Role of TRPM8 in switching between fever and hypothermia in adult mice during endotoxin- induced inflammation


 Transient receptor potential melastatin 8 (TRPM8) functions in the sensing of noxious and innocuous colds; however, its significance in pathogen-induced thermoregulation and inflammation remains unclear. In the present study, we investigated the role of TRPM8 in the regulation of endotoxin-induced body temperature control and inflammation. The peripheral administration of low-dose LPS or zymosan generated fever in wild-type (WT) mice and hypothermia in TRPM8 knockout (KO) animals. TRPM8 KO mice exhibited severe hypothermia and sickness responses following the peripheral administration of high-dose LPS. An intracerebroventricular injection of LPS and interleukin-1ß (Il-1ß) elicited hypothermia in TRPM8 KO mice, in contrast to fever in WT animals, whereas that of prostaglandin E2 induced normal fever. Fos immunohistochemistry showed the stronger activation of hypothalamic thermoregulation-associated nuclei following the peripheral administration of low-dose LPS. Therefore, TRPM8 is necessary for switching between fever and hypothermia during endotoxin-induced inflammation.


Introduction
The recognition of pathogen infections is one of the most important mechanisms of innate immunity for eliminating pathogens and the first line of host defense against infection 1 . Several families of pattern recognition receptors are responsible for detecting the pathogen-associated molecular patterns (PAMPs) of many pathogens 2 . Toll-like receptors (TLRs) are the most important and widely studied pattern recognition receptors 2,3 . The infection-induced activation of immune cells via TLRs leads to a broad spectrum of sickness symptoms, such as body temperature changes, nausea, decreased appetite, malaise, and fatigue, which are adaptive responses to promote animal survival 4 .
Fever is generated by integrated physiological and neuronal responses that confer a survival benefit against infection. The inhibition of fever with antipyretic drugs was previously reported to increase mortality in humans infected with influenza virus 5 . Fever enhances immunoprotective mechanisms in both the innate and adaptive immune systems, such as the acceleration of cytokine production and the cellular cytotoxic activity of dendritic cells, natural killer cells, macrophages as well as signaling by and the differentiation of T cells 6 . Fever also results in an increased metabolic rate, with only a 1°C increase in body temperature elevating the metabolic rate by approximately 10% 7 .
Hypothermia occurs in the most severe cases of systemic inflammatory syndromes instead of fever and is regarded as an adaptive host strategy to attenuate the harmful effects caused by the overshoot of inflammatory responses 8 . Hypothermia suppresses endotoxin increases, abdominal organ dysfunction, hypotension, and mortality in animals injected with Escherichia coli 9 . More specifically, the intraperitoneal administration of the TLR4 agonist, lipopolysaccharide (LPS), was shown to induce fever or hypothermia in a dose-dependent manner: a low dose of 50~100 μg/kg induced fever, whereas a high dose of more than 2.5 mg/kg caused hypothermia 10, 11 . The intraperitoneal administration of a low dose of the TLR2 agonist zymosan of 1~10 mg/kg induced fever 12 , whereas a high dose of more than 50 mg/kg caused Scientific Reports 4 hypothermia 13 . Therefore, the two thermoregulatory responses are complementary strategies of survival in infection-induced inflammation; however, limited information is currently available on the switching mechanism between fever and hypothermia.
Prostaglandin E2 (PGE2), a lipid mediator, plays a key role in eliciting fever or hypothermia by acting on prostaglandin EP receptors in the preoptic area (POA) of the hypothalamus 11 .
