Human heat sensation: A randomized crossover trial

Abstract

Sensing of noxious heat has been reported to be mediated by TRPV1, TRPA1, TRPM3, and ANO1 in mice, and this is redundant so that the loss of one receptor is at least partially compensated for by others. We have established an infusion-based human heat pain model. Heat-induced pain probed with antagonists for the four receptors did not match the redundancy found in mice. In healthy participants, only TRPV1 contributes to the detection of noxious heat; none of the other three receptors are involved. TRPV1 inhibition reduced the pain at all noxious temperatures, which can also be seen as an increase in the temperature that causes a particular level of pain. However, even if the TRPV1-dependent shift in heat detection is about 1°C, at the end of the temperature ramp to 52°C, most heat-induced pain remains unexplained. This difference between species reopens the quest for the molecular safety net for the detection of noxious heat in humans.

Fig. 1.. Study design and experimental heat pain model.

(A) Overview for both study visits. Visit 1 established the infusion heat pain model, including modulation by coinjection of lidocaine. Visit 2 consisted of eight sequential injections, performed double-blinded according to 1 of 16 predetermined sequences. (B) From a syringe filled with synthetic interstitial fluid ± test substances, the solution is passed through a heat block with constant temperature and equilibrates to that temperature. The injection rate determines the time for cooling toward room temperature before the fluid reaches the intradermally positioned 27-gauge needle. Nonlinear increase in injection rate (gray) allowed to obtain a largely linear temperature ramp at the outlet in the range of 44° to 52°C. The total injection volume is 812 μl. (C) Time-dependent numerical pain rating in single-blinded experiments every 5 s on a 0 to 100 scale. There was minimal reported pain for the three injections at room temperature. (D) In contrast, the three injections with increasing temperature generated substantial pain, which rapidly subsided at the end of the 2.5-min injection. (E) Lidocaine (2 mM) served as a positive control for pain reduction by a substance added to the injection. The distribution is visualized with the median as a solid line and decreasing gray or color shades for percentiles more distant from the median (in 10% percentile steps, as indicated by the scale bar). (F) Pain AUC for injections at room temperature, heated injections, and heated injections with 2 mM lidocaine. Violin plot, indicating the median, interquartile range, and distribution, was overlaid with a spaghetti plot reflecting the 48 individual results. The HPI by lidocaine was calculated; its median (circle) is plotted with the 95% CI. (G) Pain induced by heated injections without and (H) with lidocaine plotted against the temperature at the needle tip.

Fig. 2.. Heat-induced pain ratings in response to pharmacological inhibition of TRPV1, TRPA1, TRPM3, and ANO1.

Each panel shows the distribution of pain ratings in the range of 44° to 52°C in visit 2 obtained from 24 participants. (A) Pain ratings without antagonists. (B to P) TRPA1 inhibitor A-967079 (10 μM) is present in panels of columns 2 and 4 and can be compared to the panels to the left of it. TRPV1 inhibitor BCTC (1 μM) is present in columns 3 and 4. Similarly, the TRPM3 inhibitor naringenin (20 μM) is present in rows 2 and 4 and can be compared to the panels above it. The ANO1 inhibitor Ani9 (10 μM) is present in rows 3 and 4. The red dashed line is the median of the control experiment shown in (A). The experimental design derives statistical efficiency from four measurements with and without every antagonist in every participant. Direct statistical pairwise comparisons of panels were formally not justified based on the prespecified design and due to the nonsignificant interaction terms.

Fig. 3.. Heat pain is reduced by inhibition of TRPV1 but not TRPA1, TRPM3, and ANO1.

Each panel shows the distribution of pain ratings plotted against the temperature. Data are from visit 2 and all 48 participants. The left column shows injections without the respective antagonist but with or without the other antagonists. The middle column consists of all injections including the respective antagonist. (A) Pain ratings of all heated injections without and (B) with 1 μM BCTC. The red dashed lines indicate the median of all injections without the respective substance. (C) HPI due to BCTC over the whole time course (contrast estimate with 95% CI). (D) Pain ratings of all injections without and (E) with 10 μM A-967079, resulting in (F) no relevant HPI. (G) Pain ratings of all injections without and (H) with 20 μM naringenin, resulting in (I) no relevant HPI. (J) Pain ratings of all injections without and (K) with 10 μM Ani9, resulting in (L) no relevant HPI.

Fig. 4.. Heat pain threshold and temperature-dependent fractional inhibition.

(A to D) Shift in temperature required to induce the same pain rating. (E to H) Temperature-dependent fractional inhibition of the heat-induced pain by the antagonist. (I to L) Maximum pain ratings in injections without the antagonist (red) versus injections with the antagonist (ochre). Data are estimates with 95% CIs. (M to P) Modeled probability to distinguish the heated injections from the individual three injections at room temperature in the presence and absence of the respective antagonist. Each participant had four injections without the antagonist (red) versus injections with the antagonist (ochre). Data are estimates with 95% CIs.