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

Background and objectives Hyperpolarization-activation cyclic nucleotide-gated (HCN) channels contribute to the effects of lidocaine. Capsazepine (CPZ), a competitive inhibitor of capsaicin of transient receptor potential vanilloid-1 channel, has also been found to inhibit HCN channel currents (Ih). This study was designed to investigate whether CPZ could prolong durations of lidocaine in regional anesthesia. Methods Mouse HCN1 and HCN2 channels were expressed in human embryonic kidney 293 (HEK 293) cells. The effect of CPZ on Ih was measured by whole-cell patch-clamping recording. Sciatic nerve block model in mice was used for the study in vivo. The mice were randomly divided into seven groups, respectively, receiving lidocaine, CPZ, ZD7288 (HCN channel blocker), CPZ + lidocaine, ZD7288 + lidocaine, ZD7288 + CPZ + lidocaine, forskolin (an activator of adenylyl cyclase) + CPZ + lidocaine. Regional anesthetic durations of lidocaine were determined. Voltage-gated sodium channel currents (INa) and Ih were recorded in dorsal root ganglion neurons of mice. The effects of CPZ on INa and Ih with or without Cyclic adenosine monophosphate (cAMP) were assessed. Isolated mice sciatic nerve was prepared to evaluate the effect of CPZ on the compound action potentials (CAP). Results Capsazepine non-selectively inhibited transfected mHCN1 and mHCN2 channel currents in HEK 293 cells. In sciatic nerve block in vivo, compared to lidocaine alone, adding CPZ extended the durations of lidocaine for noxious sensory block (35.1 ± 3.3 vs. 20.3 ± 1.7 min), tactile sensory block (25.5 ± 4.4 vs. 20.0 ± 3.7 min), thermal sensory block (39.6 ± 6.6 vs. 26.8 ± 5.5 min), and motor function block (28.6 ± 4.1 vs. 20.9 ± 4.2 min). Duration of thermal sensory block was longer in CPZ + lidocaine group than that of ZD7288 + lidocaine group (39.6 ± 6.6 vs. 33.4 ± 4.5 min). Forskolin reversed the prolongation by CPZ on lidocaine durations. CPZ or ZD7288 alone did not produce typical regional anesthetic effects. Increased intracellular concentration of cAMP reversed the inhibition of CPZ on Ih. Although CPZ alone inhibited INa at the concentration more than 30 μM, it did not inhibit the CAP amplitudes in isolated sciatic nerves. CPZ dose-dependently enhanced the inhibitory effect of 1% lidocaine on the CAP amplitudes. Conclusions Capsazepine may prolong durations of lidocaine in peripheral nerve block by modulation of HCN channel currents.


INTRODUCTION
Although local anesthetics such as lidocaine have been widely used in clinics for decades, there are still many disadvantages for currently used local anesthetics including toxicities and unsatisfied durations (Mather, 2010). To prolong the duration of local anesthetics, many adjuvants have been added in clinical regional anesthesia (Kirksey et al., 2015). However, the enhancement of most adjuvants are unsatisfied and some of the adjuvants can even induce unacceptable complications (Lundblad & Lonnqvist, 2016).
Voltage-gated sodium channel is the principal molecular target for local anesthetics (Becker & Reed, 2012). Recent studies have identified that hyperpolarization-activation cyclic nucleotide-gated (HCN) channel might also contribute to the effects of local anesthetics in regional anesthesia (Meng et al., 2011;Zhou et al., 2015). HCN channel is a non-selective cation channel activated by hyperpolarization of membrane potential (Benarroch, 2013). HCN channel is the principal pacemaker for the rhythmical regulator and determinant for resting membrane potential in the nervous system and sinoatrial node for its distinct electrophysiological properties (Benarroch, 2013;Biel et al., 2009). HCN channel also plays an important role in neuropathic pain (Cao, Pang & Zhou, 2016). Our previous study indicates that the duration of lidocaine is significantly shortened in HCN1 knockout mice (Zhou et al., 2015). Cyclic adenosine monophosphate (cAMP) is the enhancer for HCN channel currents (I h ) (Wainger et al., 2001). Increased level of cAMP by activator of adenylyl cyclase (e.g., forskolin) can shorten the durations of lidocaine in vivo (Kroin et al., 2004;Zhou et al., 2015). In a previous study, maximal inhibition of lidocaine on I h is ∼50% (Meng et al., 2011). Therefore, adding more potent HCN channel modulators (e.g., ZD7288, clonidine and dexmedetomidine) to local anesthetics have been found to extend the durations of lidocaine and ropivacaine in vivo (Brummett et al., 2011;Harris & Constanti, 1995;Kroin et al., 2004).
Capsazepine (CPZ) is a competitive inhibitor of capsaicin activation of transient receptor potential vanilloid-1 (TRPV-1) channel and has been reported as an emerging class of novel analgesic (Menendez et al., 2006). TRPV-1 channel is mainly expressed on nociceptor and involved in sensory processing (Cui et al., 2006). TRPV-1 channel can be activated by high temperature, acid, capsaicin and other chemicals (Szallasi et al., 2007), by which to trigger action potentials and release of neuropeptides and/or transmitters (Cui et al., 2006;Peters et al., 2010). Therefore, TRPV-1 channel blocker might be potent in the therapy of pain. CPZ has also been found to inhibit HCN channel currents (I h ) (Zuo et al., 2013).
In the present study, we hypothesized that CPZ, a potential adjuvant for local anesthetics, might extend the durations of lidocaine by modulation of HCN channels. This study was designed to test this hypothesis.

