Introduction

Modeling human physiology in animals has been essential to understand most biological processes and diseases, and to develop many therapeutics. Due to genetic conservation and ease of breeding, the laboratory mouse (Mus musculus) undeniably stands out as a preferred biological system for studying metabolic diseases such as obesity1,2,3. Researchers can easily control and standardize their genetic background and environmental conditions. However, with over 80% of candidate drugs ultimately failing in human trials4, many challenges exist in translating preclinical observations to clinical applications. While interspecies differences are often cited to explain this failure, emerging factors include unsuitable housing conditions2,3,5 and biased strain selection6,7,8.

Ambient temperature (Ta) is a critical environmental variable that profoundly influences metabolic phenotypes3,5,9. Standard facility temperatures (20–22 °C), chosen for human comfort and cost-efficiency, impose chronic cold stress on mice, thereby confounding experimental outcomes3,5,10. Consequently, there is increasing interest in conducting preclinical studies at thermoneutrality to better model human physiology. However, the exact thermoneutral zone in mice remains debated11,12. Thermoneutrality is defined as the range of Ta where minimal energy is expended to maintain core body temperature (Tb)9,13. Below the thermoneutral point (TNP), thermogenesis is activated via vasoconstriction, shivering, and non-shivering thermogenesis5,14. Above TNP, heat-dissipating processes, including vasodilatation, panting and sweating, dominate15. While ~ 30 °C is often cited as thermoneutral for mice14, this temperature may impair reproduction and other physiological functions12. Furthermore, TNP varies with sex, age, circadian phase, housing (group vs. single), cage type, and genetic background16,17,18,19,20.

At sub-thermoneutral temperatures, rodents rely on brown adipose tissue (BAT)-mediated non-shivering thermogenesis, triggered by norepinephrine (NE) acting on β-adrenergic receptors to stimulate uncoupling protein 1 (UCP1)21,22. Under cold or adrenergic stimulation, white adipose tissue (WAT) can also undergo browning, forming beige adipocytes with thermogenic capacity23,24,25. Cold exposure and β-adrenergic stimulation improve metabolic parameters in obese animals26. The C57BL/6 J (C57) strain has been extensively used to model human metabolic diseases6,27. However, the poor clinical translation of ADRB3 β3-adrenergic receptor (ADRB3) agonists highlights key species differences28,29,30. Human BAT thermogenesis appears to rely more on ADRB1 and ADRB2, with limited ADRB3 expression, limiting the relevance of ADRB3-targeted therapies31,32.

Beyond UCP1-dependent pathways, recent studies have identified UCP1-independent thermogenic mechanisms. These include creatine futile cycling 33,34, SERCA2b-mediated calcium cycling35, and mitochondrial uncoupling via N-acyl amino acids catalysed by PMD20D136,37. These alternative pathways also enhance energy expenditure (EE) and improve metabolic health, highlighting the therapeutic potential of BAT activation. Despite this promise, the translational relevance of such findings remains uncertain. Notably, C3H/HeJ (C3H) mice show distinct marrow adipose remodelling in response to cold and ADRB3 agonism38,39, but their systemic metabolic response to cold exposure has not been characterized.

Here we compared the metabolic adaptations to different Ta in male C57 and C3H mice. Despite similar TNPs, C3H mice exhibited greater EE and thermogenic gene induction during cold exposure. We observed divergent adrenergic receptor profiles in adipose tissues, and a blunted response to ADRB3 agonism in C3H mice, suggesting reliance on alternative pathways. These findings position C3H mice as a useful model for studying thermoregulation and metabolic adaptation with greater translational relevance to human physiology.

Methods

Animals

All experimental protocols were approved by the Animal Ethics Committee of Université Laval (CPAUL) and followed the guidelines of the Canadian Council on Animal Care. Seven-week-old male C57BL/6 J (C57, #000,664) and C3H/HeJ (C3H, #000,659) mice were obtained from The Jackson Laboratory. Mice were individually housed under a 12:12-h light/dark cycle (lights on 06:00 AM) with ad libitum access to chow (Teklad 2918) and water. Environmental enrichment was provided with nesting material (Bed-r’Nest®) and a Nylabone.

