An adamantane-based ligand as a novel chemical tool for thermosensory TRPM8 channel therapeutic modulation

Abstract

Transient receptor potential cation channel subfamily M member 8 (TRPM8) is a nonselective thermosensory cation channel expressed in peripheral nociceptor terminals where it transduces cold temperatures and cooling agents such as menthol. TRPM8 dysfunction has been involved in disabling sensory symptoms, such as cold allodynia. In addition, its widespread expression has signaled this channel as a pivotal therapeutic target for a variety of diseases, from peripheral neuropathies to cancer. Thus, the design and therapeutic validation of TRPM8 antagonists is an important endeavor in biomedicine. To address this, we used the multicomponent Passerini and Ugi reactions to design a novel family of TRPM8 modulators using as a scaffold the adamantane ring that exhibits drug-like qualities. These green chemistry transformations are ideal for the fast synthesis of libraries of medium complexity with minimal or no generation of waste by-products. We report the identification of a family of TRPM8 agonists and antagonists. Among them, 2-((3S,5S,7S)-adamantan-1-ylamino)-2-oxoethyl [1,1'-biphenyl]-2-carboxylate (referred to as compound 23) is a potent and selective antagonist that reduces TRPM8-induced neuronal firing in primary nociceptor cultures. Compound 23 exhibits 10-fold higher potency for human TRPM8 (hTRPM8) than for hTRPV1 and hTRPA1 channels. Notably, local administration of compound 23 significantly attenuated oxaliplatin-induced peripheral cold allodynia by modulating epidermal TRPM8 sensory endings. Thus, α-acyloxy carboxamide 23 appears as a promising therapeutic candidate to topically intervene on TRPM8-mediated peripheral neuropathies.

Keywords: cold allodynia; drug discovery; ion channel; medicinal chemistry; neuropathy.

Fig. 1

Preparation reaction of 1‐isocyanoadamantane 3 and chemical structure of carboxylic acids 4–14 (red color) and amines 15–19, 1 (green color) used in this study.

Fig. 2

Passerini and Ugi reactions. (A) Preparation of α‐acyloxy carboxamides 20–29 via the Passerini reaction and compounds obtained. (B) Synthesis of α‐aminoacyl amides 30–41 and α‐acylpiperazino amides 42–45 exploiting the Ugi and split‐Ugi reactions, respectively, and compounds obtained (black color refers to isocyanide component, blue color refers to aldehyde component, red color refers to carboxylic acid component, and green color refers to amine component).

Fig. 3

Compound 23 is a hTRPM8 competitive blocker with nanomolar potency. (A) Representative traces of WS‐12‐evoked ionic currents recorded at a holding potential of −60 mV in HEK293 cells heterologously expressing hTRPM8. Control cells (upper panel) were exposed to two pulses (1 μm) of WS‐12, interspaced by a washing period whereas compound 23‐treated cells (middle panel) were exposed before to 1 min of compound 23 (1 μm) and then to 30s of compound 23 during the second pulse of WS‐12. In the lower panel, the recovery of WS‐12‐evoked currents after compound 23 treatment is shown, indicating a reversible effect of the compound on hTRPM8 activity. (B) Dose–response curve for hTRPM8 current blockade at a holding voltage of −60 mV. The line represents the fit of the experimental data to the Hill equation: $Y=\text{Bottom}+\left(\mathrm{Top}‐\text{Bottom}\right)/\left(1+{10}^{\left(X-{\text{LogIC}}_{50}\right)}\right)$ with a standard slope of 1.0 (Hill coefficient) and a restriction Top (= 100). The fitted value for IC₅₀ was 0.08 ± 0.07 μm. Each point is the mean ± SEM of N = 3, n = 10. (C) WS‐12 dose–response curve in the absence (black) or presence of 0.08 μm compound 23 (blue). The best fitted values for WS‐12 EC₅₀ value were 3.0 μm (95% CI 2.5–3.68 μm) (n = 18) in the absence of 0.1 μm compound 23 and 15 μm (95% CI 12.8–16.8 μm) in its presence (blue curve, n = 16, Top = 100). Hill coefficient in the absence of compound 23 was 1.17 (0.9–1.4) and in its presence was 0.95 (95% CI 0.8–1.0). All data are expressed as mean ± SEM. N = 3. (D) Schild plot derived from the interaction between agonist (WS12) and antagonist (compound 23) at four different concentrations: 0.02, 0.08, 0.1, 0.5 μm. Equation: Y = 1003 × X + 1469; Slope: 1.003; 1/Slope: 0.9973; Std. error: 0.1935. R square: 0.9307; P value: 0.0353. N = 3, n = 10. (E) Representative current density (J)‐V curves elicited by a protocol of voltage steps from −120 to 120 mV in steps of 20 mV in the absence (control) or presence of 0.08 μm compound 23 (n = 10) or 10 μm AMTB (n = 8). In the control condition, an additional pulse of WS‐12 was used to consider channel desensitization. Data were analyzed with two‐way ANOVA, N = 3. Each point is the mean ± SEM. Data were analyzed using a two‐way ANOVA followed by a Sidak post hoc test when appropriate, **P value = 0.0069; ****P value < 0.0001.

