Thermoneutral housing is a critical factor for immune function and diet-induced obesity in C57BL/6 nude mice

Abstract

OBJECTIVES

Obesity-related cancers represent public health burdens of the first order. Nevertheless, suitable mouse models to unravel molecular mechanisms linking obesity to human cancer are still not available. One translational model is the immunocompromised Foxn1 (winged-helix/forkead transcription factor) nude mouse transplanted with human tumor xenografts. However, most xenograft studies are conducted in nude mice on an in-bred BALB/c background that entails protection from diet-induced obesity. To overcome such resistance to obesity and its sequelae, we here propose the dual strategy of utilizing Foxn1 nude mice on a C57BL/6 background and housing them at their thermoneutral zone.

METHODS

C57BL/6 nude and corresponding wild-type mice, housed at 23 or 33 °C, were subjected to either low-fat diet or high-fat diet (HFD). Energy expenditure, locomotor activity, body core temperature, respiratory quotient as well as food and water intake were analyzed using indirect calorimetry. Immune function at different housing temperatures was assessed by using an in vivo cytokine capture assay.

RESULTS

Our data clearly demonstrate that conventional housing protects C57BL/6 nude mice from HFD-induced obesity, potentially via increased energy expenditure. In contrast, HFD-fed C57BL/6 nude mice housed at thermoneutral conditions develop adiposity, increased hepatic triglyceride accumulation, adipose tissue inflammation and glucose intolerance. Moreover, increased circulating levels of lipopolysaccharide-driven cytokines suggest a greatly enhanced immune response in C57BL/6 nude mice housed at thermoneutrality.

CONCLUSION

Our data reveals mild cold stress as a major modulator for energy and body weight homeostasis as well as immune function in C57BL/6 nude mice. Adjusting housing temperatures to the thermoneutral zone may ultimately be key to successfully study growth and progression of human tumors in a diet-induced obese environment.

INTRODUCTION

Obesity is one of the most prevalent chronic disorders worldwide and the second most preventable cause of death in the developed countries.¹ It is characterized by lipid accumulation in both adipose and non-adipose tissues and accompanied by a variety of comorbidities, such as insulin resistance and cardiovascular disease.² Recent epidemiological studies have linked obesity with an increased risk of developing numerous types of cancers, including esophageal, pancreatic, colon, breast, endometrial, kidney and liver.^3–5 Studies in rodents have suggested that elevated circulating levels of obesity-related hormones and pro-inflammatory cytokines (for example, insulin, insulin growth factor-1, leptin, interleukin-6 (IL-6) and tumor necrosis factor (TNF)) have a major role in cancer promotion and propagation.^5,6 However, the exact molecular mechanisms of action still remain elusive.

A major challenge is the choice of the appropriate research model—a model suitable for studying the underlying mechanisms of both cancer and obesity. In cancer research, numerous murine models, of genetically engineered mice lacking or overexpressing cancer-associated genes, have been developed to study tumor initiation, promotion, progression and response to cancer therapies. However, despite providing useful tools to study the role of specific genes on tumor development, such genetic murine models do not fully represent the genetic and epigenetic complexity of a human tumor.⁷ Of note, human xenograft models, where human tumor cells or biopsies are either transplanted under the skin (subcutaneous models) or into the organ of tumor origin (orthotopic models) of immunocompromised mice, provide attractive tools for translational cancer research. The most widely used murine model for human tumor xenografts utilizes nude mice carrying a single spontaneous mutation in the Foxn1 (winged-helix/forkead transcription factor) gene that leads to an abnormal thymus morphology, a lack of thymus-derived T cells, and as such represent an immunocompromised state.⁸ In fact, xenografts of human tumors into nude mice represent a commonly used model to interrogate the impact of various environmental cues on human tumors or to evaluate the clinically efficacy of anti-cancer agents.^9–11

Human tumor xenografts in the nude mouse model may also allow for analysis on the impact of diet and obesity on human tumor growth and metastasis. However, nude mice used in cancer-related studies are typically bred on an obesity-resistant BALB/c background.^12,13 In contrast, obesity and diabetes research are mostly conducted in mice on a C57BL/6 background, which facilitates diet-induced obesity and obesity-associated comorbidities, including glucose intolerance and hepatic steatosis.^13–15 Although nude mice have been backcrossed to a pure C57BL/6 background, their susceptibility to diet-induced obesity and comorbid conditions has not been studied in detail.

