pH‐dependent modulation of TRPV1 by modality‐selective antagonists

Antagonists of TRPV1 that inhibit all activation modes cause hyperthermia, hampering their medical use as novel analgesics. TRPV1 antagonists that do not (fully) inhibit responses to low pH do not cause hyperthermia, but it remains incompletely understood how such antagonists affect channel gating. We tested the hypothesis that pH‐sparing antagonists act in a modality‐selective manner on TRPV1, differentially affecting channel activation by protons and capsaicin.

Current treatment options for chronic pain all have their drawbacks, including limited efficacy or the propensity of abuse and addiction (Skolnick, 2018).Extensive research in the last few decades has focused on peripheral targets implicated in chronic pain states, yielding potential novel, effective and safe avenues for pain relief (Ciotu & Fischer, 2020).In particular, transient receptor potential channel vanilloid 1 (TRPV1), a heat-, proton-and capsaicin-activated channel on nociceptor neurons has been at the centre of attention following the discovery that its genetic ablation or pharmacological inhibition alleviates pathological pain in various animal models (Bamps et al., 2021;Caterina et al., 2000Caterina et al., , 1997;;Davis et al., 2000;Ghilardi et al., 2005;Walker et al., 2003).However, despite two decades of research and development, not a single TRPV1 antagonist has been approved for clinical use, even though many drug candidates reached clinical trials (Bamps et al., 2021;Walker et al., 2003).
The primary reason for discontinuation of the clinical development of the majority of first-generation TRPV1 antagonists was hyperthermia (a rise in core body temperature), which was observed in both animal models and humans upon dosing small-molecule antagonists such as AMG-517 (Gavva et al., 2008), mavatrep (JNJ-39439335) (Manitpisitkul et al., 2018), ABT-102 (Othman et al., 2013), or AZD1386 (Krarup et al., 2011).Hypothermia is an ontarget effect, because TRPV1-deficient mice do not experience TRPV1 antagonist-induced changes in body core temperature (Gavva, 2008).The exact mechanisms driving this effect remain incompletely understood.Despite these obstacles, the pursuit of TRPV1 antagonists yielded some fruits in the form of compounds that have no or reduced effects on core body temperature, at least in laboratory animals.These compounds are often referred to as modality-selective TRPV1 antagonists, to describe their differential effect on different modes of TRPV1 activation (Arsenault et al., 2018;Reilly et al., 2012) as compared to the polymodal first generation antagonists.For instance, compounds such as A-1165442 (Reilly et al., 2012), JYL1421 (Gavva et al., 2007), or SB-366791 (Sakakibara et al., 2019) fully block TRPV1 responses evoked by vanilloids and heat, while (at least partly) preserving channel activity evoked by low pH and do not cause hyperthermia (Garami et al., 2010).Furthermore, antagonists that inhibit heat and vanilloid responses but potentiate responses to low pH, such as A-1165901 or AMG7905 (Garami et al., 2018), evoke hypothermia in rats.Notably, both hypothermic and hyperthermic responses to TRPV1 antagonists are abolished by localized, intra-abdominal TRPV1 desensitization or selective elimination of TRPV1 in sensory neurons (Garami et al., 2018;Steiner et al., 2007;Yue et al., 2022).Taken together, these findings have led to the general view that TRPV1 antagonistinduced changes in body core temperature are caused by modulation of low pH-activated TRPV1 activity in sensory neurons innervating the abdominal cavity, such as the intra-abdominal viscera or muscles, providing important guidance for the development of safe treatments based on modality-selective TRPV1 antagonism (Bamps et al., 2021;Garami et al., 2018Garami et al., , 2020;;Steiner et al., 2007;Yue et al., 2022).
At this point, however, the mechanisms whereby modalityselective TRPV1 antagonists interact with channel gating remain poorly understood.It has been well established that heat, vanilloids and low pH all activate TRPV1 by shifting the voltage dependence of channel activation towards more negative voltages (Aneiros et al., 2011;Voets et al., 2004), but it is unclear whether polymodal and modality-selective TRPV1 antagonists differentially interfere with the channel's voltage dependence.Moreover, it is generally assumed that activation of TRPV1 by vanilloids, heat and low pH occurs via partially separate, parallel gating modes that converge on channel opening (Aneiros et al., 2011;Jordt et al., 2000;Jordt & Julius, 2002) and that modality-selective antagonists specifically leave the pH-sensitive gating mode intact (Garami et al., 2018(Garami et al., , 2010)).However, the latter assumption has never been experimentally tested.
To address these issues, we evaluated the effect of a selection of TRPV1 antagonists, including hyperthermia-inducing mavatrep (AMG-517), thermoneutral A-1165442 and hypothermia-inducing AMG7905 on wild type TRPV1 and on the pH-insensitive F660A mutant at different pH values.Our data reveal that A-1165442 and AMG7905 modulate capsaicin-induced responses in a pH-dependent manner.These findings oppose a strictly modality-selective action of these compounds.Instead, they support a model in which low pH affects the activity of these antagonists at TRPV1 for different activation modalities.