Fig. 4

Compound 23 modestly inhibits currents of hTRPV1 and hTRPA1 expressed in HEK293 cells. (A) Representative capsaicin (0.5 μm)‐evoked hTRPV1 inward currents recorded at a holding potential of −60 mV, for control cells (untreated) and cells treated with 1 μm compound 23 or 10 μm of Capsazepine. Red lines represent the duration of the capsaicin pulse (top panel). Dose–response curve for hTRPV1 current blockade at a holding voltage of −60 mV. The line represents the fitted curve of the experimental data to the Hill equation. The fitted value for IC₅₀ was 1 μm ± 1.02. Each point is the mean ± SEM of n = 6 (bottom panel). (B) Representative AITC (60 μm) evoked hTRPA1 inward current recorded at a holding potential of −60 mV, for control cells (untreated) and cells treated with 5 μm compound 23 or 5 μm of HC030031. Orange lines represent the duration of the AITC pulse (top panel). Dose–response curve for hTRPA1 current blockade at a holding voltage of −60 mV. The line represents fits of the experimental data to the Hill equation. The fitted value for IC₅₀ was 3.04 μm ± 2.01. Each point is the mean ± SEM of n = 8 (bottom panel).

Fig. 5

Molecular docking of compound 23 in TRPM8 channels. (A) Side and top view of human TRPM8 structure (PDB ID: 8BDC) used for docking studies. Subunits are colored differently. The tetrameric nature of the channel and the pore are clearly seen. The red square roughly indicates the simulation box built around the menthol binding site to accommodate compound 23. (B) Detail of the human TRPM8 menthol binding pocket with bound compound 23 (orange color). Residues mainly involved in ligand interactions were F738 (S1), W798 (S3), R842, H845, I846, V849, L853 (S4), and L1001, Y1005 (TRP helix). Interactions are mainly hydrophobic, although a salt bridge and π‐cation interactions are also observed. (C) Detail of compound 23 bound to the mouse TRPM8 binding pocket (equivalent view to panel B). Residues interacting with compound 23 (orange) were Q785 (S2), D796, W798, N799, D802 (S3), H845 (S4), and E1004 (TRP helix). (D) Compound 23 (orange) and WS‐12 (cyan) or AMTB (green) superimposed in the human TRPM8 menthol binding site for comparison. (E) Compound 23 (orange) and WS‐12 (cyan) or AMTB (green) superimposed in the mouse TRPM8 menthol binding site for comparison. Figures were constructed using open‐source pymol v3.0 (https://pymol.org/).

Fig. 6

Compound 23 blocks WS‐12‐induced TRPM8 activation in murine nociceptors. (A) Representative WS‐12‐evoked currents in DRG neurons of neonatal rats recorded at a holding potential of −60 mV. For control conditions (untreated) cells were exposed to two pulses (1 μm) of WS‐12, interspaced by a washing period, whereas treated cells were exposed before to 1 min of compound 23 (1 μm) and then to 30 s of compound 23 during the second pulse of WS‐12. (B) Dose–response curve for mTRPM8 current blockade at a holding voltage of −60 mV. The line represents fits of the experimental data to the Hill equation. The fitted value for IC₅₀ was 0.43 μm ± 0.75. Each point is the mean ± SEM of N = 3, n = 8.

