Cold exposure elicits a range of physiological effects in overall body processes in order to maintain body temperature (Kizaki et al., 1996; Sonna et al., 2002; Cannon and Nedergaard, 2004; Karp, 2012). Notably, in small mammals, cold exposure induces non-shivering thermogenesis is brown adipose tissue (BAT) and browning white adipose tissue (WAT), which results in fat mobilization and weight loss (Himms-Hagen, 1990; Palou et al., 1998; Nedergaard et al., 2007; Fuster et al., 2009; Reynés et al., 2015). Cold exposure and its resulting fat loss produce severe hypoleptinemia, which is accompanied by increased compensatory hyperphagia (Bing et al., 1998a; Lau et al., 2016).

The key regulator of food intake control is the hypothalamus, particularly the arcuate nucleus, which plays a crucial role in the integration and coordination of peripheral and central orexigenic and anorexigenic signals (Sahu, 2004; Palou et al., 2009). Briefly, leptin and other satiety signals inhibit the orexigenic neurons that produce agouti related peptide (AgRP) and neuropeptide Y (NPY), which are key enhancers of appetite (Morton and Schwartz, 2001; Valassi et al., 2008). Orexigenic signals such as ghrelin, or the absence of insulin and leptin, induce the contrary response (Ellacott and Cone, 2004). Accordingly, some studies revealed that Agrp gene expression increases due to the low leptin levels related to cold exposure (Tang et al., 2009; Lau et al., 2016). Nevertheless, in spite of the hypoleptinemia resulting from cold exposure, Npy regulation by cold is controversial (Mercer et al., 1997; Bing et al., 1998a,b; Park et al., 2007; Tang et al., 2009; Lau et al., 2016), since besides its orexigenic role, it also acts as a hypothermic peptide with potent central anti-thermogenic actions (Egawa et al., 1991; Billington et al., 1994). At the other end of the scale, the anorexigenic signals induce the production of pro-opiomelanocortin (POMC) and cocaine and amphetamine regulated transcript peptide (CART) in other neuron populations, also located in the arcuate nucleus, which induce satiety (Seeley and Woods, 2003; Porte et al., 2005; Valassi et al., 2008). While there is no evidence regarding the regulation of POMC by cold, it has been shown that CART signaling is required for maintenance of energy homeostasis under cold exposure (Lau et al., 2016). The long form of the leptin receptor (LEPR/ObRb) and the suppressor of cytokine signaling 3 (SOCS3), expressed in arcuate hypothalamic neurons, are commonly regarded as key hypothalamic factors mediating leptin regulation of feeding control (Blundell et al., 2001). However, the effects of cold exposure on leptin signal transduction are not fully elucidated, since one study indicates increased hypothalamic Lepr gene expression after cold exposure (Mercer et al., 1997) while another shows no changes in hypothalamic leptin transduction, suggesting that leptin sensitivity remains unchanged during cold acclimation (Tang et al., 2009).

Aging is a process with an important impact on the metabolism, including cold response. In rodents, BAT thermogenesis activation by cold exposure decreases with age (Reynés et al., 2015). Accordingly, in rabbits, the responses of sympathetic neurons to cold stress have been described to be reduced (Andrews et al., 1993). This suggests that food intake control in response to cold could be affected by aging. Thus, we considered it useful to study the effects of cold exposure on neuronal hypothalamic regulation at different ages of rat development, paying attention to the effects upon leptin signaling.

Neuronal hypothalamic regulation of the cold induced increase in food intake is difficult to study in humans due to inaccessibility of hypothalamic samples. Rodents provide a realistic alternative, but in rodents, in contrast to humans, BAT thermogenesis is a major contributor to cold induced energy expenditure, which is thought to induce the hyperphagic response. Ferrets (Mustela putorius furo) provide an alternative animal model, closer to humans in terms of adipose tissue organization (Murano et al., 2005; Liew et al., 2006; Wolfgang et al., 2006) as well as the thermogenic contribution of BAT to energy expenditure in response to cold exposure (Fuster et al., 2009).

In summary, the effects of cold exposure on the hypothalamic leptin signaling cascade, and on the regulation of food intake are not clear and may be affected by age. Studies in rats should be complemented by studies in other species, due to the different role of BAT in cold induced energy expenditure in rats compared to humans, in order to strengthen the data. Ideally, the responses should be analyzed in humans, but the hypothalamus is inaccessible for molecular analysis. Therefore, it is worthwhile to examine whether hypothalamic responses are reflected by responses in other tissues that are more easily accessible and amenable to molecular analyses, such as blood cells. Our previous work shows that peripheral blood mononuclear cells (PBMC) can be a surrogate tissue to study metabolic responses of internal tissues, including hypothalamus (Caimari et al., 2010b; Reynés et al., 2014, 2015), leading us to examine whether PBMC reflect hypothalamic adaptations to cold exposure.

