Introduction

Stroke is the leading cause of long-term disability and the fifth leading cause of death in the United States1. Despite major advances in the treatment of ischemic stroke (i.e. thrombolysis and thrombectomy), a large number of patients are still left with significant brain injury and lasting neurological deficits2,3,4,5,6,7,8. Thus, additional treatment options are still needed for patients with ischemic stroke.

Therapeutic hypothermia (TH) has been shown to be effective in other conditions that lead to brain hypoxia, including cardiac arrest and neonatal hypoxic-ischemic encephalopathy (nHIE)9. Indeed, the use of TH for the treatment of cardiac arrest and nHIE has been standard clinical care for over a decade9. In experimental stroke models, TH has been repeatedly shown to provide significant neuroprotection and improved functional outcome10. Unfortunately, efforts to translate this benefit to human stroke patients have faced multiple challenges, largely related to the counterproductive activation of robust cold defense mechanisms11. The current methods of inducing therapeutic hypothermia involve either forced external cooling (cold blankets or ice baths) or cooling via intravascular infusion of ice-cold saline or placement of an intravascular heat exchanging device12,13,14,15. While these cooling methods are effective in cardiac arrest patients with heavy sedation and paralytics, they are much less effective in the conscious stroke patient. In particular, current cooling methods for conscious patients suffer from variable efficacy in achieving and maintaining hypothermic temperatures as well as their potential to induce intense shivering, hypotension, or other clinical complications16. The major source of these complications is due to the incomplete suppression of the body’s cold-defense mechanisms in current cooling protocols. In humans and other mammals, even a slight drop in the core body temperature triggers the cold defense mechanisms, including increased thermogenesis (e.g. shivering) and actions to decrease heat loss (e.g. cutaneous vasoconstriction)11,17. Therefore, to quickly and effectively achieve target temperatures while also minimizing physiological stress in conscious subjects, new cooling approaches are needed that incorporate reliable suppression of cold-defense mechanisms.

Pharmacological activation of transient receptor potential vanilloid 1 (TRPV1) channels with capsaicin or dihydrocapsaicin (DHC) has been shown to produce a hypothermic response in rodents and larger mammals without triggering shivering or tachycardia11,18,19. Pharmacologically inducing hypothermia in this manner has potential to be used in place of current TH protocols, or in combination with current TH protocols to provide effective suppression of cold defense mechanisms in conscious subjects. However, one concern about the use of traditional TRPV1 agonists (e.g. capsaicin or DHC) is that they might activate pain pathways and produce a burning sensation (“pungency”) that could limit patient tolerance. Additionally, non-regionally specific TRPV1 activation can lead to unwanted off-target effects such as activating inflammation pathways or other adverse effects20,21,22. However, there is a class of capsaicin-related TRPV1 agonists (capsinoids) that are described as “non-pungent”. The capsinoids (capsiate, dihydrocapsiate, nordihydrocapsiate) are similar in molecular structure to capsaicinoids, but have an ester functional group in place of an amide, making the capsinoids vulnerable to breakdown by ubiquitously expressed esterases23. Thus, capsinoids get broken down before they are able to reach TRPV1-containing nerves of the tongue, cannot pass as far along the gastrointestinal tract as capsaicin, and do not appear in the blood following oral ingestion (unlike capsaicin)24,25.This chemical lability severely limits the ability of capsinoids to diffuse through tissues intact, thereby limiting the TRPV1 activating potential to the region of delivery. This “regional restriction” of activity also reduces potential for unwanted on-target (TRPV1-mediated) effects in other body regions.

The purpose of this study was to determine the efficacy of a > 97% purified capsinoid mixture (capsiate/dihydrocapsiate) to induce a therapeutically relevant hypothermia to reduce stroke in conscious mice. Herein, we demonstrate that capsinoids, when delivered to the peritoneal cavity, can provide a TRPV1-mediated induction of mild hypothermia in conscious mice that is suitable for application to post-stroke neuroprotection.

Results

Capsinoids are non-pungent TRPV1 agonists that are structurally related to capsaicinoids (e.g. dihydrocapsaicin (DHC) and capsaicin) differing only by an ester group in place of an amide group (Fig. 1A). Several studies have shown that capsaicinoids (DHC) can induce mild hypothermia in conscious mice, but the capacity of capsinoids to induce a similar response has not been explored18,26,27,28,29. To investigate this potential, we compared the effects of IP injection of capsinoids on body temperature (Tb) to the known effect of DHC in awake young male C57BL6 mice (Fig. 1B). Corn oil was injected as a vehicle control. Pre-injection baseline Tb values were within a small range (0.4 °C) between experimental groups (Vehicle: 37.5 ± 0.1 °C; DHC: 37.4 ± 0.1 °C; Capsinoid: 37.8 ± 0.1 °C). Though baseline Tb were statistically different (one way ANOVA, p = 0.0214, n = 4), the small starting value differences were minor compared with the subsequent TRPV1 agonist-induced changes. The response of Tb to IP injection is presented as change from baseline for DHC (−1.1 ± 0.7 °C), capsinoids (−2.5 ± 0.6 °C), and vehicle (+ 0.2 ± 0.2 °C). The drop in Tb was statistically significantly compared to baseline for DHC and capsinoids (two-way RM ANOVA, p < 0.05, n = 3 per group). This initial finding provided sufficient basis to further explore capsinoids as an effective means of pharmacologically inducing mild hypothermia.

Fig. 1
figure 1

Capsinoids induce an administration- and TRPV1-dependent hypothermia in young mice. a: Chemical structures of capsaicinoids and capsinoids. Black arrows represent amide groups and red arrows represent ester groups. b: Decrease in body temperature (Tb) in young male mice due to IP administration of DHC (2 mg/kg) and capsinoids (40 mg/kg) compared to vehicle. c: Tb does not change after OG administration of capsinoids. d: SC administration of capsinoids (40 mg/kg) does not alter Tb compared to vehicle control. e: SC administration of DHC (2 mg/kg) to same mice as in panel d elicits a drop in Tb. f: Capsinoid-induced hypothermia is TRPV1 dependent. (*: p < 0.05, **: p < 0.01, all drug administration points indicated by blue dashed lines).

