TRPM8 Is a Noncanonical Target of 7-Nitrobenzodiazepines

Summary:

Benzodiazepines are widely prescribed therapeutics whose actions are primarily attributed to modulation of GABA_A receptors. Here, we identify a previously unrecognized target of a subset of benzodiazepines: the cold- and menthol-sensitive ion channel TRPM8. Clonazepam and related 7-nitrobenzodiazepines act as potent TRPM8 agonists, eliciting calcium influx and membrane depolarization in heterologous expression systems, native human cells, and trigeminal sensory neurons. Pharmacological inhibition, genetic knockdown, and gene deletion of TRPM8 abolish clonazepam-evoked responses, establishing TRPM8 as the relevant molecular target. Mechanistically, clonazepam shares pharmacological features with icilin, a canonical TRPM8 agonist, and engages conserved determinants of channel activation. These findings expand the pharmacology of benzodiazepines beyond GABA_A receptors and suggest that TRPM8 activation may contribute to the clinical efficacy of clonazepam in burning mouth disorder and other sensory neuropathic conditions

eTOC Blurb

Benzodiazepines are a major class of drugs that work by activating GABA_A receptors. Elkin et al. show that clonazepam, a 7-nitrobenzodiazepine, also robustly activates TRPM8 ion channels. The engagement of TRPM8 by clonazepam may clonazepam’s efficacy in treating the painful condition known as burning mouth disorder.

Graphical Abstract

Introduction

Benzodiazepines are a very well-studied class of drugs. They exert inhibitory effects on central nervous system excitability underpinning their clinical use as anxiolytics, hypnotics and anti-seizure medications. Their mechanism of action through allosteric modulation of central GABA_A receptors to enhance GABAergic tone^1,2 has been well established over 50 years of study^3,4. However, small molecule selectivity for a single class of target is rarely absolute, with the broader polypharmacological profile of any drug being critical for understanding physiological activity as well as unwanted side effects.

Here, we examined the polypharmacology of benzodiazepines in the context of engagement of transient receptor potential (TRP) ion channels. This interest was prompted by our recent discovery that clonazepam, a ‘first generation’ benzodiazepine⁵, activates a schistosome transient receptor potential melastatin (TRPM) ion channel^6,7. Schistosomes are parasitic flatworms and the causative agents of the neglected tropical disease schistosomiasis (‘Bilharzia’). Despite the evolutionary distance between humans and platyhelminths, we were curious whether a similar profile of pharmacological regulation could exist in humans, underpinning the rationale for this study.

Results

Clonazepam and meclonazepam (methyl-clonazepam) were recently shown to activate a schistosome transient receptor potential melastatin ion channel named TRPM_MCLZ^6,7. TRPM_MCLZ harbors a transmembrane ligand binding pocket that closely resembles the menthol binding pocket in TRPM8^6,7. On the basis of this binding pocket homology, we screened a benzodiazepine library against human TRPM8 (hTRPM8) using an intracellular Ca²⁺ reporter assay. Whereas none of the screened benzodiazepines triggered changes in cytoplasmic Ca²⁺ in untransfected HEK293 cells, multiple analogs evoked Ca²⁺ signals in hTRPM8 expressing HEK293 cells (Figure 1A). The top three ‘hits’ in terms of peak amplitude were all 7-nitrobenzodiazepines: clonazepam, flunitrazepam and N-desmethylflunitrazepam (a metabolite of flunitrazepam, Supplementary Table 1).

Figure 1. Benzodiazepine activation of human TRPM8.

(A) Heatmap of Ca²⁺ signal amplitude (RFU, relative fluorescence units) evoked by 53 different benzodiazepines (A1 through to E9) together with icilin (E11, asterisked). Each benzodiazepine ligand was tested at ~20 μM concentration and values are averaged from two biological replicates. A list of all the tested ligands is provided in Supplementary Table 1. (B&C) Example traces of fluo-4 NW fluorescence representing Ca²⁺ signals evoked by various concentrations (10nM to 100μM) of (B) clonazepam and (C) icilin in hTRPM8-expressing HEK293 cells. Data are plotted over same ranges. Also see Figure S1 & S2. (D) Concentration response curves for clonazepam, resourced from a different vendor, together with the canonical hTRPM8 agonists (icilin, WS-12 and menthol). Data represent mean ± SEM from n=3 independent experiments. Data are normalized to the peak response seen with icilin. (E) Plot of IC₅₀ values for four different hTRPM8 antagonists in response to stimulation by clonazepam (y-axis) or icilin (x-axis). Also see Figure S1 & S2.

This same assay was then repeated with different lots of repurchased clonazepam to confirm activity at hTRPM8 (Figure 1B). Clonazepam elicited a concentration dependent activation of hTRPM8 (EC₅₀=828±84nM), with the evoked Ca²⁺ signals displaying a greater peak magnitude than observed with icilin (Figure 1C), a Ca²⁺ dependent hTRPM8 agonist⁸, and other canonical hTRPM8 activators (Figure 1D). Robust Ca²⁺ signals with clonazepam were also apparent using Ca²⁺ reporter dyes of lower affinity (see Figure S1A & S1B). Finally, different hTRPM8 blockers were examined for effectiveness at attenuating clonazepam-evoked Ca²⁺ signals. Four hTRPM8 antagonists (TCI-2014, RQ00203078, AMTB and AMG-333) blocked clonazepam and icilin-evoked Ca²⁺ signals, with a similar rank order of potency (Figure 1E). In contrast, flumazenil, a benzodiazepine blocker of GABA_A responses⁹, had no effect on clonazepam-evoked Ca²⁺ signals (≤100μM, Figure S1C).