Thus, to better understand the hypothalamic response to cold exposure, we analyzed molecular satiety and leptin signaling gene expression in rats of different ages, validated these responses in ferrets, and examined the validity of PBMC as a surrogate tissue.

Prolonged cold exposure sympathetically stimulates energy expenditure by increasing basal metabolism and BAT thermogenesis (Himms-Hagen, 1990; Nedergaard et al., 2007; Reynés et al., 2015). To counteract increased energy expenditure, cold exposure also induces hyperphagia, (Mercer et al., 1997; Bing et al., 1998a,b; Park et al., 2007; Tang et al., 2009; Lau et al., 2016). The anorexigenic hormone leptin, mainly expressed in the adipose tissue, seems to play a key role in energy intake regulation during cold acclimation; however, the molecular mechanisms involved are not yet clear (Bing et al., 1998a,b; Ricci et al., 2000; Zhang and Wang, 2007; Tang et al., 2009). Moreover, as reported recently, in mice, leptin treatment induces brain pyrexic mechanisms to prevent heat loss during cold exposure, but does not directly affect the thermogenic process; thus, thermogenesis is unlikely to be involved in changes in body weight of cold-exposed animals (Fischer et al., 2016), giving more relevance to the potential effects of leptin affecting food intake control.

The results obtained here demonstrate that leptin levels in rats decrease after cold acclimation regardless of the age of animals, and show a strong negative correlation between circulating leptin levels and food intake, suggesting a critical role of this hormone in the hyperphagia observed after cold exposure. Another important hormone with a role in food intake is ghrelin, which is mainly expressed in the stomach and increases in conditions of negative energy balance (Wren et al., 2001). Contrary to what was expected (Tomasik et al., 2005), our study showed that circulating serum ghrelin levels were not affected by cold exposure at any of the ages studied.

The molecular effectors of leptin action in the hypothalamus are the anorexigenic peptides POMC and CART and the orexigenic peptides NPY and AGRP; however, their involvement in increased food intake in cold-exposed rodents is not clear, and most of the differences observed could be related to differences between rodent species and to the length and magnitude (°C) of cold exposure. For instance, while 28 day-cold exposure (4°C) increases Agrp gene expression in Brandt's voles (Tang et al., 2009), chronic (9–16 weeks) mild cold exposure (22°C vs. 30°C) in mice reduces Cart gene expression (Lau et al., 2016). Regarding NPY, different effects have been observed depending on the time of cold exposure (Mercer et al., 1997; Park et al., 2007; Tang et al., 2009; Lau et al., 2016). Our data show that the expression of genes coding for anorexigenic peptides, especially Cart, is decreased in the hypothalamus after 1 week of cold exposure in rats, as has been reported in mice (Lau et al., 2016). The differential expression of this gene is dependent on age, and was only observed in the youngest (1 month) and surprisingly in the oldest animals (6 months of age) studied. On the other hand, 1 week of cold exposure did not induce the expression of the orexigenic genes Agrp and Npy in rat hypothalamus, at any of the ages studied. These two neuropeptides have been accredited with sympathetic thermogenic suppression activity, and an absence of regulation may be interpreted as prevention of their hypothermic activity (Egawa et al., 1991; Ruan et al., 2014). Interestingly, when we analyzed a ratio computed for orexigenic vs. anorexigenic gene expression, we only found a high increase at the age of 1 month, suggesting that the neuronal adaptations to cold exposure are of special relevance in young animals. These results are consistent with our previous study with the same cohort of animals which show that cold acclimation for 1 week increases lipid metabolism, BAT thermogenesis, and WAT browning, especially in the youngest animals (Reynés et al., 2015). Thus, even if increased food intake were to be observed at different ages, our data suggest stronger hypothalamic adaptations at early ages to counteract grater energy expenditure.

In spite of the role of leptin in the gene expression regulation of the neuropeptides involved in food intake control, the results of effects of cold exposure on leptin signaling in rodents are controversial. While cold exposure in mice resulted in increased Lepr gene expression (Mercer et al., 1997), it did not result in alterations in the regulation of hypothalamic leptin signaling in Brandt's voles (Tang et al., 2009). Surprisingly, our rat study revealed a transient peak for Socs3 mRNA expression in the hypothalamus of cold exposed animals only at 4 months of age, accompanied by increased mRNA expression levels of Lepr. Taking into account that SOCS3 is involved in the inhibition of leptin receptor signaling (Bjørbaek et al., 1999; Münzberg and Myers, 2005), the observed increase in Lepr mRNA expression could be a compensatory response to overcome the increased expression of Socs3. These results may be related to the smaller increase in food intake that is only observed in 4-month-old rats and suggest that specific factors may play a role in gene expression regulation at this particular age.