Full size image

To determine the effectiveness of different delivery methods to alter Tb, we also investigated delivery of capsinoids by oral gavage (OG) and subcutaneous injection (SC). The baseline Tb for all experimental groups were similar prior to injection: vehicle (capsinoid): 38.1 ± 0.2 °C, capsinoid: 38.0 ± 0.2 °C, vehicle (DHC): 38.1 ± 0.3 °C, DHC: 37.8 ± 0.2 °C (one way ANOVA, p = 0.8854, n = 4). With delivery by OG, capsinoids did not promote a hypothermic effect, even at a dose of 100 mg/kg (Fig. 1C, male C57BL6, n = 3). Similarly, delivery of capsinoids by SC injection (40 mg/kg) failed to elicit a greater drop in Tb compared to vehicle control (Fig. 1D, two-way RM ANOVA, p = 0.5556, n = 4 per group). In contrast to the lack of hypothermic effect of capsinoids, SC administration of DHC (2 mg/kg) within the same mice induced a rapid and significant decrease in Tb (Fig. 1E, two-way RM ANOVA, p = 0.0054, n = 4 per group). Since capsinoid administration by OG and SC delivery failed to promote significant drop of Tb, all subsequent studies were performed with IP delivery.

Given that the hypothermic response to traditional TRPV1 agonists (e.g. capsaicin and DHC) is dependent on TRPV1 activation29,30,31, we sought to determine if capsinoid-induced hypothermia was similarly TRPV1-dependent. To this end, we administered capsinoids to young male mice with global knockout of TRPV1 (TRPV1 del/del) and TRPV1 fl/fl (floxed only) controls (Fig. 1F). Pre-injection baseline Tb was not different between the experimental groups (TRPV1 del/del: 36.7 ± 0.1 °C, TRPV1 fl/fl: 36.9 ± 0.2 °C; t-test, p = 0.4009, n = 4). Capsinoid injection induced a hypothermic response at 30 min post injection in the TRPV1 fl/fl control mice (−1.6 ± 0.3 °C), but not in the TRPV1 deletion mice (−0.2 ± 0.3 °C) (two-way RM ANOVA, p < 0.05, n = 4 per group). This finding indicates that the capsinoid mechanism of action is also highly TRPV1-dependent.

To further define the capacity of capsinoids to promote a drop in Tb, we examined the dose-responsiveness of the hypothermic response. We employed a rotating dose paradigm of 0 mg/kg (vehicle control – corn oil), 20 mg/kg capsinoids, and 40 mg/kg capsinoids, to young C57BL6 male (n = 7) and female mice (n = 8). Mice were allowed to recover for 3 days between injections to reduce the risk of confounding factors associated with repeated administrations. With this strategy, each mouse received all experimental doses, thus allowing them to serve as their own dose-to-dose controls. The baseline Tb for the sex-combined experimental groups were 37.5 ± 0.1 °C for vehicle, 37.6 ± 0.1 °C for 20 mg/kg capsinoids, and 37.5 ± 0.1 °C for 40 mg/kg capsinoids (one way ANOVA, p = 0.7763, n = 15). There was a significant dose effect of capsinoids in the sex combined group, with reductions in Tb corresponding to −1.6 ± 0.1 and − 2.5 ± 0.2 °C, for 20 and 40 mg/kg, respectively (Two-way RM ANOVA, p < 0.0001, n = 15 in all groups) (Fig. 2A). Note that there was a slight progressive decline in temperature in the vehicle-treated group over the two-hour period (−0.8 ± 0.1 °C). The cause of the slight vehicle effect on Tb was not determined, though the consistent response in multiple experiments may indicate a small hypothermic effect related to the vehicle solution. A similar small drop in Tb following intraperitoneal vehicle injection has also been observed in other labs with similar timecourse32. To determine if sex-differences exist in the hypothermic responses to capsinoid administration, we separated and analyzed the data by sex. The dose-dependency remained significant for both sexes (Fig. 2B/C, Two-way RM ANOVA, F: n = 8, M: n = 7, all doses, M: p < 0.0001, F: p < 0.0001). The temperature responses were then further compared by amplitude and kinetics. The baseline Tb were similar between experimental groups and sexes: female vehicle: 37.5 ± 0.2 °C, female 20 mg/kg capsinoids: 37.7 ± 0.1 °C, female 40 mg/kg capsinoids: 37.5 ± 0.1 °C, male vehicle: 37.5 ± 0.1 °C, male 20 mg/kg capsinoids: 37.5 ± 0.2 °C, and male 40 mg/kg capsinoids: 37.6 ± 0.2 °C (one way ANOVA, p = 0.9672, n = 7/8). Direct comparisons of the 20 and 40 mg/kg capsinoid Tb responses showed no apparent differences by sex (Figs. 2D and 20 mg/kg capsinoid: two-way ANOVA, F: n = 8, M: n = 7, p = 0.9620; Figs. 2G and 40 mg/kg capsinoid: two-way ANOVA, F: n = 8, M: n = 7, p = 0.3586). The averaged Tb response (Average Delta Tb) also showed no statistical difference between the sexes at either dose (Figs. 2E and 20 mg/kg capsinoid: unpaired t-test F: n = 8, M: n = 7, p = 0.9306; Figs. 2H and 40 mg/kg capsinoid: unpaired two-tailed t-test, F: n = 8, M: n = 7, p = 0.3699). Lastly, comparison of the capsinoid-induced maximum temperature drop (Max Delta Tb) showed no difference by sex (Figs. 2F and 20 mg/kg capsinoid: unpaired two-tailed t-test, F: n = 8, M: n = 7, p = 0.1998; Figs. 2I and 40 mg/kg capsinoid: unpaired two-tailed t-test, F: n = 8, M: n = 7, p = 0.4229).

Fig. 2
figure 2

Capsinoids induce equivalent dose-dependent hypothermia in young mice of both sexes. a-c: Change in Tb relative to baseline (ΔTb) of young mice following IP administration of 0 mg/kg (vehicle), 20 mg/kg capsinoid, and 40 mg/kg capsinoid with both sexes combined (a), females only (b), and males only (c). d: Comparison of sexes based on ΔTb response to 20 mg/kg capsinoids. e-f: Comparison of ΔTb response between young female and male mice to IP administration of 20 mg/kg capsinoid. Bar graph values represent the average ΔTb recorded over 0–120 min after IP capsinoid administration (e) and max temperature drop after IP capsinoid administration (f). g: Comparison of ΔTb response between young female and male mice to IP administration of 40 mg/kg capsinoid. h-i: Summary of average ΔTb (0–120 min) (h) and max Tb drop (i) after administration of 40 mg/kg capsinoid. (*: p < 0.05, **: p < 0.01, ***: p < 0.001, #: p < 0.0001, all drug administration points indicated by blue dashed lines).