Electrophysiology was used as an orthologous method to investigate clonazepam action at hTRPM8. ‘Whole cell’ currents in HEK293 cells overexpressing hTRPM8 showed outward rectification in response to clonazepam and other hTRPM8 activators (see Figure S2A). The clonazepam-evoked inward current density was greater with clonazepam than with icilin or menthol (see Figure S2B). In response to cooling, clonazepam caused a shift of hTRPM8 responsiveness toward warmer temperatures, with half-maximal open probability occurring at 26.3±2.1⁰C in the presence of clonazepam, compared with ~15.9±1.8⁰C in the absence of ligand (see Figure S2C & S2D). This behavior was similar to that seen with other hTRPM8 agonists, including WS-12 (Figure S2C & S2D) and menthol¹⁰.

Activity at other vertebrate TRPM8 channels was then examined. Clonazepam activated mouse TRPM8 (Mm.TRPM8, EC₅₀=367±55nM) and rat TRPM8 (Rn.TRPM8, EC₅₀=357±55nM), underscoring retention of benzodiazepine sensitivity at TRPM8 in commonly used animal models (Figure 2A). However, clonazepam did not activate an avian TRPM8 ortholog (Ficedula albicollis TRPM8, Fa.TRPM8, Figure 2A). These results parallel data obtained with the hTRPM8 agonist icilin, which lacks activity at avian TRPM8⁸. Responsiveness of avian TRPM8 to icilin is however restored when a human amino acid variant found within the third transmembrane helix (TM3) is engineered into the avian sequence⁸. Reciprocally, icilin sensitivity of hTRPM8 is lost when the avian variant is introduced into the hTRPM8 sequence⁸. Reciprocal mutants of the human and flycatcher channel were therefore profiled against clonazepam. Clonazepam activated the mutant avian construct, Fa.TPRM8[A796G] which harbors the mammalian amino acid variant at this position (Figure 2B), but not the hTRPM8 mutant, hTRPM8[G805A], where the avian amino acid variant was introduced into the hTRPM8 channel (Figure 2B). These data suggest commonalities in the binding mode of both ligands to the TRPM8 orthologs. Indeed, cheminformatic profiling of hTRPM8 ligands clustered clonazepam together with icilin relative to other commonly used hTRPM8 activators (Figure 2C). Finally, TRP channel selectivity was examined by screening clonazepam against other human TRP channels. No activation of other human TRP channels by clonazepam was observed (Figure 2D). Collectively, these data demonstrate clonazepam potently and robustly activates mammalian TRPM8.

Figure 2. Activity of clonazepam at different TRPM8 ortholog and paralogs.

(A) Concentration response curves in a Ca²⁺ reporter assay showing activation of mouse TRPM8 (Mm.TRPM8), rat TRPM8 (Rn.TRPM8), collared flycatcher TRPM8 (Fa.TRPM8) by clonazepam. (B) Reciprocal mutagenesis of human (hTRPM8) and avian (Fa.TRPM8) constructs at a key TM3 residue controls clonazepam sensitivity. Data are shown for hTRPM8[G805A] (loss of sensitivity) and Fa.TRPM8[A796G] (gain of sensitivity). (C) Cheminformatic analysis of five different hTRPM8 agonist chemotypes visualized using a multidimensional scaling plot underscores the similarity between clonazepam and icilin. (D) Responses to clonazepam (100 μM) were resolved across a panel of eighteen human TRP channels. Individual assay details for the different channels are detailed in the methods. Data are normalized to the positive control ligand for each channel, and are represented as mean ± SEM.

To study the effects of clonazepam on native hTRPM8, assays were performed in G-402 cells, a kidney epithelial cell line that endogenously expresses hTRPM8. In Ca²⁺ reporter assays in G-402 cells, clonazepam triggered Ca²⁺ signals (EC₅₀=597±64nM) which again were larger in magnitude than responses elicited by icilin or WS-12 (Figure 3A). In ‘cell-attached’ recordings, clonazepam activated an outward rectifying, non-selective cation current in G-402 cells, similar to responses evoked by canonical hTRPM8 agonists (Figure 3B). These currents were blocked by hTRPM8 antagonists (see Figure S3A). Single channel responses evoked by clonazepam displayed a unitary conductance of ~70pS (see Figure S3B, C & D) consistent with prior studies of hTRPM8¹¹.

Figure 3. Clonazepam activation of native hTRPM8 channels.