This study was designed to provide the first full metabolic characterization of C57BL/6 nude mice. We further aimed to assess their propensity for diet-induced obesity and its sequelae. Overall, our data address the question whether C57BL/6 nude mice can serve as a suitable translational model to study the link between obesity and cancer. For the first time, we show that nude mice on a C57BL/6 background are prone to the development of high-fat diet (HFD)-induced obesity and glucose intolerance. However, we also show that such propensity is dependent on housing C57BL/6 nude mice at thermoneutral temperatures. At conventional, ambient housing temperatures, nude mice display increased energy expenditure and full protection from HFD-induced obesity, compared with wild-type (WT) controls. Notably, thermoneutral housing significantly enhanced lipopolysaccharide (LPS)-driven pro- and anti-inflammtory cytokine production in C57BL/6 nude mice. Together, our data indicate that conventional housing-driven cold stress has broad effects on the development of obesity and its sequelae, as well as the immune system, in C57BL/6 nude mice.

MATERIALS AND METHODS

Mice

Male nude mice on a C57BL/6 background (B6.Cg-Foxn1nu/J) and corresponding C57BL/6 WT mice were purchased from The Jackson Laboratory (Bar Harbour, ME, USA) at 6 weeks of age. Mice were housed in high-efficiency particulate-filtered laminar flow hoods at the Cincinnati Children’s Hospital Medical Center or at the animal facility of the Metabolic Diseases Institute of the University of Cincinnati, with free access to irradiated food and autoclaved water. Four independent experimental groups were used for the studies. In a first set of experiments, 10-weeks-old nude mice and WT controls (N = 5) were ad libitum fed with a low-fat diet (LFD), 10% kcal from fat (D12450B, Research Diets Inc., New Brunswick, NJ, USA), or a HFD, 60% kcal from fat (D12492, Research Diets Inc.), for 35 days to compare the level of diet-induced obesity in both strains. Body weights were measured weekly. Body composition was analyzed prior to, 2 and 4 weeks after feeding LFD or HFD using a whole-body composition analyzer (EchoMedical Systems, Houston, TX, USA).

With the second experimental cohort (N = 8), we aimed to compare the effects of temperature on metabolic performance in nude and corresponding WT mice. Ten-weeks-old mice were ad libitum fed with HFD. Metabolic performance (indirect calorimetry, locomoter activity, body core temperature, respiratory quotient (RQ) and caloric intake during light and dark phases) was analyzed as described below. The third cohort of C57BL/6 nude mice (N = 8) were kept at an increasing temperature gradient ranging from 30 to 33 °C and were ad libitum fed with either LFD or HFD for 12 weeks. Body weights and food intake were measured weekly. Intraperitoneal (i.p.) glucose tolerance tests were performed 1 week before killing. In a last set of experiments, we investigated the ability of LFD-fed C57BL/6 nude mice (N = 8) housed either at 23 or 33 °C to respond to an inflammatory stimulus (LPS challenge; see below). Animal care was provided in accordance with the procedures outlined in the Guide for the Care and Use of Laboratory Animals under animal study proposals approved by the Cincinnati Children’s Hospital Medical Center or by the University of Cincinnati Institutional Animal Care and Use Committee.

I.p. glucose tolerance test

Mice (N = 6) were fasted for 6 h, starting 1 h after the initiation of the light phase. Baseline glucose levels (0 min) were measured in duplicate from tail blood using hand-held glucometers. Mice then received a 2.0 mg kg⁻¹ bolus of dextrose (Phoenix Pharmaceutical, St Joseph, MO, USA) via i.p. injection, and glucose concentrations were measured after 15, 30, 45, 60 and 120 min.