Cold exposure studies

Three independent cohorts of mice were studied:

Cohort 1 (short-term cold adaptation in metabolic cages)

After a 2-week acclimation at 30 °C, mice (n = 8 per group, initial body weight: C57, 25.96 ± 0.60 g; C3H, 25.50 ± 0.38 g) were gradually cooled (-1 °C/h for 20 h) to 10 °C and maintained at this temperature for 3 days in metabolic cages to assess energy metabolism.

Cohort 2 (long-term cold adaptation in metabolic cages)

A separate group of mice (n = 8 per group, initial body weight: C57, 27.00 ± 0.46 g; C3H, 25.25 ± 0.44 g) underwent the same 2-week acclimation at 30 °C and gradual cooling to 10 °C but were maintained at 10 °C for 14 days in metabolic cages.

Cohort 3 (acute cold exposure for molecular analyses)

To evaluate gene and protein expression, mice (n = 6–12 per group, initial body weight: C57, 24.65 ± 0.36 g; C3H, 25.52 ± 0.32 g) were acclimated at 30 °C for 2 weeks and then exposed to 10 °C for either 6 h (ZT2-ZT8) or 24 h (starting at ZT8) using conventional cages. Control mice remained at 30 °C.

Tissues including brown adipose tissue (BAT), inguinal white adipose tissue (iWAT), and epididymal white adipose tissue (eWAT) were collected at ZT8, snap-frozen, and stored at -80 °C. All mice were euthanized at ZT8 by cervical dislocation under isoflurane anaesthesia. The timing of tissue collection was chosen to ensure consistency across experiments and to control for circadian variability. Experiments with mice were performed in accordance with relevant guidelines and regulations, including the with ARRIVE guidelines (https://arriveguidelines.org).

Metabolic cage experiments

Oxygen consumption (VO2), carbon dioxide production (VCO2), respiratory exchange ratio (RER, computed as the ratio of CO2 production to O2 consumption) and energy expenditure were assessed by indirect calorimetry using a computer-controlled system (Promethion Metabolic Screening, Sable Systems International, Las Vegas, NV). The measurement of respiratory gases was conducted using an integrated setup comprising a fuel cell oxygen analyzer, a spectrophotometric CO2 analyzer, and a capacitive water vapor partial pressure analyzer (Sable Systems International). VO2 and VCO2 were measured at 7-min intervals. Food intake was monitored simultaneously.

ADRB3 agonism study

Mice acclimated at 30 °C for 2 weeks (n = 8 per group) were studied using a crossover design. Each animal received an intraperitoneal (i.p.) injection of either the selective β3-adrenergic receptor agonist CL 316,243 (0.1 mg/kg, dissolved in saline) or vehicle (saline), with a washout period of 72 h between treatments. Tissues including BAT, iWAT, and eWAT were collected 4 h post-injection at ZT8.

Noradrenaline (NE) and propranolol study

Mice were housed at 30 °C for 2 weeks prior to experimentation and studied using a crossover design (n = 8 per group). In the first condition, mice received a single subcutaneous injection into the interscapular region of either noradrenaline (NE; 0.1 mg/mL in 200 µL, corresponding to 0.8 mg/kg for a 25 g mouse, dissolved in saline) or vehicle (saline). This dose and route have been previously validated to induce robust BAT activation 40. Following a 2-week washout period, the same mice were treated with propranolol (10 mg/kg, i.p., dissolved in saline) or vehicle (saline), administered 30 min prior to NE injection (0.1 mg/mL in 200 µL, corresponding to 0.8 mg/kg for a 25 g mouse, dissolved in saline) or vehicle (saline). Tissues including BAT, iWAT, and eWAT were collected 4 h post-injection at ZT8.