(A) Concentration response curves for clonazepam, icilin, WS-12 and menthol in wildtype G-402 cells resolved using a Ca²⁺ reporter assay. Data are normalized to the peak response seen with WS-12. (B) Current-voltage (I-V) plot in G-402 cells recorded in the ‘cell-attached’ mode in the absence (DMSO vehicle, 0.1%) or presence of various hTRPM8 agonists (each at 10 μM) in the pipette solution. Also see Figure S3. (C) Concentration response curves for clonazepam in G-402 knockdown cells resolved using a Ca²⁺ reporter assay and normalized to a ‘no RNA’ control (Control). The same assay is shown are shown for cells treated with dual control (scramble, PPIB) or hTRPM8-targeting siRNA constructs (hTRPM8–1 or hTRPM8–2). Also see Figure S4. (D) Current-voltage (I-V) plot of G-402 cells in the ‘cell-attached’ mode that had previously been treated with the same siRNA constructs. Clonazepam (10 μM) was present in the pipette solution. (E) Concentration response curves for clonazepam evoked responses in G-402 wildtype and two clonal hTRPM8 knockout cell lines (hTRPM8-KO-1 or hTRPM8-KO-2). Responses were normalized to the peak WS-12 response of G-402 wildtype cells. (F) Current-voltage (I-V) plot of G-402 knockout cell lines (hTRPM8-KO-1 and hTRPM8-KO-2) recorded in ‘cell-attached’ mode in the presence of clonazepam (10 μM) in the pipette solution.

Endogenous hTRPM8 channels in G-402 cells were then targeted for knockdown using two different siRNA constructs (hTRPM8–1, hTRPM8–2, Supplementary Table S2). Knockdown of hTRPM8 using either construct inhibited Ca²⁺ signals evoked by clonazepam (Figure 3C). Clonazepam-evoked Ca²⁺ signals however persisted in G-402 cells treated with either of two control siRNA constructs (scramble or PPIB; Figure 3C). Similar results were seen in electrophysiological assays (Figure 3D). The effectiveness of each siRNA construct used for knockdown was confirmed using real-time PCR (see Figure S4). Finally, hTRPM8 knockout G-402 lines (hTRPM8-KO-1, hTRPM8-KO-2) were generated by CRISPR-Cas9 editing. In both these hTRPM8-KO cell lines, responses to clonazepam were not apparent, assessed either by Ca²⁺ reporter assay (Figure 3E) or electrophysiology (Figure 3F).

Clinically, clonazepam functions as an anti-seizure and anxiolytic drug via GABA_A receptor activation². Plasma concentrations considered therapeutic for these effects (20–70ng/ml¹², ≤200nM) are lower than those shown here to activate hTRPM8 (EC₅₀ for clonazepam at hTRPM8 ~600nM, Figure 3). However, the interaction of clonazepam with hTRPM8 could nevertheless have relevance for specific conditions. Most obvious is the disease known as burning mouth disorder (BMD¹³), an extremely painful chronic oral disorder hypothesized to result from sensory neuropathy^14,15. Topical clonazepam yields relief from symptoms in most patients^16–18, although there is no correlation between the efficacy of various benzodiazepines in treating BMD, systemic plasma concentration, and alleviation of anxiety or depressive symptoms^17,18. The high, local concentration of clonazepam effective clinically for treating BMD (60–300μM in mouthwash, or similarly high local concentrations attained by sucking a sublingual tablet) surpasses the concentration range for clonazepam-evoked hTRPM8 activation. This suggests that the therapeutic action of clonazepam in BMD is likely realized by localized activation of hTRPM8 present within the sensory nerve pathways that innervate the mouth and tounge^19–21. Assignment of hTRPM8 as the oral target of clonazepam would catalyze development of improved therapeutics for this condition.

To explore this idea, responses to clonazepam were examined in a primary culture of mouse trigeminal neurons. Trigeminal sensory neurons, which innervate the mouth and tongue, are functionally heterogeneous with multiple classes of cold sensing neurons defined physiologically or molecularly^22–24. In Ca²⁺ imaging assays, a subset of isolated neurons exhibited responses to the TRPM8 agonist WS-12 (~24% of total neurons, Figure 4A). Almost all of the WS-12 responsive neurons co-responded to clonazepam (~95%, Figure 4A), with a Ca²⁺ signal that kinetically resembled the WS-12 response. To mitigate any desensitization in this dual agonist application assay, responses were compared between trigeminal neurons exposed to a single application of either clonazepam or WS-12 in either wild type (Figure 4B), or TRPM8 knock out (TRPM8-KO) mice (Figure 4C). Ca²⁺ signals evoked by clonazepam were observed in wild type mice but were absent from TRPM8-KO mice (Figure 4B & 4C). Analysis of these Ca²⁺ signals revealed responses to clonazepam were significantly higher in magnitude and prevalence than with WS-12 (Figure 4D & 4E). Overall, these data demonstrate that clonazepam evokes TRPM8-dependent Ca²⁺ signals in trigeminal mouse neurons.

Figure 4. Clonazepam evokes Ca²⁺ signals in TRPM8-positive trigeminal neurons.