Combined indirect calorimetry

A 32-cage combined indirect calorimetry system (PhenoMaster, TSE Systems GmbH, Bad Homburg, Germany) was used to assess energy expenditure, locomotor activity, body core temperature, respiratory quotient as well as food and water intake. Mice were allowed to acclimatize to the air-tight cages for 16 h. Subsequently, volumes of oxygen consumption (Δvol%O₂) and volumes of CO₂ production (ΔVol% CO₂) were measured every 10 min for a total of six light and six dark phases (144 h) to determine the respiratory quotient (RQ = VCO₂/VO₂) and energy expenditure (EE = VO₂ × (3.815+(1.232 × (VCO₂/VO₂))) × 4.1868).¹⁶ Home-cage locomotor activity was determined by a multidimensional infrared light beam system. Stationary locomotor activity was defined as consecutive infrared light beam breaks of one single light beam and ambulatory movement as consecutive breaks of two different light beams. Scales integrated into the sealed cage environment continuously measured cumulative food intake. Mean body core temperature was monitored via i.p. E-mitter telemetry devices (E-mitters; Mini Mitter, Bend, OR, USA) which had been implanted 1 week previously, using isoflurane anesthesia and buprenorphine hydrochloride analgesia (Buprenex, single dose of 0.05 mg kg⁻¹ bodyweight). The mice were kept at a constant temperature of 23 °C for 72 h, housing temperatures were then increased to 33 °C until the end of the experiment.

Killing and tissue collection

Mice were killed by CO₂ anesthesia followed by heart puncture and blood collection. White and brown adipose tissue samples were excised, weighed and frozen on dry ice for subsequent sample processing.

Liver triglycerides quantification

Lipids were isolated from 50 mg of frozen livers of HFD- and LFD-fed nude mice using a chloroform/methanol (2:1) extraction procedure as described previously.^17,18 After evaporation of the organic solvent, total tryglyceride contents were determined by an enzymatic method (Infinity Triglycerides Reagent, Thermo Fisher Scientific, Middletown, VA, USA), according to the manufacturer’s instructions.

Gene expression analyses

RNA from epididymal white adipose tissue was isolated using the RNeasy Lipid Tissue mini kit (Qiagen, Valencia, CA, USA) and transcribed into cDNA by using the Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions. Quantitative real-time PCR reactions were performed in a ViiA 7 System (Applied Biosystems, Foster City, CA, USA) by using murine-specific TaqMan probe sets for leptin (Lep, Mm00434759_m1) and chemokine (C-C motif) ligand 2 (Mm00441242_m1). Relative expression levels were normalized to the housekeeping gene hypoxanthine guanine phosphoribosyl transferase 1 (HPRT1; Mm01545399_m1); expression changes were evaluated using the delta-delta Ct (ΔΔCt) method.

In vivo cytokine capture assay

Systemic TNF, IL-6, IL-17A, interferon (IFN)-γ and IL-10 levels were detected using the in vivo cytokine capture assay, an assay that integrates in vivo cytokine production over the 24-h period of challenge, as described before.^19,20 Briefly, systemic systemic cytokine levels were quantified by using biotinylated capture antibodies, detection antibodies and recombinant protein murine standards. Biotinylated capture mouse antibodies (TNF, clone TN3-19; IL-6, clone MP5-32C11; IL-17A, clone eBio17B7; IFN-γ, clone R4-6A3; and IL-10, clone JES5-16E3; all eBioscience, San Diego, CA, USA) were injected via tail vein 3 h prior to Toll-like receptor 4-specific LPS challenge (25 mg per mouse, ultrapure LPS, E. coli 0111:B4; InvivoGen, San Diego, CA, USA), and terminal serum collection was performed 24 h later. Detection antibodies (TNF, clone G281-2626; IL-6, clone MP5-20F3; IL-17A, clone eBio17CK15A5; IFN-γ, clone AN-18; and IL-10, clone JES5-2A5; all eBioscience) along with the appropriate standards were used for cytokine level quantification.

