One major factor affecting physiology often overlooked when comparing data from animal models and humans is the effect of ambient temperature. The majority of rodent housing is maintained at ~22°C, the thermoneutral temperature for lightly clothed humans. However, mice have a much higher thermoneutral temperature of ~30°C, consequently data collected at 22°C in mice could be influenced by animals being exposed to a chronic cold stress. The aim of this study was to investigate the effect of housing temperature on glucose homeostasis and energy metabolism of mice fed normal chow or a high-fat, obesogenic diet (HFD). Male C57BL/6J(Arc) mice were housed at standard temperature (22°C) or at thermoneutrality (29°C) and fed either chow or a 60% HFD for 13 weeks. The HFD increased fat mass and produced glucose intolerance as expected but this was not exacerbated in mice housed at thermoneutrality. Changing the ambient temperature, however, did alter energy expenditure, food intake, lipid content and glucose metabolism in skeletal muscle, liver and brown adipose tissue. Collectively, these findings demonstrate that mice regulate energy balance at different housing temperatures to maintain whole-body glucose tolerance and adiposity irrespective of the diet. Despite this, metabolic differences in individual tissues were apparent. In conclusion, dietary intervention in mice has a greater impact on adiposity and glucose metabolism than housing temperature although temperature is still a significant factor in regulating metabolic parameters in individual tissues.
Thermoneutrality refers to the temperature at which the energy expenditure required to maintain core body temperature is the lowest (Lodhi & Semenkovich 2009). In clothed humans, this temperature is between 15°C and 25°C (Kingma et al. 2014) and most humans live in environments where engineering solutions (e.g. heating or cooling) have been used to achieve this thermoneutral temperature. Because of this, it is not unexpected that the standard housing temperature for animal research facilities is usually 20–24°C, a temperature range where clothed humans are most comfortable. This environment however, may subject the rodents to mild but chronic cold stress because the thermoneutral temperature for rodents is closer to 30°C (Cannon & Nedergaard 2011). Thus, the translation of the results from rodent studies to human situations could be compromised if the results from animal models actually reflect a state of thermal stress.
The housing temperature of mice has been reported to affect a wide range of physiological parameters including heart rate and blood pressure (Swoap et al. 2008, Overton 2010), recruitment of immune cells (Tian et al. 2016, Giles et al. 2017) and potentially tumour biology (Hylander & Repasky 2016). It has also been shown that mice housed at 22°C have substantially increased energy expenditure (Cannon & Nedergaard 2011) and increased food intake (Fregly et al. 1957) compared to mice housed at temperatures between 29°C and 32°C. Interestingly, although the relationship between energy expenditure and intake is very clear (Cannon & Nedergaard 2011), there have been a wide range of reported effects of housing temperature on fat mass and glucose metabolism. These include reports that HFD-fed mice housed at slightly different temperatures defined as thermoneutrality have increased adiposity (Cui et al. 2016, Giles et al. 2016), redistribution of adipose tissue between adipose depots (Uchida et al. 2010) or no change in adiposity (Tian et al. 2016, Giles et al. 2017). Similarly, investigations into how housing temperature affects glucose tolerance in mouse models have yielded variable results. These include reported improvements in glucose tolerance of mice held at 25°C compared to 20°C (Uchida et al. 2010), thought to be due to an increase in insulin secretion, as well as no change in mice held at thermoneutrality compared to standard housing temperatures (Rippe et al. 2000, Cui et al. 2016, Tian et al. 2016).
Because of the inconsistencies in the literature, the current study was conducted to investigate the effects of thermoneutral housing on energy balance and glucose metabolism at a whole-body and tissue level in mice in the context of both a standard chow diet and a HFD. In particular, we were interested in the energy metabolism pathways of the liver and muscle, the two largest contributors to whole-body glucose homeostasis (Kowalski & Bruce 2014), as well as the brown adipose tissue, known to be important for thermogenesis in mice. We hypothesised that housing mice at thermoneutrality would increase lipid content in liver and skeletal muscle and increase adiposity and glucose intolerance in mice.