(A) Representative traces of fura-2 fluorescence ratio (F₃₄₀/F₃₈₀) over time in WS-12 responding neurons that were exposed first to WS-12 (10 μM) and then to clonazepam (10 μM). Neurons were classified as single responders (to WS-12 alone, pink) or as dual responders (WS-12 and clonazepam, orange). The proportion (percentage of total neurons in each imaging field) of each class of neuron is shown on the right. (B & C) Comparison of responses to application of WS-12 (10 μM) or clonazepam (10 μM) in (B) wildtype or (C) TRPM8-KO neurons. These data are quantified in terms of (D) signal amplitude and (E) number of cells responding. A total of ten (n=2 males, n=8 females) C57BL/6 mice and four (n=2 males, n=2 females) TRPM8 knockout mice were used for these experiments, with 22–55 neurons imaged per mouse.

In conclusion, this report demonstrates that clonazepam, as well as several other 7-nitrobenzodiazepines, activate hTRPM8. This discovery provides new insight into benzodiazepine polypharmacology and highlights opportunity to explore the benzodiazepine scaffold for hTRPM8 modulation. Further, this result identifies a local mechanism of clonazepam action to mitigate the painful symptoms of BMD.

Limitations of the study:

The proposed contribution of TRPM8 activity to the therapeutic efficacy of clonazepam in burning mouth disorder remains inferential. Although the concentrations of clonazepam used clinically for topical treatment are sufficient to activate TRPM8, direct evidence linking TRPM8 engagement to symptom relief in vivo is not presented here. In addition, the broader contribution of TRPM8 activation to other physiological effects of benzodiazepines remains unclear. Finally, the structural basis of clonazepam binding to TRPM8 was not resolved in this study and will require future investigation.

RESOURCE AVAILABILITY.

Lead Contact:

requests for further information and resources should be directed to the lead contact, Jonathan Marchant (JMarchant@mcw.edu).

Materials Availability:

all unique reagents generated in this study are available from the lead contact through a materials transfer agreement.

Data & Code Availability:

• Data reported in this paper are available from the lead contact upon request.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR Methods

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mouse breeding and housing

C57BL/6 mice (JAX: 000664) and TRPM8 knockout mice (B6.129P2-Trpm8tm1Jul/J, JAX: 008198) were procured from Jackson Labs. All mice were housed within Medical College of Wisconsin (MCW) facilities according to NIH stipulated guidelines. All animal protocols were approved by the Institutional Animal Care and Use Committee at MCW.

Cell lines and reagents

HEK293 cells (ATCC CRL-1573, female) were cultured in high glucose, L-glutamine, Phenol Red, Sodium pyruvate Dulbecco’s modified Eagle’s medium supplemented (DMEM, Thermo Fisher Scientific, Cat#11995073) with 10% fetal bovine serum (FBS, Thermo Fisher Scientific, Cat#A5256701), penicillin (100 units/ml), streptomycin (100 μg/ml), and L-glutamine (290 μg/ml) (Thermo Fisher Scientific, Cat#10378016). G-402 cells (ATCC CRL-1440, female) were cultured in McCoy’s 5A Medium (Thermo Fisher Scientific, Cat#16600082) supplemented with 10% FBS, penicillin (100 units/ml), streptomycin (100 μg/ml), and amphotericin B (0.25 μg/ml) (Thermo Fisher Scientific, Cat#15240062). All cell lines were cultured in a humid environment at 37°C and 5% CO₂. All lines were tested monthly for Mycoplasma contamination via PCR (Millipore Sigma, Cat#MP0035) and quarterly via antigen (Lonza Bioscience, Cat# LT07–318). All cell lines, including engineered lines, were authenticated by short tandem repeat DNA profiling (ATCC).

Ex vivo trigeminal nerve harvest and culture

A total of ten C57BL/6 (n=2 males, n=8 females) mice and four TRPM8 knockout (n=2 males, n=2 females) mice aged 8–10 weeks were used for these experiments. To isolate trigeminal ganglion (TG) neurons, mice were deeply anesthetized by isoflurane inhalation prior to euthanasia by decapitation. Whole TG were then incubated in collagenase IV (1 mg/mL) in DMEM/F12 (Thermo Fisher Scientific, Cat#21331020) for 45 minutes at 37^oC. Ganglia were then incubated in 0.05% trypsin (Thermo Fisher Scientific, Cat# 25300062) in DMEM for 45 minutes at 37^oC, which was immediately thereafter neutralized with horse serum (Thermo Fisher Scientific, Cat#26050088). Ganglia were then washed and resuspended in complete media (DMEM/F12, 10% inactivated fetal horse serum, 1% glucose, 2 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin). Ganglia were progressively mechanically dissociated with a p1000 and p200 micropipette prior to plating on laminin-coated 12 mm round glass coverslips. Dissociated neurons were fed with complete media 1–2 hours after plating. Dissociated TG neuronal somata were used for calcium imaging 18–24 hours after plating.

METHOD DETAILS

Chemical preparation

hTRPM8 control ligands (icilin, WS-12, and menthol) were sourced from Cayman Chemical and solubilized in 100% DMSO (Sigma-Aldrich). The benzodiazepine library was also procured from Cayman Chemical, solubilized in 135μL of DMSO (final concentration ~20 μM). All chemicals were used at a purity of ≥98%.