Statistics

All data are expressed as mean ± s.e.m. Body weight, food intake, metabolic performance and glucose excursions following the i.p. glucose tolerance test were analyzed via two-way analysis of variance (variables: treatment and time) with a Bonferroni multiple comparisons test. Energy expenditure was analyzed by analysis of covariance (ANCOVA) using body weight as covariate. Statistical differences between the surgical groups of all other measurements were analyzed using one-way analysis of variance followed by a Bonferroni post hoc test. Analyses were done using the GraphPad Prism 6.0 software (GraphPad Prism, San Diego, CA, USA). A P-value<0.05 was considered significant.

RESULTS

Nude mice on a C57BL/6 background are protected from diet-induced obesity when exposed to mild cold stress

Thirty-five days of HFD feeding at ambient temperature (23 °C) resulted in a significantly higher body weight gain in C57BL/6 WT mice compared with LFD-fed control mice (Figures 1a–b). In contrast, C57BL/6 nude mice were protected from HFD-induced obesity as they gained similar body weight compared with LFD-fed C57BL/6 nude mice (Figures 1c–d). Further, HFD-fed WT mice, but not nude mice, showed a significant increase in body fat compared with the LFD-fed controls (Figure 1e). Notably, lean mass, initially lower in nude mice, was not affected by diet in both strains (Figure 1f).

Figure 1.

Lack of diet-induced obesity in nude mice kept at room temperature. Comparative body weight data in C57BL/6 WT mice (a and b) and nude littermates (c and d) kept on 10% LFD or a 60% HFD, respectively. Total body weights in gram (a and c). Body weight gain in percentages (b and d). Comparative fat mass (e) and lean mass (f) measurements in C57BL/6 WT and nude mice at weeks 0, 2 and 4 of LFD or HFD feeding. N =5; mean ±s.e.m.; two-way analysis of variance with Bonferroni’s correction, *P<0.05; **P<0.01, ***P<0.001.

Increased energy expenditure in nude mice on a C57BL/6 background is restricted to normal housing temperatures

We next aimed to delineate whether the protection from diet-induced weight gain in nude mice at conventional housing conditions of 23 °C was the result of a decreased caloric intake, an increased energy dissipation due to an increased locomotor activity or thermogenesis or both.

Indirect calorimetry, followed by ANCOVA with body weight as covariate (Figures 2a and b), revealed a higher energy expenditure in HFD-fed nude mice compared with the WT controls when housed at 23 °C (ANCOVA, F = 6.093, P = 0.028). Such changes in energy expenditure were observed during both light (ANCOVA, F = 8.267, P = 0.013) and dark phases (ANCOVA, F = 4.645, P = 0.05) (Figure 2b). Notably, the difference in energy expenditure between WT and nude mice was lost after housing mice at a thermoneutral temperature of 33 °C (ANCOVA, F = 0.625, P = 0.443) (Figures 2a) and was not influenced by either the light phase (ANCOVA, F = 0.685, P = 0.423) or the dark phase (ANCOVA, F = 0.342, P = 0.568) (Figure 2b). We corroborated our finding of enhanced energy expenditure at a standard housing temperature of 23 °C in an inversed experimental set up, using an additional cohort of nude mice (Supplementary Material and Methods, Supplementary Figure S1A and Supplementary Table 1). Specifically, after an initial acclimatization period of at least 2 weeks at thermoneutrality, we conducted indirect calorimetry measurements in LFD-fed nude mice and four differentially aged groups of C57BL/6J WT mice. We neither observed differences in energy expenditure (Supplementary Figure S1A and Supplementary Table 1), food intake (Supplementary Figure S1B and Supplementary Table 1) or respiratory quotients (Supplementary Figure S1C and Supplementary Table 1) in mice fed LFD at 33 °C nor later in the same mice fed HFD at 33 °C. However, at mild cold exposure of 23 °C, a significant increase in energy expenditure (ANCOVA F = 5.58, P = 0.019) even after correction for the covariates body weight (F = 0.268, P = 0.619), fat mass (F = 0.454, P = 0.52), lean mass (F = 0.116, P = 0.742), body length (F = 0.962, P = 0.355) and age (F = 1.217, P = 0.302) was observed. Notably, such enhancement of energy expenditure was also apparent in linear regression plots of energy expenditure against lean mass, body weight and body length (Supplementary Figure S2).