What is already known
• TRPV1 antagonists are analgesic but cause hyperthermia, however modality-selective TRPV1 antagonists do not cause hyperthermia.
• These antagonists are thought to block TRPV1activation by capsaicin but not that by low pH.

What this study adds
• Modality-selective TRPV1 antagonists, A-1165442 and AMG7905 affect TRPV1 activation by capsaicin in a pHdependent manner.
• These findings challenge the concept of modalityselective of these compounds

What is the clinical significance
• Such insights may aid the development of efficient and safe of TRPV1-based treatment for pain.
Fluorescence was measured during alternating illumination at 340 and 380 nm using either a Cell M (Olympus) or Eclipse Ti (Nikon) fluorescence microscopy system, and absolute calcium concentrations were calculated from the ratio of the fluorescence signals at these two where K m , R min and R max were estimated from in vitro calibration experiments with known calcium concentrations (Vandewauw et al., 2018).The bath solution contained (in mM): 150 NaCl, 4 KCl, 2 CaCl 2 , 2 MgCl 2 and 10 HEPES (pH 7.4 with NaOH) or 10 MES (pH 5.5, 5.0, 4.5 and 4.0 with NaOH).

| Whole-cell patch-clamp
Membrane currents in TRPV1-expressing HEK293T cells were measured in the whole-cell configuration using an EPC-10 amplifier and PatchMaster Software (HEKA Elektronik).Currents were measured in response to either a ramp protocol, consisting of a 200-ms voltage ramp from À100 to 160 mV applied at 0.5 Hz or a step protocol, consisting of 100-ms voltage steps ranging from À120 to 160 mV, 20 mV spaced.Holding potential was 0 mV.Voltage-dependent activation curves represent the normalized conductance, G/G max , at the end of the voltage steps, in function of the applied voltage.Activation curves were fitted to a Boltzmann function of the following form: where V 1/2 represents the midpoint of the activation curve, V s the slope factor and G max the maximal conductance.We performed a global fit to all activation curves obtained within a single cell under different conditions (i.e., extracellular pH 7.4 or 5.5; absence or presence of capsaicin or TRPV1 antagonists), assuming constant G max and maximally 10% variation in V s , which allowed us to estimate V 1/2 values even under conditions where the conductance did not saturate at the with NaOH).GFP fluorescence upon excitation at 480 nm was used to identify transfected cells (Vandewauw et al., 2018).