Plasmid preparation

cDNA constructs for TRPM8 orthologs were cloned in a pcDNA3.1 backbone sourced as: Homo sapiens TRPM8 (hTRPM8; NCBI: NM_024080.5, UniProt: Q7Z2W7), Mus musculus TRPM8 (Mm.TRPM8; NCBI: NP_599013.1, UniProt: Q8R4D5), and Ficedula albicollis (Fa.TRPM8; NCBI: XP_005049003.1) by GenScript. Rattus norvegicus TRPM8 (Rn.TRPM8) tagged with GFP²⁵ was a gift from Sebastián Brauchi (Addgene plasmid #64879; RRID:Addgene_64879).

Over-expression studies

Transient transfections of HEK293 cells were performed using Lipofectamine 2000, following the manufacturer’s protocol (Thermo Fisher Scientific). For a 60mm petri dish, cells (seeded at 1×10⁶ cells) were incubated with OptiMEM (800μl) containing a mixture of lipofectamine 2000 (20μl) and DNA (8μg).

Knockdown studies

Endogenous expression of hTRPM8 in G-402 cells was ablated using gene-specific siRNA constructs transfected using Lipofectamine 3000 (Thermo Fisher) following the manufacturer’s protocol. The initial seeding densities of wildtype cells were 1.5×10⁵, 5×10⁵, 8.5×10⁵ cells in 35mm, 60mm, and 100mm dishes, respectively. Dual siRNA constructs targeting either exon 5 (Silencer^® siRNAs, Cat No. AM16708, ID 104796, “hTRPM8–1”) or exon 26 (Silencer^® siRNAs, Cat No. AM16708, ID 30886, “hTRPM8–2”) were used to knockdown hTRPM8 (Supplementary Table S2). An anti-cyclophilin B (PPIB) siRNA (Cat. No. AM16078, ID: 110793), Silencer^™ Negative Control No. 1 siRNA (Scramble) (Cat. No. AM4611, Invitrogen), and water were used as controls. Dishes were incubated in siRNA-containing medium (33pmol siRNA/mL) for no more than 4.5 hours, after which the medium was aspirated and replaced with growth medium containing no antibiotics. Cells were re-plated on glass chips for electrophysiology or directly into FLIPR plates for the high-throughput Ca²⁺ reporter assay.

Generation of TRPM8 knockout cell lines

The G-402 hTRPM8-knockout cell lines were generated by Applied Biological Materials Inc. (Vancouver, Canada). A lentiviral and CRISPR-Cas-9 package with three unique hTRPM8 sgRNA targets (hTRPM8 sgRNA#1–3) were transfected into wildtype cells. Cells containing the lentiviral construct were selected using puromycin treatment and serially diluted to select single clones for expansion. Validation of knockout was confirmed by Sanger sequencing using the following forward and reverse primers (TRPM8#1F and TRPM8#1R). Two clones hTRPM8-KO-1 and hTRPM8-KO-2) were confirmed to harbor biallelic heterozygous frameshift mutations.

Quantitative PCR

RNA was isolated using the RNeasy Mini Kit according to the manufacturer’s protocol (Qiagen). Equivalent amounts of cDNA (1μg) were generated using a High-Capacity RNA-to-cDNA^™ Kit (Applied Biosystems) and diluted 1:5 v/v for PCR analyses. Primers for qPCR were synthesized by Integrated DNA Technologies. Quantitative PCR reactions were performed as 10μL reactions in 96-well plates using 8μL of PowerUp^™ SYBR^™ Green Master Mix supplemented with 3μM of gene specific primers (3μM) and 2 μL of diluted cDNA. Plates were sealed and briefly spun down prior to PCR (CFX Duet Real-Time PCR System, BioRad). The PCR protocol was as follows: uracil-DNA glycosylase degradation 50°C for 2 mins, hot-start activation at 95°C for 2 mins, denature 95°C for 15 s, anneal 59°C for 15s, and 72°C for 1 min, repeated for 40 cycles. Bio-Rad CFX manager was used to track individual well RFUs. A quantification cycle (Cq) threshold of 200 RFUs was set for all experiments. Values were normalized to GAPDH expression using the ∆∆Ct method. Data were visualized using GraphPad Prism (v.10.6.1).