Recent studies reported that the medial and lateral POA is an important thermoregulatory center in the brain [14][15][16] . The specialized population of these POA neurons elicits a long-lasting hypothermic and low metabolic state by acting on the dorsomedial hypothalamic nucleus (DMH) [14][15][16] . The organum vasculosum of the lamina terminalis (OVLT) is one of the circumventricular organs (CVOs), which are characterized by fenestrate capillaries or the lack of the blood-brain barrier (BBB) 17 . The OVLT has been reported to express receptors for PAMPs and cytokines 18,19 and is located adjacent to the medial and lateral POA. Moreover, TLR2 and TLR4 are expressed at microglia/macrophages and astrocytes/tanycytes, respectively, in the OVLT [20][21][22] . The selective deletion of cyclooxygenase-2 in brain endothelial cells has been shown to attenuate the fever response 23 . Interleukin-1β (IL-1β) signaling in blood vessels, particularly fenestrated capillaries in the CVOs, acts as a critical signaling relay between the immune and nervous systems to cause sickness responses 24 .
In the present study, we aimed to elucidate whether transient receptor potential melastatin 8 (TRPM8) is involved in endotoxin-induced body temperature regulation and inflammation.
TRPM8, a cold receptor for menthol, is activated by innocuous cooling to less than approximately 26℃ 25 . TRPM8 is necessary for sensing skin cooling and, hence, TRPM8 knockout (KO) mice are deficient for noxious and innocuous cold-evoked behaviors 26,27 .
TRPM8 may respond to alterations in temperature decreases regardless of whether the temperature is above the activation threshold 28,29 . In contrast to peripheral functions, the significance of central TRPM8 in endotoxin-induced body temperature regulation and inflammation remains completely unknown; however, a recent study showed that TRPM8-expressing neurons were present in the POA and lateral septal nucleus (LS) of the mouse brain 30 .
In the present study, we found that hypothermia was induced in TRPM8 KO mice, in contrast to fever in wild-type (WT) animals, by the peripheral administration of low-dose LPS and zymosan. A similar hypothermic response was observed in TRPM8 KO mice following the intracerebroventricular (i.c.v.) injection of LPS and IL-1ß despite normal fever generation with the i.c.v. injection of PGE2. Moreover, the peripheral administration of high-dose LPS induced prolonged severe hypothermia in TRPM8 KO mice. The number of Fos-expressing neurons in hypothalamic thermoregulation-associated regions was significantly higher in TRPM8 KO mice than in WT animals after the peripheral administration of low-dose LPS. These results indicate that TRPM8 plays a crucial function in switching between fever and hypothermia during endotoxin-induced inflammation.

Peripheral administration of low-dose LPS induced hypothermia in TRPM8 KO mice.
LPS, the most potent TLR4 agonist, is used as in vitro and in vivo experimental neuroinflammation models of bacterial infection. The intraperitoneal administration of a low dose of LPS of 50~100 μg/kg has been shown to generate fever in mice, whereas a high dose of more than 2.5 mg/kg induces hypothermia 10 . To elucidate whether TRPM8 KO mice exhibit LPS-induced fever, the body temperature of WT and TRPM8 KO mice was measured after the intraperitoneal administration of low-dose (50 μg/kg) LPS. A two-way repeated measures ANOVA showed a significant effect of group (F3,243 = 20.84, p < 0.001) and time (F7,1944 = 4.56, p < 0.001), but not their interaction (F21,1940 = 0.90, p > 0.05) on changes in core body temperature (Fig. 1a). In comparisons with mice administered saline, the body temperature of WT mice significantly increased and peaked (1.03 ± 0.35 °C) 40 min after the administration of LPS, whereas that of TRPM8 KO mice markedly decreased with a nadir (-2.12 ± 0.53 °C) at Scientific Reports 6 120 min. Mice in all experiments exhibited an initial stress-induced fever as a result of handling during the administration procedure, regardless of whether the endotoxin, cytokine, or saline was administered. The temperature index was significantly higher in WT mice (5.89 ± 1.62 Δ o C×hr) after the administration of LPS than in those administered saline (-1.18 ± 1.51), whereas it was significantly lower in TRPM8 KO animals (-6.33 ± 3.26) than in those administered saline (1.15 ± 1.74) (Fig. 1b). The one-way ANOVA showed no significant changes in the cumulative locomotor activity of WT and TRPM8 KO mice following the administration of LPS ( Fig. 1c and Fig. S1). Therefore, TRPM8 KO mice developed hypothermia, in contrast to fever in WT mice, in response to the peripheral administration of low-dose LPS.
