Capsazepine prolongation of the duration of lidocaine block of sensory transmission in mice may be mediated by modulation of HCN channel currents

Animals

Following the ARRIVE Guidelines and with the approval of the Institutional Animal Experimental Ethics Committee of Sichuan University (Chengdu, Sichuan, China, protocol 2015014A), a total of 63 adult male C57BL/6J mice (20–30 g) for experiments of sciatic nerve block in vivo and in vitro and seven C57BL/6J mice at age of postnatal 8 days for experiment of isolated dorsal root ganglion (DRG) neurons were included. All the study mice were housed in standard condition with free access to food and water. All the mice were placed in the experimental environment for 1–2 h per day for consecutive 3 days before the formal experiments. The observers of the behavioral test were blinded to the group assignment of the study mice.

Chemicals

Lidocaine solution was diluted from 2% (w/v) lidocaine (Shanghai Fortune Zhaohui Pharmaceutical Co., Ltd., Shanghai, China) with normal saline to the final concentration of 1% (w/v, ∼35 mM). ZD7288 (Tocris Bioscience, Bristol, UK), the non-selective HCN channel blocker, and forskolin (Sigma-Aldrich, Co., St. Louis, MO, USA), the activator of adenylyl cyclase, were dissolved with normal saline. CPZ (Sigma-Aldrich, Co., St. Louis, MO, USA) was dissolved with Dimethyl sulfoxide (DMSO) (Sigma-Aldrich, Co., St. Louis, MO, USA) to a 2.5 mg/mL stock solution and then diluted with normal saline (for animal study in vivo) or bath solution (for electrophysiological recordings) to the final concentrations. For electrophysiological recordings, the final concentration of DMSO was 0.05% (v/v), which produces minimal effect in electrophysiological recordings (Liu et al., 2016; Wang et al., 2016).

Patch clamping recordings

The vectors of pcDNA3-HE3 with mHCN1 (Mouse HCN1) and mHCN2 (Mouse HCN2) were expressed in human embryonic kidney 293 (HEK 293) cells as previously described (Gill et al., 2004). HEK 293 cells were cultured under standard procedures (Meng et al., 2011). The mHCN channel constructs were co-transfected with a green fluorescent protein plasmid (pEGFP; Clontech Laboratories, Mountain View, CA, USA) by Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA). Recordings were performed 36–48 h after transfection. Whole-cell patch clamping recordings were performed at room temperature (∼24–25 °C). Patch pipettes (with resistance of three to five MΩ) were made from the fire-polished borosilicate glass capillaries (Sutter Instrument Co., Novato, CA, USA). I_h were sampled using an Axon 200B amplifier, digitized via a Digidata 1440A interface. Standard bath solution contained (mM) 118 NaCl, 25 KCl, two MgCl₂, two CaCl₂, 10 HEPES, and 10 glucose, pH = 7.4 adjusted by NaOH. Bath solution was continuously perfused at a speed of ∼2 mL/min. The pipette solution contained (mM) 120 KCH₄SO₃, four NaCl, one MgCl₂, 0.5 CaCl₂, 10 HEPES, 10 EGTA, three Mg-ATP and 0.3 GTP-Tris, pH = 7.3 adjusted by KOH (Meng et al., 2011).

Hyperpolarization-activated currents (I_h) were evoked by hyperpolarized pulses (4 s) from −40 to −130 mV with an interval of −10 mV (Gill et al., 2004). Amplitudes of steady-state currents were produced at the end of hyperpolarizing voltage steps as tail currents, which used for the analysis of voltage dependence of channel gating. Tail currents were acquired at a fixed potential of −90 mV immediately after hyperpolarized pulses and were normalized and fitted with Boltzmann curves as: G/G_max = 1/(1 + e(V_1/2a − V)/k), where G/G_max = normalized conductance; G_max = maximum conductance; V_1/2a = voltage of half-maximum activation; k = slope factor. Concentration-response of CPZ on the I_h was fitted to the Hill equation: Y = 1/(1 + 10(logIC₅₀ − X) * h), where Y = inhibitory ratio; X = concentration of CPZ; h = Hill coefficient.

