Temperature is a critical factor for the development and survival of all organisms. Thermal homeostasis requires the precise regulation of numerous physiological and behavioural mechanisms. Animal species have developed different strategies to maintain an internal temperature which may allow an efficient function of cells and organs, despite a changing external environment (1). Depending on their processes of thermoregulation, organisms are divided into ectotherms and endotherms (2). In ectotherms, internal sources of heat are negligible, and body temperature relies on environmental heat. They are usually poikilotherms, although some species can maintain a homeothermic state by using behavioural mechanisms. In contrast, endotherms use internal metabolic processes to generate heat and usually maintain a narrow range of core body temperature (Tc), homeothermy.

Endothermy, the internal production of heat, has been very important during evolution. It allowed species, like birds and mammals, to live rather independently from external sources of heat, thus enlarging their ecological niches. However, this mode of living requires the consumption of large amounts of food, providing sufficient energy to sustain a high and stable Tc. Fluctuations of environmental temperatures turn on thermoregulatory mechanisms to defend homeothermy, i.e. heat-conserving and heat producing responses, when exposed to low temperatures and heat-dissipation mechanisms when temperature increases. Extreme external temperatures that overwhelm the normal thermoregulatory mechanisms or, alternatively, disruptions in the normal functioning of the thermoregulatory system will produce hyperthermia or hypothermia, which may be life-threatening.

Increasing evidence has highlighted the intimate interlink between thermal and energy homeostasis. The maintenance of homeothermy requires a fine regulation of food intake, heat production, heat conservation or dissipation and energy expenditure, which will heavily affect energy metabolism and energy balance. Although basic mechanisms responsible for maintaining the Tc in mammals are similar in males and females, differences in anatomic, hormonal, and other sex-dependent characteristics greatly influence several aspects of thermoperception, thresholds of activation of thermoregulatory effectors and body temperature levels and set points. The present review focuses on sexual differences in thermoregulation and their implications for energy homeostasis. To do so, we will first provide a brief overview of the variables governing temperature regulation, which will serve to provide a background to evaluate potential sex differences. We finally discuss how sex-dependent characteristics in thermal responses may induce metabolic differences between males and females, with a clear physiological, clinical, and therapeutic relevance.

Tc in endothermic homeotherms is the result of an increased generation of basal heat, also called obligatory or basal thermogenesis, which is mainly dependent on thyroid hormones, and the development of regulatory processes to maintain this internal temperature around a set point(s), through thermogenic and heat dissipative processes. Temperature is detected at the skin and in the core by thermal sensors, which send afferent information to the central thermoregulatory centre in the preoptic area of the hypothalamus (POA) and to sensory and other cortical areas to orchestrate the activation of thermoregulatory mechanisms. Warm and cold sensitive neurons are also found in the POA (3, 4).

In most mammals, including humans, average Tc is around 37 °C and oscillates between day and night following a circadian rhythm with a period close to 24 hours. The circadian system modulates metabolic heat production and heat loss to generate the Tc rhythm (5). In nocturnal rodents (i.e., the great majority), the uppermost Tc levels are reached during the night (active phase) and the lowermost values are attained during the daytime (inactive phase). The opposite occurs in humans, due to our diurnal condition (6). In both, rodents and humans, there is a diurnal variation of ≈ 1 °C (5, 6). Tc circadian oscillations are mainly regulated by neurons in the suprachiasmatic nucleus of the hypothalamus (SCN), the master clock of the organism, which have been shown to project to the thermoregulatory circuits in the POA to imprint the circadian fluctuations (7, 8).

Finally, sex-specific differences are observed in Tc levels in laboratory animals (9, 10) and in humans (6). Furthermore, Tc experience cyclic changes throughout the ovarian cycle in mammalian females, suggesting that sexual hormones regulate thermal homeostasis.