TRPM8 KO mice developed prolonged severe hypothermia after the peripheral administration of high-dose LPS. To elucidate whether TRPM8 is involved in the regulation of LPS-induced hypothermia, the body temperature of WT and TRPM8 KO mice was examined after the intraperitoneal administration of high-dose (5 mg/kg) LPS. A two-way repeated measures ANOVA revealed a significant effect of group (F3,600 = 488.27, p < 0.001), time (F23,13800 = 9.33, p < 0.001), and their interaction (F69,13800 = 5.59, p < 0.001) on changes in core body temperature (Fig. 2a). Body temperature was significantly lower in WT mice with a nadir (-3.47 ± 0.79 °C) at 365 min after the LPS administration than in those administered saline.
Body temperature was also significantly lower in TRPM8 KO mice with a nadir (-7.04 ± 0.44 °C) at 565 min after the LPS administration than in those administered saline; however, LPS-induced reductions in body temperature were more prominent in TRPM8 KO mice than in WT animals. Following the administration of LPS, the temperature index was markedly lower in WT (-63.24 ± 9.43 Δ o C×hr) and TRPM8 KO (-105.37 ± 7.01) mice than in their controls (WT, 8.17 ± 2.75; TRPM8 KO, 1.07 ± 5.96) (Fig. 2b). However, the temperature index was markedly lower in LPS-treated TRPM8 KO mice than in WT mice. Locomotor activity Scientific Reports 7 during daytime and nighttime was significantly lower in LPS-treated TRPM8 KO mice than in those administered saline (Fig. 2c,d and Fig. S2). Furthermore, WT mice had slightly lower locomotor activity. These results indicate that TRPM8 KO mice developed severe hypothermia and exhibited reduced locomotor activity in response to the peripheral administration of highdose LPS.

The central injection of LPS induced hypothermia in TRPM8 KO mice in the early phase.
To elucidate whether central TRPV8 is involved in endotoxin-induced body temperature changes, the body temperature of WT and TRPM8 KO mice was measured after an i.c.v. injection of 3.6 µg/kg LPS (Fig. 3). A two-way repeated measures ANOVA showed a significant effect of group (F3,168 = 112.04, p < 0.001), time (F7,1176 = 4.69, p < 0.001), and their interaction (F21,1176 = 7.89, p < 0.001) on alterations in core body temperature (Fig. 3a). The body temperature of WT mice was significantly higher at 30-480 min with a peak (2.19 ± 0.28 °C) 255 min after the i.c.v. injection of LPS than in those administered saline. In contrast to WT mice, the body temperature of TRPM8 KO mice was markedly lower at 95-175 min with a nadir (-1.99±0.59 °C) at 150 min than in those administered saline. Thereafter, it gradually returned to control levels and then was significantly higher in the later phase than that in mice administered saline. The temperature index was significantly higher in WT (4.94 ± 0.41 Δ o C×hr) 0~3.5 hr after the central LPS injection than in those administered saline (-0.45 ± 0.51), whereas it was lower in TRPM8 KO mice (-1.99 ± 0.81) than in those administered saline (0.44 ± 0.53) (Fig. 3b). The temperature index was significantly higher in WT (7.59 ± 0.82) and TRPM8 KO (4.56 ± 0.99) mice 3.5~8.0 hr after the i.c.v. injection of LPS than in those administered saline (WT, -1.52 ± 0.69; KO, -0.94 ± 0.70). The one-way ANOVA showed no significant differences in locomotor activity in WT and TRPM8 KO mice following the i.c.v. injection ( Fig. 3c and Fig. S3). These results indicate that the i.c.v. injection of LPS caused hypothermia in TRPM8 KO mice in the early phase, in contrast to fever in WT animals.