Figure 2.

Conventional housing results in an increased energy expenditure in nude mice. (a) Significantly increased energy expenditure in nude vs WT mice at a housing temperature of 23 °C (P<0.05) but not at 33 °C. (b) Increased average energy expenditure in nude mice was present during both the dark and the light phase (P<0.05). (c) Similar locomotor activity in nude and WT mice at both housing temperatures. (d) Food intake was unchanged in both groups regardless of housing temperature. (e) Unchanged food intake following normalization to the significantly different body weights (f). (g) Similar respiratory quotient (RQ) after HFD feeding in both strains irrespective of housing temperatures. (h) Similar mean body core temperature during the light and dark phases at housing temperatures of 23 °C (left panel) and 33 ° C (right panel). (i) Higher percentage of brown adipose tissue (BAT) mass (normalized to their total body weights) in nude vs WT mice. N =8; mean ±s.e.m.; ANCOVA (a–g), Student’s t-test (f–i), *P<0.05; **P<0.01, ***P<0.001.

Differences in energy expenditure were independent of any changes in locomotor activity, as both WT and nude mice exhibited similar activity at both temperatures (Figure 2c). Thermoneutral housing reduced calorie intake in both genotypes to a similar degree (Figure 2d). WT and nude mice consumed 2.8 ± 0.41 and 2.3 ± 0.57 g of HFD at 23 °C (ANCOVA, F = 1.829, P = 0.206, body weight as covariate), and 1.9 ± 0.33 and 1.9 ± 0.34 g of HFD at 33 °C (ANCOVA, F = 0.733, P = 0.503, body weight as covariate), respectively. Similarly, no difference in food intake was observed following normalization to the respective body weights (Figure 2e), to correct for the significantly lower body weight of nude vs WT mice (Figure 2f). The respiratory quotient (RQ) after HFD feeding was 0.7 in both strains and at both temperatures (Figure 2g), suggesting that both nude and WT mice mainly used fat as a source of energy. Continuous body core temperature measurements at both housing temperatures demonstrate that both strains are able to effectively maintain homeothermia (Figure 2h). However, nude mice exhibited significant increase in the percentage of thermogenic brown adipose tissue compared with WT mice (Figure 2i). Together these data indicate that nude mice have to expend more energy to keep their body core temperature constant—something that correlates also with protection from weight gain.

Housing at thermoneutrality promotes diet-induced obesity in nude mice on a C57BL/6 background

The comparable energy expenditure at thermoneutrality between nude and WT mice suggests that exposure of C57BL/6 nude mice to HFD should lead to weight gain when housed at higher temperatures. To test this, nude mice were fed either LFD or HFD and housed at an increasing temperature gradient (30 or 33 °C). Body weights were similar in HFD and LFD mice during a feeding period of 30 days at 30 °C (Figure 3a). After increasing the temperature to 33 °C, HFD-fed nude mice showed a significant increase in total body weight compared with the LFD-fed controls (Figure 3a). Such increase in body weight directly correlated with an increase in visceral white adipose tissue consisting of epididymal and perirenal fat pads and subcutaneous inguinal fat pad (Figure 3b). Increased obesity in nude mice was also characterized by a higher leptin expression and adipose tissue inflammation, as defined by increased chemokine (C-C motif) ligand 2 expression (Figure 3c). The obese phenotype in HFD-fed nude mice was further accompanied by a small but significant decrease in their insulin tolerance (Figures 3d–e). Further, HFD-fed nude mice exhibited higher liver weights (Figure 3f), which was accompanied by a slight, however, not yet significant increase of hepatic triglycerides (Figure 3g), indicative of a development of an early stage of non-alcoholic fatty liver disease.¹⁹ Food intake was similar at both temperatures (23 and 33 °C), with HFD feeding leading to a significant increase in the cumulative caloric intake (Figure 3h). However, the higher caloric intake only translated into an obese phenotype at thermoneutrality.