All experimental procedures performed were approved by the Garvan Institute/St Vincent’s Hospital Animal Ethics Committee and were in accordance with the National Health and Medical Research Council of Australia’s guidelines on animal experimentation. Male C57BL/6J(Arc) mice (8 weeks of age) were sourced from Animal Resource Centre (Perth, Australia) and were randomly allocated to two separate, adjacent rooms (one at 22°C and one at 29°C) in the same animal facility and were acclimatised to the different temperature rooms for 2 weeks prior to dietary intervention. Mice were communally housed (4 per cage) in temperature-controlled (22°C ± 0.5°C standard housing or 29°C ± 0.5°C for thermoneutral studies) and light-controlled (12 h light:12 h darkness cycle, 07:00–19:00 light) rooms with corn cob bedding and received either a standard chow diet (6% fat, 23% protein and 71% carbohydrate, by calories; Gordon Specialty feeds, Sydney Australia) or a lard-based, high-fat diet (HFD; 60% fat, 20% carbohydrate, 20% protein, by calories) made inhouse (based on Research Diets formula #D12492), which was available ad libitum for a period of 13 weeks.
Assessment of body composition, respirometry and energy intake
Lean and fat mass were measured using an EchoMRI-500 (EchoMRI LLC, Houston, USA) according to the manufacturers’ instructions, excluding body water, after 3, 7 and 12 weeks of diet. Whole-body respirometry was performed on mice after 7 weeks of diet utilising the Promethion metabolic system (Sable Systems International, North Las Vegas, USA). Mice were individually housed and acclimatised for 2 days in Promethion cages at either 22°C or 29°C. Food and water were available ad libitum and bedding was the same as in the home cage. The airflow of each chamber was 2 L/min and O2 and CO2 measurements were taken every 5 min across a 48- to 72-h period and was averaged for every half hour. Activity measurements were calculated by adding x and y axis beam breaks for each cage for a 24-h period. Data were analysed by the ExpeData software package (Sable Systems International). Energy intake measurements were performed on mice after 7 weeks of diet by the daily weighing of food hoppers and food spillage in communally housed cages and was averaged to account for multiple mice per cage. Energy density of the food was determined as 13.0 kJ/g for chow and 14.5 kJ/g for HFD.
Glucose tolerance test
Oral glucose tolerance tests (oGTT) were carried out after 8 and 13 weeks of dietary intervention. Mice were fasted for 6 h (food removed at 8 am, oGTT started at 2 pm), as suggested by Andrikopoulos et al. (Andrikopoulos et al. 2008). A fixed dose of 50 mg of glucose (200 µL of 25% glucose solution in water) was gavaged by laryngeal cannula. This corresponded to a dose of 2 g glucose/kg lean mass for a mouse with 25 g of lean mass. Blood glucose levels were monitored from the tail-tip using a hand-held glucometer (Accuchek Performa, Roche) before, and for 90 min following glucose administration. Insulin levels during the oGTT were measured using a mouse ultra-sensitive, ELISA kit (Crystal Chem, Elk Grove Village, USA) in samples of whole blood collected from the tail.
In vivo 2-deoxyglucose accumulation into tissue
Mice in the postprandial period (08:00) were administered an intraperitoneal bolus of 5 µCi [1,2 3H]2-deoxyglucose ([3H]2-DG, Perkin Elmer) in 0.1 g/kg lean mass glucose while they were at their normal housing temperatures. These animals were unrestrained and in the conscious state. Ninety minutes after trace glucose administration, mice were euthanized by cervical dislocation and tissues were snap frozen and stored at −70°C for later analysis. Relative tissue glucose uptake was estimated by measuring the amount of [3H]2-DG tracer that had accumulated in tissues as [3H]2-DG-6-phosphate at cull. Powdered tissue was homogenised in 1 mL of water and then centrifuged at 17,000 g for 10 min at 4°C. 400 µL of supernatant was added to scintillation fluid (Ultima Gold, Perkin Elmer) for measurement of total counts and another 400 µL was deproteinised with the addition of 200 µL of 0.3N ZnSO4 and 0.3N Ba(OH)2 and then vortexed and centrifuged at 17,000 g for 10 min at 4°C to precipitate [3H]2-DG-P. Four hundred microliters of supernatant were then separated into another scintillation vial for measurement of free counts on a liquid scintillation analyser (Tri-Carb 2800TR, Perkin Elmer). The difference between ‘total’ and ‘free’ disintegrations per minute (dpm) was calculated by subtraction and corrected for volume and weight of tissue to determine [3H] 2-deoxy-glucose-6-phosphate accumulation as dpm/mg tissue.
