Gabapentin Modulates HCN4 Channel Voltage-Dependence

Abstract

Gabapentin (GBP) is widely used to treat epilepsy and neuropathic pain. There is evidence that GBP can act on hyperpolarization-activated cation (HCN) channel-mediated I_h in brain slice experiments. However, evidence showing that GBP directly modulates HCN channels is lacking. The effect of GBP was tested using two-electrode voltage clamp recordings from human HCN1, HCN2, and HCN4 channels expressed in Xenopus oocytes. Whole-cell recordings were also made from mouse spinal cord slices targeting either parvalbumin positive (PV⁺) or calretinin positive (CR⁺) inhibitory neurons. The effect of GBP on I_h was measured in each inhibitory neuron population. HCN4 expression was assessed in the spinal cord using immunohistochemistry. When applied to HCN4 channels, GBP (100 μM) caused a hyperpolarizing shift in the voltage of half activation (V_1/2) thereby reducing the currents. Gabapentin had no impact on the V_1/2 of HCN1 or HCN2 channels. There was a robust increase in the time to half activation for HCN4 channels with only a small increase noted for HCN1 channels. Gabapentin also caused a hyperpolarizing shift in the V_1/2 of I_h measured from HCN4-expressing PV⁺ inhibitory neurons in the spinal dorsal horn. Gabapentin had minimal effect on I_h recorded from CR⁺ neurons. Consistent with this, immunohistochemical analysis revealed that the majority of CR⁺ inhibitory neurons do not express somatic HCN4 channels. In conclusion, GBP reduces HCN4 channel-mediated currents through a hyperpolarized shift in the V_1/2. The HCN channel subtype selectivity of GBP provides a unique tool for investigating HCN4 channel function in the central nervous system. The HCN4 channel is a candidate molecular target for the acute analgesic and anticonvulsant actions of GBP.

Keywords: HCN4; epilepsy; gabapentin; pain; spinal cord.

FIGURE 1

HCN4 channel function is reduced by GBP. (A) Raw HCN4 channel-mediated steady-state (left) and tail currents (right) before and after GBP. (B) Average steady-state current before and after GBP. (C) Average 100 μM GBP to vehicle steady-state current ratio at various voltages. (D) Average 100 μM GBP to vehicle time to half-activation ratio at various voltages. (E) Normalized conductance–voltage relationship constructed from tail currents. (F) Average shift (Δ) in the voltage of half-activation and slope of HCN4 channel by 100 μM GBP and vehicle control, relative to baseline measurements. ^∗∗p < 0.0001.

FIGURE 2

HCN1 channels are minimally affected by GBP. (A) Raw HCN1 channel-mediated steady-state (left) and tail currents (right) before and after 100 μM GBP. (B) Average steady-state current before and after GBP. (C) Average 100 μM GBP to vehicle steady-state current ratio at various voltages. (D) Average 100 μM GBP to vehicle time to half-activation ratio at various voltages. (E) Normalized conductance–voltage relationship constructed from tail currents. (F) Average shift (Δ) in the voltage of half-activation and slope of HCN1 channel by 100 μM GBP and vehicle control, relative to baseline measurements.^∗p < 0.05.

FIGURE 3

HCN2 channels are unaffected by GBP. (A) Raw HCN2 channel-mediated steady-state (left) and tail currents (right) before and after 100 μM GBP. (B) Average steady-state current before and after GBP. (C) Average 100 μM GBP to vehicle steady-state current ratio at various voltages. (D) Average 100 μM GBP to vehicle time to half-activation ratio at various voltages. (E) Normalized conductance–voltage relationship constructed from tail currents. (F) Average shift (Δ) in the voltage of half-activation and slope of HCN2 and HCN2+4 channels by 100 μM GBP and vehicle control, relative to baseline measurements.

FIGURE 4

HCN4 expression in mouse dorsal horn populations. (A) Immunolabeling for calretinin (CR) (green) and HCN4 (red) overlaps substantially in the superficial laminae of the spinal dorsal horn, but rarely in the same cells (asterisk and inset shows the lack of HCN4 labeling in a CR neuron). (B and C) HCN4 immunolabeling is absent in Pax2-expressing (blue) CR⁺ neurons (arrowheads), however, HCN4 is detected in many neurons lacking Pax2 (arrows). Scale bar = 10 μm.

FIGURE 5

GBP reduces I_h recorded from PV⁺ inhibitory neurons. (A) Raw steady-state I_h (left) and tail currents (right) before and after application of 100 μM GBP recorded from PV⁺ neurons. (B) Average steady-state current before and after GBP recorded from PV⁺ neurons. (C) Average 100 μM GBP to vehicle steady-state current ratio at various voltages. (D) Average 100 μM GBP to vehicle time to half-activation ratio at various voltages. (E) Normalized conductance–voltage relationship constructed from tail currents recorded from PV⁺ neurons. (F) Average shift (Δ) in the voltage of half-activation and slope of PV⁺ neurons by 100 μM GBP and vehicle control, relative to baseline measurements. ^∗p < 0.05.

FIGURE 6

GBP minimally affects I_h recorded from CR⁺ inhibitory neurons. (A) Raw steady-state I_h (left) and tail currents (right) before and after application of 100 μM GBP recorded from CR⁺ neurons. (B) Average steady-state current before and after GBP recorded from CR⁺ neurons. (C) Average 100 μM GBP to vehicle steady-state current ratio at various voltages. (D) Average 100 μM GBP to vehicle time to half-activation ratio at various voltages. (E) Normalized conductance–voltage relationship constructed from tail currents of CR⁺ neurons. (F) Average shift (Δ) in the voltage of half-activation and slope of CR⁺ neurons by 100 μM GBP and vehicle control, relative to baseline measurements. ^∗p < 0.05.