Scientific Reports 8
The peripheral administration of the TLR2 agonist zymosan only induced hypothermia in TRPM8 KO mice. To elucidate whether TRPV8 is involved in TLR2-induced body temperature regulation, the body temperature of WT and TRPM8 KO mice was examined after the intraperitoneal administration of 10 mg/kg zymosan (Fig. 4) However, zymosan-induced reductions in body temperature were markedly smaller in TRPM8 KO mice than in WT mice. The body temperature of WT mice then increased to generate fever with a peak (1.10 ± 0.34 °C) at 140 min, whereas TRPM8 KO animals did not exhibit significant fever. The temperature index was significantly lower in TRPM8 KO mice (-3.26 ± 0.86 Δ o C×hr) 0~1.5 hr after the administration of zymosan than in those administered saline (0.25 ± 0.20), whereas zymosan did not change the temperature index (control, -0.88 ± 0.32; zymosan, 0.01 ± 0.21) in WT mice (Fig. 4b). The temperature index was significantly lower in TRPM8 KO mice (-0.59 ± 0.73) 1.5~3.0 hr after the administration of zymosan than in those administered saline (-0.32 ± 0.26), but was significantly higher in WT animals (saline, -0.53 ± 0.24; zymosan, 1.03 ± 0.51). The one-way ANOVA showed no significant changes in the cumulative locomotor activity of WT and TRPM8 KO mice administered zymosan ( Fig. 4c and Fig. S4). Therefore, TRPM8 KO mice only exhibited hypothermia in the early phase, in contrast to mild hypothermia in the early phase and fever in the later phase in WT animals, in response to the peripheral administration of zymosan.
The central injection of IL-1β may cause hypothermia in the early phase and fever in the late phase in TRPM8 KO mice. To elucidate whether central TRPV8 is involved in IL-1βinduced fever, the body temperature of WT and TRPM8 KO mice was measured after an i.c.v.
injection of 400 ng/kg IL-1β (Fig. 5). A two-way repeated measures ANOVA showed a significant effect of group (F3,152 = 89.65, p < 0.001), time (F7,456 = 2.54, p < 0.05), and their interaction (F21,456 = 5.88, p < 0.001) on alterations in core body temperature (Fig. 5a). Body temperature was significantly higher in WT mice at 80-205 and 440-465 min with a peak (1.71 ± 0.13 °C) 200 min after the i.c.v. IL-1β injection than in those administered saline. In contrast to WT mice, TRPM8 KO mice did not show any significant fever within 3 hr, but appeared to develop hypothermia. In comparisons with mice administered saline, the body temperature of TRPM8 KO mice gradually increased to generate significant fever at 240-480 min with a peak (1.63 ± 0.27 °C) at 370 min. The temperature index was significantly higher in WT mice (3.37 ± 0.40 Δ o C×hr) 0~3.0 hr after the i.c.v administration of IL-1β than in those administered saline (-0.02 ± 0.47), whereas no significant differences were observed in the temperature index of TRPM8 KO mice between those administered IL-1β (0.18 ± 0.72) and saline (0.69 ± 0.49) (Fig.   5b). The temperature index was significantly higher in WT (5.84 ± 0.50) and TRPM8 KO (5.48 ± 1.41) mice 3.0~8.0 hr after the i.c.v. administration of IL-1β than in those administered saline (WT, -1.66 ± 0.73; KO, -0.94 ± 0.77). The one-way ANOVA revealed no significant differences in locomotor activity in WT and TRPM8 KO mice following the i.c.v. IL-1β injection ( Fig. 5c   and Fig. S5). These results indicate that TRPM8 KO mice were more likely to develop hypothermia in the early phase and delayed fever generation in the later phase in response to The colocalization of TRPM8 