Full size image

After characterizing the capsinoid-mediated hypothermic response in young mice, we next sought to determine the response in aged C57BL6 mice (18–20 m), as a more translational model for stroke or other aging-driven pathologies. We chose a dose of 40 mg/kg based on the capsinoid dosage that promoted a drop into the mild hypothermia range in young mice. The pre-injection baseline Tb for vehicle and capsinoid-treated groups were similar: vehicle: 36.7 ± 0.1 °C, capsinoid: 36.9 ± 0.1 °C (two tailed t test, p = 0.1778, n = 10). In this aged cohort, capsinoids promoted a significant reduction in Tb (Two-way RM ANOVA, p < 0.0001, n = 10 for each group) with a peak change of −3.2 ± 0.2 °C (Fig. 3A). As in the young mice, the vehicle group also showed a small drop in Tb by the end of the two-hour period (−1.0 ± 0.3 °C). When separated by sex, the statistical significance of the Tb decrease in capsinoid-treated mice was maintained (Fig. 3B/C, Two-way RM ANOVA, M: p = 0.0220, F: p = 0.0006, n = 5 for each group). Baseline Tb in sex-specific groupings were not statistically different: female vehicle: 36.6 ± 0.1 °C, female capsinoid: 36.8 ± 0.2 °C, male vehicle: 36.7 ± 0.2 °C, male capsinoid: 37.1 ± 0.2 °C (one way ANOVA p = 0.3714, n = 5). Similar to what was observed in the young mice, there was no significant difference in the Tb response to capsinoids between the sexes (Fig. 3D, Two-way RM ANOVA, p = 0.3622, n = 5 for both sexes) or in the respective characteristics of average delta from baseline Tb (Fig. 3E, unpaired two-tailed t-test, p = 0.3880, n = 5 for both sexes) or maximum delta from baseline Tb (Fig. 3F, unpaired two-tailed t-test, p = 0.1265, n = 5 for both sexes). Comparison of the 40 mg/kg capsinoid response between young and aged mice revealed that the aged cohort responded with a significantly greater drop in Tb compared with the young cohort (Fig. 3G, two-way RM ANOVA, p = 0.0076, n = 15/10).

Fig. 3
figure 3

Capsinoids induce mild hypothermia in aged mice of both sexes, supporting translational relevance to a more typical stroke population. a-c: Measured ΔTb in response to IP capsinoid administration (40 mg/kg) in aged mice with sexes combined (a), grouped by males only (b), and grouped by females only (c). d: Comparison of capsinoid induced hypothermia in aged males compared to aged females. e-f: Summary and comparison of the aged female and aged male responses to IP capsinoid as average ΔTb (0–120 min) (e) and the max drop in Tb (f). g: Comparison of ΔTb response to capsinoid (40 mg/kg) between aged and young mice (both sexes combined). The ΔTb response for young mice is from Fig. 2a. (*: p < 0.05, **: p < 0.01, ***: p < 0.001, #: p < 0.0001, all drug administration points indicated by blue dashed lines).

Full size image

For practical application of hypothermia to ischemic stroke, hypothermia needs to be sustainable for several hours. We therefore determined the potential for successive deliveries of capsinoids to promote prolonged hypothermia. Due to the higher prevalence of stroke in the aged population, we again used aged (18–20 m) C57BL6 mice of both sexes. Based on the single administration time course, we selected 90 min as an injection interval predicted to sustain Tb in the mild-to-moderate hypothermia range (32–34 °C)11. The pre-injection baseline Tb was recorded for 30 min prior to the first administration and averaged. The sex-combined baseline Tb were: 37.3 ± 0.1 °C (vehicle) and 37.4 ± 0.1 °C (capsinoids) (two tailed t-test, p = 0.3567, n = 10). After the initial dose (time 0 min), 3 additional administrations were given at 90 min intervals (90, 180, and 270 min) to provide an average drop of Tb by −3.5 ± 0.9 °C over 6 h (Fig. 4A). This extended hypothermia was statistically significant compared with vehicle treatment (two-way RM ANOVA, p < 0.0001, n = 10 for each group), despite a gradual decline in Tb with vehicle treatment alone (average drop of Tb in vehicle-treated mice as −1.1 ± 0.2 °C over 6 h). Following completion of the injections, Tb of the capsinoid-treated mice gradually and spontaneously returned to vehicle levels (within 3 h) without any supplemental heat support. After separating the responses by sex, baseline Tb were still similar (female vehicle: 37.0 ± 0.1 °C, female capsinoid: 37.3 ± 0.1 °C, male vehicle: 37.4 ± 0.2 °C, male capsinoid: 37.6 ± 0.3 °C) (one way ANOVA, p = 0.2062, n = 5). Additionally, as with the single injections, the significance of the hypothermic response was maintained in sex-specific analyses (Fig. 4B/C, Two-way RM ANOVA, M: p = 0.0012, F: p = 0.0018, n = 5 for each group) and no sex difference was evident by analysis of amplitude or kinetics of the Tb response to repeated capsinoid delivery (Fig. 4D, Two-way RM ANOVA, p = 0.5352; Fig. 4E, average delta Tb from baseline: unpaired two-tailed t-test, p = 0.5367; Fig. 4F, max delta Tb from baseline: unpaired two-tailed t-test, p = 0.6717; n = 5 for both sexes and all analyses).

Fig. 4
figure 4

Repeated administration of capsinoids can promote sustained hypothermia in conscious aged mice. a-c: Measured ΔTb of aged mice to repeated IP administration of capsinoids (40 mg/kg; indicated by blue dashed vertical lines) presented as the combined group of males and females (a), females only (b), and males only (c). d: Sex-specific comparison of ΔTb response to repeated IP administration of capsinoids. e-f: Summary and comparison of the female and male responses to repeated IP capsinoid administration: average ΔTb after administration (0–420 min) (e), and max drop in Tb (f). (*: p < 0.05, **: p < 0.01, ***: p < 0.001, #: p < 0.0001).