Plasma insulin, adiponectin and leptin levels were determined by ELISA (mouse, ultra-sensitive, Crystal Chem). Plasma non-esterified fatty acids (NEFA) were determined by NEFA kit (Wako Diagnostics, Mountain View, USA). Plasma alanine aminotransferase (ALT) and aspartate transaminase (AST) were measured spectrophotometrically at 30°C using commercial enzyme/substrate reagents (Infinity ALT reagent, Infinity AST reagent, Thermo Fisher Scientific). Tissue triglyceride and glycogen content were measured as described previously (Hoy et al. 2009). Enzymatic assay of β-hydroxyl-CoA dehydrogenase (β-HAD) in muscle was performed as previously described (Turner et al. 2007).
Immunoblotting was conducted as previously described (Brandon et al. 2015). Antibodies for acetyl-CoA carboxylase (ACC), fatty acid synthase (FAS) and stearoyl-CoA desaturase 1 (SCD1) were from Cell Signalling Technologies, anti β-actin-HRP and anti-pan 14-3-3 were from Santaz Cruz, anti-phosphoenolpyruvate carboxykinase (PEPCK) was from Cayman Chemical Company, anti-UCP1 was from Alpha Diagnostics, anti-UCP3 was from Affinity BioReagents (Thermo Fisher Scientific) and total OXPHOS Rodent WB Antibody Cocktail (Complex I-V) was from Abcam. Protein content was determined by the Bradford assay and compared to a BSA standard.
A piece of liver and interscapular brown adipose tissue (BAT) were fixed in 10% formalin solution for 24 h and then transferred to a 70% v/v ethanol solution for storage. Fixed sections were embedded in paraffin blocks and transverse 4 μm sections were cut and stained with haematoxylin and eosin (H & E). Embedding, cutting and staining were performed by the Garvan Institute Histopathology Core service. Pictures were taken using the Leica DMIL light microscope with a MC120 HD camera (Leica Microsystems) at 40× magnification for BAT and 20× magnification for liver. Pictures were processed using Leica application suite software (version 4).
Data are expressed as means ± s.e.m. Results were analysed by 2-way ANOVA for a main effect of temperature and a main effect of diet. Sidak’s post hoc test was used to determine statistical difference between temperatures for each diet. Statistical analysis was performed in GraphPad Prism software (Prism 7). Statistical significance was set at P < 0.05.
Thermoneutral housing decreases energy intake and energy expenditure without changing body mass or composition
After 13 weeks of feeding, HFD mice were significantly heavier than mice fed a standard chow diet, however, housing temperature did not alter the final body mass or the rate at which mice gained weight (Fig. 1A). Similarly, HFD animals had a substantially larger fat mass than chow animals measured by EchoMRI, (Fig. 1B), as well as larger fat depot weights after euthanasia (Table 1). Again, there was no effect of housing temperature on fat mass. Lean mass was not affected by temperature or diet (Fig. 1C). Oxygen consumption of mice housed at 29°C was significantly lower than 22°C mice (~30%, Fig. 1D and E); however, there was no significant effects of diet on oxygen consumption. Conversely, RER, a measure of substrate oxidation, was significantly lower in HFD mice compared to chow, indicating increased oxidation of fatty acids at the expense of carbohydrate, and housing temperature had no effect on RER (Fig. 1F and G). Food intake measurements showed that both diet and temperature influenced energy intake (Fig. 1H). HFD mice displayed a higher energy intake compared to chow-fed mice, in line with expectations (Fig. 1H). Mice that were housed at 29°C had a lower energy intake compared to 22°C, irrespective of their diet. The activity of mice was similar in animals housed at the two different temperatures; however, HFD mice had significantly reduced activity compared to chow controls (Fig. 1I).
