In this study, we found that genetic deletion of TRPV3 alone or simultaneous deletion of TRPV3 and TRPV4, had minimal impact on innocuous or noxious heat sensation under naïve conditions, following inflammation with CFA, or when TRPV1 was masked.
Our findings on TRPV3 knockout mice differ from those of a previous study [19
], even though both reports examined mice from the same founder line. Several experimental differences between these two studies may have contributed to the discrepant results. The animals used in the Moqrich et al. study were the progeny of intercrossed C57BL6/129J N1 littermates, and their thermosensory phenotypes might therefore have been influenced by inhomogeneous inheritance of non-TRPV3 determinants. In contrast, the mice used in our study were on more homogeneous C57BL6 and 129S6 genetic backgrounds. Large differences in heat nociceptive behavior have been demonstrated among inbred mice of different genetic backgrounds [21
]. At least some of this strain dependence has been linked to the calcitonin gene-related peptide locus [19
]. Our data indicate that such strain dependence may also extend to innocuous thermosensation. Wild-type C57BL6 mice selected thermal gradient temperatures rather consistently, forming a single occupancy peak at ~32°C, whereas wild-type 129S6 mice distributed themselves more heterogeneously between one peak at ~22°C and another at ~32°C. On the 129S6 background, the absence of TRPV3 did have a modest effect on thermal selection behavior. On the thermal gradient, TRPV3 knockout 129S6 mice more consistently favored a single narrow range around ~22°C. TRPV3 knockout 129S6 mice exhibited a similarly altered pattern of thermal preference in the two-temperature selection task, although the differences were not statistically significant at any individual time point. In this assay, they showed less preference for 32°C over 22°C and a greater preference for 32°C over 37°C, compared with wild-type controls. The fact that these knockout mice also exhibited an apparently stronger preference for 32°C over 19°C than wild-type controls might at first appear paradoxical. However, closer inspection of Figure reveals that, even in the thermal gradient assay, the "shoulder" of the ~22°C peak of the wild-type 129S6 distribution appears to extend to cooler temperatures than that of the TRPV3 knockouts. While such a trend is certainly within the range of variability of the assay, it is tempting to speculate that it might reflect a generally more restrictive thermal preference of TRPV3 knockout 129S6 mice with a peak ~22°C. It is unclear whether this apparent phenotype arises from one or more specific roles of TRPV3 in governing the breadth of thermal tolerance as a sensor of warm temperatures, as opposed to less specific compensatory adaptations. Given the reported link between TRPV3 and heat-evoked keratinocyte release of ATP [20
], however, it is interesting to note that C57BL6 mice lacking the adenosine triphosphate-gated ion channel, P2X3, also exhibit a more restricted distribution than wild-type controls on the thermal gradient, although their occupancy, like that of their wild-type controls, is centered in the 32-34°C range [4
]. Still, it is worth emphasizing that we saw no restriction of distribution in TRPV3 knockout C57BL6 mice, compared with wild-type C57BL6 controls. Thus, background strain appears to be a key modifier of mouse thermal selection behavior.
The subtle patterns described above emphasize another possible reason for differences between our results and those of Moqrich et al. [19
], namely the use of different assay parameters. The thermal gradient extended from 15°C to 55°C in their study, versus 1°C to 49°C in the present study. For the two-temperature discrimination assay, the Moqrich et al. study employed 3 temperature pairs (room temperature vs. room temperature, room temperature vs. 35°C, room temperature vs. 15°C; room temperature = 22-24°C), whereas in our study, the temperature choices were 33°C vs. 37°C, 34°C vs. 28°C, or 35°C vs. 24°C for C57BL6 mice and 33°C vs. 37°C, 32°C vs. 22°C, or 32°C vs. 19°C for 129S6 mice. Moreover, mice were assayed for 10 min in the Moqrich et al. study but for 60 minutes in the present study. Non-thermal (e.g. spatial) cues in the testing environments may also have contributed to the apparently disparate results. Finally, differences in acute nociception assay criteria might also have contributed to the lack of agreement between studies. For example, in the 55°C hot plate test, wild-type mice exhibited an average withdrawal latency of ~13 seconds in the Moqrich et al. study, versus ~7 to 8 seconds in our study. Despite these differences in assay conditions and mouse populations, both studies found, at most, modest phenotypes in thermal selection and heat nociception behaviors in mice lacking TRPV3.