Determination of the thermoneutral point (TNP)

TNP was defined as the ambient temperature (Ta) at which EE was minimized. EE values were plotted against Ta, and linear regressions were used to identify the intersection point for each mouse (Figure S1). Group means were then calculated. This analysis was performed using data from Cohorts 1 and 2, as both groups underwent a 2-week acclimation at 30 °C followed by a gradual decrease in Ta to 10 °C (n = 16 per strain).

Quantitative real-time PCR

Total RNA was extracted from BAT, iWAT and eWAT using the Monarch Total RNA Miniprep Kit (Bio-Rad, T2010S), according to the manufacturer’s instructions, including an on-column DNAse I digestion step to remove genomic DNA contamination. RNA concentration and purity were assessed by BioDrop Touch Duo spectrophotometry (Cambridge, UK). For each sample, 700 ng of total RNA was reverse transcribed into cDNA using qScript cDNA SuperMix (Quantabio, 95,047–100) in a 20 µL reaction, following manufacturer’s protocol. The resulting cDNA was diluted (1:15) in nuclease-free water before amplification. qPCR reactions were performed in duplicate in a 10 µL final volume using PerfeCTa SYBR® Green FastMix (Quantabio, 95,072–012) on a Bio-Rad CFX384 real-time PCR detection system. Each reaction contained 50 nM of forward and reverse primers (sequences provided in Supplementary Table 1). Amplification specificity was confirmed by melt curve analysis, agarose gel electrophoresis, and sequencing. No-template controls (NTC) were included for each primer pair. Gene expression was normalized to Hypoxanthine Phosphoribosyltransferase1 (Hprt), which was validated as stable across experimental conditions. Relative expression levels were calculated using the ΔΔCt method and are presented as fold change relative to C57BL/6 J mice housed at 30 °C.

Histology

BAT was fixed in 10% formalin (48 h, 4 °C), embedded in paraffin, and sectioned at 10 µm. Sections were stained with hematoxylin and eosin (H&E) and imaged at 40X magnification using an Olympus BX60 microscope.

Western blot

Adipose tissues were homogenized in lysis buffer containing (50 mM HEPES, pH 7.4, 2 mM EDTA, 10 mM sodium pyrophosphate, 10 mM sodium glycerophosphate, 40 mM NaCl, 50 mM NaF, 2 mM sodium orthovanadate, 1% Triton-X 100, 0.1% sodium lauryl sulfate, 1% sodium deoxycholate and one tablet of EDTA-free protease inhibitors Roche per 25 ml). Protein concentrations were measured by Bradford assay. Equal amounts of total protein (10 μg) for BAT and eWAT/iWAT (5 μg) per sample were denatured (95 °C, 10 min), separated on SDS-PAGE, and transferred to PVDF membranes. Blots were probed with primary antibodies against AKT (CST, 4691S), phospho-Ser473-AKT (CST, 4060), HSL (CST, 18,381), phospho-Ser563-HSL (CST, 4139), and UCP1 (CST, 14,670), at a dilution of 1:1000 according to the manufacturer’s recommendations, followed by HRP-conjugated secondary antibodies. Signal was detected using Amersham ECL reagent and band intensities were quantified using Image Lab software (Bio-Rad). Protein levels were normalized to total protein or AKT. Each lane represents an individual mouse. Uncropped and unprocessed images of the gels and western blots corresponding to Figs. 2C, 3C, and 3G are provided in Supplementary Fig. 6.

Plasma metabolite measurement

Blood glucose levels were measured from the tail vein using a glucometer (Roche, Accu-Chek Performa). Blood was collected by cardiac puncture in EDTA-conditioned syringes. Plasma was isolated by centrifugation (4000 g × 10 min at 4 °C) and was stored at -80 °C for further biochemical analyses. Non-esterified fatty acid (NEFA) (Fujifilm, NC9587, 999–34,691, 995–34,791, 991–34,891, 993–35,191) and glycerol (Sigma-Aldrich, F6428) levels were quantified in accordance with the manufacturer’s recommendations.