Full size image

We speculated that the mechanism of capsinoid-mediated hypothermia was through pharmacological stimulation of TRPV1-containing heat-responsive afferent fibers in the peritoneal cavity, which then triggers whole body heat reduction through centrally driven heat defense mechanisms (e.g. increased cutaneous vasodilation and decreased thermogenesis)11,30,31,32,33. TRPV1 channels are located within the vagal afferents within the deep body core region of the peritoneal cavity and contribute to body temperature control26,30,34,35. Additionally, TRPV1 channels are classically known to be activated by elevated temperature (i.e. ≥ 43 °C)36. Thus, we predicted that pharmacological activation of TRPV1 channels in the peritoneal cavity should mimic the state of elevated core temperature and thus trigger protective heat defense mechanisms to lower body temperature. A major component of the heat defense mechanism is an increase in cutaneous blood flow to dissipate excess heat11. In mice, this increase in cutaneous flow is particularly evident in the tail and thus abrupt increase in tail surface temperature can be used to detect activation of heat defense mechanisms37. We measured tail surface skin temperature (Ttail) by FLIR camera in young WT and TRPV1 KO (TRPV1 del/del) mice of both sexes before and after injection of capsinoids (40 mg/kg, i.p.). The pre-injection Ttail values were well below Tb, reflecting a generally vasoconstricted baseline state consistent with an ambient temperature below that of the mouse thermoneutral zone (TNZ): (WT Female: 24.3 ± 0.1 °C, WT Male: 24.4 ± 0.2 °C, TRVP1del/del Female: 24.0 ± 0.1 °C, TRPV1del/del Male: 24.0 ± 0.1 °C). The TRPV1 KO mice exhibited a 0.3–0.4 °C lower Ttail, consistent with prior studies37 (one way ANOVA, p = 0.0228, n = 4/3). In response to capsinoid injection, there was a marked increase in Ttail of WT mice, beginning within the first measurement at two minutes following injection (Fig. 5A/B). When delivered to TRPV1 del/del mice, the Ttail response was nearly abolished, indicating specificity of the tail blood flow response to TRPV1 channel activation. At 2 min post injection, Ttail of male mice increased an average of 2.7 ± 0.3 °C, compared to 0.7 ± 0.1 °C of their TRPV1 del/del counterparts (Two-way RM ANOVA, p = 0.0158, n = 4/3). Similarly, at 2 min post injection, C57BL6 female mice showed an increase of 3.1 ± 0.6 °C in Ttail compared to 0.4 ± 0.3 °C in TRPV1 del/del mice (Two-way RM ANOVA, p = 0.0076, n = 4/3). The Tb of the mice was also recorded to confirm that the capsinoid administration yielded an effective hypothermia and to allow for a temporal correlation between changes in Ttail and Tb. The pre-injection baseline Tb showed no differences by genotype or sex (WT Female: 37.9 ± 0.1 °C, WT Male: 37.9 ± 0.1 °C, TRPV1del/del Female: 38.2 ± 0.4 °C, TRPV1del/del Male: 37.7 ± 0.1 °C) (one way ANOVA, p = 0.3703, n = 4/3). However, as would be predicted for a mechanism involved in driving the initial heat loss, the increase in Ttail peaked at or before the first tail recording (at 2 min) and preceded the maximal drop in Tb. The WT mice did not demonstrate a statistical drop in Tb (compared to TRPV1 del/del response) until 4 to 8 min after capsinoid injection (Fig. 5C, Two-way RM ANOVA, p = 0.0004, n = 4/3). It is noteworthy that the Tb continued to drop even after the Ttail returned to baseline. Together, these data suggest that IP delivery of capsinoids induces a rapid initial drop in Tb through cutaneous heat loss, with subsequent Tb reduction through other components of heat defense.

Fig. 5
figure 5

Capsinoids activate heat defense mechanisms prior to drop in core body temperature. a: Representative FLIR images captured 2 min before and 2 min after IP capsinoid administration (40 mg/kg) in C57BL6 and TRPV1 del/del mice. b-c: Summary of change in infrared Tail temperature (ΔTtail) (b) and ΔTb (c) in C57BL6 and TRPV1 del/del of both sexes. (*: p < 0.05, **: p < 0.01, all administrations indicated by blue dashed lines; sex of the wild type (WT) and TRPV1 del/del mice are indicated by M and F).

Full size image

The vagus nerve has TRPV1 containing sensory neurons/fibers and projects to thermoregulatory regions of the brain33,38,39. We therefore sought to determine if signaling via peritoneal capsinoids is conveyed to the brain by vagal pathways. To test the involvement of vagal signaling, we administered capsinoids or vehicle by IP injection 4 times at 30-minute interval to awake adult male (7–8 m) C57BL6 mice. Two hours after the first injection, the mice were euthanized and the nodose ganglia were extracted. The nodose ganglion contains the cell bodies of the visceral vagal afferents conveying visceral sensory signals heading to the brain. To determine if these neurons were activated by capsinoid administration, each ganglion was sectioned (5 sections at 50 μm interval) to encompass the majority of the ganglionic volume. Sections were stained with NeuN (neurons), DAPI (all nuclei), and c-Fos, an immediate early gene used to indicate neuronal activity (Fig. 6A-C). In capsinoid-treated mice, there was a 3.1-fold increase in c-Fos positive neurons (Fig. 6C1,2) compared with vehicle-treated mice (Fig. 6C8) (Capsinoid: 2.41%, Vehicle: 0.77% c-Fos positive neurons; two-tailed Welch’s t-test, p = 0.0079, n = 5 for both conditions) (Fig. 6D). For these analyses, c-Fos was only counted if it was located in a neuron (NeuN) and in the nucleus (DAPI). Note that some non-neuronal c-Fos positive cells were evident in both the capsinoid and vehicle treated mice but were not counted in this analysis (Fig. 6C3-7).

Fig. 6
figure 6

Capsinoid administration increases neuronal activation in the nodose ganglion. a-b: Nodose ganglia from mice given IP administration of capsinoids (a) or vehicle (b). c: Increased magnification of the numbered circles in a and b. Yellow circles represent example c-Fos positive neurons, red circles represent example c-fos positive non-neurons, white circle represents example c-fos negative neurons. d: Summary and statistical comparison of total c-fos positive neurons between capsinoid- and vehicle-administered mice. (**: p < 0.01).