Calcium reporter assay

A Ca²⁺ reporter assay using a Fluorescence Live Imaging Plate Reader assay system (Molecular Devices, FLIPR^TETRA) was used for functionally profiling TRPM8. HEK293 cells were seeded (20,000 cells/well) in a black wall, clear bottom, poly-L-lysine coated 384-well plate and incubated (16–24 hours) in DMEM supplemented with 10% FBS. G-402 cells lines were seeded (13,500 cells/well) and incubated (40–48 hours) in McCoy’s 5A medium + 10% FBS. Prior to the assay, cells were incubated in 20 μL assay buffer (HBSS + 20 mM HEPES at pH 7.4, supplemented with probenecid (2.5 mM) and the fluorescent Ca²⁺ indicator Fluo-4NW (Thermo Fisher Scientific, Cat#F36206)) for 45–50 min at 37°C and 15 min at room temperature. Drug plates for screening the benzodiazepine library were prepared using assay buffer but without inclusion of probenecid or fluorescent dye. Individual benzodiazepine ligands were tested at ~20 μM concentration and values are averaged from two biological replicates. A list of all the tested benzodiazepine ligands and screening data is provided in Supplementary Table 1. After dye loading, this screen was performed by measuring basal fluorescence of each well for 20s prior to compound addition (5 μL of 5x stock). Raw-fluorescence was captured for another 250–600s following drug addition. Screens in ‘agonist’ mode were performed using the same approach. For antagonist assays, antagonists (5 μL of 6x drug stock) were added with fluorescence monitored for 250s prior to subsequent injection of 1μM icilin or CLZ (5μL of 6μM stock) with fluorescence continually monitored for a further 250–600s. For assays with lower Ca²⁺ affinity fluorescence dyes (Fluo-4FF, AM and Fluo-5F, AM, Thermo Fisher Scientific, Cat#F23980 and Cat#F14221), an additional wash step (HBSS with 20 mM HEPES, pH = 7.4) was included after dye loading. All assays were performed at room-temperature and included vehicle, positive (icilin and/or WS-12), and negative (buffer, and/or non-transfected cells) controls. Peak fluorescence signals (relative fluorescence units, RFU) were corrected for baseline and normalized to relevant positive controls (100μM icilin or WS-12). For concentration response curves, zero was defined as the smallest value within a given biological replicate.

Cheminformatic analysis

Cheminformatic properties of different TRPM8 agonists icilin, WS-12, (−)-menthol, cryosim-3, and CLZ were obtained from the PubChem database. To evaluate chemical structural similarity, a multidimensional scaling (MDS) clustering algorithm was employed from the ChemMine Tools web server²⁶. Simplified Molecular-Input-Line-Entry System (SMILES) was used for chemical structure representation in data entry and compound similarity scored with a 2D MDS algorithm using a Tanimoto coefficient equal to 0.4. MDS clustering values were calculated by all-against-all comparisons of compounds using atom pair similarity measures (Tanimoto coefficient equal to 0.4) with the generated similarity scores transformed into distance values, and represented using the CanvasXpress javascript library which provides an easy graphical interface representing dissimilarity as distance. Similar methods have also been used by Mebrat et al²⁷ for analysis of TRPM8 ligands.

Human TRP channel profiling

Clonazepam was profiled against a panel of 18 human TRP channels (SB Drug Discovery, Glasgow). In these assays, HEK293 cells expressing specific human TRP channels were challenged with clonazepam and responses resolved using either a membrane potential dye (TRPC1, TRPC3, TRPC4, TRPC5, TRPC6, TRPC7, TRPM4 and TRPM5; Molecular Devices, R8126) or a Ca²⁺ reporter (TRPA1, TRPV1, TRPV2, TRPV3, TRPV4, TRPV5, TRPV6, TRPM2, TRPM3 and TRPM8; Calcium 5, Molecular Devices, R8186) based on the ionic permeability of each channel. Responses were evaluated relative to peak responses obtained with reference compounds (ASP7663, TRPA1; Englerin A, TRPC1, TRPC4; GSK-1702934A, TRPC3, TRPC6; Carbachol, TRPC5, TRPC7; H₂O₂, TRPM2; CIM0216, TRPM3; ionomycin, TRPM4, TRPM5; icilin, TRPM8; capsaicin, TRPV1; cannabidiol, TRPV2; 2-APB, TRPV3; RN1747, TRPV4; CaCl₂, TRPV5, TRPV6). For each assay, cells were trypsinised, counted and seeded in black, clear bottomed 384 well plates (10,000 cells per well) and incubated overnight. Media was then removed from cell plates and assay buffer (30 μl) supplemented with reporter dye (10 μl) added prior to incubation (37⁰C, 40 mins). Plates were then placed in a FLIPR (Molecular Devices, serial no: FX01138), and fluorescence monitored (488nm/510–570nm) every second prior to addition of clonazepam (10μM) after which recordings continued for 5 minutes at room temperature