and PGE2 receptor 3 (EP3R) in the POA and PGE2-induced fever. Double labeling immunohistochemistry for TRPM8 and EP3R was performed to clarify whether TRPV8 is present in the EP3R-expressing neurons of the POA (Fig. 6). TRPM8 expression was prominent in the POA and LS of the hypothalamus (Fig. 6a,c). A 3D analysis revealed the colocalization of TRPM8 and EP3R in a POA neuron (Fig. 6b) (Fig. 6d). The i.c.v. injection of 160 nmol/kg PGE2 induced rapid and robust fever in TRPM8 KO mice that was indistinguishable from that in WT mice. On the other hand, the i.c.v. injection of 16 nmol/kg PGE2 induced significant fever in TRPM8 KO mice only. The injection of 160 nmol/kg PGE2 resulted in a significantly higher temperature index in WT (2.38 ± 0.40 Δ o C×hr) and TRPM8 KO (1.52 ± 0.45) mice than in those administered saline (WT, 0.26 ± 0.21; TRPM8 KO, 0.21 ± 0.13) (Fig. 6e). In comparisons with saline-treated mice, the temperature index was significantly higher in TRPM8 KO mice administered 16 nmol/kg PGE2 (0.62 ± 0.30 Δ o C×hr), but not in WT (1.67 ± 0.20). The temperature index was not significantly different between WT and TRPM8 KO mice following the injection of 160 nmol/kg PGE2. Furthermore, no significant differences were observed in locomotor activity between WT and TRPM8 KO mice following the PGE2 injection ( Fig. 6f and Fig. S6). These results indicate that PGE2-dependent fever generation in the brain did not significantly differ between WT and TRPM8 KO.
Augmentation of Fos expression in the hypothalamus in TRPM8 KO mice after the peripheral administration of low-dose LPS. As shown in Fig. 1, the intraperitoneal administration of low-dose LPS induced opposite thermal responses between WT and TRPM8 Scientific Reports 11 KO mice: fever in WT mice and hypothermia in TRPM8 KO animals. Fos immunohistochemistry was performed to clarify whether severe brain inflammation occurs in TRPM8 KO mice (Fig. 7, 8). Immunohistochemistry revealed that many Fos + nuclei were present in the ventral part of the LS (LSV), bed nucleus of the stria terminalis (BST), lateral preoptic area (LPO), median preoptic area (MnPO), and medial preoptic area (MPA) in LPStreated WT mice (Fig. 7a,a'c,c', Fig. S7, 8). Similarly, Fos + nuclei were frequently observed in the dorsal part (LSD) and intermediate part (LSI) of the LS, LSV, BST, LPO, MnPO, and MPA of LPS-treated TRPM8 KO mice (Fig. 7b,b'd,d', Fig. S7, 8). The quantitative analysis showed that the number of Fos + nuclei was significantly increased in the LSV, BST, LPO, MnPO, and MPA of WT mice by the administration of LPS (Fig. 7e). In TRPM8 KO mice, the number of Fos + nuclei was increased by the LPS treatment in all hypothalamic regions examined. More importantly, the number of Fos + nuclei was significantly higher in the LSD, LSI, BST, and MPA of TRPM8 KO mice than in those of WT mice.
The number of Fos + nuclei in the paraventricular nucleus (PVN) and DMH was higher in LPS-treated WT mice (Fig. 8a,a',c,c') than in saline-treated mice (Fig. S9). Fos + nuclei were more frequently observed in the PVN and DMH of LPS-treated TRPM8 KO mice than in WT mice. The number of Fos + nuclei was significantly increased in the PVN and DMH of WT and TRPM8 KO animals by the administration of LPS, but was significantly higher in LPS-treated TRPM8 KO mice than in WT mice (Fig. 8c). These results indicate that Fos expression in the hypothalamus was stronger in TRPM8 KO mice than in WT animals following the peripheral administration of low-dose LPS.