Statistical analysis

Data are expressed as mean ± SEM. All statistical analyses were performed using GraphPad Prism 10. Data normality was evaluated using the Shapiro–Wilk test. No outliers were identified, and all data points were included in the analyses. Depending on the dataset, statistical significance was assessed using paired t-tests, two- or three-way ANOVA with Bonferroni post hoc tests, or ANCOVA/regression analyses, as appropriate. A p value ≤ 0.05 was considered statistically significant. For metabolic parameters, a two-way ANOVA was performed with strain and cycle (light/dark) as factors within each temperature conditions, followed by multiple comparisons. For gene expression, two-away ANOVA was conducted with strain and temperature as factors was conducted, followed by multiple comparisons. For pharmacological interventions (ADRB3 agonist, NE, and propranolol followed by NE), two-way ANOVA was performed with strain and time as factors for each treatment.

Results

Metabolic adaptation to cold exposure differs between male C57BL/6 J and C3H/HeJ mice

To assess strain-specific metabolic adaptation, two independent cohorts of male C57 and C3H mice were acclimated at 30 °C for two weeks, followed by gradual (-1 °C/h) to 10 °C in Promethion metabolic cages (Fig. 1A). Consistent with previous reports 21, both strains gained weight after the acclimation period (C57: 25.96 ± 0.60 g at baseline vs. 28.35 ± 0.76 g post-acclimation; C3H: 25.50 ± 0.37 g vs. 28.63 ± 0.44 g; p < 0.0001 for time effect) (Fig. 1B). After 2 weeks at 10 °C, C3H mice weighted significantly less than C57 mice (C57: 28.25 ± 0.31 g vs. C3H: 24.91 ± 0.55 g; p = 0.0011, Fig. 1B).

Fig. 1
Fig. 1
Full size image

Metabolic adaptation to cold exposure differs between male C57BL/6 J and C3H/HeJ mice. (A) Ambient temperature (Ta). (B) Body weight at baseline and after cold exposure. (C) Food intake (g/h). (D) Average food intake (FI; g/cycle). (E) Respiratory exchange ratio (RER). (F) Average RER during the light and dark phases. (G) Energy expenditure (EE). (H) Total EE during the light and dark phases. (I) ANCOVA analysis of EE versus body weight. (J) Thermoneutral points (TNP, °C) for each strain. Data are presented as mean ± SEM. Statistical analyses were performed using two-way ANOVA with Bonferroni post-hoc tests, t-tests for pairwise comparisons, and ANCOVA. *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.0001; n = 16/group (30 °C and 10 °C), n = 8/group (10 °C, 2 weeks).

Food intake increased during the dark phase, reflecting the nocturnal behavior of mice, but no differences were observed between strains at 30 °C or during acute and chronic cold exposure (Fig. 1C-D). No significant differences in respiratory exchange ratio (RER) were observed between strains at 30 °C or 10 °C chronically. During acute cold exposure, RER decreased significantly in C3H compared to C57 (p = 0.0234), although C3H mice trended toward lower RER during the dark phase (p = 0.0612, Fig. 1E-F). EE increased in both strains under cold exposure, but with distinct patterns. At 30 °C, C3H mice had lower EE in the dark phase; however, during acute and chronic cold exposure, their EE became significantly higher in the light phase (Fig. 1G-H).

We then plotted individual EE and body weight data and performed an analysis of covariance 41. This revealed no EE difference at 30 °C, but C3H mice exhibited significantly higher EE during acute (p = 0.0002) and chronic cold exposure (p < 0.0001; Fig. 1I). TNP determination via linear regression yielded similar values for both strains (C57: 29.26 ± 0.28 °C; C3H: 29.46 ± 0.17 °C; Fig. 1J). Thus, despite similar TNPs, C3H and C57 mice display distinct cold-induced metabolic adaptations.