| Statistical analysis and sample sizes
The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology (Curtis et al., 2022)  3 | RESULTS

| A-1165442 fully reverts proton-evoked
TRPV1 current activation at pH 5.5 We performed whole-cell patch-clamp experiments to investigate the effects of the thermoneutral antagonist A-1165442 on human TRPV1 currents activated by protons at pH 5.5 and to compare its activity with that of the modality-independent, hyperthermia-inducing antagonists mavatrep and AMG-517.We used a voltage step protocol ranging from À120 to 160 mV to address voltage-dependent modulation (Voets et al., 2004).In line with earlier work (Aneiros et al., 2011) we evoked by the A-1165442 was attributed to this acid-sparing property (Reilly et al., 2012;Voight et al., 2014).We therefore compared the concentration dependence of the effect of A-1165442 on TRPV1 currents activated by capsaicin (50 nM).In line with earlier work (Voets et al., 2004) and with the results using low pH as a stimulus, we found that capsaicin causes a robust shift of the voltage dependence of activation towards more negative potentials, with a mean change in V 1/2 of À234 ± 10 mV, and that A-1165442 caused a concentration-dependent shift in the voltage-dependent activation curve towards positive potentials.Overall, the concentration dependence of the A-1165442-induced changes in V 1/2 for capsaicinactivated currents was not different from the results obtained for currents activated by low pH (Figure 1d).Cumulatively, these data demonstrate that A-1165442 is equipotent at inhibiting TRPV1 currents activated by capsaicin or pH 5.5.

| Reduced inhibitory efficacy of A-1165442 to acid activation at pH values <5.5
In search for an explanation for the apparent discrepancy between our findings and the previously reported modality-selective properties of A-1165442 (Reilly et al., 2012), we noted that our patch-clamp experiments were performed at pH 5.5 whereas these earlier studies used a slightly lower pH 5.0 as stimulus in their Ca 2+ influx assays. We

| Reduced inhibitory efficacy of A-1165442 at low pH in the pH-insensitive mutant F660A
The pH-dependent effects of A-1165442 on acid-evoked TRPV1 responses seem at odds with the proposed modality-selective mode of action of the compound.Indeed, if A-1165442 would selectively spare the proton-activated gating mode while fully inhibiting the vanilloidactivated mode, the prediction would be that A-1165442 would only partially inhibit acid-activated responses at any acidic pH.We therefore considered an alternative model for the mode of action of this compound, where A-1165442 is a modality-independent antagonist, similar to mavatrep and AMG-517, but with antagonistic efficacy that is compromised at very acidic pH.If this model were correct, we would expect that A-1165442 is less effective in inhibiting TRPV1 activated by other stimuli (e.g.vanilloids) when tested at pH values ≤5.0.
To directly test this prediction, we used the pH-insensitive TRPV1 mutant F660A, which renders the channel fully insensitive to pHinduced channel activation or potentiation, while fully preserving activation by capsaicin (Aneiros et al., 2011) Notably, the inhibitory effect of A-1165442 on the capsaicin-evoked responses was markedly attenuated when tested at pH 4.5.Full inhibition was only obtained at 100 nM and the IC 50 shifted to 11.4 ± 1.0 nM.A mildly reduced potency was observed at pH 5.5, with an IC 50 value of 2.96 ± 0.18 nM.
Cumulatively, our data indicate that A-1165442 is not a strictly modality-selective antagonist.Instead, our findings suggest that A-1165442 inhibits both modes of TRPV1 activation, similar to modality-independent antagonists such as mavatrep and AMG-517, but with antagonistic efficacy that is compromised at very low pH values.

| AMG7905 potentiates acid-evoked TRPV1 responses
AMG7905 is an example of a TRPV1 antagonist that causes hypothermia, which has been attributed to a modality-selective mode of action whereby it enhances acid-activated TRPV1 responses while potently inhibiting responses to capsaicin (Garami et al., 2018).Using Fura-2-based intracellular calcium imaging experiments we indeed found that responses of WT TRPV1 to acid stimulation at pH 5.5 were significantly (by 32 ± 9%) and reversibly increased in the presence of 1 μM AMG7905, a concentration at which the compound fully inhibits capsaicin-evoked responses (Figure 5a,b).