Full size image

As a final proof-of-principle experiment, we sought to determine if capsinoid-induced hypothermia could be neuroprotective in an experimental stroke model. Young (11w) C57BL6 mice of both sexes underwent permanent distal middle cerebral artery occlusion (pdMCAO)40,41,42. Mice then recovered in a warming cage for 2 h to ensure that they returned to normothermia before treatment and to increase translational relevance by delaying the start of treatment. At the 2-hour post-stroke point, pre-injection baseline Tb were similar between groups (vehicle: 37.1 ± 0.1 °C, capsinoid: 37.2 ± 0.2 °C) (two-tailed t-test, p = 0.3117, n = 6/7). At this 2-hour post-stroke point, injection of capsinoids (40 mg/kg) or vehicle was initiated at 90 min intervals over 6 h. Capsinoid treated mice demonstrated an average Tb drop of −2.5 ± 0.1 °C that was significantly greater than the vehicle treated mice (Fig. 7A, Two-way RM ANOVA, p < 0.0001, n = 6/7). At 3 days post-stroke, brains were evaluated for infarct volume by TTC staining. Mice that underwent capsinoid-induced hypothermia demonstrated a significant reduction of infarct volume (> 60% reduction) (Fig. 7B/C, vehicle: 16.73 ± 2.65mm3, capsinoid: 6.65 ± 0.91mm3, two-tailed Welch’s t-test, p = 0.0108, n = 6/7).

Fig. 7
figure 7

Capsinoid-induced hypothermia is neuroprotective when initiated two hours after ischemic stroke. a: ΔTb of combined young male and female mice after pdMCAO stroke and subsequent treatment with capsinoids or vehicle control (repeated administration represented by blue dashed vertical lines). Time 0 reflects the initial of capsinoid or vehicle injections at two hours after induction of stroke. b: Representative TTC stained images of male and female mice 3 days after pdMCAO stroke comparing capsinoid and vehicle treatments. c: Summary and statistical comparison of total infarct volume at 3 days post pdMCAO measured by TTC quantification from six 1 mm thick brain sections in male (blue circles) and female mice (red circles).

Full size image

Discussion

Therapeutic hypothermia has shown significant benefit at improving short- and long-term outcomes in the clinical setting for various ischemic injuries including cardiac arrest and nHIE43,44,45,46,47,48,49,50,51,52,53,54,55,56,57. Unfortunately, this therapeutic benefit has not yet translated to application in adult stroke patients12,13,14,15. One potential reason for the lack proven benefit in clinical stroke relates to practical limitations of current hypothermia protocols to induce hypothermia in the patients. Current means of inducing hypothermia include external and intravenous cooling approaches, which are very effective in cooling unconscious patients, but require heavy sedatives or paralytics to prevent violent shivering58,59,60,61. Unfortunately, in the conscious patient, these approaches can often provide inconsistent depth of hypothermia, unsuppressed shivering, and long times to reach hypothermia12,13,14,15. Pharmacological approaches to induce hypothermia that also provide simultaneous suppression of cold defense mechanisms (e.g. shivering, cutaneous vasoconstriction) would offer new and promising options to promote mild hypothermia in the conscious stroke patient.

The present study explored the potential of capsinoids, non-pungent TRPV1 agonists, to promote mild hypothermia in conscious mice and their application to provide neuroprotection following stroke. We demonstrate that 1) capsinoid delivery promotes consistent hypothermia into the mild/moderate range, 2) that capsinoid-dependent hypothermia is mediated selectively through TRPV1 channel activation, 3) that capsinoid administration can provide prolonged hypothermia in aged mice of both sexes, and 4) that activation TRPV1 channels within the peritoneal cavity signals in part via the vagus nerve and triggers rapid activation of heat defense mechanisms. We further demonstrate proof-of-principle that capsinoid-mediated hypothermia can provide neuroprotection in an experimental stroke model. We elaborate on these findings below.

Capsinoid delivery promotes consistent body cooling to the level of mild/moderate hypothermia

TRPV1 has long been speculated to be a thermosensor as, in vitro, TRPV1 containing neurons are activated in noxious heat environments36. However, the temperature required to activate the channel (> 43 °C) is very rarely seen even in febrile subjects; as such, it is unlikely that TRPV1 channels act as a thermosensor within the range of normal body temperature26,32. However, delivery of TRPV1 agonists does promote hypothermic effects, indicating that activation of TRPV1 dependent pathways can modulate thermoregulatory pathways11,18,19. Among these, delivery of capsaicin and related capsaicinoids have already been shown to promote decreases in core body temperature. We now show that capsinoids, delivered to the peritoneal cavity, also provide TRPV1-dependent hypothermia in conscious mice. We believe this is the first report demonstrating that ‘non-pungent’ capsinoids can be used to promote pharmacological hypothermia.

Molecular properties of capsinoids offer distinct opportunities for regionally targeting TRPV1 activation

Unlike with capsaicinoids, capsinoid-induced hypothermia is highly dependent on the administration route, as the hypothermic response is only observed by IP injection and not by OG or SC delivery. The high dependence of the effect on delivery method is likely due to the rapid breakdown of the capsinoids into vanillyl alcohol and fatty acids by esterases23,24,25,62,63. Since esterases are ubiquitously expressed throughout the body, the enzymatic breakdown of this TRPV1 agonist leads to its site of action being effectively limited to the site of delivery. Previous studies show that after oral delivery of capsinoids, plasma concentration of capsinoids or their metabolites were too low for quantification23,24. Our data further showed no significant effect on Tb with oral gavage of high dose capsinoids. Additionally, our data revealed the inability of capsinoids to be absorbed and distributed via the blood, even when delivered directly to the subcutaneous space. Subcutaneous delivery of capsinoids failed to elicit a hypothermic response whereas delivery of DHC to the same site elicited a temperature drop. The less labile properties of capsaicinoids results in their ability to be distributed throughout the body, being absorbed into the blood following oral, intraperitoneal, or subcutaneous delivery64,65,66,67. Indeed, oral gavage of capsaicin can promote a TRPV1-dependent hypothermic response68. The more labile property of capsinoids (versus capsaicinoids) provides a fortuitous means of restricting action of these TRPV1 agonists to the region of delivery and thus reducing unwanted distal on-target or off-target effects. This limited capacity for uptake and biodistribution of capsinoids was exploited in the current study, as it allowed for targeting TRPV1 channels in the peritoneal cavity without the potential activation of distant TRPV1 targets.

Capsinoid-dependent hypothermia is mediated selectively through TRPV1 channel activation

We used TRPV1 KO mice to determine the selectivity of capsinoid-mediated hypothermia. The induction of heat defense mechanisms (evident by increase of tail surface temperature) and the resultant drop in Tb were both significantly attenuated in TRPV1 KO mice. Furthermore, due to the limited potential for capsinoids to distribute to other body regions, the TRPV1 channels involved are likely located in the peritoneal cavity. Vagal afferents are known to express TRPV1 channels39, and therefore represent a presumed site of action for the capsinoids.