Fat pad weights and blood and plasma metabolites.
|Housing temperature||Chow||HFD||Effect of temp||Effect of diet|
|Fat pad weights|
|EpiWAT (g)||0.46 ± 0.05||0.49 ± 0.04||1.66 ± 0.15||1.57 ± 0.14||ns||***|
|SubcutWAT (g)||0.30 ± 0.03||0.34 ± 0.03||1.12 ± 0.11||1.05 ± 0.12||ns||***|
|BAT (g)||0.098 ± 0.007||0.157 ± 0.012||0.158 ± 0.010||0.193 ± 0.025||***||***|
|Blood and plasma metabolites|
|Fasting blood glucose (mM)||8.1 ± 0.5||7.9 ± 0.2||10.9 ± 0.3||9.8 ± 0.5||ns||***|
|Plasma TAG (mM)||1.12 ± 0.11||0.98 ± 0.09||0.77 ± 0.07||0.77 ± 0.04||ns||**|
|Plasma NEFA (mM)||0.25 ± 0.03||0.30 ± 0.04||0.18 ± 0.01||0.29 ± 0.03||**||ns|
|Plasma insulin (mU/L)||10.3 ± 0.7||14.7 ± 0.9||36.1 ± 8.6||53.3 ± 8.7||ns||***|
|Plasma adiponectin (µg/mL)||5.5 ± 0.3||6.7 ± 0.3||7.4 ± 0.3||8.3 ± 0.4||**||***|
|Plasma leptin (ng/mL)||5.3 ± 1.11||6.3 ± 3.1||46.2 ± 4.4||45.5 ± 5.4||ns||***|
|Plasma AST (U/L)||27.8 ± 2.3||28.9 ± 3.8||28.8 ± 3.2||24.2 ± 3.3||ns||ns|
|Plasma ALT (U/L)||13.3 ± 1.4||13.5 ± 2.1||13.0 ± 2.3||12.7 ± 1.5||ns||ns|
Data are means ± s.e.m. Fasting blood glucose was determined after a 6 h fast. Plasma parameters were measured from a heart puncture after sacrifice, in the postprandial period. Analysed by 2-way ANOVA for an effect of temperature and an effect of diet. n = 12. **P<0.005, ***P<0.0005.
No effect of temperature on glucose tolerance, fasting glucose and insulin
HFD mice displayed an elevated glucose excursion during an oGTT after both 8 (Fig. 2A and B) and 13 (Fig. 2E and F) weeks of diet, indicating glucose intolerance compared to chow controls. This occurred despite an increase in plasma insulin levels in HFD mice during the oGTT (Fig. 2C, D, G and H). Housing temperature did not affect glucose tolerance or plasma insulin levels in either chow or HFD mice during the oGTT. HFD mice had increased fasting blood glucose, decreased plasma TAG and increased levels of insulin, adiponectin and leptin (Table 1). Mice housed at 29°C had significantly increased plasma adiponectin levels compared to 22°C mice as well as significantly increased plasma NEFA. There was no significant difference in the activity of plasma AST or ALT between diets or housing temperatures (Table 1).
Thermoneutral housing decreased UCP1 abundance and basal glucose uptake in BAT
Further investigations were carried out to assess alterations in energy metabolism in metabolically important tissues (BAT, liver and muscle). As expected, UCP1 abundance was decreased in the BAT of mice housed at thermoneutrality (Fig. 3A and B). Interestingly, HFD mice had increased UCP1 abundance in BAT compared to chow mice. 2-DG accumulation into the BAT mirrored UCP1 abundance and was significantly decreased in thermoneutral mice (Fig. 3C). Abundance of the subunits of the mitochondrial electron transport chain in BAT were not substantially different with either temperature or diet; however, there was a small but significant decrease in Complex II and increase in Complex IV in mice housed at 29°C (Fig. 3D, E, F, G, H and I). There was an expected decrease in protein content in the BAT of HFD-fed mice and mice held at 29°C, which is likely due to an increase in stored lipid (Fig. 3J). H & E staining of the BAT displayed classical brown adipocytes with multilobular lipid droplets in the 22°C mice with evidence of some larger lipid droplets in the HFD mice (Fig. 3K). The interscapular BAT depot was larger in fat-fed mice and in mice housed at 29°C (Table 1).