Acute isolated DRG neurons were prepared for recordings of sodium channel currents (I_Na) and I_h. Briefly, mice were anesthetized by ketamine/xylazine (40/15 mg/kg). Then the DRGs of mice were removed and digested respectively in 1% papain for 45 min and in 1% collagenase for 30 min. DRG neurons were recorded after 2-h cell attachment to the coverslip. I_h was evoked by hyperpolarized pulse (4 s) to −120 mV with the same solutions used for recordings on HEK 293 cells. To test the effect of intracellular cAMP, 200 μM cAMP was added to the patch pipette. I_Na was recorded by depolarized pulse (20 ms) to 10 mV. Bath solution contained (in mM): 125 NaCl, 25 NaHCO₃, 1.25 NaH₂PO₄, 2.5 KCl, one MgCl₂, two CaCl₂, 15 glucose, 25 TEA-Cl. Internal solution contained (in mM): 110 CsF, nine NaCl, 1.8 MgCl₂, four Mg-ATP, 0.3 Na-GTP, 0.09 EGTA, 0.018 CaCl₂, nine HEPES, 10 TEA-Cl. CsOH was used to adjust to pH = 7.35 (290–310 mOsm).

Sciatic nerve block model

Anesthetized with 2% isoflurane (Abbott Pharmaceutical Co. Ltd., Shanghai, China), total volume of 50 μL study solution was injected into the popliteal fossa of the mice and the needle was removed 5 s after injection. The injected site was then pressed for at least 5 s to prevent exudation of study solution. A total of 56 mice were randomly divided into seven groups (n = 8/group): respectively receiving lidocaine alone, CPZ alone, ZD7288 (HCN channel blocker) alone, CPZ + lidocaine, ZD7288 + lidocaine, ZD7288 + CPZ + lidocaine, forskolin (an activator of adenylyl cyclase) + CPZ + lidocaine. The concentrations of lidocaine, ZD7288, CPZ and forskolin were 1% (w/v, ∼35 mM), five mM, 53 and 153 μM, respectively (Kroin et al., 2004; Zuo et al., 2013). For the groups that received two or three drugs, ZD7288 and/or forskolin were injected 20 min before lidocaine and CPZ was injected 10 min before lidocaine (Ando et al., 2004; Kroin et al., 2004).

After the mice recovered to consciousness (∼5 min after injection), a 25-G needle (Terumo Medical Corp., Tokyo, Japan) tied to a plastic fiber (∼3 g equivalent to the von-frey fiber) was used to evaluate noxious sensory function of the injected limb. Noxious sensory block was defined as no purposeful aversive movements to the pinprick stimulus. Von-frey filaments (37450-277; Ugo Basile, Comerio, Italy) was used to test tactile sensation of the injected limb and tactile sensory block was defined as no purposeful reactions to increased von-frey fibers more than two g compared to baselines. The specific site for application of pinprick and von-frey fiber was lateral plantar surfaces innervated by sciatic nerve. Thermal sensory function was determined by heat stimulus (Plantar test 37370; Ugo Basile, intensity of infrared radiation set at 60) applied to the injected limb. Thermal sensory block was defined as the withdraw latency of the mice increased more than 50% of the baseline latency (Becker & Reed, 2012). Cut-off time of thermal stimulus was 10 s to avoid injury. Motor function of mice was evaluated by the ability of hanging upside down with the injected paws (Yamada et al., 2016). Before injection, baselines of each function were measured. Because TRPV-1 channel blockers have been reported to induce dysregulation of body temperature (Yang et al., 2015), rectal temperature of the study mice was observed. For all the studied mice, rectal temperatures were normal (36–38 °C) during the regional anesthesia. All the mice completely recovered from regional anesthesia without any adverse events.