Capsinoid administration can provide prolonged hypothermia in aged mice of both sexes

As stroke disproportionately affects older individuals, we sought to determine if capsinoid-induced hypothermia was effective in aged mice of both sexes. Our studies found that capsinoids promoted hypothermia effectively in aged mice and without significant difference by sex. The resulting hypothermia was able to be maintained for several hours with repeated capsinoid administration, demonstrating the potential to provide prolonged and reversible body cooling. Interestingly, compared to young mice, the aged mice exhibited a more significant hypothermic response to 40 mg/kg capsinoids. The cause of this difference was not determined; however, it could be due to the greater amount injected into the aged mice. Since dosing was based on body weight, the aged mice received a larger amount of capsinoids into the peritoneal cavity. Given the limited capacity of the capsinoids to be taken into the blood and distribute throughout the body, this dosing strategy might result in higher capsinoid concentration in the peritoneal cavity of aged mice. Alternatively, the greater response in aged mice could be a result of age having an as-yet unknown impact on the activation of TRPV1 by capsinoids, as has been shown for other activators such as LPS69. Regardless, these studies demonstrated the potential for capsinoid-mediated hypothermia in aged mice and support the translational potential of this pharmacological intervention to induce hypothermia in pathologies that disproportionately affect aged subjects, such as ischemic stroke.

Activating TRPV1 channels within the peritoneal cavity triggers heat defense mechanisms

We provide evidence that capsinoids delivered to the peritoneal space induces mild hypothermia by modulating endogenous thermoeffector mechanisms through TRPV1 receptor stimulation. We further speculate that the resulting hypothermia may be partially mediated through TRPV1-induced activation of vagal afferent pathways similar to what has been demonstrated by Mohammed et al. in the context of inhibiting thermogenesis triggered by skin cooling33. Our proposed model is further supported by other studies showing that pharmacological hypothermia induced by IP administration can be abolished by specifically desensitizing sensory abdominal nerves37. Our studies showed that capsinoid delivery can rapidly trigger mechanisms used in heat defense, as tail temperature increased within two minutes of capsinoid delivery. Increased Ttail was used as an indicator of heat defense activation, as mice regulate blood flow to the tail as a means of thermoregulation37,70. In times of heat stress, mice increase blood flow to the tail to dissipate body heat across the skin surface (similar to flushing response in humans)11. The increase in Ttail preceded the drop in Tb, as would be expected for an early component of heat loss. We additionally found that the percentage of c-Fos positive neurons in the nodose ganglion of the vagus nerve is three-fold higher after capsinoid administration, indicating an involvement of vagal activation. Approximately 40% of mouse nodose neurons express TRPV1, and while the vast majority of them are also afferents, only a subset would specifically receive sensory signals from the region of capsinoid delivery (i.e. gut/viscera)39. Our finding that 2.4% of total nodose neurons were activated is likely a reflection of our selective activation of one branch of the vagus.

As alluded to earlier, TRPV1 receptors are also expressed in temperature regulatory regions of the brain and their local activation can promote mild systemic hypothermia30,31,33,71. While this central action on TRPV1 channels may contribute to the hypothermic effect of peritoneal capsaicin and DHC, capsinoids lack similar biodistribution capacity and are unlikely to reach central TRPV1 channels23,24,72. Thus, the hypothermic effect from IP capsinoids is more likely solely a reflection of TRPV1 channels activated within the peritoneal space.

Capsinoid-administration promotes activation of heat defense mechanisms and suppression of cold defense mechanisms

We propose that capsinoid administration promotes mild hypothermia through the pharmacological activation of the body’s heat defense mechanisms, which provides additional translational benefit of suppressing the body’s aggressive response to cooling (i.e. cold defense mechanisms). Current best practices for inducing therapeutic hypothermia involve either forced external cooling or intravascular cooling73,74,75. While recent advances in intravascular cooling can provide more rapid cooling, both forced cooling approaches require heavy sedation or paralytics in order to avoid complications from violent shivering and other consequences of triggering cold defense mechanisms. Note that heat defense activation also promotes suppression of cold defense mechanisms and thus pharmacological activation of heat defense mechanisms (without actually increasing heat) might also be an effective strategy to suppress shivering and other responses to forced cooling19. With respect to the overall effect of capsinoid administration on the drop in Tb, our data suggests that multiple thermomodulatory components contribute at different points in the time course of cooling. The increase in tail temperature occurs abruptly after capsinoid administration, consistent with the initial drop in Tb resulting from heat dissipation via cutaneous vasodilation. However, the duration of elevated Ttail is relatively brief (~ 10 min) compared to the falling phase in Tb (> 20 min). These results are consistent with capsinoid treatment engaging additional TRPV1-dependent thermoeffector mechanisms, such as reduced BAT thermogenesis, that can augment and sustain the hypothermic response33.

Capsinoid-induced hypothermia is neuroprotective in an experimental model of ischemic stroke

Stroke patients are typically conscious at the time of treatment and there are many benefits to maintaining their conscious state throughout treatment (e.g. monitoring neurofunction, evaluating therapeutic intervention). Current options to induce hypothermia in this patient population are limited, largely related to practical issues related to poor suppression of cold defense mechanisms or additional clinical risks with intravascular cooling devices12,13,14,15. We reason that induction of hypothermia that is driven by the patient’s thermoregulatory system (i.e. capsinoid-induced hypothermia) should provide a drop in temperature that is both more rapid and less physiologically stressful due to the absence of cold defense activation. As an initial step to test this idea, we provide proof-of-principle that capsinoid-induced hypothermia can be neuroprotective in an experimental model of stroke. The pdMCAO model was selected for its consistent infarct volume and low mortality rate in aged mice. Capsinoid-mediated hypothermia was initiated two hours after induction of stroke and maintained for 6 h. At 72 h, capsinoid/hypothermic mice demonstrated a > 60% reduction in infarct volume compared with vehicle/normothermic mice. These findings provide support for feasibility and effectiveness of this strategy of providing therapeutic hypothermia in an experimental stroke model.

Study limitations and future directions

In some instances, we used TRPV1 del/del mice to determine the specificity of capsinoid actions. It should be noted that certain studies have shown TRPV1 KO mice to be hypometabolic, have slightly lower tail temperature (~−0.5 °C), have a higher thermoneutral zone (TNZ) range (+ 1 to 1.5 °C), lower ambient temperature preference, and higher activity compared with WT mice37. Our studies also showed a slightly lower Ttail (−0.4 °C), although we did not measure metabolic activity, physical activity, or TNZ. However, our conclusions from the use of the TRPV1 del/del mice demonstrated the near abolishment of the measured response compared to WT mice, suggesting that the slight differences described above do not account for the response.