Elevated triglyceride and abundance of lipogenic proteins in the liver of chow-fed mice housed at thermoneutrality
Liver triglyceride content was increased in HFD mice as expected; however, there was also a temperature effect with triglyceride levels in the liver increased in mice housed at 29°C (Fig. 4A). This reached significance in chow-fed mice. Liver glycogen content was not different with diet or temperature (Fig. 4B). Corresponding to the elevated triglyceride content, the enzymes of the lipogenesis pathway ACC (Fig. 4D), FAS (Fig. 4E) and SCD1 (Fig. 4F) were elevated in the livers of mice housed at 29°C, due largely to a significant effect in chow-fed animals. The abundance of these enzymes was suppressed in HFD liver. The abundance of PEPCK, a rate-limiting enzyme of the gluconeogenesis pathway, was not altered either by temperature or diet (Fig. 4G). Histology of the liver revealed the presence of unilocular lipid droplets in HFD mice (Fig. 4H); however, these were not visible in chow-fed mice housed at either temperature, despite differences in liver triglyceride.
Muscle from thermoneutral mice had increased 2-DG accumulation and decreased triglyceride content compared to 22°C mice
As expected, there was increased triglyceride content in the quadriceps muscle of HFD mice compared to chow controls (Fig. 5A). However, in contrast to the observations with liver triglyceride, mice housed at 29°C had significantly reduced muscle triglyceride compared with mice housed at 22°C. This difference reached significance in HFD animals (Fig. 5A). The activity of β-HAD, the enzyme catalysing the third step in the β-oxidation pathway, was increased in HFD muscle and interestingly, was further increased in the quadriceps of HFD mice housed at 29°C (Fig. 5B). 2-DG accumulation in the quadriceps (Fig. 5C) and gastrocnemius muscles (Table 2) of HFD mice housed at 22°C was significantly lower compared to chow mice, however, at 29°C, this difference was much less substantial. Comparably to the 2-DG uptake, glycogen content in the muscle of HFD mice at 22°C tended to be lower compared to chow controls (P = 0.051); however, this was not as apparent in mice housed at 29°C (Fig. 5D). There was a significant negative correlation between triglyceride content and 2-DG uptake in the quadriceps of mice, independently of temperature or diet (Fig. 5E). Abundance of mitochondrial uncoupling protein 3 (UCP3) was not different between the quadriceps muscle of 29°C and 22°C housed mice; however, UCP3 was slightly elevated in HFD mice (Fig. 5F and G). Basal glucose uptake into the brain was higher with both HFD-feeding and thermoneutrality (29°C), while heart glucose uptake was lower in HFD mice (Table 2). 2-DG accumulation into the epididymal and subcutaneous adipose depots was not different with either temperature or diet (Table 2).
Tissue accumulation of 2-deoxyglucose radioactivity.
|Housing temperature||Chow||HFD||Effect of temp||Effect of diet|
|2-DG Uptake (dpm/mg tissue)|
|Brain||120 ± 5||149 ± 13||148 ± 8||170 ± 11||*||*|
|EpiWAT||12.7 ± 2.8||7.3 ± 1.2||5.9 ± 0.7||8.7 ± 1.0||ns||ns|
|SubcutWAT||8.7 ± 1.0||9.6 ± 1.6||9.0 ± 0.9||9.6 ± 0.6||ns||ns|
|Heart||743 ± 81||615 ± 115||536 ± 63||472 ± 62||ns||*|
|Gastrocnemius||151 ± 21||173 ± 16||100 ± 9||159 ± 14||*||ns|
Data are means ± s.e.m. 2-DG uptake was calculated by the measurement of phosphorylated [3H]2-deoxyglucose. Analysed by 2-way ANOVA for an effect of temperature and an effect of diet. n = 7–12. *P<0.05.