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.
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(logIC 50 -X) Ã 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 mM 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 NaHCO 3 , 1.25 NaH 2 PO 4 , 2.5 KCl, one MgCl 2 , two CaCl 2 , 15 glucose, 25 TEA-Cl. Internal solution contained (in mM): 110 CsF, nine NaCl, 1.8 MgCl 2 , four Mg-ATP, 0.3 Na-GTP, 0.09 EGTA, 0.018 CaCl 2 , 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 mL 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 mM, 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 CaCl 2 , 1.3 mM Na 2 HPO 4 and 0.7 mM NaH 2 PO 4 , 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 (a = 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 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 mM) produced thermal sensory block with duration of 24.6 ± 2.7 min (Table 1). 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.

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 mM 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 mM inhibited I h density to 1.9 ± 0.7 pA/pF. Intracellular cAMP at 200 mM reversed I h density to 7.1 ± 1.6 pA/pF even with perfusion of 10 mM CPZ (P < 0.05, Fig. 6D). Motor function evaluation. Regional anesthetic durations of CPZ + Lido group and ZD7288 + Lido group were significantly longer than lido alone group, while forskolin reversed the prolongation by CPZ. In pin-prick test (A), ZD7288 + CPZ + Lido group did not produce any longer block than neither CPZ + Lido group nor ZD7288 + Lido group. For foot-thermal test (C), block duration of CPZ + Lido group was significantly longer than ZD7288 + Lido group. CPZ, capsazepine; Lido, lidocaine; Ã P < 0.05 vs. Lido alone group; ÃÃ P < 0.01 vs. Lido alone group; ## P < 0.01 vs. CPZ + Lido group. Sample size = 8.