Electrophysiological recording

Cells were plated onto glass coverslips 24–48 h prior to being secured in an Olympus BX51WI inverted microscope recording chamber. A multiClamp 700B amplifier and Digidata 1440A digitizer (Molecular Devices, San Jose, CA) were used for electrophysiological recordings. Signals were passed through an 8-pole Bessel low pass filter at 1 kHz and sampled at 10 kHz. Data analysis was performed using Clampfit 11 software (Molecular Devices). Pipettes were pulled from borosilicate glass (BF150–110-10, Sutter Instruments, Novato, CA) to a resistance of 6–8 MΩ. For ‘cell-attached’ recordings, both pipette and bath solutions were Hank’s buffered saline solution (HBSS) supplemented with HEPES (20mM). In ‘whole cell’ and ‘outside-out’ recordings the pipette solution comprised CsCH₃SO₃ (140mM), CsCl (5mM), HEPES (10mM), EGTA (1mM), at pH 7.4 (with CsOH). Holding membrane voltage (V_m) was −80 mV. All recordings were performed at room temperature except recordings in 'outside-out’ mode that were performed across a temperature range ~8–42⁰C. To estimate hTRPM8 temperature dependence 160 ms long voltage ramps from −80 to 80 mV were applied to assess channel activity at different temperatures. Ligands were applied to the bath solution at 10 μM and DMSO (0.1%) used as a control. Temperature was controlled by a temperature controller TC-324B (Warner Inst.). After formation of an ‘outside-out’ patch, the recording chamber was perfused with cooled solution (~6–8⁰C) and then heated up to ~40–42⁰C. Open probability (P_open) was estimated as normalized peak current at 80 mV (I/I_max) plotted against temperature. Typically, P_open was steady at ~1 at cooler temperatures and approached 0 at warmer temperatures. The temperature at which the P_open= 0.5 was defined as ET₅₀. Tukey’s test was used to compare ET₅₀ values between control and ligand treated conditions. Single channel hTRPM8 activity in G-402 cell was recorded in ‘cell-attached’ mode. To quantify the properties of the step-like fluctuations, an all-points histogram (Clampfit) for ~500ms of recording was plotted and analyzed with a Gaussian multipeak protocol (Origin) at several membrane voltages. The difference between the center value of the converged peaks was used for determining the single channel current (I_SC, pA).

Calcium imaging of primary mouse neurons

Somata were washed with extracellular normal HEPES (ENH) buffer (in mM: 150 NaCl, 10 HEPES, 8 glucose, 5.6 KCl, 2 CaCl₂, 1 MgCl₂; pH 7.4 +/− 0.03, mOsm 320 +/−3) for 1–2 minutes, incubated with 2.5 mg/mL Fura-2AM (Life Technologies, Cat#F1221) in 2% bovine serum albumin for 45 minutes at room temperature, and washed again in ENH at room temperature for 30 minutes. Coverslips were then mounted to an imaging stage and perfused with ENH (6 mL/min). Fura-2 fluorescence was measured with excitation at 340 nm and 380 nm and captured by a cooled DS-Qi1MC monochrome camera (Nikon). NIS Elements software (Nikon) was used to analyze signals following stimulation with 340 nm or 380 nm and calculated bound/unbound calcium ratio (340/380). To induce intracellular calcium transients, neuronal somata were exposed to WS-12 (10 μM, 1 minute), or clonazepam (10 μM, 1 minute) followed by a 2–6 minute washout in ENH depending on the experiment. Cell viability was assessed by transient (20 second) exposure to 50 mM KCl followed by washout with ENH. Somata exhibiting at least 20% increase in 340/380 ratio relative to most recent baseline or washout were considered positive responders. Cells were only included for analysis if they exhibited no positive response to initial ENH exposure and exhibited a positive response to at least one test compound or 50 mM KCl.

QUANTIFICATION AND STATISTICAL ANALYSIS

All HEK293 experiments comprised at least three biological replicates (independent transfections). Within each of these assays, individual analogs or constructs were screened in triplicate (technical replicates). All G-402 experiments shown contained at least three replicates, with biological replicates defined as unique plates. Data in concentration response curves were normalized with zero defined as the smallest value within a replicate and 100 was defined as the peak icilin or WS-12 response. Curve fitting was performed using non-linear dose-response algorithm in GraphPad Prism (v.10.6.1). All electrophysiology experiments shown were from a minimum of three unique recordings. Data were analyzed using Clampfit 11 (v.4) and visualized using Origin (v.10.2.0.188). No significant difference between male and female mice were observed. Statistical analyses for TG calcium imaging experiments were performed using the R programming language (version 4.3.2) and the RStudio environment. Comparisons of response and co-response frequency considered results from individual mice as samples, treated test-compound exposure as a within-sample independent variable, and thus used a paired t-test. Response magnitude was compared at an individual cell level, precluding the use of within-sample statistics. As these results violated the assumption of normality, they were compared with the Mann-Whitney U test. For Figure 2E, a 2-way ANOVA with repeated measures was used, treating “Agonist” as the within-subjects independent variable and “Genotype” as the between-subjects independent variable.