C3H mice show enhanced thermogenesis under cold exposure in the brown adipose tissue

We next analyzed thermogenic responses in BAT. After acclimation at 30 °C, mice were sacrificed either at 30 °C or following 6 h or 24 h at 10 °C. Gene expression analysis revealed that C3H mice displayed stronger cold-induced thermogenic program than C57 mice. In particular, Ucp1, Dio2, and Ppargc1a were more strongly induced in C3H mice (Fig. 2A). For example, Dio2 increased 334-fold in C3H versus 147-fold in C57 after 6 h of cold exposure. At baseline (30 °C), lipolytic genes Atgl and Hsl were higher in C3H, although these differences disappeared under cold conditions (Fig. 2B). In contrast, Mgll expression remained higher in C3H across temperatures (Fig. 2B). Protein analysis confirmed these transcriptional changes. UCP1 and phosphorylated AKT (Ser473) were significantly different in C3H mice after 6 h at 10 °C, whereas phosphorylated HSL (Ser563) did not differ between strains (Fig. 2C-D).

Fig. 2
Fig. 2
Full size image

Strain-specific brown fat activation in response to acute cold. (A) Expression of thermogenic genes (Ucp1, Dio2, Ppargc1a). (B) Expression of lipolytic genes (Atgl, Hsl, and Mgll). (C) Protein expression in BAT after 6 h of cold exposure. (D) Quantification of UCP1, pAKT and p-HSL. (E) Expression of creatine pathway gene (Ckb, Ckm and Gatm). (F) Expression of Pmd201 and Fgf21. (G) H&E-stained BAT sections from both strains. Data are presented as mean ± SEM. Statistical analyses were performed using two-way ANOVA with Bonferroni post hoc tests. *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.0001; n = 6–12/group.

To explore UCP1-independent mechanisms, we measured genes involved in creatine cycling and alternative thermogenesis. Ckb and Ckm were higher in C57 mice, while Gatm and Pm20d1 were higher in C3H at 30 °C (Fig. 2E-F). Notably, Fgf21 was consistently higher in C3H mice (Fig. 2F). Morphological differences in BAT further supported enhanced thermogenesis in C3H (Fig. 2G). Despite this, C3H mice had significantly less BAT mass than C57, independent of temperature (p < 0.0001, Figure S2A). Expression of Cpt1b increased in both strains upon cold exposure (p = 0.0043) but remained significantly higher in C3H overall (p < 0.0001, Figure S2B). Prdm16 and Cidea were also elevated in C3H compared to C57, with no additional effect of temperature (Figure S2C-D).

Circulating glycerol levels were higher in C3H than in C57 (p = 0.0024, Figure S3A), while non-esterified fatty acids (NEFA) levels in plasma did not differ between strains (Figure S3B). Blood glucose increased significantly in both strains during cold exposure (Figure S3C).

Strain-specific activation of white fat in response to acute cold

Both strains displayed similar iWAT mass, with no significant effect of temperature (Figure S4A). Cold exposure increased Ucp1 in both strains, but Dio2 and Ppargc1a were more strongly induced in C57 (Fig. 3A). Atgl increased after 6 h of cold in both strains, while Hsl remained unchanged (Fig. 3B). Mgll was higher in C3H under cold conditions (Fig. 3B). At the protein level, UCP1 was significantly higher in C3H after 3 days at 10 °C, while p-HSL tended to be higher in C57 (Fig. 3C-D). Additional markers of oxidative capacity were also induced. Cpt1b expression increased after 24 h of cold exposure in both strains (p = 0.0001, Figure S4B), while Prdm16 and Cidea were elevated after 24 h of cold exposure (p = 0.0386 and p = 0.0001, respectively; Figure S4C-D).