Discussion
TRPM8 functions as a neuronal sensor of cold temperatures and is activated by moderate cooling and cooling agents, such as menthol and icilin 25 . TRPM8-deficient mice show no preference for warm temperatures over cold temperatures and have impaired cold avoidance behavior 26,31 . Besides its role in cold sensation, TRPM8 is also involved in the suppression of inflammatory responses in colitis 32 and experimental autoimmune encephalomyelitis 33 and the augmentation of airway inflammatory diseases 34 . However, limited information is currently available on the functional significance of TRPM8 in endotoxin-induced body temperature alterations and brain inflammation. In the present study, we found that the peripheral administration of low-dose LPS induced hypothermia in TRPM8 KO mice, in contrast to fever in WT animals. Moreover, hypothermia was more severe in TRPM8 KO mice than in WT mice after the peripheral administration of high-dose LPS. Similar to TLR4 signaling, the peripheral administration of the TLR2 agonist zymosan only induced hypothermia in TRPM8 KO mice, but caused hypothermia in the early phase and fever in the late phase in WT mice. Therefore, the present study is the first to show that TRPM8 is important for switching between fever and hypothermia under endotoxin-induced inflammation.
PGE2 generates fever by acting on EP3R-expressing neurons in the POA of the hypothalamus 35,36 . The present study showed that the i.c.v. injection of PGE2 generated fever in TRPM8 KO mice in a similar manner to WT mice. TRPM8 KO mice showed normal fever generation by brown adipose tissue with β3-adrenergic receptor activation 37 and cold-induced shivering and tachycardia 38 . These results indicate that heat-generating thermoregulatory effectors are basically normal in TRPM8 KO mice. On the other hand, the i.c.v. injection of LPS induced hypothermia in TRPM8 KO mice, in contrast to fever in WT animals, in the early phase. This hypothermic response following the central injection of LPS was similar to that after the peripheral administration of low-dose LPS. Similarly, the i.c.v. injection of IL-1β appeared to cause hypothermia and did not induce fever in the early phase. The selective deletion of PGE2-synthesizing enzymes in brain endothelial cells strongly attenuated fever responses to peripheral LPS and IL-1ß 23,39,40 . A central injection of LPS induced STAT3 signaling in the OVLT and POA to generate fever 41 . Furthermore, a central injection of IL-1β induced NF-κB signaling in endothelial cells, which is necessary for inducing fever, with Scientific Reports 13 endothelial cells in the fenestrate capillaries of the OVLT playing the most potent role 24 . The OVLT lacks the BBB and is located close to the central thermoregulatory center, the POA, and, thus, is important for immune-to-brain signaling to initiate systemic inflammation 17,42 .
Peripherally administered LPS and zymosan reach TLR4 on astrocytes and TLR2 on microglia and/or macrophages in the OVLT 13,43 . TRPM8-expressing neurons are specifically localized at the POA and LS 30 . Therefore, TRPM8 in the brain may be involved in switching between fever and hypothermia in the early phase of systemic inflammation.
Fever and hypothermia are controlled alterations in body temperature and correspond to mild and severe forms of systemic inflammation, respectively 11 . Infection-induced hypothermia is accompanied by decreases in the metabolic rate, heat production, and locomotor activity and, thus, hypothermia is a hallmark of severe systemic inflammation 44 . PGE2 and cytokine levels in plasma and POA are higher in mice exhibiting hypothermia than in fever-generating animals 45 . Fos expression in the PVN and DMH was previously reported to be stronger in mice with hypothermia than in fever-generating animals after the administration of LPS 46 . In the present study, the number of Fos + nuclei was higher in the POA, LS, PVN, and DMH of TRPM8 KO mice than in WT animals after the peripheral administration of low-dose LPS. The activation of the POA-DMH axis was previously shown to be crucial for inducing hypothermia and hypo-metabolism 16 . The PVN has been implicated in sickness behaviors, such as decreases in locomotor activity and food intake 47,48 . The present study showed that hypothermic responses were more severe in TRPM8 KO mice than in WT mice following the peripheral administration of high-dose LPS. Furthermore, locomotor activity was significantly lower in TRPM8 KO mice than in WT mice after the administration of high-dose LPS. These results indicate that TRPM8 is necessary for suppressing brain inflammation and, thus, the lack of TRPM8 results in severe neuronal inflammation and, ultimately, hypothermia.