The ovarian cycle is essential for periodic production of mature ova through ovulation but encompasses many other cyclic events in reproductive tissues (uterus, vagina, external genitalia,…). Maintenance of the ovarian cycle mainly depends on the coordinated function of the hypothalamic-pituitary-ovarian axis. Briefly, secretion of gonadotropin-releasing hormone (GnRH) from the hypothalamus induces the synthesis and liberation of the gonadotropins, luteinizing hormone (LH) and follicle stimulating hormone (FSH) by pituitary gonadotrophs. Gonadotropins act on cells of the ovarian follicles, regulating the secretion of estrogens and progesterone. Finally, estrogens and progesterone drive the development of ovarian follicles and exert a large variety of reproductive and non-reproductive actions in the female organism (11). Significant differences in ovarian cyclicity exist between rodents and humans. In rodents, the ovarian cycle is linked to changes in reproductive behavior and lasts 4-5 days (commonly called estrous cycle). It is generally divided into four stages: proestrus, estrus, metaestrus and diestrus, with cyclic changes of sexual hormones ( Figure 1 ). Estrus is the period of the cycle during which ovulation occurs and mating behavior is displayed (12). In women, ovarian cycle lasts approximately 28 days, and is divided into a follicular phase (with high levels of estrogens), and a luteal phase (with high levels of progesterone). Cyclic ovarian hormonal changes are also responsible in women for the presence of the menstrual cycles, i.e., the regular appearance of menses due to the shedding of the endometrium (11) ( Figure 1 ).

In rodents and humans Tc levels are higher in the feminine sex (0.2-0.5 °C higher) (6, 9, 10, 13) ( Figures 2A, B ). Furthermore, Tc recordings during several days in females reflect the changes of this variable across the ovarian cycle. Thus, Tc in female mice shows sustained high levels during the dark phase of estrus versus the typical biphasic profile of the other phases of the estrous cycle. In contrast, Tc in male mice display a consistent and constant pattern (10) ( Figures 2C, D ). In women, there is a notable regulation of Tc throughout the ovarian cycle with an increase of 0.3-0.7 °C in the luteal phase, in comparison with the lower and constant levels in men (6). The physiological significance of the higher levels of Tc in females is not clear, but it may serve to create an accurate environment for implantation and survival of the embryo (6).

Although the mechanism(s) underlying the differences in Tc between sexes are still not well established, gonadal hormones are considered as main factors. Among them, progesterone seems to play the most crucial role as thermogenic signal, although, the progesterone:estrogen ratio is also important for modulating Tc in females (6, 10). Sex hormones are proposed to influence Tc by direct action on hypothalamic POA neurons. In fact, receptors for progesterone, estrogens and testosterone are expressed in this hypothalamic area and regulatory effects of these hormones have been described, especially, on temperature-sensitive neurons (15–17). However, the major effect at the central level is exerted by progesterone. This hormone shifts the thermoregulatory set point(s) in the hypothalamus towards higher temperatures. This results in a higher Tc, as it can be observed in the luteal phase of the human ovarian cycle, with very high levels of circulating progesterone (6, 18) ( Figure 1 ). In rodents the effect is not so evident because they do not have a real luteal phase, but the increase of plasma progesterone concentrations during the end of proestrus and beginning of estrus may be involved in the elevated Tc levels of estrus (19) ( Figure 1 ). Nevertheless, the cellular and molecular mechanisms and circuitry to induce progesterone’s thermogenic effects are still poorly understood. Estrogens and testosterone are proposed to participate principally in the peripheral regulation of thermoregulatory responses to changes in environmental temperature, although central effects have also been described (10, 18, 20–23). Interestingly, differences in Tc levels between sexes are mostly eliminated in gonadectomized adult mice (10), suggesting that sex effects on Tc are mainly mediated by gonadal sex hormones.