Fig. 3
Fig. 3
Full size image

Gene expression profile of white adipose tissue (WAT) in both strains. (A) Expression Ucp1, Ppargc1a, and Dio2 in iWAT. (B) Expression of Atgl, Hsl and Mgll in iWAT. (C) Protein expression in iWAT after 3 days of cold exposure. (D) Quantification of UCP1 and p-HSL protein expression in iWAT. (E) Expression of Ucp1, Ppargc1a, and Dio2 in eWAT. (F) Expression of Hsl, Atgl, and Mgll in eWAT. (G) Protein expression in eWAT. (H) Quantification of UCP1 and p-HSL protein levels in eWAT. Data are presented as mean ± SEM. Two-way ANOVA with Bonferroni post hoc tests were used. *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.0001; n = 6–12/group.

In eWAT weight, both strains displayed similar depot mass across conditions (Figure S4E), but thermogenic responses were markedly different. C3H mice exhibited a robust induction of Ucp1 after cold exposure, with increases of 47-fold at 6 h and 237-fold at 24 h, whereas no strain differences were observed in Ppargc1a or Dio2 (Fig. 3E). Hsl decreased with cold exposure, particularly in C3H, while Atgl and Mgll remained higher in C3H across conditions (Fig. 3F). Protein analyses confirmed a significant increase of UCP1 in C3H eWAT (Fig. 3G-H), a notable finding given that eWAT is generally resistant to browning compared to iWAT. Consistent with these results, Cpt1b expression was higher in C3H than in C57 mice regardless of temperature and further increased with cold exposure (Figure S4F). Similarly, Prdm16 and Cidea were elevated in C3H compared to C57 (Figure S4G-H).

C3H mice have blunted response to the ADRB3 agonist CL 316,243 and NE

To investigate adrenergic mechanisms, we first profiled β-adrenergic receptor expression across adipose depots. In BAT and iWAT, Adrb3 expression was nearly undetectable in C3H and was not altered by cold, whereas Adrb1 was modulated by cold in BAT but unchanged in WAT (Fig. 4A-B). In eWAT, Adrb1 expression was significantly higher in C3H, while Adrb3 remained undetectable (Fig. 4C). Adrb2 expression decreased significantly in BAT with cold exposure and was higher at 30 °C in C57 mice compared to C3H (p = 0.0493, Figure S5A). In iWAT and eWAT, Adrb2 expression also declined with temperature, with no significant strain differences (Figure S5B-C). Together, these findings suggest a limited contribution of ADRB3 to cold-induced thermogenesis in C3H mice.

Fig. 4
Fig. 4
Full size image

Pharmacological stimulation of β-adrenergic pathways in C57 and C3H mice. (A-C) Expression of Adrb1 and Adrb3 in BAT (A), iWAT (B), and eWAT (C). (D) Experimental design for crossover administration of CL 316,243. (E) EE following CL 316,243 injection (0.1 mg/kg, i.p.). (F) Average EE over 4 h. (G) RER following CL 316,243 injection. (H) Experimental design for crossover NE and propranolol treatment. (I) EE following NE or propranolol + NE injection. (J) Average EE over 4 h. (K) RER in all treatment groups. Data are shown as mean ± SEM. Statistical analyses were performed using two-way or three-way ANOVA with Bonferroni post hoc tests. *p ≤ 0.05, **p ≤ 0.01 and ***p ≤ 0.0001; n = 6–12/group for A-C; n = 8/group for D-K.

To test this hypothesis, we conducted a pharmacological experiment using the selective β3-adrenergic agonist CL 316,243. Mice acclimated to 30 °C received CL 316,243 (0.1 mg/kg) or saline in a crossover design (Fig. 4D). In C57 mice, CL 316,243 induced a sustained increase in EE, whereas in C3H the effect was modest and transient (Fig. 4E-F). RER decreased in both strains following injection but the reduction was more pronounced in C57 (p = 0.0181; Fig. 4G).

We next assessed the thermogenic response to NE, a non-selective β1/β2/β3-adrenergic receptor agonist (Fig. 4H). NE injection increased EE in both strains, but the effect was stronger in C57 (p = 0.0281; Fig. 4I-J). Pre-treatment with propranolol, a non-selective β-adrenergic antagonist with higher affinity for β1/β2, attenuated the EE response in both strains and abolished the difference between them (p = 0.1185; Fig. 4J). Under these conditions, no differences in RER were observed (Fig. 4K).

