Rational Design of a Modality-Specific Inhibitor of TRPM8 Channel against Oxaliplatin-Induced Cold Allodynia

Abstract

Platinum-based compounds in chemotherapy such as oxaliplatin often induce peripheral neuropathy and neuropathic pain such as cold allodynia in patients. Transient Receptor Potential Melastatin 8 (TRPM8) ion channel is a nociceptor critically involved in such pathological processes. Direct blockade of TRPM8 exhibits significant analgesic effects but also incurs severe side effects such as hypothermia. To selectively target TRPM8 channels against cold allodynia, a cyclic peptide DeC-1.2 is de novo designed with the optimized hot-spot centric approach. DeC-1.2 modality specifically inhibited the ligand activation of TRPM8 but not the cold activation as measured in single-channel patch clamp recordings. It is further demonstrated that DeC-1.2 abolishes cold allodynia in oxaliplatin treated mice without altering body temperature, indicating DeC-1.2 has the potential for further development as a novel analgesic against oxaliplatin-induced neuropathic pain.

Keywords: TRPM8; designed protein; ion channels, oxaliplatin-induced cold allodynia; pain.

Figure 1

Rational design of cyclic peptides DeC1. a) Sideview of Cryo‐EM structure of TRPM8 in the apo state (PDB ID: 6O6A). b) Top view of TRPM8 outer pore region in the apo, desensitized (PDB ID: 6O77), and our computational model of cold activated state. c) An arginine (colored in purple) stably docked to the outer pore of TRPM8 as the optimal hotspot. d) The designed peptide DeC‐1.1 with two cysteine residues forming a disulfide bond (colored in yellow) bound to outer pore of TRPM8. e) Representative whole‐cell recording of DeC‐1.1 inhibiting menthol activation of mouse TRPM8 channel. f) In silico affinity maturation of DeC‐1.1 improving both the binding energy (Rosetta energy term ddg) and peptide stability (Rosetta energy term score) to get DeC‐1.2.

Figure 2

Functional validation of DeC‐1.2. a) Sequence logo of the amino acid sequence alignment among designed cyclic peptides with top ten binding energy. The height of a letter is proportional to the relative frequency of that residue at a particular site. The sequence of DeC‐1.2 is CRRDRARHYRQRC. b) DeC‐1.2 with the largest binding energy was shown in cyan. c) Average fluorescence response and d) representative calcium imaging of mTRPM8 in response to 1 × 10⁻³ m menthol, 500 × 10⁻⁹ m DeC‐1.2, 500 × 10⁻⁹ m S‐DeC‐1.2, and 10 × 10⁻⁶ m ionomycin, respectively (n = 20). Data in (d) are represented as average ± SEM. Scale bars, 100 × 10⁻⁶ m. e) Representative whole‐cell recording of DeC‐1.2 inhibiting menthol activation of mouse TRPM8 channel recorded at −80 mV. f) The concentration–response curves of DeC‐1.1 and DeC‐1.2 were measured with whole‐cell patch‐clamp recordings (n = 3). g) Representative whole‐cell current response of mTRPM8 to the scrambled peptide (S‐DeC‐1.2) and DeC‐1.2 (both at 500 × 10⁻⁹ m). The holding potential was −80 mV.

Figure 3

Modality‐specific inhibition of TRPM8 activation by DeC‐1.2. a) Representative single‐channel recordings of TRPM8 in outside‐out configuration at +80 mV. Saturating menthol (1 × 10⁻³ m) and the mixture of menthol (1 × 10⁻³ m) and DeC‐1.2 (100 × 10⁻⁶ m) were perfused to activate and inhibit the channel, respectively. These four representative recordings were from the same membrane patch. b) All‐point histograms of the representative single‐channel recordings shown in (a). Histograms were fitted to a double Gauss function (solid line in red), where the difference in two fitted peak values was used to calculate the single‐channel conductance. c) Open probability of TRPM8 was significantly decreased by DeC‐1.2 (100 × 10⁻⁶ m) in the presence of saturating menthol (1 × 10⁻³ m). Data were shown as Mean ± SEM of five cells for each group. **** indicates p < 0.0001. d) Single‐channel conductance of TRPM8 was significantly decreased when DeC‐1.2(100 × 10⁻⁶ m) was applied in the presence of saturating menthol (1 × 10⁻³ m). Data were shown as Mean ± SEM of three cells for each group. *** indicates p < 0.001. e) Representative single‐channel traces of TRPM8 cold activation recorded from inside‐out patches at +80 mV with or without DeC‐1.2 in the pipette solution (100 × 10⁻⁶ m). f) All‐point histograms of the representative single‐channel recordings shown in (e). Histograms were fitted to a double Gauss function. g) Open probability of TRPM8 was not significantly changed when DeC‐1.2 (100 × 10⁻⁶ m) was applied at 16 °C. Data were shown as Mean ± SEM of five cells for each group. N.S. indicates No significance. h) Single‐channel conductance of TRPM8 which showed no significant changes when DeC‐1.2(100 × 10⁻⁶ m) was applied at 16 °C. Data were shown as Mean ± SEM of three cells for each group. N.S. indicates No significance.