Another consideration for the study is that our temperature probes were implanted behind the right scapula, a region known for its proximity to a brown adipose tissue (BAT) site. We considered the measured temperature as a surrogate for deep body temperature (Tb) based on our previous finding that temperature from probes in this location correlated well with a simultaneously implanted intraperitoneal temperature probe76. Nevertheless, due to the location of our temperature probes near BAT, our temperature measurements could have been influenced by changes in BAT thermogenesis33. Regarding this point, mice in our study experienced an ambient temperature of 23 °C, a temperature about 2–6 °C below the lower end of the TNZ range typically measured for laboratory mice37,70. Thus, mice would be considered slightly cold-stressed and relying partially on BAT thermogenesis at the time of experiments. Cold-activated BAT thermogenesis in rats has been shown to be suppressed by TRPV1-dependent pathways originating from both central (nucleus tractus solitarius) and peripheral locations (sub-diaphragmatic vagus)33. To evaluate the potential influence of local BAT temperature on our presumed Tb measurement, we simultaneously implanted subcutaneous temperature probes in both the interscapular region and at the lateral abdominal region. Baseline temperature was 0.4 °C higher at the interscapular probe (consistent with some baseline BAT thermogenesis), however, the measured drop in temperature was nearly identical from both measurement locations (supplemental figure). Thus, our measured Tb was not a simple reflection of BAT region temperature, but rather a measure of body temperature reflective of a balance of all thermogenesis and heat loss mechanisms.

Our studies identified a subpopulation of activated neurons in the nodose ganglion following capsinoid administration. From this finding and other studies linking central and vagal TRPV1 signaling to regulation of thermoeffectors33, we concluded that intraperitoneal capsinoid administration signals in part through the vagus. However, our studies did not co-stain for TRPV1 to determine if the c-Fos positive neurons were additionally positive for TRPV1. Future studies would be required to confirm co-localization of c-Fos with TRPV1 mRNA or protein. It should also be noted that, while we found a role for signaling via the nodose ganglion, spinal nerves also contribute TRPV1-containing afferents to the abdominal viscera77,78,79, which could serve as additional capsinoid-responsive pathways. TRPV1 is also not exclusively activated by heat but can also be activated by protons and other agonists (besides capsaicinoids and capsinoids) or modulators26,31,32,80,81,82. As such, these other activators can affect the response of subjects to capsinoids administration and should be explored in future studies.

With regard to our stroke studies, future studies will be required to define the long-term benefits of capsinoid-mediated hypothermia and determine the effectiveness in aged mice. Of additional translational consideration is that mice have a greater surface area-to-body mass ratio than humans and therefore a higher metabolism and more rapid heat loss70. These biological characteristics would assuredly change the characteristics of capsinoid-induced hypothermia in humans, and thus additional studies would be warranted to better define the timing and depth of the hypothermic response in larger mammals.

Conclusion

In summary, we have shown that capsinoids administered to the peritoneal cavity provides a promising means of pharmacologically inducing mild hypothermia in young and aged conscious mice of both sexes. Based on our observations and existing literature, it is likely that in our model, capsinoid-induced temperature change is dependent on peritoneal TRPV1 and induces components of the heat defense mechanism to elicit a drop in Tb within the mild hypothermia range. Importantly, the hypothermic effect could be maintained for multiple hours, supporting the translational potential in the context of ischemic stroke. Together, these data highlight the potential for capsinoids to be used to regionally target TRPV1 channels in the peritoneal space and thereby induce a neuroprotective therapeutic hypothermia to reduce brain injury following ischemic stroke.

Materials and methods

Animals. Experiments were performed with male and female mice of two age groups. “Young” adult mice were 10–16 weeks of age. “Aged” mice were 18–20 months of age. Young mice (C57BL/6) were obtained directly from Jackson Labs or bred in house from Jackson Labs lineage mice. The TRPV1 deletion (TRPV1 del/del) or TRPV1 floxed only controls (TRPV1 fl/fl) were generated through the BCM mouse core and crossed onto a C57BL/6 background. The TRPV1 deletion mice were generated during the CRISPR-Cas9 based creation of the TRPV1 fl/fl mice due to occasional DNA sequence deletion between targeted loxP sites. Aged mice (C57BL/6) were obtained from the NIA aging colony and housed for at least one month locally before experimental use. Euthanasia for all experiments was carried out in the following manner: mice were deeply anesthetized with 2.5% avertin (1% by body weight, IP) and then perfused transcardially with heparinized PBS (10 U/mL) followed by 4% paraformaldehyde (PFA). All animal studies were approved by the Institutional Animal Care and Use Committee at the University of Texas Health Science Center at Houston.

Temperature Probe Implant and Recording. Two days prior to any planned experiments, mice were very briefly anesthetized with isoflurane (3–4% induction in 70%/30% oxygen/nitrogen) to implant temperature probes. Miniature wireless Implantable Programmable Temperature Transponder (IPTT-300, BioMedic Data Systems, Seaford, DE) were implanted beneath the skin behind the right scapula. This location of transponder placement agreed well with simultaneously implanted subcutaneous transponders at the lateral abdomen (see Supplemental Figure) and earlier studies comparing the interscapular placement with intraperitoneal placement76. Temperatures were read by a wireless reader and then logged. All recordings began with at least 30 min of baseline prior to first administration and recorded every 10 min, with the exception of the tail temperature experiment where a shorter interval was needed. In this experiment, baseline was recorded for 10 min prior and every 2 min until the conclusion of the experiment. All temperature transponders were given a unique, non-group identifying moniker to allow for unbiased blinded data analysis.

Ambient Temperature Maintenance. Throughout all experiments, mice were housed in their home cages in a temperature-controlled room. The internal cage temperature (ambient temperature) was measured in a sample cohort across 8 h and on 3 different days and was found to average to 23.1 ± 0.05 °C.

Capsinoid Extraction. Capsinoids were isolated from Capsicum species fruit by first using a non-polar organic solvent (pentane), second liquid/liquid partitioning using acetonitrile (ACN)/pentane, and lastly a step gradient C-4 purification using H2O/ACN as the mobile phase83. This method yields a > 97% purity ACN fraction consisting of capsiate (56%) and dihydrocapsiate (41%) measured by high performance liquid chromatography (HPLC).