We previously reported that TRPV4 knockout [16
] mice prefer slightly warmer floor temperatures in both thermal gradient and two-temperature selection assays, and exhibit longer tail withdrawal latencies to moderately hot temperatures (45°C and 46°C) but not at higher temperatures. Other investigators reported evidence of reduced inflammatory heat hyperalgesia in these same mice [15
]. Based on these data, one might have expected TRPV3/TRPV4 double knockout C57BL6 mice to exhibit deficits in innocuous thermosensation and heat nociception that were at least as pronounced as those reported in TRPV4 knockout mice, if not more so. Surprisingly, thermal selection behavior on the thermal gradient was indistinguishable between wild-type and TRPV3/TRPV4 double knockout mice. Furthermore, thermal nociception was relatively unimpaired in these mice, with the exception of a small increase in response latency in the radiant paw heating assay. Although this difference persisted following paw inflammation with CFA, the reduction in latency produced by inflammation was comparable between genotypes. One possible explanation for these discrepancies with prior findings is that double knockout mice have undergone compensatory changes in non-TRPV3/TRPV4 thermosensory mechanisms, either as a consequence of lifelong absence of these channels or as these mutant alleles were propagated across generations. Variability in such compensation, the small magnitude of the knockout phenotype, and/or subtle differences between assay conditions might explain why we were able to measure changes in the radiant paw heating response latency in some cohorts of mice but not others. In this light, it would be interesting to assess the effects of TRPV3- and/or TRPV4-selective antagonists in wild-type mice, or the acute, inducible disruption of both genes.
Our data also indicate that a TRPV1 does not mask major thermosensory roles for endogenous TRPV3 and TRPV4. We previously used a TRPV1 antagonist to unmask thermal hyperalgesia in transgenic mice overexpressiong TRPV3 in keratinocytes [23
]. In contrast, TRPV1 antagonism resulted in increases in heat response latencies to comparable levels between wild-type and TRPV3/TRPV4 double knockout mice in the present study, both under naïve conditions and following inflammation with CFA. Taken together, these findings support the notion that endogenous TRPV3 and TRPV4 are largely expendable for heat nociception. Another candidate contributor to TRPV1-independent thermosensation is TRPV2, which is strongly expressed in a subpopulation of medium-to-large diameter sensory neurons and which can be activated by heat at temperatures > 52°C. Although the thermosensory characterization of TRPV2 knockout mice has not been published, the relatively high threshold for activation of this channel by heat argues for consideration of other mechanisms for heat detection at lower temperatures. Regardless of the signaling mechanism(s) involved, residual heat-evoked responses are most likely mediated by neurons that normally express TRPV1, since chemical ablation of the central terminals of these neurons with intrathecal resiniferatoxin [24
] has been reported to ablate the vast majority of heat avoidance behaviors in mice.
It remains possible that endogenous TRPV4 or TRPV3 contribute more substantially to temperature sensation in some capacity, but that our behavioral assays are inadequate to evaluate these contributions. It might therefore be worthwhile to devise operant-based behavioral assays or employ non-behavioral readouts to identify any finer influence TRPV3 and TRPV4 might have on thermosensory physiology. For example, electrophysiological recordings from secondary nociceptive neurons or peripheral nerve fibers may be able to tease out subtle deficits in thermosensory coding. It might be best to perform such studies without the confounding influence of TRPV1. Unfortunately, the very close linkage between TRPV1 and TRPV3 loci [25
] precludes the generation of TRPV1/TRPV3 double knockouts using standard crosses between single knockout lines, while the hyperthermic effects of TRPV1 antagonists [23
] would impose experimental constraints on a pharmacologically-based strategy. Moreover, classically-defined warmth receptive peripheral neurons have been best-studied in cats, nonhuman primates, and humans [28
], but have been exceedingly difficult to identify and characterize neurophysiologically in rodents [30
]. Another consideration is that since our behavioral evaluations were restricted to males, we cannot exclude the possible influence of gender on TRPV channel thermosensory contribution.
Although TRPV3 and TRPV4 are heat-sensitive, the major endogenous functions of TRPV3 and TRPV4 may involve processes other than temperature perception. For example, it has been demonstrated that both TRPV3 and TRPV4 are involved in different aspects of forming or maintaining the skin permeability barrier, and in the case of TRPV4, temperature has been shown to modulate this process [31
]. Multiple studies have also implicated endogenous TRPV4 in mechanical hyperalgesia or hypotonicity-induced pain, particularly under inflammatory or nerve-injured conditions [17
Our results support the notion that TRPV3 and TRPV4 likely make limited and strain-dependent contributions to innocuous warm temperature perception or noxious heat sensation, even when TRPV1 is masked. These findings imply the existence of other significant mechanisms for heat perception.