The mechanisms by which TRPM8 controls switching between fever and hypothermia have not yet been elucidated. One explanation is that TRPM8 is necessary for the proper functioning

Measurement of body temperature and locomotor activity. Under anesthesia with chloral
hydrate, the G2 E-mitter transponder was implanted intraperitoneally to record changes in core body temperature and gross locomotor activity. Mice were then maintained at an ambient temperature of 25°C ± 0.5 C under a 12-hr light/dark cycle (lights on at 7:00 and lights off at 19:00) for at least 1 week after surgery. Intraperitoneal and i.c.v. administration protocols were initiated at 11:00. Abdominal temperature was measured by biotelemetry at 5-min intervals, and at 1-min intervals for the PGE2 treatment, over a period of 12 hr before and 24 hr after the treatment. Data were acquired and fed to a computer using VitalView software (VitalView series 4000). The baseline temperature was calculated as the mean core body temperature in each group at 10:45-11:00. The fever index (℃ × hr) was calculated as the area under the temperature curve according to the baseline temperature before the treatment. DMH (between the bregma -1.46 and -2.06 mm) according to the mouse brain atlas (Paxinos & Franklin, 2001). To perform a quantitative analysis, fluorescent images were obtained using a DS-Fi3 digital microscope camera (Nikon, Tokyo, Japan) under the same pinhole size, brightness, and contrast settings. Images (1,440 × 1,024 pixels) were saved as TIF files (1,440 × 1,024 pixels) by employing NIS Elements BR (Nikon) and arranged using Photoshop CC (Adobe Systems Incorporated, San Jose, CA). In quantitative analyses, the total area of each brain region was measured using Image J 52 . The numbers of Fos + nuclei were counted using ImageJ, the threshold intensity of which was set to include measurement profiles by visual inspections and was kept constant. An analysis of all images was performed such that the experimenter was blind to the treatment group.

Declaration of interest
The authors declare that there is no conflict of interest that may be perceived as prejudicing the impartiality of the research reported.     The body temperature of WT and TRPM8 KO mice intraperitoneally administered 5 mg/kg LPS was significantly lower than that of mice administered saline, whereas the core body temperature of TRPM8 KO mice was significantly lower than that of WT animals following the administration of LPS. b: The temperature index was significantly lower in LPS-treated TRPM8 KO mice than in LPS-treated WT mice. c,d: The cumulative locomotor activity of LPS-treated TRPM8 KO mice during all periods examined was markedly lower than that of the saline controls, whereas that of LPS-treated WT animals did not show any significant decrease. Data (WT saline, n = 9; WT LPS, n = 6; TRPM8 KO saline, n = 5; TRPM8 KO LPS, n = 9) are expressed as the mean (± s.e.m.). Lines in a indicate a significant period (p < 0.05). * p < 0.05, ** p < 0.01, ***p < 0.001 among groups by a one-way ANOVA with Tukey's post hoc test.  of both LPS-injected WT and TRPM8 KO mice was negative in the early phase, whereas that of LPS-injected WT and TRPM8 KO animals was positive and negative, respectively, in the late phase. c: Cumulative locomotor activity was not significantly different among groups. Data (WT saline, n = 9; WT zymosan, n = 5; TRPM8 KO saline, n = 5; TRPM8 KO zymosan, n = 4) are expressed as the mean (± s.e.m.). Lines in a indicate a significant duration (p < 0.05). * p < 0.05, ** p < 0.01 among groups by a one-way ANOVA with Tukey's post hoc test.   were expressed as the mean ± s.e.m. *: P < 0.05, **: P < 0.01, ***: P < 0.01 vs control WT mice. Statistical analyses were performed using the Student's t-test.