| Reduced inhibitory efficacy and partial agonism of AMG7905 at low pH in the pH-insensitive mutant F660A
To further evaluate whether the enhanced responses reflect a shown in Figure 6a, AMG7905 at a concentration of 10 nM caused full inhibition of the capsaicin response at pH 7.4, whereas a substantial residual capsaicin response was observed at pH 5.5.Full inhibition of the capsaicin response at pH 5.5 was observed at an AMG7905 concentration of 1 μM.
Notably, at pH 5.5 we also observed a small but sizable and concentration-dependent increase in intracellular calcium concentration upon application of AMG7905, which was never observed at pH 7.4 or in non-transfected HEK293 cells.In contrast, at pH 7.4, AMG7905 caused a small decrease in intracellular calcium concentration, possibly reflecting inhibition of basal TRPV1 activity.Taken together, these data indicate that the inhibitory potency of AMG7905 on capsaicin-evoked responses is attenuated at pH 5.5 and that the compound by itself has a partial agonistic effect at pH 5.5.

| DISCUSSION
Based on extensive pharmacological and genetic studies, TRPV1 represents a validated therapeutic target in several preclinical models of chronic pain, including inflammatory, neuropathic and cancer-related pain (Caterina et al., 2000(Caterina et al., , 1997;;Davis et al., 2000;Engler et al., 2007;Ghilardi et al., 2005;Gouin et al., 2017;Morgan et al., 2005;Rigoni et al., 2003).However, the clinical development of TRPV1 antagonists has been hampered by important on-target side effects that were observed for the first generation of polymodal antagonists, most prominently the occurrence of pronounced hyperthermia (Gavva, 2008).This fuelled the search for next-generation TRPV1 antagonists that do not increase core body temperature (Garami et al., 2018;Lehto et al., 2008;Voight et al., 2014).Important progress was made with the discovery that, for several structurally unrelated antagonists, changes in core body temperature correlate with the effect of the antagonists on acid-evoked TRPV1 responses (Garami et al., 2010(Garami et al., , 2020)).Antagonists that potently inhibit capsaicin-evoked TRPV1 responses but do not (fully) block acidevoked responses do not cause hyperthermia and antagonists that potentiate acid-evoked TRPV1 responses even cause hypothermia (Garami et al., 2018(Garami et al., , 2010(Garami et al., , 2020;;Lehto et al., 2008).These findings provided new opportunities for the rational development of therapies using tailored TRPV1 antagonists that do not cause unwanted changes in core body temperature.Yet, the precise mechanisms whereby modality-selective antagonists interact with TRPV1 remain poorly understood.The general tenet is that vanilloids, heat and low pH act via separate, parallel gating processes that converge on channel opening, and that pH-sparing antagonists differentially affect these gating modes (Garami et al., 2020).Our current results indicate that this view may need to be revised, at least for the two prototype compounds we tested, A-1165442 and AMG7905.
A-1165442 was reported to be a modality-selective antagonist that fully inhibits capsaicin-evoked TRPV1 responses, while only partly blocking TRPV1 responses evoked by acid (Reilly et al., 2012).
However, in our patch-clamp experiments we found that A-1165442 caused a concentration-dependent inhibition of acid-evoked responses at pH 5.5, with full inhibition at sub-micromolar concentrations.Similar to the polymodal antagonists mavatrep and AMG517, A-1165442 acted by shifting the voltage dependence of activation towards positive potentials, thereby reverting the negative shift caused by low pH (Aneiros et al., 2011;Voets et al., 2004).Moreover, the concentration dependence of the effect of A-1165442 on channel activation was identical for TRPV1 activation by pH 5.5 or capsaicin, arguing against a modality-selective mode of action.Nevertheless, we observed that the inhibitory potency of A-1165442 on pH-evoked responses was reduced at pH values ≤ 5.0.To investigate the cause of this pH-dependent effect, we used the TRPV1 F660A mutant, which is fully insensitive to acid-induced activation while showing normal sensitivity to capsaicin (Aneiros et al., 2011).Interestingly, we found that the potency of A-1165442 to inhibit capsaicin-evoked responses in this mutant was also reduced at more acidic pH.Taken together, these data indicate that the inhibitory potency of A-1165442 is reduced at acidic pH for both capsaicin and acid activation, and thus do not support a strictly modality-selective mode of action.