In all mammals there is a range of environmental temperatures within which basal metabolism generates enough heat to maintain Tc. This temperature interval is known as the thermoneutral zone (TNZ) ( Figure 3 ). When environmental temperatures either exceed or fall below the TNZ, the metabolic rate increases by activation of thermoregulatory mechanisms. Above the TNZ, (above the upper critical temperature, UCT), there is an increase of the metabolic rate because of the energy cost of the mechanisms for cooling. Conversely, at the lower bound of the TNZ (lower critical temperature, LCT), the animal activates facultative heat production mechanisms to defend its body temperature. For most research animals (and for naked humans), LCT lies close to 30 °C. The increase in heat production required to balance heat lost is inversely proportional to environmental temperature ( Figure 3 ). The slope of the line is a measure of the thermal conductance, i.e., steep for poorly insulated animals (e.g., mice, naked human) and a shallow slope for good insulation (e.g., arctic species, dressed humans). The TNZ will therefore depend principally on basal thermogenesis, body size and insulation. Greater basal thermogenesis, greater insulation and greater sizes produce larger TNZs. In contrast, smaller animals present a greater surface area-to-volume ratio and are more susceptible to heat loss.

Sex also influences the TNZ in rodents and humans. Experiments with rodents have demonstrated a higher threshold temperature for cold-induced thermogenic response in females compared to males (27). Likewise, in humans, activation of thermoregulatory mechanisms in response to cold and heat are shifted to higher temperatures in women in relation to men (28–31). Thus, it was shown that in lightly dressed women, metabolic heat production increases when air temperature decreases below 31 °C, while in men the LCT was found at 28.5 °C (32). More research is warranted to determine whether the width of the TNZ is also different between sexes. Notice, that in many human studies, data refers to the range of internal body temperature between thermoregulatory thresholds (Core interthreshold zone, CIZ), and not to the ambient temperature range, as it defines the concept of TNZ.

Sex-dependent anthropometric parameters contribute to the observed differences in TNZ between sexes. In general, female mammals are smaller than males and have a larger surface-to-mass ratio, which favours heat loss. In contrast, females have more fat content than males (both whole body and subcutaneous), which increases insulation. Differences in obligatory thermogenesis have also been described. Women present lower levels of basal metabolic rate (BMR) than men (a reduction of about a 23%) (13, 33), even at a given body mass, although the difference is much smaller (about 3%) (33). But the most important factor for the higher threshold temperatures of thermoeffector responses in females is progesterone. Accordingly, the elevated levels of this sexual hormone in the luteal phase in women ( Figure 1 ) are responsible for the shift of most thermoregulatory response thresholds towards warmer temperatures during this phase in relation to the follicular phase, as we will detail below (6, 18, 31, 34, 35).

Because we often use mice to model human physiology, we need to be aware that mice and human exhibit large differences in body size, shape and composition, which impact on thermal homeostasis. In mice, thermal conductance is higher, Tc is more variable and unstable, temperature preference is set at warmer temperatures and metabolic rate is greatly enhanced when compared to humans (24). In typical housing conditions of 21-24 °C, mice are subjected to a situation of chronic cold stress, even more intense in the case of females. This will affect the experimental outcomes and the possible interpretation of the results. Considering that humans live most of their time close to or in thermoneutral conditions (due to clothing and to the control of housing temperature), numerous authors have recently claimed that mice should be housed at thermoneutrality, to develop a more humanized experimental model. Nevertheless, while humans can maintain a constant Tc across the range of temperatures which are considered as the TNZ (between the LCT and UCT) ( Figure 3 ), Tc rises in mice, especially over 30 °C (24, 25), with a circadian variation. This near absence of thermoneutrality in mice together with the higher values of LCT in females than in males make very complex to establish standard conditions of ambient temperatures for housing mice. Setting an external temperature of 26-28 °C, incorporating abundant bedding and nesting material in mouse cages (especially for females) and controlling housing density will help mice to have a more active role in their thermoregulation under a thermal situation more similar to humans.

Body weight and body composition result from the equilibrium between food intake and energy expenditure. Energy homeostasis, energy metabolism and energy deposition appear, at first sight, very simple and direct processes, but they constitute entangled conditions. Numerous factors (i.e., genetic determinants, sexual hormones, environment, nutritional characteristics, hedonic behaviour, grade of activity, metabolic efficiency, social and economic factors, etc.) influence and regulate the different components of this balance. Although research has focused primarily on either increased palatable food intake and the effects of sedentary lifestyles, this paradigm may not fully address the complexity of interactions that modulate both the food that individuals eat, nor the processes that account for energy expenditure and energy storage. In this regard, it has been suggested that multiple, undervalued, or subtle factors may be additionally contributing to the pandemic increase in fat gain. As discussed below, one of these may be ambient temperature and the thermal mechanisms involved in maintaining homeothermy (169–171). Furthermore, sexual differences in thermal physiology influence energy homeostasis.