Figure 3.

Diet-induced obesity in nude mice is determined by the housing temperature. (a) Body weights of LFD- and HFD-fed nude mice at two different housing temperatures (30 and 33 °C). Similar body weight gain during 30 days of 30 °C exposure, while HFD feeding resulted in a significant body weight gain at 33 °C. (b) Increased wet tissue mass of epididymal (e), retroperitoneal (RP) and subcutaneous (SC) fat in diet-induced obese nude mice housed at 33 °C. (c) Increased leptin and chemokine (C-C motif) ligand 2 expression in the epididymal fat mass of obese nude mice, normalized to the housekeeping gene HPRT1. (d) Blood glucose levels before (time 0) and after an i.p. glucose tolerance test indicate a significantly higher glucose excursions in obese vs lean mice. (e) Area under the curve for plasma glucose across the 2-h glucose tolerance test showed a significant increase in HFD-fed nude mice compared with the LFD-fed controls. (f) Significantly increased liver wet tissue mass in HFD-fed nude mice. (g) Increased liver triglycerides in HFD-fed nude mice. (h) Higher cumulative caloric intake in HFD- vs LFD-fed nude mice. N = 6–8; mean+s.e.m.; two-way analysis of variance (a, d, h), Student’s t-test (b, e–g). *P<0.05; **P<0.01, ***P<0.001.

The housing temperature affects systemic cytokine production in nude mice on a C57BL/6 background

Previous studies have shown that housing temperature can lead to broad (patho-) physiological changes in mice—something associated with significant modulation of immune responses.²¹ To elucidate possible effects of the housing temperature on the immune response in nude mice, LFD-fed nude mice were kept at housing temperatures of 23 and 33 °C for 12 weeks, before they were challenged with Toll-like receptor 4-specific LPS. Systemic cytokine production in nude mice was significantly higher, when housed at 33 °C than at 23 °C. Specifically, nude mice displayed higher production of pro-inflammatory cytokines, including TNF-α, IL-6, IL-17A and IFN-γ, (Figures 4a–d) as well as anti-inflammatory cytokine IL-10 (Figure 4e).

Figure 4.

Thermoneutral housing exacerbates LPS-driven cytokine production in nude mice. Significantly increased circulating levels of (a) TNF-α, (b) IL-6 (P<0.001), (c) IL-17A, (d) IFN-γ and (e) IL-10 in nude mice housed at 33 °C compared with 23 °C. (a–e) N =5–7; Student’s t-test; *P<0.05; **P<0.01; ***P<0.001.

DISCUSSION

The discovery of the Foxn1 nude mouse in 1962 and the subsequent development of human subcutaneous and orthotopic tumor xenograft models represented major breakthroughs in cancer research. Until today, nude mice remain powerful and clinically relevant animal models that allow for efficient screening for anti-cancer therapeutics.^9–11 We here aimed to assess whether nude mice on a C57BL/6 background could also be a valuable model to evaluate a causal relationship between diet-induced obesity and cancer. Foxn1 nude mice lack mature T cells and are unable to mount most CD4⁺ or CD8⁺ T cell-dependent immune responses, including graft rejection. The mutation further leads to an abnormal hair follicle morphology, and thus to a nude phenotype. Previous findings have demonstrated that even fur-carrying mice constantly loose heat at normal housing temperatures of 20–23 °C,²² which requires a continuous compensatory heat production in order to maintain euthermia.²³ Indeed, hair loss and reduced insulation in Foxn1 nude mice has been associated with higher energy dissipation compared with WT mice.²⁴ We here corroborate and expand those findings by showing that C57BL/6 nude mice, in contrast to WT C57BL/6 controls, demonstrate a significant increased energy expenditure and protection from diet-induced obesity when kept at normal housing temperatures of 23 °C. The increased energy expenditure was independent from several co-variants, such as body weight, fat mass, lean mass, body length and age. Accordingly, our data suggest that the different energy homeostasis depicted in C57BL/6 nude mice at 23 °C is a specific consequence of the Foxn1 mutation and not an artifact of normalization methods or differential body morphometry.