AMG7905 is an example of a TRPV1 antagonist that potentiates acid-evoked responses and causes hypothermia (Garami et al., 2018).
Our calcium imaging confirmed potentiation of the TRPV1 response to pH 5.5 by AMG7905, at a concentration (1 μM) that fully inhibited responses to capsaicin.Using the acid-insensitive TRPV1 F660A mutant (Aneiros et al., 2011), we again found that the potency of AMG7905 to inhibit capsaicin-evoked responses is reduced at pH 5.5.
Strikingly, at this low pH level, we observed that AMG7905 by itself acts as a partial agonist, evoking a small but sizeable and concentration-dependent calcium response.Taken together, these data confirm the differential effect of AMG7905 on the capsaicinevoked and acid-evoked modes of TRPV1 activation.However, they In support of this hypothesis is the observation that structural analogues of A-1165442 that lack the protonable isoquinoline moiety, as exemplified by A-1106625 (Reilly et al., 2012), cause full inhibition of acid responses evoked at pH 5.0.A similar protonation effect could underlie the reduced inhibitory potency and partial agonistic action at pH 5.5 of AMG7905, which contains several protonable aryl amines (Lehto et al., 2008).Alternatively, the electrostatic properties of key interaction sites for these antagonists on the TRPV1 protein may change at lower pH due to protonation of residues such as histidine, with a pK a of 6.0.Clearly, further structural and structure-activity studies will be required to fully clarify the molecular basis of the differential antagonist actions at lower pH.As illustrated in this study, the acid-insensitive TRPV1 F660A mutant represents a highly instrumental tool to distinguish between true modality-selective antagonism versus pH-dependent effects on multiple modes of channel gating.The finding that low pH reduces the antagonistic effect of certain TRPV1 antagonists, irrespective of the mode of channel activation, may need to be taken into account in models explaining the differential effects of TRPV1 antagonists on core body temperature.Moreover, extracellular pH can drop to values below pH 5.0 in pathologies such as cancer-associated bone pain, gastroesophageal reflux or asthma, conditions that have been studied as potential indications for the use of TRPV1 antagonists (Bamps et al., 2021;Choi et al., 2018;Ghilardi et al., 2005;Hunt et al., 2000;Yoneda et al., 2015).Therefore, reduced inhibitory potency and partial agonism (as seen for AMG7905) of TRPV1 antagonists at these low pH values may translate into reduced analgesic efficacy.
highest tested voltage.The extracellular solution contained (in mM):-150 NaCl, 2 MgCl 2 and 10 HEPES (pH 7.4, adjusted with NaOH) or 10 MES (5.5, or 4.5, adjusted with NaOH).The internal solution contained (in mM):-150 NaCl, 1 MgCl 2 , 1 EGTA and 10 HEPES (pH 7.2 on experimental design and analysis in pharmacology.Data analysis was performed using Origin software (v8.6-9.0;OriginLab).Data are represented as mean ± SEM from n cells.Statistical tests, one-way ANOVA with Tukey's post hoc test along with Student's t test were used were appropriate and are indicated in the figure legends.Post hoc tests were conducted only if F in ANOVA achieved P < 0.05, and P < 0.05 was considered as statistical significance of the difference.Based on previous experience and similar studies, the minimal number of cells for each condition in the patch-clamp experiments was set at n = 6, recorded on minimally two independent measurement days.Additional cells successfully recorded on those days were also included in the analyses.For calcium imaging, we set a minimum of five independently measured coverslips on minimally two measurement days.In each of these individually imaged coverslips, between 5 and 250 cells were recorded simultaneously.Representative time courses of changes in intracellular calcium concentration are shown as mean ± SEM from individual cells from a single coverslip.Drug responses, quantified as percentage inhibition, shift in voltage dependence or change in intracellular calcium concentration (Δ[Ca 2+ ] i ), were determined relative to the control condition within the same cell, reducing variability due to cell-to-cell differences in TRPV1 channel expression levels.2.5 | MaterialsAMG-517, mavatrep and A-1165442 (MedChemExpress) were added to the extracellular solutions from 1-mM stock solutions made in DMSO, capsaicin (Sigma) from a 5-mM stock solution made in ethanol, ionomycin (Alfa Aesar) from a 2-mM stock solution made in DMSO, and phorbol 12-myristate 13-acetate (PMA; Sigma) and 4αphorbol 12-myristate 13-acetate (4α-PMA; Santa Cruz Technology) from a 1-mM stock in DMSO.Details of other materials and suppliers are provided in the specific sections.2.6 | Nomenclature of targets and ligandsKey protein targets and ligands in this article are hyperlinked to corresponding entries in the IUPHAR/BPS Guide to PHARMACOLOGY in the Concise Guide to PHARMACOLOGY 2021/22(Alexander et al., 2021).