Figure 4

The subtype selectivity of DeC‐1.2. a–e) The representative whole‐cell current recording of DeC‐1.2 inhibition of TRPV1, TRPV2, TRPV3, TRPM2, and TRPM4 ligand activation, respectively. TRPM4‐K1045A mutant was used to remove PIP2 dependence. f) Comparison of normalized inhibition of DeC‐1.2 at 500 × 10⁻⁹ m against TRP channels and Nav channels in TRPV channel. The normalized inhibition of ion channels was determined by first measuring and calculating the fraction of current inhibition by 500 × 10⁻⁹ m DeC‐1.2 in each channel, and then the inhibition fraction for each channel was normalized to that of TRPM8. For each channel, n = 3, **** indicates p < 0.0001.

Figure 5

Critical residue(s) on DeC‐1.2 for its specific inhibition of TRPM8. a–d) Representative whole‐cell current recording of 5 × 10⁻⁹ m wild type DeC‐1.2, mutant R2A, A6G and R12A inhibition of TRPM8 ligand activation, respectively. e–g) Representative whole‐cell current recording of 5 × 10⁻⁹ m wild type DeC‐1.1, DeC‐1.2 and mutant R3N, D4S, and R12K inhibition of TRPM8 ligand activation, respectively. h) Normalized inhibition of the wild type DeC‐1.2 and its mutants against TRPM8 ligand activation at 5 × 10⁻⁹ m, *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively (n = 3). i) Normalized inhibition of the wild type DeC‐1.1, DeC‐1.2 and its mutants against TRPM8 ligand activation at 5 × 10⁻⁹ m, ** and **** indicate p < 0.01 and p < 0.0001, respectively. n = 3–18. j) The amino acid residues essential for TRPM8 inhibition as determined in alanine scan mapped onto the designed structure of DeC‐1.2. Residues eliminated or largely reduced channel inhibition were colored in red or orange, respectively.

Figure 6

Effects of TRPM8 inhibition by DeC‐1.2 in vivo and ex vivo. a) DeC‐1.2 significantly and dose dependently reduced icilin induced wet dog shake (WDS) behavior in mice. Data were shown as Mean ± SEM of 6–9 animals for each group. *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively. b) Schematic illustration of the procedures to establish oxaliplatin‐induced cold allodynia in mice and in vivo tests of DeC‐1.2 against cold allodynia. c) Representative ratiometric calcium images of DRG neurons isolated from vehicle treated mice perfused with extracellular solution (Vehicle), oxaliplatin treated mice perfused with extracellular solution (Oxaliplatin) and oxaliplatin treated mice perfused with DeC‐1.2 (10 × 10⁻⁶ m) in extracellular solution (Oxaliplatin + DeC‐1.2). Increased calcium influx was observed upon TRPM8 channel activation by menthol application (100 × 10⁻⁶ m). Capsaicin (1 × 10⁻⁶ m) that activated TRPV1 channels in DRG neurons served as the positive control. The neurons responded to the agonists were pointed out by arrows in white. d) Representative fluorescent ratio traces in response to corresponding reagents in DRG neurons from vehicle treated and oxaliplatin treated mice. e) Percentage of menthol responding DRG neurons from vehicle treated mice and oxaliplatin treated mice. Total neurons for vehicle, oxaliplatin, and oxaliplatin + DeC‐1.2 were 229, 238, and 334, respectively. * and ** indicate p < 0.05 and p < 0.01, respectively. Each open circle represented one field of view for neuron counting.

Figure 7

In vivo effect of DeC‐1.2 against oxaliplatin‐induced cold allodynia. a) Effect of DeC‐1.2 on oxaliplatin‐induced cold allodynia. The time spent in withdrawal, flinching, or licking the stimulated paw was recorded and blindly counted, as acetone score, during 5 min after acetone treatment. 3.5 µg of DeC‐1.2 or S‐DeC‐1.2 in 20 µL solution was injected. Data were shown as Mean ± SEM of 9–13 animals in each group. * indicates p < 0.05 in two‐way ANOVA. b) Effect of DeC‐1.2 on oxaliplatin‐induced mechanical allodynia. 3.5 µg of DeC‐1.2 or S‐DeC‐1.2 in 20 µL solution was injected. Data were shown as Mean ± SEM of 9–13 mice in each group. c) Effect of DeC‐1.2 on body temperature. Rectal temperatures in mice were measured before treatment (baseline, BL) and at 5 min, 15 min, 0.5, 1, 2, 6, and 24 h following DeC‐1.2 or S‐DeC‐1.2 (30 µg g⁻¹, i.v.) administration. Data were shown as Mean ± SEM of 5–9 animals. N.S. indicates No significance in one‐way ANOVA. d) Schematic diagram of the modality‐specific inhibition of TRPM8 channel by DeC‐1.2.