The vast majority of published mouse studies report experiments conducted under a standard housing temperature of 22°C (Abreu-Vieira et al. 2015). At this temperature, mice have to use a substantial amount of energy to generate enough heat to maintain body temperature and can be considered to be exposed to chronic thermal stress (Maloney et al. 2014). The ability for thermoneutral housing to alter various aspects of what has previously been considered to be normal physiology in mice (Feldmann et al. 2009, Overton 2010, Stemmer et al. 2015) suggests that there could also be an effect of housing temperature on lipid accumulation and glucose homeostasis in mice. Indeed, there have been several reports of metabolic phenotypes in mice differing with housing at thermoneutral temperatures or standard housing temperature (Liu et al. 2003, Castillo et al. 2011).
To specifically investigate any interaction between housing temperature and obesogenic diet, C57BL/6J(Arc) mice were fed either a normal chow or HFD while being housed at two different temperatures, the standard housing temperature 22°C or a thermoneutral temperature of 29°C. The study was able to demonstrate the well-documented effects that are associated with a HFD, such as the development of increased adiposity and impaired glucose homeostasis in the mouse. Also, the change in housing temperature to the thermoneutral zone of mice resulted in the expected decrease in the energy expenditure of the mice (Fig. 1E), balanced by a corresponding decrease in food intake (Fig. 1H). Interestingly, the change in temperature promoted contrasting effects on the storage of lipids and handling of glucose in the liver, muscle and BAT. Critically, the study provides important data suggesting that the thermal stress induced by housing mice at 22°C produces changes in tissue metabolism but these differences are not substantial enough to alter the development of adiposity and impaired glucose tolerance in HFD mice, a model often used to investigate the pathophysiology of insulin resistance and diabetes.
Previous reports have demonstrated increases (Hoevenaars et al. 2014, Cui et al. 2016), decreases (Tian et al. 2016) or no change (Uchida et al. 2010, Giles et al. 2017) in body mass between male HFD mice housed at thermoneutrality and standard housing temperatures. There is also one report that describes an increase in body weight in female mice fed a HFD and held at thermoneutrality (Giles et al. 2017). In the current study, male chow and HFD mice housed at thermoneutrality demonstrated no differences in body mass or body composition compared to animals housed at 22°C, despite a clear difference in adiposity between chow and HFD mice (Fig. 1). Therefore, the decrease in energy intake, combined with the significantly reduced energy expenditure (~30%) observed at thermoneutral housing, results in the maintenance of body weight and body fat in mice. Similarly, housing temperature did not affect the ratio of whole-body carbohydrate to lipid oxidation (determined by RER), although the expected difference between chow and HFD mice was observed (Fig. 1G). Follow-up studies should include female mice to investigate if this phenomenon is consistent between genders.
Systemic glucose handling, measured by oral glucose tolerance tests, demonstrated an impaired glucose tolerance in HFD mice compared to chow controls after both 8 and 13 weeks of dietary intervention at both 22°C and 29°C (Fig. 2). This was accompanied by a significant increase in plasma insulin levels during the GTT in HFD mice. Within each diet group, the effect of thermoneutral housing on glucose clearance was far less noticeable. Mice housed at 29°C exhibited very similar glucose tolerance and plasma insulin during the oGTT to mice that were housed at a temperature of 22°C, irrespective of diet. Similar to previous reports (Lee et al. 2011, Turner et al. 2013), a longer dietary intervention did not result in a further deterioration of glucose tolerance in HFD mice. The lack of effect of housing temperature on oral glucose tolerance on chow and HFD observed in this study is different to another study examining a similar question (Uchida et al. 2010). However, there are several technical differences between the two studies including the temperatures examined (20°C and 25°C for Uchida et al. vs 22°C and 29°C), the timing of the oGTT (10 days after exposure to different temperature vs 8 and 13 weeks), the duration of fast before oGTT (16 h vs 6 h) and the dose of glucose used (1 g/kg vs a fixed dose of 50 mg/mouse). It is possible that early effects of altered housing temperature on glucose tolerance were captured in the study of Uchida et al., whereas our study captures adaptive changes in tissue metabolism that enable the maintenance of similar whole-body glucose homeostasis independent of housing temperature but still dependent on diet.