Intraperitoneal Injection (IP). Vehicle (corn oil), DHC (Cayman, Ann Arbor, MI), or capsinoids, were injected into the intraperitoneal cavity of mice at a maximum volume of 1% of the mouse’s body weight. The injections were administered at concentrations of 20 mg/kg and 40 mg/kg. For repeated injections, the side of injection was swapped with each administration.

Oral Gavage (OG). Two days before planned experiments, temperature probes were implanted as described above. The capsinoid-enriched fraction (“capsinoids”) was suspended in corn oil to be administered at a concentration of 100 mg/kg and then delivered through a stainless-steel gavage needle.

Subcutaneous Injection (SC). Adult (48w) male C57BL6 mice were given injections under the skin behind the left scapula with vehicle (corn oil) or 40 mg/kg capsinoids suspended in corn oil at a maximum volume of 1% of body weight. Temperatures were measured every 10 min starting 30 min before injection and continued for 2 h post-injection. Three days later, the mice that were given vehicle were readministered vehicle and the mice that were previously administered capsinoids were administered 2 mg/kg DHC suspended in corn oil at a maximum volume of 1% of body weight.

Tail Temperature Measurement. Two days prior to planned experiments, temperature probes were implanted as described above in 16-week-old C57BL6 and TRPV1 del/del mice of both sexes to record core temperatures in conjunction with tail temperature measures. Tail temperatures were measured by first capturing forward looking infrared (FLIR) images with the TC004 thermal imager (TOPDON) and saved to a memory card. The images were then downloaded and the tail temperatures determined with the associated TDView program. Starting 10 min before capsinoid administration, baseline measurements for both core temperature and tail temperature were taken at 2-minute intervals. Next, capsinoids were injected (40 mg/kg, i.p.) and core and tail temperatures were recorded for an additional 20 min at 2-minute intervals.

Nodose Extraction and Staining. Harvesting of nodose ganglia was performed similarly to the procedures detailed by Han and Araujo84. Mice were first perfused with 20mL of phosphate buffered saline (PBS) mixed with heparin and subsequently with 10mL of formalin. The entire head was then removed and immersion fixed in 50mL formalin overnight. The next day, the muscle, skin, skull cap, and brain were removed, and the remaining skull was suspended in 30% sucrose overnight. The nodose ganglion was extracted and placed into optimal cutting temperature (OCT) medium and flash frozen over dry ice before being stored at −80 C until slicing by cryostat. The nodose ganglia were sliced at 10 μm thickness and then stored at −20 C. For immunofluorescence, 5 sections at 50 μm interval were selected per mouse. Sections were first washed with PBS with 2% triton-x (PBST) to remove OCT. The slides were then run through a permeabilization step of 45 min of PBS with 5% triton-x followed by a citric acid antigen retrieval for 10 min at 65 C and allowed to return to room temperature. Blocking buffer (3% normal donkey serum in PBST) was applied for one hour followed by primary antibody (1:4000 c-Fos: Synaptic Systems, 1:500 preconjugated NeuN) in blocking buffer for 3 days. Slides were then washed with PBST before introducing secondary antibody (AlexaFluoro 546) in blocking buffer for 90 min. Another series of washes, followed by DAPI (1:5000 in PBST) and a final set of washes were performed before mounting with Diamond ProLong Antifade mountant. Slides were imaged with a 20x objective on a Leica DMi8 microscope and the LASX software by a second investigator. To maintain blinding, samples were given nonidentifying image names (e.g. Image 1). The total number of cells that were positive for NeuN and showed c-Fos and DAPI colocalization were counted as well as the total number of NeuN + neurons and recorded before the key was revealed and comparisons made.

Permanent Distal Middle Cerebral Artery Occlusion (pdMCAO) Model and Infarct Volume Quantification. Young (11w) male and female C57BL6 mice had temperature transponders implanted 2 days prior to stroke. On the day of surgery, female mice were administered vaginal smears to determine estrus cycle (4 out 6 were determined to be in estrus stage at time of surgery, 2 out of 6 were determined to be in the proestrus). All mice were then briefly anesthetized with isoflurane (4% induction and 1.75-2% maintenance). The body temperature was maintained at 37 °C with a rectal probe and a feedback-controlled heating pad. Using a micro-drill, a small hole was generated in the skull just above the distal middle cerebral artery (MCA). A low temperature cautery was used to permanently ligate the MCA. During recovery, the mice were placed in a custom-built warming chamber76 for two hours to maintain normothermic body temperature. Over the last 30 min of recovery, the core body temperature of the mice was measured by the implanted temperature probe and registered as baseline. The mice were then removed from the recovery chambers and administered 40 mg/kg capsinoids or vehicle control at two hours after stroke induction (time 0) and every 90 min thereafter to provide 6 h of hypothermia or normothermia. During this time, temperatures were recorded every 10 min. Mice were monitored daily for 3 days, after which a lethal injection of avertin was administered and the mice were perfused with PBS mixed with heparin. The brains were then removed and sectioned into 1 mm slices, then immersed in 3% 2,3,5-triphenyltetrazolium chloride (TTC) solution in PBS at 37 °C for 5 min, at which point they were flipped and stained for another 5 min. TTC is converted into a red compound by active mitochondria, thereby staining metabolically active tissue red and leaving infarcted tissue white. The slices were then fixed in 4% paraformaldehyde and quantified using ImageJ software (National Institutes of Health, Bethesda, MD).

Data Presentation and Statistical Analysis. All data is presented as the mean ± standard error of measurement (SEM). All temperature data is represented as absolute temperature at baseline and then as Delta Temperature (ΔTb °C), the difference between every datum and the baseline temperature measurement’s average. The data was presented as delta temperature to limit potential complications and variability due to transponder-to-transponder differences as well as potential effects of minor differences in implantation sites. For comparative data, Two-way repeated measures ANOVA was used for comparing the breadth of the temperature responses, and, where appropriate, individual time point differences. For all ANOVA tests, only the experimental condition main effects were presented, without consideration for time effects. For average temperatures after administration the complete average across the entire measurement timeframe was calculated for each subject. For sex differences, unpaired two-tailed t-tests were used. For any grouped data that had unequal variance, the data was analyzed for statistical significance by Welch’s t-test. This study is performed in accordance with relevant guidelines and regulations. All methods are reported in accordance with ARRIVE guidelines.