KEY resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Bacterial and virus strains
Biological samples
Chemicals, peptides, and recombinant proteins
Clonazepam	Cayman Chemical	Cat#14263; CAS: 162261–3
WS-12	Cayman Chemical	Cat#30352; CAS: 68489–09-8
Icilin	Cayman Chemical	Cat#10137; CAS: 36945–98-9
Menthol	Cayman Chemical	Cat#25753; CAS: 8978–1
Benzodiazepine Analytical Standards Panel	Cayman Chemical	Cat#26260
Flumazenil	Santa Cruz	Cat#sc-200161; CAS: 78755–81-4
TCI-2014	Tocris Bioscience	Cat#5410; CAS: 1221349–53-6
AMTB Hydrate	Cayman Chemical	Cat#25336: CAS: 926023–82-7
AMG-333	Cayman Chemical	Cat#36825; CAS: 1416799–28-4
RQ00203078	MedChem Express	Cat#HY-18662; CAS: 1254205–52-1
DMSO	Sigma-Aldrich	Cat#D2650; CAS: 6768-5
Opti-MEM	Thermo Fisher Scientific	Cat# 31985062
Critical commercial assays
Fluo-4 NW	Thermo Fisher Scientific	Cat#F36206
Fluo-4FF, AM	Thermo Fisher Scientific	Cat#F23981
Fluo-5F, AM	Thermo Fisher Scientific	Cat#F14221
Fura-2AM	Life Technologies	Cat#F1221
RNeasy Mini Kit	Qiagen	Cat#74104
High-Capacity RNA-to-cDNA™ Kit	Thermo Fisher Scientific	Cat#4387406
PowerUp™ SYBR™ Green Master Mix	Thermo Fisher Scientific	Cat#A25742
Lipofectamine 2000	Thermo Fisher Scientific	Cat#11668019
Lipofectamine 3000	Thermo Fisher Scientific	Cat#L3000015
Deposited data
Experimental models: Cell lines
Human: HEK293T	ATCC	CRL-1573
Human: G-402	ATCC	CRL-1440
Human: G-402 KO-1	Applied Biological Materials Inc.	Cat#C208; Lot#C24C11EANB
Human: G-402 KO-2	Applied Biological Materials Inc.	Cat#C208; Lot#C24C17EANB
Experimental models: Organisms/strains
Mouse: C57BL/6J	Jackson Labs	JAX: 000664; RRID: IMSR_JAX:000664
Mouse: B6.129P2-Trpm8tm1Jul/J	Jackson Labs	JAX: 008198; RRID: IMSR_JAX:008198
Oligonucleotides
Table S2	This study	N/A
Recombinant DNA
hTRPM8	GenScript	NCBI: NM_024080.5
Mm.TRPM8	GenScript	NCBI: NP_599013.1
Rn.TRPM8	Toro et al. 25	RRID: Addgene_64879
Fa.TRPM8	GenScript	NCBI: XP_005049003.1
hTPRM8[G805A]	GenScript	NCBI: NM_024080.5
Fa.TRPM8[A796G]	GenScript	NCBI: XP_005049003.1
TRPM8+eGFP	GenScript	NCBI: NM_024080.5; NCBI: UDY80669.1
Software and algorithms
Excel v.2202	Microsoft	RRID:SCR_016137
PRISM v.10.6.1	GraphPad Software	RRID:SCR_002798
Origin v.10.2.0.188	Origin Lab	RRID:SCR_014212
R programming language v.4.3.2	R-Project	RRID:SCR_001905
ScreenWorks v.4.2.0.11	Molecular Devices LLC.	Cat#5074979
ChemMine Tools Web Server	Backman et al.25	http://chemmine.ucr.edu
Clampfit 11 v.4	Molecular Devices	RRID:SCR_011323
NIS Elements Advanced Research	Nikon	RRID:SCR_027181
CFX Maestro Software	Bio Rad Laboratories Inc.	RRID:SCR_018064; Cat#12013758
Other
FLIPR® Tetra	Molecular Devices	N/A
FLIPR® Pipettor Head Kit, 384	Molecular Devices	Cat#0200–6072
Greiner Bio-One CELLSTAR 384-well, Cell Culture-Treated, Flat-Bottom Microplate	Fisher Scientific	Cat#07–000-072
Greiner Bio-One 384 Well Polypropylene Flat Bottom Non-Sterile Microplates	Fisher Scientific	Cat#781201
Axygen® 384-well tips Device, 30μL, Black, Non-filtered, Non-sterile, SLAS Rack	Corning	Cat#FT-384-BK-R
FLIPR® pipette tips, clear, non-sterile, 384-well	Molecular Devices	Cat#9000–0763
CFX Duet Real-Time PCR System	Bio Rad Laboratories Inc.	Cat#12016265
Olympus BX51WI inverted microscope	Olympus	N/A
multiClamp 700B amplifier	Molecular Devices	Cat#MULTICLAMP 700B
Digidata 1440A digitizer	Molecular Devices	Cat#DIGIDATA 1440A
Borosilicate glass pipettes	Sutter Instruments	BF150–110-10
TC-324B Single Channel Automatic Heat Controller	Warner Institute Inc.	N/A
DS-Qi1MC monochrome camera	Nikon	N/A

Highlights.

• Human TRPM8 is robustly activated by clonazepam

• Other 7-nitrobenzodiazepines also activate TRPM8

• Clonazepam efficacy in burning mouth disorder may be mediated via engagement of TRPM8

Significance :

Benzodiazepines are among the most widely used neuroactive drugs, with clinical actions largely attributed to modulation of GABA_A receptors. This study identifies TRPM8, a cold- and menthol-sensitive ion channel expressed in sensory pathways, as a non-canonical target of clonazepam and related 7-nitrobenzodiazepines. By showing that clonazepam activates TRPM8 in heterologous systems, native human cells, and trigeminal sensory neurons, these findings expand the pharmacological landscape of benzodiazepines beyond inhibitory neurotransmission. The work provides a potential molecular explanation for the efficacy of topical clonazepam in burning mouth disorder and highlights TRPM8 engagement as a mechanism that may contribute to sensory effects of benzodiazepines not readily explained by GABA_A receptor modulation.