Other markers of metabolic health such as fasting glucose and plasma insulin were elevated in HFD mice as expected but were not affected by housing temperature (Table 1). The increase in circulating NEFA in mice maintained at 29°C may indicate a greater rate of lipolysis; however, this was not reflected in an increased rate of whole-body fatty acid oxidation (RER Fig. 1G). Of note, elevated NEFA has been shown to correlate with cardiovascular inflammation (Mas et al. 2010) and may contribute to the reported increase in atherosclerotic lesions in mice housed at thermoneutrality (Giles et al. 2016, Tian et al. 2016).
As expected, plasma leptin was substantially increased in HFD mice and was not affected by temperature. Interestingly, adiponectin levels were affected by both diet and temperature and were elevated with HFD and thermoneutral housing. Although obesity and insulin resistance in humans is strongly associated with a decrease in circulating total and high molecular weight (HMW) adiponectin (Lara-Castro et al. 2006), this observation is less clear in rodent models. Genetic (Oana et al. 2005) and dietary models (Barnea et al. 2006, Bonnard et al. 2008) of obesity and insulin resistance in mice and rats often do not report consistent differences in the total circulating adiponectin levels although the circulating amount of the more bioactive HMW form of adiponectin is rarely measured in animal models.
UCP1 abundance was substantially reduced in the BAT of mice housed at 29°C (Fig. 5A and B) and 2-DG accumulation followed a similar pattern, presumably reflecting a reduced need for glucose as a substrate for thermogenesis (Cooney et al. 1985). In mice fed a HFD, UCP1 expression was increased, which could indicate a compensatory upregulation of thermogenic capacity to offset the increase in energy intake in HFD mice as reported previously (Fromme & Klingenspor 2011). Interestingly, after normalising for protein, the abundance of mitochondrial oxidative phosphorylation (OXPHOS) complexes in BAT was relatively unchanged between temperatures (Fig. 3D, E, F, G, H and I). However, there was a small decrease in Complex II and increase in Complex IV abundance in 29°C mice, as previously reported in BAT from mice housed at 30°C compared to 4°C (Shabalina et al. 2013). Because of its role in FADH2 oxidation, the decrease in Complex II (SDHB) may be due to a decreased requirement for fatty acid oxidation in BAT of mice housed at 29°C. Histology of interscapular BAT depot from mice from both diet groups that were housed at 29°C revealed a substantial increase in lipid droplet size indicating that triglyceride was accumulating in BAT rather than being used as a substrate for thermogenesis in this tissue (Fig. 3K).
Although the data in the current study showed no significant difference in whole-body glucose handling in mice housed at different housing temperatures, differences in lipid storage and glucose metabolism (2-DG accumulation) were observed in metabolically important tissues. Interestingly, housing mice at 29°C resulted in elevated liver triglyceride in both chow and HFD mice, similar to a previous report that found elevated liver lipid in HFD mice housed at thermoneutrality (Giles et al. 2017). This was accompanied by an increase in the abundance of the key lipogenic enzymes ACC, FAS and SCD1, suggesting an increased rate of de novo lipogenesis in 29°C chow-fed mice (Fig. 4). The abundance of these lipogenic enzymes was suppressed in the liver from HFD mice, presumably due to higher dietary lipid availability. Interestingly, there was no indication of visible lipid droplets in the liver of 29°C chow-fed mice, despite elevated triglyceride content compared to 22°C controls, which may indicate microvesicular lipid storage. Lipid accumulation in the liver of mice housed at 29°C was most likely the result of a reduced energy requirement for thermogenesis but no difference in markers of liver injury, plasma AST and ALT with either temperature or diet (Table 1) was observed. This suggested that 13 weeks of HFD-feeding at either housing temperature was insufficient to produce overt liver injury even though injury may occur after a longer period of feeding (Giles et al. 2017).