found that lowering the pH to 5.5 caused a robust shift of the voltage dependence of activation towards more negative potentials, with a change in the midpoint of activation (V 1/2 ) of À234 ± 10 mV (Figure1a-d).Application of mavatrep, AMG-517, or A-1165442 resulted in a concentration-dependent shift of the activation curve towards positive potentials (Figure1a-c).Notably, at the highest concentration tested, activation curves were fully shifted back to or beyond the basal level (average values for the change in V 1/2 (ΔV 1/2 ) were 254 ± 11 mV for mavatrep (10 nM), 252 ± 9 mV for AMG-517 (10 nM) and 247 ± 11 mV for A-1165442 (100 nM); Figure1d).Likewise, inward currents activated by pH 5.5 at À60 mV were concentration-dependently inhibited by all three antagonists, with full inhibition at the highest tested concentrations (Figure1e).Taken together, these data demonstrate that mavatrep, AMG-517 and A-1165442 inhibit TRPV1 currents activated at pH 5.5 by shifting the voltage dependence of activation towards positive potentials.F I G U R E 1 Mavatrep, AMG517 and A-1165442 inhibit acid-activated wild type TRPV1 currents by shifting the voltage dependence of activation.(a-c) Representative whole-cell TRPV1 currents in response to 100-ms voltage steps ranging from À120 to +160 mV, measured at pH 7.4 and at pH 5.5 in the absence or presence of the indicated concentrations of mavatrep (a), AMG517 (b) or A-1165442 (c).Each panel represents currents measured from one TRPV1-expressing cell (left) and the corresponding activation curves (right) from that cell.Solid lines represent the best fit of a Boltzmann function, assuming that the antagonists cause a shift in the midpoint of activation (V 1/2 ), without affecting G max (see Section 2 for more details).(d) Concentration dependence of the antagonistinduced rightward shift in V 1/2 (ΔV 1/2 ).The dotted line represents the mean leftward shift in V 1/2 upon switching from pH 7.4 to pH 5.5 (214 mV).For A-1165442, data are included for channel activation by both low pH and capsaicin (50 nM).(e) Concentration dependence of inhibition of inward current (at À60 mV) by the three antagonists.Each data point in (d, e) represents the mean ± SEM from six to nine cells.Whereas the full reversal of low pH-activated TRPV1 currents by mavatrep and AMG-517 was expected in the light of earlier studies, the full reversal with the reportedly modality-selective antagonist A-1165442 was unexpected.Indeed, earlier studies had reported that A-1165442 fully inhibits capsaicin-induced TRPV1 responses but only partially inhibits TRPV1-mediated responses activated by acid, even at micromolar concentrations.The absence of a hyperthermic effect hypothesized that A-1165442 may show reduced potency at lower pH values and therefore, conducted whole-cell patch-clamp recordings to compare the inhibitory effect A-1165442 (100 nM) at pH 5.5 and 4.5 (Figure2).Because the stability of whole-cell patchclamp recordings is often compromised at pH 4.5, precluding stringent protocols using voltage steps, we initially used a voltage ramp protocol ranging from À100 to 160 mV, applied every 2 s.In line with the experiments shown in Figure1, 100 nM A-1165442 fully reverted acidactivated inward and outward currents at pH 5.5 (Figure2a, b, e, f).However, at pH 4.5 the same concentration of A-1165442 caused only a partial inhibition, leaving a residual current of $30% of the maximally activated inward current (Figure2c-f).The pH dependence of this antagonistic effect of A-1165442 was further investigated using Fura-2-based intracellular calcium imaging experiments, which allow prolonged measurements at low pH values (Figure3a).In line with the patch-clamp recordings, these experiments showed that 100 nM A-1165442 caused a full inhibition of calcium responses evoked at pH 5.5.However, responses to acid stimulation at lower pH values (5.0, 4.5 and 4.0) were only partially inhibited by 100 nM A-1165442 (Figure3a,b).Note that under these conditions we did not observe any pH-induced calcium responses in non-transfected HEK293 cells (Figure3a), demonstrating the TRPV1-specificity of the responses.Protein kinase C (PKC) activation potentiates responses of TRPV1 to activating stimuli, including low pH(Bhave et al., 2003;Premkumar & Ahern, 2000;Vellani et al., 2001), and this potentiation contributes to increased TRPV1-mediated pain signalling under pathological conditions(Joseph et al., 2019).To test whether PKC-induced potentiation affects the inhibitory effects of A-1165442 on low pHactivated TRPV1 responses, we performed Fura-2-based intracellular calcium imaging experiments in TRPV1-expressing HEK293 cells treated with the PKC-activating phorbol ester phorbol 12-myristate 13-acetate (PMA) or the inactive analogue 4α-phorbol 12-myristate 13-acetate (4α-PMA).In line with earlier work, responses to acidic pH were enhanced in cells treated with PMA (Figure S1A,B).Notably, the inhibitory effect of 100 nM A-1165442 on pH-evoked responses was attenuated following PMA treatment, with only partial inhibition of the responses at pH 5.5 and very limited inhibition at pH 4.5 (Figure S1A,D).In contrast, treatment with 4α-phorbol 12-myristate 13-acetate (4α-PMA) did not affect the inhibitory effect of A-1165442 (Figure S1B,D).Taken together, the cumulative results from patch-clamp recordings and calcium imaging indicate that the reportedly modalityselective TRPV1 antagonist A-1165442 inhibits low pH-induced responses in a pH-dependent manner, with reduced inhibition at more acidic pH values.
. The lack of pH-induced responses in the mutant channel allowed us to directly evaluate the inhibitory effect of A-1165442 on capsaicin-evoked responses at different environmental pH values.As expected, A-1165442 caused a potent and concentration-dependent inhibition of capsaicin-evoked calcium responses at pH 7.4.In line with the patch-clamp recordings in wild type TRPV1 (Figure 1), we measured a full inhibition at concentrations of >3 nM and an IC 50 value of 1.34 ± 0.10 nM (Figure 4a,b).
modality-selective potentiation by AMG7905 of pH-evoked responses, we tested the compound's effect on capsaicin responses in the pH-insensitive F660A mutant, both at pH 7.4 and pH 5.5.As F I G U R E 2 pH dependence of the inhibition of acid-activated wild type TRPV1 currents by A-1165442.(a) Representative time course of whole-cell TRPV1 currents in a single cell at À80 and +80 mV upon stimulation with acidic solution at pH 5.5 and subsequent inhibition by A-1165442 (100 nM).(b) Current-voltage relationships measured during voltage ramps ranging from À100 to +160 mV, taken at the time points indicated in (a).(c, d) Same as (a, b), but upon stimulation at pH 4.5.(e, f) Residual current at +80 and À80 mV following inhibition by A-1165442 at pH 5.5 and 4.5.Data points in (e, f) represent the mean ± SEM from six to seven cells.Groups were compared using two-tailed Students t-test.*P < 0.05.
dependence of the inhibition of acid-activated wild type TRPV1 calcium responses by A-1165442.(a) Time course of the intracellular calcium concentration in TRPV1-expressing (red) and non-transfected (black) human embryonic kidney 293 (HEK293) cells upon repetitive stimulation with acidic solution at the indicated pH values.Before the second acid stimulus, cells were preincubated with A-1165442 (100 nM).Traces indicate the mean ± SEM from 10 to 20 cells within a single imaging experiment.(b) pH dependence of the response to acid stimulation in the presence of A-1165442 (100 nM).Responses were normalized to the mean of the responses to the first and third acid stimulation.Data points represent the mean ± SEM from between 147 and 650 cells in five independently imaged coverslips.Groups were compared using one-way ANOVA with Tukey's post hoc test.*P < 0.05.F I G U R E 4 pH dependence of the inhibition of capsaicin-activated calcium responses by A-1165442 in the pH-insensitive F660A TRPV1 mutant.(a) Representative time course of the intracellular calcium concentration in human embryonic kidney 293 (HEK293) cells expressing the F660A TRPV1 mutant upon stimulation with capsaicin in the absence or presence of the in indicated concentrations of A-1165442, at pH 7.4 and pH 4.5.Responses were normalized to the response the calcium ionophore ionomycin (2 μM).Traces indicate the mean ± SEM from 178 to 285 cells within a single imaging experiment.(b) Concentration dependence of the inhibition of capsaicin responses by A-1165442, measured at pH 7.4, pH 5.5 and pH 4.5.Data points represent the mean ± SEM from between 178 and 285 cells from five to eight independently imaged coverslips.