Compound action potential recording on isolated sciatic nerves

Adult C57BL/6J mice were used. Bilateral sciatic nerves were obtained from the lumbar plexus to the knee. Then the sciatic nerves were put into a Ringer’s solution (115.5 mM NaCl, two mM KCl, 1.8 mM CaCl₂, 1.3 mM Na₂HPO₄ and 0.7 mM NaH₂PO₄, pH = 7.35) for 20 min. BL-420N biological signal acquisition and analysis system (Techman Software Co. LTD, Chengdu, China) was used. Baseline values of compound action potentials (CAPs) were recorded every 5 min until 20 min to obtain a stable baseline. Then, the CAPs of each group were obtained after the sciatic nerve incubating into the drug solutions for 5 min. The parameters of the system were set as: voltage = one V; frequency = 100 Hz; duration of rectangular pulse = 0.1 ms (Li et al., 2010).

Statistical analysis

SPSS 22.0 (SPSS Inc., Chicago, IL, USA) was used to analyze the data. The sample size (n = 8/group) of the animal study in vivo was based on our preliminary test (n = 4, not included in formal data). The duration of lidocaine for noxious sensory block was compared between the groups of lidocaine alone and lidocaine + CPZ (20.5 vs. 30.5 min). Then the sample size was calculated as 5.0 (α = 0.05; β = 0.20). Duration data were calculated by averaging the duration of each mouse in the same group and expressed as mean ± SEM. The homogeneity of variance tests indicated that the duration data were normaly distributed. One-way analysis of variance (ANOVA) followed by post hoc of Bonferroni was applied to compare the durations between groups. Kaplan–Meier followed by pair-wise over strata of Log-rank was applied to compare the recovery curves of motor function block and noxious sensory block. For the results of electrophysiological recordings, all the data were presented as mean ± SEM. Voltage-dependent activation curves of I_h were fit to the Boltzmann equation. Two-way ANOVA or student’s t-test was applied to compare I_h and channel gating as indicated. In all cases, P < 0.05 was considered as statistically significant.

CPZ inhibited transfected mHCN channel currents and hyperpolarized voltage-dependent activation

Capsazepine at concentration of 10 μM potently inhibited I–V curves of both mHCN1 and mHCN2 channels (Fig. 1). Median inhibitory concentration (IC₅₀) of CPZ on peak I_h were 8.2 ± 0.3 and 9.6 ± 0.4 μM, respectively for mHCN1 and mHCN2 (P > 0.05, Fig. 2). CPZ at concentration of 10 μM significantly caused a hyperpolarizing shift in voltage-dependence of channel activation for both mHCN1 (V_1/2a of −98.4 ± 3.8 vs. −88.6 ± 3.6 mV, P < 0.01, Fig. 3) and mHCN2 (V_1/2a of −102.7 ± 13.5 mV vs. −114.6 ± 12.9 mV, P < 0.01, Fig. 3).