These results demonstrate that C3H mice have a reduced thermogenic response to ADRB3 agonism, consistent with their low Adrb3 expression. Their response to NE is less robust than in C57 mice and appears to rely on β1/β2 receptors. This pharmacological profile aligns with our earlier findings that C3H mice activate BAT and especially eWAT through alternative, ADRB3-independent pathways, distinguishing their thermogenic strategy from that of C57 mice.

Discussion

Targeting metabolic processes in adipose tissue using sympathomimetics has been actively pursued for many years as a potential treatment for obesity. Since the discovery of ADRB342,43,44,45, many compounds were developed with the hope to cure metabolic diseases. However, despite promising results in rodents, ADRB3-targeted therapies have largely failed in clinical trials. But what is to blame for this translation failure?

It is now clear that ambient housing temperature can significantly alter metabolic phenotypes in mice, with standard facility temperatures (~ 22 °C) imposing chronic cold stress and elevating basal metabolic rate2,3,5,12,17,19. To better reflect human physiology, thermoneutral housing (~ 30 °C) has been recommended3,5,9. Another key limitation is the widespread use of C57 mice, which differ markedly from humans in adipose adrenergic receptor expression. In humans, BAT thermogenesis is predominantly mediated via ADRB1 and ADRB2, not ADRB331,32, questioning the relevance of C57 mice in modeling human brown fat function.

Here we compared the metabolic responses to cold exposure in C57 and C3H mice. Metabolic comparisons have previously been done between different strains of mice. While previous studies have documented metabolic differences between other strains (e.g., 129/Sv, FVB/NJ, A/J)46,47, this is the first to characterize how C3H and C57 mice respond to varying Ta. Although C57 mice are widely used in metabolic research due to their susceptibility to diet-induced obesity48, C3H mice are more often used in cardiovascular and infectious disease studies. Notably, both strains exhibited similar TNPs during the light phase (~ 29.3 °C), consistent with previous reports19. The theoretical value of thermoneutrality has been highly debated over the past years11,12. We previously adopted the 30 °C value in our experiments aimed at reducing the baseline thermogenic activity of BAT to study its response to cold exposure and sympathomimetics in mice24,25,49,50. The current study confirms that both C57 and C3H mice have minimal EE when housed at 29–30 °C in the daytime. However, our experimental design could not determine the thermoneutral point during the dark phase, which was suggested to be ~ 4 °C higher in C57 mice19.

Despite similar TNPs, C3H mice displayed markedly different cold-induced adaptations. EE increased ~ 2.5-fold in both strains at 10 °C, but C3H mice showed significantly higher induction of thermogenic genes (Ucp1, Dio2, Ppargc1a) and UCP1 protein in BAT. Phosphorylated AKT levels also increased in C3H, consistent with prior reports linking AKT activation to BAT thermogenesis50. Despite this enhanced thermogenic activation, C3H mice had significantly less BAT mass than C57 mice across all temperatures. Markers of oxidative capacity were also elevated in C3H BAT: Cpt1b expression was higher overall and increased with cold exposure, while Prdm16 and Cidea were significantly elevated in C3H compared to C57, independent of temperature. After 24h of cold exposure, Cpt1b, Prdm16, and Cidea remained significantly induced, indicating sustained activation of oxidative and thermogenic pathways. These findings suggest an enhanced thermogenic capacity in C3H mice, corroborating observations in marrow adipose tissue remodeling39, although the role of marrow fat in thermogenesis remains uncertain51. Thus, our study provides the first direct evidence of robust BAT activation in C3H mice during cold exposure.