Fig. 7

Compound 23 reduces WS‐12‐induced excitability of murine nociceptors. (A) Multielectrode array (MEA) recordings, with representative traces showing WS‐12‐evoked action potential (AP) firing under different conditions: control, 0.43 μm compound 23, 0.43 μm AMTB, and 10 μm AMTB. WS‐12 (1 μm) was applied in two consecutive pulses, with compounds added 1 min before and during the second pulse. The protocol concluded with a 15 s pulse of 40 mm KCl to ensure neuronal viability. (B) Normalized WS12‐induced firing (P2/P1) in the absence (vehicle) and presence of antagonists (0.43 μm compound 23, 0.43 μm AMTB and 10 μm AMTB) were compared. Data were analyzed with one‐way ANOVA followed by Dunnett's test; P values for statistical differences are indicated. The number of independent experiments (N) was 3, with 94 electrodes per condition. The data points are plotted with error bars representing the standard deviation (SD). (C) Multielectrode array (MEA) recordings, with representative traces showing Cap‐evoked action potential (AP) firing under different conditions: control and 1 μm compound 23. Cap (0.5 μm) was applied in two consecutive pulses, with compounds added 1 min before and during the second pulse. The protocol concluded with a 15 s pulse of 40 mm KCl to ensure neuronal culture viability. (D) Normalized capsaicin‐induced firing (P2/P1) in the absence (vehicle) and presence of antagonist (1 μm compound 23) were compared. The data points are plotted with error bars representing the standard deviation (SD). The number of independent experiments (N) was 3, with 63 electrodes per condition. Data were analyzed with Mann–Whitney test; P values for statistical differences are indicated. (E) Representative traces showing AITC‐evoked action potential (AP) firing under different conditions: control and 5 μm compound 23. AITC (100 μm) was applied in two consecutive pulses, with compounds added 1 min before and during the second pulse. The protocol concluded with a 15 s pulse of 40 mm KCl to ensure neuronal culture viability. (F) Normalized AITC‐induced firing (P2/P1) in the absence (vehicle) and presence of antagonist (5 μm compound 23) were compared. Data were analyzed with Mann–Whitney test; P values for statistical differences are indicated. The data points are plotted with error bars representing the standard deviation (SD). The number of independent experiments (N) was 3, with 67 electrodes per condition.

Fig. 8

Local application of 23 alleviates oxaliplatin‐induced sensitivity to cold. (A) Repeated administration of oxaliplatin increased the duration of responses to cold induced by acetone application (A, ^# P < 0.05 Before vs After OXP, two‐way ANOVA). However, no significant effect of systemic 23 was found 30 or 90 min after its intraperitoneal (i.p.) administration. (B) Oxaliplatin induced a significant decrease in the withdrawal latency to cold induced by dry ice application (B, ^# P < 0.05 Before vs After OXP, two‐way ANOVA). Vehicle‐treated mice retained significant cold sensitization after i.p. treatment (B, ^$,# P < 0.05, Before OXP vs 30 min after Vehicle, two‐way ANOVA), whereas this difference was not evident in 23‐treated mice. However, 23 did not exhibit cold antinociceptive effects following systemic treatment. (C) Oxaliplatin increased the duration of the responses to acetone‐induced cold (C, ^## P < 0.01 Before vs After OXP, two‐way ANOVA) and 23 alleviated oxaliplatin‐induced cold sensitivity 30 and 90 min after its local subcutaneous administration in the paw (i.pl., *P < 0.05 After OXP vs. 30 min after 23, **P < 0.01 After OXP vs. 90 min after 23). Two‐way ANOVA followed by Tukey post hoc test when appropriate. Bars represent average values and error bars represent SEM. Dots are the individual values of each mouse (N = 6–7).

Fig. 9

Proposed in vitro phase I metabolic scheme for the compound 23.