CPZ prolonged duration of lidocaine in sciatic nerve block model in vivo

ZD7288 alone (five mM) did not produce any sciatic nerve block in vivo. CPZ alone (53 μM) produced thermal sensory block with duration of 24.6 ± 2.7 min (Table 1). Compared with lidocaine alone, adding ZD7288 prolonged durations of lidocaine including noxious sensory block (34.3 ± 5.2 vs. 20.3 ± 1.7 min, P < 0.001, Fig. 4A), tactile sensory block (25.0 ± 4.3 vs. 20.0 ± 3.7 min, P = 0.036, Fig. 4B), thermal sensory block (33.4 ± 4.5 vs. 26.8 ± 5.5 min, P = 0.027, Fig. 4C) and motor function block (27.9 ± 4.3 vs. 20.9 ± 4.2 min, P = 0.008, Fig. 4D). Compared with ZD7288 + lidocaine group, CPZ similarly extended the durations of lidocaine including noxious sensory block, tactile sensory block and motor block. For noxious sensory block, CPZ prolonged the duration from 20.3 ± 1.7 to 35.1 ± 3.3 min (P < 0.001 vs. lidocaine alone, Fig. 4A). For tactile sensory block, CPZ prolonged the duration from 20.0 ± 3.7 to 25.5 ± 4.4 min (P = 0.025 vs. lidocaine alone, Fig. 4B). For motor function block, CPZ prolonged the duration from 20.9 ± 4.2 to 28.6 ± 4.1 min (P = 0.003 vs. lidocaine alone, Fig. 4D). For thermal sensory block, CPZ + lidocaine group was significantly longer than that of ZD7288 + lidocaine group (39.6 ± 6.6 vs. 33.4 ± 4.5 min, P = 0.017, Fig. 4C). Adding ZD7288 and CPZ to lidocaine (ZD7288 + CPZ + Lido group) did not further prolong the duration of lidocaine for noxious sensory block compared to ZD7288 + Lido group or CPZ + Lido group (Fig. 4A). Of note, the durations of tactile sensory block, thermal sensory block and motor block in ZD7288 + CPZ + Lido group were not determined because slight sedation was found in some mice, which might affect the accuracy of the results. The pinprick stimulus (puncturing the skin by a 25-G needle) was much stronger than other stimulus like von-frey and in the manner of “all or none” compared with other quantitative measurements (e.g., von-frey and heat). No redness or swelling was found in the test limb during the whole experiment. All mice recovered completely after regional anesthesia. The time courses of all the regional anesthetic effects were shown in Fig. 5.

Pre-injection of forskolin reversed the prolongation of CPZ in regional anesthesia

Pre-injection of forskolin (153 μM) diminished the prolongation of CPZ on durations of lidocaine (Figs. 4 and 5). The anesthetic durations were shorter in Forskolin + CPZ + Lido group than that in CPZ + Lido group including noxious sensory block (35.1 ± 3.3 vs. 18.9 ± 3.0 min, P < 0.001, Fig. 4A), tactile sensory block (25.5 ± 4.4 vs. 18.5 ± 5.0 min, P = 0.015, Fig. 4B), thermal sensory blockade (39.6 ± 6.6 vs. 24.8 ± 5.2 min, P < 0.001, Fig. 4C) and motor blockade (28.6 ± 4.1 vs. 19.5 ± 3.0 min, P = 0.001, Fig. 4D). All the data were shown in Table 1. Of note, pre-injection of forskolin was prone to shorten the durations of lidocaine compared with lidocaine alone group, indicating the contribution of HCN channels in the effects of lidocaine (Figs. 5A, 5B and 5D).

cAMP reversed the inhibitory effect of CPZ on I_h in DRG neurons

The average current density of I_h was 2.3 ± 0.4 pA/pF in control condition, while intracellular cAMP at 200 μM significantly increased the current density of I_h to 7.6 ± 2.3 pA/pF (P < 0.05, Fig. 6B). CPZ at concentration of 10 μM inhibited I_h density to 1.9 ± 0.7 pA/pF. Intracellular cAMP at 200 μM reversed I_h density to 7.1 ± 1.6 pA/pF even with perfusion of 10 μM CPZ (P < 0.05, Fig. 6D).

CPZ enhanced inhibition of lidocaine on action potential in sciatic nerves

Capsazepine at concentration of 10 μM did not inhibit I_Na in isolated mice DRG neurons (Fig. 7A), while 30 μM CPZ significantly inhibited the current density of I_Na from 105.9 ± 23.3 to 68.2 ± 20.3 pA/pF (P < 0.05, Fig. 7C). However, neither 10 nor 30 μM CPZ inhibited the CAP amplitudes in isolated sciatic nerves compared to baseline (Figs. 7E–7H). The CAP amplitudes were inhibited by 43.3% ± 1.5% after incubation in 1% lidocaine for 5 min. CPZ dose-dependently enhanced the inhibitory effect of 1% lidocaine on the CAP. For 1% lidocaine + 10 μM CPZ group and/or 1% lidocaine + 30 μM CPZ group, the CAP amplitudes were decreased to 20.5% ± 3.0% and/or 12.6% ± 0.8% compared to baseline, respectively (P < 0.01 vs. 1% lidocaine, Fig. 7I).