The increase in energy expenditure in nude mice kept at 23 °C was not the result of a higher locomotor activity, which was similar in both strains. Stable body core temperatures in nude and WT mice suggest that nude mice have a higher energy demand for thermogenesis at regular housing temperatures.

Importantly, at temperatures in the thermoneutral zone (33 °C), nude mice and WT controls significantly decreased their energy expenditure to a similar level. Consequently, nude mice housed at thermoneutrality were prone to the effects of HFD exposure and displayed a clear obese phenotype as well as glucose intolerance. A trend towards increased liver triglycerides further points toward an early stage of development of hepatic steatosis. Overall, our data clearly show that housing temperature is a major denominator for metabolic homeostasis in nude C57BL/6 mice. Housing mice at thermoneutral conditions may in fact reflect the human situations, who usually live at thermoneutral conditions.

The thermoneutral zone of a fur-bearing mouse is about 30 °C. Our nude mice were still protected from diet-induced obesity at this temperature. Only a temperature increase to 33 °C resulted in a significant weight gain due to HFD feeding, indicating that the thermoneutral zone in nude mice is even increased. Housing mice at 23 °C represents a constant mild cold stress, which may affect numerous physiological and metabolic processes.^25,26 Together, our data suggests that future studies—utilizing nude C57BL/6 mice and focused on the interplay between obesity and cancer—should be conducted at thermoneutrality.

Cold exposure normally leads to an increased food intake to compensate for the thermogenic loss of calories.^25,27 Nude mice, which suffer from cold stress even at regular housing temperatures, should hence display increased food intake at room temperature, compared with WT mice. However, even when food intake was normalized to the significantly lower body weights of nude mice, nude mice were not hyperphagic compared with WT controls. Such lack of compensatory food intake during mild cold stress may provide an additional explanation as to why nude mice but not WT mice are resistant to HFD-induced obesity at 23 °C but not at 33 °C.

Molecular roles for Foxn1 in food intake control have not been reported to date. Our data nevertheless point toward a central—potentially hypothalamic²⁸—role of Foxn1 in controlling ingestive behavior upon cold stimulation. Recent data showed higher neurotrophin and noradrenaline concentrations and increased density of noradrenergic fibers in the hypothalami of BALB/c nude mice housed at regular housing temperatures.²⁹ Activation of noradrenergic circuitry and enhanced hypothalamic–pituitary–adrenal (HPA) axis were previously linked with chronic cold stress in rats.³⁰ Accordingly, nude mice housed at mild cold stress of 23 °C may fail to compensate for the enhanced energy loss due a chronic activation of the HPA axis, and a subsequent dysregulation of physiological eating patterns. Future studies should elaborate on this potential link between cold stress-induced hyperactivation of the HPA axis and nude mouse metabolism and clarify whether Foxn1 has a functional role in feeding circuitry in the hypothalamus.

Recent findings demonstrated that antitumor immunity is significantly increased in murine allograft tumor models of a BALB/c and a C57BL/6 background when kept at housing temperatures of 30 °C compared with 22 °C.³¹ In contrast, increased housing temperature did not affect immunity against xenograft tumors in nude mice.³¹ In our study, C57BL/6 nude mice displayed HFD-induced weight gain only after raising housing temperatures to 33 °C. At 30 °C, our C57BL/6 nude mouse model remained resistant to detrimental effects of HFD exposure, indicating the essential role of correctly assessing the thermo-neutral zone for a specific animal model. Thermoneutral housing further augmented microbial ligand-driven activation of immune cell-driven cytokine production in our C57BL/6 nude mouse model. Specifically, C57BL/6 nude mice housed at 33 °C displayed significantly higher production of pro- and anti-inflammatory cytokines after an acute LPS challenge, compared with LPS-challenged nude mice housed at 33 °C. These data suggest that cold stress induces suppression of the LPS-driven cytokine production and that such effects are independent of T-cell response. Notably, LPS, a cell wall component of Gram-negative bacteria, is not only found in the blood stream of mice with acute or chronic bacterial infection but has also been implicated in other inflammation-related diseases, such as obesity and cancer. HFD feeding was shown to increase intestinal permeability and circulating LPS levels.¹⁹ The resulting increase in circulating LPS has been hypothesized to contribute to the development of body adiposity and insulin resistance.^32,33 On the other hand, pro-inflammatory LPS, —depending on the tumor type—can lead to a tumor-promoting³⁴ or -suppressing^35,36 effects. Importantly, cold stress-induced repression of immune cell responsiveness to LPS may therefore result in an underestimation of LPS effects on tumorigenesis and obesity.