F
I G U R E 5 Potentiation of calcium responses of wild type TRPV1 to acid stimulation by AMG7905.(a) Time course of the intracellular calcium concentration in TRPV1-expressing cells upon repetitive stimulation with acidic solution at pH 5.5.Before the second acid stimulus, cells were preincubated with AMG7905 (1 μM).Trace indicates the mean ± SEM from 56 cells within a single imaging experiment.(b) Mean normalized amplitude of the response to pH 5.5.Responses were normalized to the mean of the responses to the first and third acid stimulation.Data points represent the mean ± SEM from 270 cells in five independently imaged coverslips.Groups were compared using one-way repeated measures ANOVA with Tukey's post hoc test.*P < 0.05.also reveal that low pH alters the modulation of TRPV1 by AMG7905 independently of the acid-evoked mode of activation, reducing the inhibitory potency towards capsaicin activation and uncovering a partial agonistic effect.How could low pH influence the modulation of TRPV1 by these two compounds?The chemical structures of A-1165442 and of its analogues provide a hint that protonation of the compound may contribute to its reduced potency at lower pH values.A-1165442 and several structural analogues that incompletely inhibit acid responses at pH 5.0 all contain a protonable isoquinoline moiety(Voight et al., 2014).Notably, the pK a value of unsubstituted isoquinoline is $5.1, suggesting that A-1165442 and the other isoquinolinecontaining structural analogues are primarily uncharged at pH 7.4 and 5.5, whereas the protonated form becomes predominant at pH 5.0 and lower.We hypothesize that the protonated form has reduced inhibitory potency, which would explain the reduced inhibitory activity of both acid-evoked responses in wild type TRPV1 and of capsaicin-evoked responses of the F660A mutant at lower pH values.

F
I G U R E 6 pH dependence of the effects of AMG7905 on capsaicin-evoked responses in the pH-insensitive F660A mutant.(a) Time course of the intracellular calcium concentration in TRPV1-expressing cells upon stimulation with capsaicin (50 nM) following preincubation with AMG7905 (10 nM) at pH 7.4 or pH 5.5.Traces represent the mean ± SEM from 263 cells (pH 5.5) and 212 cells (pH 7.4) from single representative imaging experiments.(b) Zoom-in on the boxed area in (a), illustrating the determination of the calcium response to AMG7905 (ΔCa 2+ AMG7905 ) and to capsaicin (ΔCa 2+ caps ).(c) pH dependence of the inhibitory effect of AMG7905 on capsaicin-evoked responses.(d) pH dependence of the direct effect of AMG7905 on the cellular calcium concentration.Individual data points in (c, d) represent mean ± SEM from between 121 and 925 cells in five independently imaged coverslips.Groups were compared using two-way repeated measures ANOVA with Tukey's post hoc test.*P < 0.05.