DISCUSSION
Long-lasting local anesthetics are demanded in clinic especially for post-operative analgesia and treatment of chronic pain. To prolong duration of local anesthetics and reduce their The adjuvants for local anesthetics mainly include epinephrine, ketamine, opioids and NSAIDs (Bailard, Ortiz & Flores, 2014). However, the enhancement of most adjuvants are unsatisfied and some adjuvants can even induce unacceptable side effects themselves, such as poor selectivity, undesired systematic sedation or thermal irritation (Brummett & Williams, 2011). In the present study, CPZ significantly extended the durations of lidocaine for sciatic nerve block without obvious complications. Therefore, CPZ may be a potential adjuvant for local anesthetics. Capsazepine is a competitive inhibitor of capsaicin activation of TRPV-1 channel (Yang et al., 2015). At micromolar concentrations, CPZ can also block voltage-dependent potassium channel, calcium channel and sodium channel (Kuenzi & Dale, 1996;Yamamura et al., 2004;Lundbaek et al., 2005). In the present study, to test whether voltage-gated sodium channels are involved in the enhancement of CPZ on regional anesthesia of lidocaine in vivo, acute DRG neurons and isolated sciatic nerves were used for recordings of I Na and CAP, respectively. CPZ alone at a concentration of 30 mM indeed inhibited I Na in DRG neurons. However, CPZ did not affect amplitudes and/or conductions of CAP in isolated sciatic nerves at concentration of 30 mM. A previous study indicates that inhibition of CPZ on I Na is possibly by altering lipid bilayer elasticity of the cellular membrane (Lundbaek et al., 2005). Therefore, the differential results of CPZ between voltage-gated I Na and action potential might result from the myelin sheath around nerves that preventing direct exposure of CPZ to the cellular membrane. In spite of no effect was found for CPZ alone on CAP, it can enhance the inhibitory effect of lidocaine on CAP amplitudes since the concentration of 10 mM. Consistent with the observation in vivo, CPZ alone did not produce typical sciatic nerve block, but prolonged the durations of lidocaine. Although CPZ may not inhibit action potential alone by inhibition to sodium channels, it cannot completely exclude the possible contribution of sodium channel block by CPZ for its enhancement in regional anesthesia. Even a small additional inhibition of sodium channel might prolong the effects of lidocaine. Capsazepine can inhibit transfected human HCN channel currents (Zuo et al., 2013). HCN channel is widely expressed in the peripheral nervous system (Acosta et al., 2012;Cao, Pang & Zhou, 2016). HCN channels may contribute to the regional anesthetic effects of lidocaine in vivo, both for the sensory and motor function block (Zhou et al., 2015). By electrophysiological recordings, lidocaine was found to inhibit I h (Meng et al., 2011). However, lidocaine at clinical concentrations only inhibits I h by ∼30-50% (Meng et al., 2011). Thus potent HCN channel blocker might enhance the effects of local anesthetics (Brummett et al., 2011). HCN channel modulators (e.g., ZD7288, clonidine and dexmedetomidine) have been found to enhance regional anesthetic effects of lidocaine and/or ropivacaine in vivo (Brummett et al., 2011;Harris & Constanti, 1995;Kroin et al., 2004). Although CPZ can inhibit transfected human HCN channel currents (Zuo et al., 2013), it is unclear whether CPZ can enhance regional anesthesia in vivo. In the present study, CPZ inhibits transfected mice HCN channel currents without selectivity between HCN1 and HCN2. The IC 50 of CPZ measured in this study (∼10 mM) is similar to the potency on human I h (Gill et al., 2004;Zuo et al., 2013), indicating the enhancement of CPZ for regional anesthesia might be conserved between species.
ZD7288 and CPZ similarly prolong the durations of lidocaine. Adding both ZD7288 and CPZ to lidocaine did not further prolong the duration of noxious sensory block compared with ZD7288 + Lido or CPZ + Lido group. Therefore, enhancement of ZD7288 and CPZ in regional anesthesia might be both mediated through modulation of I h . For DRG neurons, intracellular cAMP reversed the inhibitory effect of CPZ on I h . The enhancement by CPZ for lidocaine in sciatic nerve block was eliminated by pre-injection of forskolin. Forskolin can increase the intracellular level of cAMP (Brummett et al., 2011) and is widely used to enhance I h (Wainger et al., 2001;Kroin et al., 2004). These results provide better evidence to support the hypothesis that modulation of CPZ on I h may contribute to its enhancement in regional anesthesia. Forskolin will also likely increase local blood flow and enhance drug clearance. However, the concentration of forskolin used in this study was 153 mM, which is lower than the blood concentration (∼400 mM) to influence the blood flow (Wysham, Brotherton & Heistad, 1986). Forskolin reversed the prolongation of CPZ most potently in von-frey test than pin-prick test, foot-thermal test and motor function evaluation (Fig. 5); thus the selective effect of forskolin might not be explained by increasing the blood flow (no selectivity). Interestingly, the enhancement of cAMP between HCN1 and HCN2 is different (Wainger et al., 2001), which might contribute to the differential reversion by forskolin between different functions.
Capsazepine alone can produce thermal sensory block without lidocaine. The thermal sensory block duration in CPZ + lidocaine group is longer than that of ZD7288 + lidocaine group. The selective enhancement of CPZ for thermal sensory might result from the inhibition of TRPV-1 channels. TRPV-1 channel blocker is a potential treatment for various types of pain (Walker et al., 2003). However, because of the role of TRPV-1 channel as a temperature sensor, systemic administration of TRPV-1 channel blocker can induce serious hyperpyrexia (Menendez et al., 2006), which limits the clinical application of TRPV-1 channel blocker. The systemic dosage of CPZ to induce hyperpyrexia is unknown, but other TRPV-1 antagonists such as AMG517, GRC6211 and NGD8243 have been dropped out of clinical trials due to hyperpyrexia (Pareek et al., 2007). In this study, CPZ is perineurally injected and no obvious change was found for mice body temperature.
There are some limitations in the present study: Firstly, forskolin, an activator of adenylyl cyclase, is not a selective agonist of HCN channels. Forskolin can increase the level of cAMP, which is an enhancer of HCN channels (Wainger et al., 2001;Kroin et al., 2004). Secondly, ZD7288 and CPZ are non-selective blockers between HCN1 and HCN2 subtypes. Thus, we cannot evaluate the specific role of HCN1 and HCN2. Further studies with HCN selective blockers such as MEL57A and/or EC18 might shed light on this question (Del Lungo et al., 2012). In addition, the potential effect of CPZ on HCN4 (cardiac subtype) should be tested further to explore the potential influence of CPZ on the heart. Co-injection of vasoconstrictor and CPZ may not only prolong the durations of local anesthetics, but also reduce the systemic impact of CPZ in regional anesthesia.

CONCLUSIONS
In summary, CPZ, a widely known competitive inhibitor of capsaicin activation of TRPV-1 channel, can inhibit I h and prolong durations of lidocaine in vivo. The enhancement of CPZ in regional anesthesia may be mediated by modulation of I h . Therefore, CPZ is a potential regional anesthetic adjuvant in the clinic.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
This work was supported by the grant 81771456 (to Cheng Zhou) and 81571353 (to Jin Liu) from the National Natural Science Foundation of China. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.