In WAT, C3H mice also displayed enhanced thermogenic responses, particularly in eWAT, where UCP1 protein expression increased significantly. While iWAT is typically the site of cold-induced browning, our data reveal that eWAT in C3H mice exhibits a remarkable thermogenic potential, even after short-term cold exposure. Consistent with this enhanced browning response, Cpt1b expression was higher in C3H eWAT across conditions and further increased with cold exposure, while Prdm16 and Cidea were also significantly elevated in C3H regardless of temperature. CIDEA is classically associated with browning and thermogenic programming in rodents52 and studies in human adipocytes demonstrate that CIDEA regulates UCP1 and promotes a beige-like thermogenic phenotype53. Nonetheless, despite evidence of browning in eWAT, UCP1 protein abundance remains substantially higher in BAT than in any white adipose depot, consistent with previous reports25.

Cold exposure also altered circulating metabolites. In particular, plasma glycerol levels were significantly higher in C3H mice than in C57, whereas NEFA levels did not differ between strains. Blood glucose increased significantly in both strains during cold exposure. These results suggest differential mobilization of lipid substrates between strains and may reflect the greater thermogenic demand of C3H mice.

One of the most striking findings was the differential expression and responsiveness of β-adrenergic receptors. In C3H mice, Adrb3 was virtually undetectable in BAT and WAT, and unaffected by cold exposure. This is surprising given its known role in driving thermogenesis in mice54,55,56. Conversely, Adrb1 was upregulated in certain depots, suggesting a potential compensatory mechanism. Adrb2 expression also changed with temperature. While in BAT, Adrb2 decreased with cold exposure and was higher at 30 °C in C57 compared to C3H, its expression in iWAT and eWAT similarly declined with decreasing temperature, with no significant strain differences. These results align with human studies showing minimal ADRB3 involvement in brown adipocyte metabolism32,57, positioning C3H mice as a potentially more relevant model for studying human-like thermogenic mechanisms.

Pharmacological validation with CL 316,243, a selective ADRB3 agonist, revealed strong and sustained EE induction in C57 mice, but only a modest, transient effect in C3H mice. This further supports the limited role of ADRB3 in C3H thermogenesis. NE, which activates all β-adrenergic receptors58, elicited EE responses in both strains, but the effect was weaker in C3H. Consistent with previous reports59, we also observed that the induction of EE was higher following CL 316,243 compared to NE. Importantly, pre-treatment with propranolol (a non-selective β1/β2 antagonist) abolished strain differences, indicating that these responses are mediated through β- β -adrenergic signaling. Taken together with the blunted CL 316,243 response and lower Adrb3 expression in C3H mice, these findings suggest that the enhanced EE observed in C3H during cold exposure is not driven by classical β3-adrenergic signaling, but rather by β1/β2-dependent mechanisms and/or additional non-canonical thermogenic pathways. This shift in receptor reliance may reflect differences in adrenergic receptor expression and transcriptional programs of thermogenic genes, highlighting that mechanisms beyond β3 stimulation sustain strain-dependent thermogenic responses. Although the use of selective β1 or β2 agonists would better delineate the specific pathways involved, such agents lack the specificity and widespread validation that CL 316,243 provides for ADRB3 in rodent studies. Nonetheless, our findings strongly suggest that C3H mice rely on alternative β-adrenergic pathways, more similar to those in humans31,32, for cold-induced thermogenesis.

Conclusion

In conclusion, our findings suggest that metabolic phenotyping in male C57BL/6 J and C3H/HeJ mice should be conducted at ~ 29–30 °C during the light phase to accurately reflect thermoneutral conditions. Although both strains exhibit similar TNP, they display markedly different metabolic responses to cold exposure. These differences are driven in part by strain-specific variations in adrenergic receptor expression in BAT and WAT. C3H mice exhibited a blunted thermogenic response to the selective ADRB3 agonist CL 316,243 and low Adrb3 expression, yet maintained robust cold-induced thermogenesis. This phenotype was abolished by β1/β2 blockade, suggesting reliance on alternative adrenergic signaling pathways. These findings identify C3H mice as a valuable model for investigating ADRB3-independent, human-relevant mechanisms of thermoregulation and metabolic adaptation.