Mice have been described as relatively insensitive towards bacterial endotoxins compared with humans.^37,38 Our findings suggest that such reduction in LPS sensitivity may not solely be explained by a species-specific difference but that cold stress-mediated impairment of myeloid cell activation may have a major role. Indeed, and in contrast to the in vivo situation, monocytes isolated from the blood of mice and humans exhibit similar in vitro responsiveness to LPS.^37,38

In summary, we demonstrate that thermoneutrality is a prerequisite to study metabolism in the cancer Foxn1 nude mouse model even when kept on the obesity-prone C57BL/6 background. In conventional housing, C57BL/6 nude mice, compared with C57BL/6 WT controls, are protected from diet-induced obesity and obesity-associated comorbidities due to increased energy expenditure and impaired feeding behavior. At thermoneutrality, however, C57BL/6 nude mice are prone to the detrimental effects of HFD exposure, making them a superior xenograft model to dissect potential mechanisms for obesity-induced carcinogenesis. Further augmentation of immune function in C57BL/6 nude mice housed at thermoneutrality may lead to novel discoveries on the role of immune system in obesity and cancer. In fact, future studies should directly and comprehensively elaborate on this potential impact of housing temperature on murine immune function and determine whether immune responsiveness in other mutant mouse models should also be examined at thermoneutrality.

Supplementary Material

Supplementary Material

Supplemental Figure 1: Conventional housing results in an increased energy expenditure in an additional cohort of nude mice with inverse experimental setup.

Indirect calorimetry measurements in LFD-fed C57BL/6J nude mice and four differentially aged groups of C57BL/6J WT mice (age: 3, 6, 9 and 12 months). We neither observed differences in energy expenditure (Suppl. Fig. 1A), food intake (Suppl. Fig. 1B) nor respiratory quotients (Suppl. Fig. 1C) in mice fed LFD at 33°C, nor later in the same mice fed HFD at 33°C. Decreasing housing temperature to 23°C revealed a superior increase in nude mice compared to the WT cohorts.

Supplemental Figure 2: Differential energy expenditure in HFD -fed nude mice at 23°C.

Average daily energy expenditure (EE) values were plotted against body weight, lean mass and body length in LFD-fed C57BL/6J nude mice and four differentially aged groups of C57BL/6J WT mice (age: 3, 6, 9 and 12 months). Regression lines indicate whether EE at different dietary conditions and temperatures correlates to the covariates lean mass (Suppl. Fig. 2A, D, G), body weight (Suppl. Fig. 2B, E, H) or body length (Suppl. Fig. 2C, F, I). Notably, all EE values, regardless of dietary condition, genotype, weight/age, lean mass and size, fit a linear regression at 33°C. At 23°C, EE values of nude mice fit a different regression compared to EE values of WT cohorts.

Suppl. Table 1: Body morphometry, age and energy expenditure: Average values ± SEM for age, body morphometry, energy expenditure (EE) and food intake of C57BL/6 Nude and WT mice. Statistical differences in body morphometry and age between Nude or WT mice were assessed by One-Way ANOVA and Dunnet’s post-hoc test with Nude mice as controls, and indicated as *p<0.05, #p<0.01, and §p<0.001.