Ambient temperature and thermoregulation have a major role on energy balance by regulating food intake, heat production and energy expenditure. In fact, mammals dedicate 40-50% of their total energy expenditure to sustain a constant Tc at 37 °C (171). Among the thermoregulatory processes, BAT thermogenesis plays a central role in the regulation of both, body temperature and energy homeostasis. Thus, a diminished response of BAT to food intake and to cold temperature has been observed in practically all genetic forms of obesity (172). In recent years, BAT has received renewed interest due to its lipid oxidizing capacity. BAT and browning of WAT have become attractive therapeutic targets to increase energy expenditure and clearing excessive fat from circulation. Thus, activating thermogenic metabolism in BAT and WAT appears to exert a beneficial impact in obesity, type-2 diabetes, dyslipidaemia and cardiovascular pathologies, improving the glucose and lipid homeostasis (136, 173–177).

Interestingly, most mammals spend most of their time in ambient temperatures lower than their TNZ, which implies that they activate energy-costly thermoregulatory mechanisms. In contrast, in the last decades, humans have greatly increased the time spent indoors, with a widespread access to central heating and air conditioning and higher expectations of thermal comfort. All these factors contribute to reduce mild thermal stress. More time spent within the TNZ has been associated with increased obesity rates (169, 170), presumably through reductions of energy expenditure due to a decrease in thermogenesis, including a diminished efficiency in diet-induced thermogenesis.

When analyzing the effect of ambient temperature on energy homeostasis, the other variable to consider is energy intake. Food intake is required to meet physiological needs of nutrients, but also as a source of energy, including what is needed for thermal homeostasis. An inverse relationship exists between ambient temperature and food intake across a wide range of temperatures in multiple species, including humans. During short exposures to cold, an almost perfect fit is shown, with intake adjustment compensating accurately for temperature-induced changes in expenditure. In contrast, during long-term cold exposures, the rise in food intake is insufficient to compensate for the increased metabolic output (169, 178, 179), resulting in a progressive reduction of fat mass. Conversely, when environmental temperatures get warmer and close to the TNZ, intake is progressively suppressed, but this reduction is lower than the decrease observed in energy expenditure, particularly in the case of highly tasty foods (169). This has important implications in our ‘obesogenic’ society, living most/all our time in our TNZ range of temperatures and surrounded by food, which is always accessible and highly palatable. The contribution of the environmental comfort of life to the current obesity pandemic merits further investigation. Furthermore, the study of the metabolic effects of cold ambient temperatures, including the activation of cold thermosensors is an emerging field with great physiological and medical interest. Thus, accumulating evidence indicate that mice exposition to cold temperatures induce loss of fat mass in animals fed a standard or a high fat diet, with a significant activation of BAT and WAT browning (178, 179). Contrarily, mice deficient in TRPM8 (the main cold sensor) develop obesity when housed at mild temperatures, exhibiting diurnal hyperphagia, reduced lipid utilization (70) and an altered circadian physiology (14). Furthermore, oral or topic TRPM8 agonists have proven to be effective in reducing WAT in mice (180–183). Therefore, targeting temperature-induced signalling pathways may open new therapeutic avenues in obesity and its metabolic comorbidities. Furthermore, sustained alterations in the function of certain thermoTRP channels (e.g.,TRPM8 polymorphisms) may play an unrecognized role in obesity pathogenesis.