Most of the studies based on the hypothalamic contribution to food intake regulation by cold exposure are performed in rodents (Park et al., 2007; Lau et al., 2016). However, although extremely useful, rodents are not the best model to extrapolate cold exposure studies to a human situation (Bonet et al., 2013). This is because of the substantial contribution of the thermogenic activation of BAT to energy expenditure in rodents that is not the case in humans (Bonet et al., 2013). Thus, in addition to rodents, in the present study we newly used ferrets as an alternative approach to study the effects of cold exposure on the hypothalamus. We chose ferrets due to their similarities to humans in terms of adipose tissue organization and cold response (Murano et al., 2005; Fuster et al., 2009). In ferrets, Cart and Pomc mRNA levels remained stable, while both Agrp and Npy expression increased after 1 week of cold exposure (4°C). This was in contrast to the rat study, where gene expression changes were observed for anorexigenic genes, (mainly decreased expression of Cart), while expression of orexigenic genes (Agrp and Npy) was not affected. Other than the regulation of neuropeptide genes, hypothalamic leptin signaling was also modulated in ferrets as a result of cold acclimation. Interestingly, as observed in 4-month-old rats, we found increased Socs3 expression. In this case, probably as a compensatory mechanism, we also observed increased expression of Stat3, which is a positive transducer of leptin signaling. A limitation of our work is that we were not able to measure food intake or circulating leptin levels in the ferret study; the former because food intake control was not the main aim of ferret experimental design, and the latter due to the lack of specific commercial kits. Moreover, our study was performed on the hypothalamus as a whole, not using specific hypothalamic regions, such as the arcuate nucleus that plays a key role in food intake control, which may mask some additional hypothalamic cold responses. Finally, we used animals of different sex (female and male animals in the rat and ferret experiments, respectively) because the data were generated from previous experimental designs. Thus, some of the differences between species observed in the regulatory pattern in hypothalamus and PBMC may be due to sex. However, this fact, apart from being a limitation of the experimental design, could be considered as a strength, because it gives extra translational value and robustness to those results which follow the same direction in the different animal models.

PBMC are increasingly being used for identification of practical molecular biomarkers. Here, we show that these cells can potentially be used as a surrogate tissue for hypothalamic studies. Rat PBMC expressed detectable levels of the hypothalamic neuropeptide genes Npy, Pomc and Agrp, but not Cart, as well as the key leptin signaling genes Lepr, Socs3, and Stat3. Remarkably, gene expression of these genes reflected the regulatory pattern observed in rat hypothalamus. As seen in the hypothalamus, Npy, Pomc, and Agrp gene expression was not affected by cold exposure at any of the ages studied, while the increased Lepr and Socs3 gene expression observed in the hypothalamus of 4-month-old cold-exposed rats was seen in PBMC. In fact, a strong, positive correlation was found at this age between Socs3 mRNA in PBMC and hypothalamus. Accordingly, in ferret PBMC, even though these cells did not express any of the hypothalamic neuropeptides studied, they reflected the increased Socs3 mRNA levels observed in ferret hypothalamus. These results suggest that in both animal models PBMC are able to reproduce key effects of cold exposure on hypothalamic leptin signaling. The use of PBMC as a surrogate tissue to examine specific hypothalamic responses is supported by previous studies from our group. In particular, the correspondence of the Npy gene expression response in PBMC and hypothalamus to dietary interventions (Caimari et al., 2010a; Oliver et al., 2013) as well as weight gain and weight loss (Reynés et al., 2014) has been reported, validating the data of the present study.

This work demonstrates that cold exposure induces an increase in orexigenic/anorexigenic gene expression ratio in rat hypothalamus, particularly at an early age (1-month-old animals), and mainly due to decreased expression of anorexigenic genes (Cart and Pomc). Despite the key role of leptin in neuropeptide regulation, expression of leptin signaling-related genes were affected by cold exposure only at the age of 4 months, indicating that some specific factors could affect leptin sensitivity at this particular age in rat hypothalamus. Furthermore, our study in adult ferrets revealed an increased expression of orexigenic peptide genes (Agrp and Npy), which could be related to hypothalamic food intake control. Moreover, cold-exposed ferrets also manifested increased hypothalamic expression of the leptin signaling related genes Socs3 and Stat3. Finally, we feel that the most original and relevant data from the present research is that PBMC were able to reflect some of the characteristic hypothalamic gene expression features of cold exposure (mainly related to leptin signaling) observed in our study, both in rodents and ferrets. This suggests that these cells could be useful to perform studies aimed at the elucidation of hypothalamic mechanisms related to food intake control and energy metabolism regulation induced by exposure to low temperatures. Further studies to confirm the usefulness of PBMC to reflect hypothalamic responses, especially to cold exposure, are needed and can pave the way to faithfully use this easily obtainable biological material.

Bioethics Committee of the University of the Balearic Islands; approval: July 21, 2011.

AP and PO conceived the experiment; BR, MKH, and EGR did most of the experimental work; BR, PO, JK, and AP wrote the manuscript; and all the authors participated in critical revising of the manuscript.