treated with Pitstop2 (20 µM). (I) Summary plots of current density–voltage relationship and (J) peak current density. Values for biophysical parameters in SI Appendix, Table S3. n = 9 to 15 cells. (K) Traces of spontaneous excitatory postsynaptic currents (sEPSC) from rat substantia gelatinosa (SG) neurons transduced with AAV-SCR (blue) or AAV-Na_V1.7-CRS (green). (L) Cumulative distribution and bar graph of sEPSC interevent intervals from Na_V1.7-CRS transduced neurons. (M) Cumulative distribution and bar graph showing decreased sEPSC amplitude in AAV-NaV1.7-CRS-treated rat slices compared to control. Data expressed as means ± SEM. Mann–Whitney U test and Kruskal–Wallis tests (details in Dataset S1).

Fig. 6.

AAV9 plasmid encoding Na_V1.7-CRS reverses and prevents mechanical allodynia in a mouse model of neuropathic pain. (A) Experimental paradigm for testing AAV9 vector expressing CRS domain in mice. AAV9-NaV1.7-CRS virus was administered intrathecally following establishment of SNI-induced neuropathic pain (SI Appendix, Fig. S10). GFP fluorescence indicates successful injection in DRGs (B) and spinal cord (C). (D) Paw withdrawal thresholds of male mice injected with AAV9-NaV1.7-CRS or control peptide 7 d after SNI. (E) Area under the curve showing pain reversal by AAV-NaV1.7-CRS. (F) Locomotor activity and (G) anxiety-like behavior in open field test. Results replicated in female mice for paw withdrawal threshold (H), area under the curve (I), locomotor activity (J), and anxiety (K), indicating pain reversal in both sexes. Intrathecal injection of AAV9 performed 7 d before SNI surgery (SI Appendix, Fig. S10). Prevention of chronic allodynia in male mice injected with AAV9-NaV1.7-CRS shown by paw withdrawal thresholds (L) and area under the curve (M). No effect on locomotor activity (N) and anxiety-like behavior (O). Data replicated in female mice for paw withdrawal threshold (P), area under the curve (Q), locomotor activity (R), and anxiety (S). AAV9-NaV1.7-CRS prevented chronic pain in both male and female mice. n = 5 to 9 animals; error bars indicate mean ± SEM; detailed statistical analysis in Dataset S1. Experiments were conducted by investigators blinded to treatments.

Fig. 7.

AAV9-Na_V1.7-CRS reverses paclitaxel-induced mechanical and cold allodynia. (A) The paw withdrawal threshold of male and female mice was measured at different time points [Baseline: time 0 (pre injection); and Day 7, 14, 16, and 21] after 4 intraperitoneal injections of paclitaxel (8 mg/kg, i.p.). Mice were treated intrathecally (i.t.) with AAV9-SCR or AAV9-Na_V1.7-CRS (1 × 10¹⁰ viral particles in 5 µL) as indicated. Results were compared using three-way ANOVA with time, treatment, and virus as factors and post hoc Holm-Šídák test; *P < 0.05 Pac-AAV9-SCR vs. Veh + AAV9-SCR; ^&P < 0.05 Pac-AAV9-Na_V1.7-CRS vs. Veh + AAV9-Na_V1.7-CRS; ^#P < 0.05 Pac-AAV9-Na_V1.7-CRS vs. Pac-AAV9-SCR). Detailed statistical analysis is in Dataset S1. (B) Area under the curve in A. Statistical significance as indicated (One-way ANOVA with Tukey’s post hoc test). Taken together, the data show that AAV9-Na_V1.7-CRS reversed CIPN-induced mechanical allodynia. (C) Time of aversive responses for male and female mice was measured at different time points [Baseline: time 0 (before any injection); and Day 8, 15, 17, and 22] after 4 intraperitoneal injections of paclitaxel (8 mg/kg, i.p.). Mice were treated intrathecally (i.t.) with AAV9-SCR or AAV9-Na_V1.7-CRS (1 × 10¹⁰ viral particles in 5 µL) as indicated. Results were compared using three-way ANOVA with time, treatment, and virus as factors and post hoc Holm-Šídák test; *P < 0.05 Pac-AAV9-SCR vs. Veh + AAV9-SCR; ^&P < 0.05 Pac-AAV9-Na_V1.7-CRS vs. Veh + AAV9-Na_V1.7-CRS; ^#P < 0.05 Pac-AAV9-Na_V1.7-CRS vs. Pac-AAV9-SCR). Detailed statistical analysis is in Dataset S1. (D) Area under the curve in C. Statistical significance as indicated in the figure (one-way ANOVA with Tukey’s post hoc test). Taken together, the data show that AAV9-Na_V1.7-CRS reversed CIPN-induced cold allodynia (acetone test) in both male and female mice. n = 8 animals; error bars indicate mean ± SEM; detailed statistical analysis is in Dataset S1. The experiments were conducted by investigators blinded to treatments.

Fig. 8.

Macaque DRG neurons transduced by AAV9-Na_V1.7-CRS show reduced Na_V1.7 currents with no effect on TTX-R currents. (A) Fluorescent intensity of CRMP2 binding to peptide array from macaque spinal cord lysate (n = 2). Highlighted in red is the broadest peak with the highest CRMP2 binding, corresponding to peptide #140. (B) Immunoblots (Left) and summary (Right) of mean relative Na_V1.7 binding to CRMP2 in spinal cord lysates treated with indicated peptides (n = 4). Error bars show mean ± SEM; P values as indicated; Mann–Whitney U test. (C) Current traces from small-sized DRGs transduced with AAV9-Na_V1.7 or AAV9-SCR, coapplying 5 nM ProTx-II, a NaV1.7 inhibitor. (D) Boltzmann fits for current density–voltage curves. (E) Summary of peak current densities (pA/pF). (F) Boltzmann fits of voltage-dependent activation. Half-maximal activation potential and slope values in SI Appendix, Table S4. N = 12 to 17 cells; error bars indicate mean ± SEM; P values as indicated; One-way ANOVA with Tukey’s post hoc test. (G) Current traces from small-sized DRGs transduced with AAV9-NaV1.7 or AAV9-SCR, coapplying 300 nM TTX. The remaining current is exclusively TTX-R (NaV1.8 and NaV1.9) sodium channels. (H) Boltzmann fits for current density–voltage curves. (I) Summary of peak current densities (pA/pF). (J) Boltzmann fits of voltage-dependent activation. Half-maximal activation potential and slope values in SI Appendix, Table S3. N = 11 cells; error bars indicate mean ± SEM; P values as indicated; One-way ANOVA with Tukey’s post hoc test.