Sex differences in thermoregulation have an impact on energy homeostasis, that needs to be considered. As discussed above, the neutral interval of temperatures is shifted to higher temperatures in female mammals (including women). Furthermore, in the feminine sex the TNZ experiences cyclic changes according to the progesterone:estrogens plasma levels, with higher temperature set-point(s) in the luteal phase. A shift of the TNZ towards higher temperatures implies that cold-induced mechanisms will be activated at warmer temperatures, as it has been widely demonstrated in rodents and women. Consequently, the metabolic impact of cold temperature on energy balance will be observed at higher ambient temperatures in females than in males. Indeed, a recent study showed that housing female mice for two weeks below 25 °C induced a decrease in fat mass, while only temperatures below 18 °C produced a reduction of white adipose tissue in males (179). Further research is warranted to investigate in humans the effectiveness of chronic stimulation of cold temperature signalling pathways on fat loss, glucose and lipid homeostasis and evaluate the threshold differences between both sexes.

Although obesity is increasing in both sexes, obesity rates and associated comorbidities are different between women and men. Data from the Centers for Disease control and prevention (CDC, USA) and from the World Health Organization (WHO) reveal that women have less prevalence of overweight than men, lower/similar rates of obesity, but double the risk to develop extreme obesity, especially after menopause (https://www.cdc.gov/obesity, https://www.who.int/health-topics/obesity). Genetic factors and sexual hormones determine significant differences in fat mass and distribution between sexes. Thus, women have a greater proportion of body fat mass than men, intimately linked to reproductive function, with larger depots of subcutaneous fat (with thermoregulatory value). In contrast, men tend to gain more visceral fat, which is associated with an increased risk of cardiometabolic complications (184). Interestingly, it has been described that adipose tissue has higher metabolic rate per kilogram in women, which agrees with the initiation of cold-induced thermoregulatory mechanisms at warmer temperatures and the increased expression of genes involved in mitochondrial function (168). Although the contribution of fat to the basal metabolic rate (BMR) is low, because the metabolic rate of adipose tissue is much lower than other internal organs and the skeletal muscle, a small increase in fat metabolism may have important long-term implications for conditions and diseases that develop slowly over several decades, like obesity and its health consequences (Figure 4). Lifestyle changes affecting thermoregulation including increasing time spent in outdoor activities, a greater exposure to seasonal cold, choosing walking or cycling instead of temperature-controlled car transportation may cooperate with a healthy food intake and exercise in preventing and reducing obesity, especially in women. In addition, the potential use of cold-mimetic compounds may override the problems associated to the discomfort of cold exposition protocols. More investigation is needed to develop new TRPM8 agonists, and study in humans their long-term effects on body fat and metabolism, paying attention to differences between sexes. Finally, although indoor temperatures are often tightly set at high fixed values, some recent studies have approached the potential application of moderate temperature drifts or ramps towards colder values to increase energy expenditure, while maintaining thermal comfort. The combination of adding more temperature variability in buildings with the use of personal comfort systems by selecting heating or cooling individual body parts to improve whole body thermal comfort is under active investigation (185, 186). Differences in the thresholds of TNZ and thermal comfort between sexes need to be considered in this type of studies (Figure 4).

Increasing evidence shows that differences in thermoregulation between sexes may play an important role in the observed sex differences in energy regulation, body weight and body composition. Although our knowledge of the interrelationship between thermal and energy homeostasis has greatly increased, most experimental studies in mice and humans examine only individuals of one sex, mainly males, to reduce the total number of subjects and the cost. Males are preferred for simplicity. There are, however, obvious downsides by omitting the feminine sex, from lack of knowledge of the physiology of females (50% of total population) to implications for optimal therapeutic impact and implementation of the findings depending on the sex. Understanding how ambient temperature and cold sensors, like TRPM8, impact on our metabolism is very relevant from a scientific, pharmaceutical, and medical point of view. It opens the opportunity for novel public health initiatives to address obesity and new therapeutic strategies. To establish in a rigorous way the influence of thermoregulation on energy homeostasis and the biomedical basis for better personalized treatment options, it will become crucial to carry out detailed and more standardized studies on molecular and functional sex differences, considering the ovarian cycle and the menopause.

RS: designed review; CF-P and AR: performed investigation; RS, CF-P, AR, FV, and VA: writing, reviewing and editing; CF-P, AR, VA, and RS: Visualization; RS and FV: Funding acquisition. All authors contributed to the article and approved the submitted version.