Although HFD-fed mice displayed an increased muscle triglyceride content, triglyceride was lower in muscle of HFD mice housed at thermoneutrality (Fig. 5A). Interestingly, in HFD mice housed at 29°C the activity of β-HAD, one of the enzymes in the β-oxidation pathway was increased in muscle indicating that the lower triglyceride content could reflect increased fatty acid oxidation (Fig. 5B). It is unclear why lipid deposition would be higher in the BAT and liver of 29°C mice and lower in the muscle compared to 22°C controls. However, it is possible that a reduced requirement of energy for thermogenesis in mice held at 29°C could promote de novo lipogenesis in the liver and BAT, increasing circulating lipids and upregulating fatty acid oxidation pathways in the muscle. Independently of temperature or diet, 2-DG accumulation inversely correlated with triglyceride content in the quadriceps muscle (Fig. 5E). This is consistent with previous reports that glucose uptake correlates inversely with lipid content in muscle (Storlien et al. 1996). UCP3, part of the same family as UCP1, which has been postulated to have a role in thermogenesis, was measured in the quadriceps and showed no significant difference with temperature. However, there was a slight but significant increase in UCP3 abundance in HFD muscle, consistent with previous reports, which supports the idea that UCP3 is more relevant for fatty acid metabolism than thermogenesis (Schrauwen et al. 2006).
Differences in 2-DG accumulation in specific tissues but no obvious change in glucose tolerance between mice held at 22°C and 29°C most likely reflect changes in the way individual tissues are competing against each other for disposal of the same glucose pool (Table 2). Therefore, the higher 2-DG accumulation in the whole brain in both HFD mice and mice housed at thermoneutrality could be explained by a decreased competition with insulin-sensitive tissues such as the muscle and heart, which had decreased 2-DG accumulation.
In conclusion, the results from this study suggest that housing mice at thermoneutral temperatures results in important changes to energy balance and thermogenesis; however, this does not translate to significant differences in glucose tolerance. However, differences in basal glucose uptake, triglyceride content and the abundance of specific proteins involved in fatty acid metabolism were evident in the BAT, skeletal muscle and liver. In light of findings showing that housing mice at thermoneutrality can alter lipid accumulation and glucose metabolism in individual tissues, this study adds to the growing body of literature in a range of disease states suggesting that it is important to consider the influences of housing conditions on experimental animals, and how this might influence results and the translatability of findings to humans.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
L S was supported by an Australian Postgraduate Award. G J C was supported by a Professorial Research Fellowship from the University of Sydney, Sydney Medical School.
Author contribution statement
L S, A E B and G J C were involved in the study design. All authors were involved in experimental data collection. Analysis of data was conducted by H G, L S and A E B. L S, G J C and A E B were involved in the drafting of the manuscript.
The authors would like to thank the staff of the Biological Testing Facility of the Garvan Institute for assistance with animal care.
BrandonAETid-AngJWrightLEStuartESuryanaEBentleyNTurnerNCooneyGJRudermanNBKraegenEW 2015 Overexpression of SIRT1 in rat skeletal muscle does not alter glucose induced insulin resistance. PLoS ONE 10 e0121959. (https://doi.org/10.1371/journal.pone.0121959)
CastilloMHallJACorrea-MedinaMUetaCKangHWCohenDEBiancoAC 2011 Disruption of thyroid hormone activation in type 2 deiodinase knockout mice causes obesity with glucose intolerance and liver steatosis only at thermoneutrality. Diabetes 60 1082–1089. (https://doi.org/10.2337/db10-0758)
CooneyGJCatersonIDNewsholmeEA 1985 The effect of insulin and noradrenaline on the uptake of 2-[1–14C]deoxyglucose in vivo by brown adipose tissue and other glucose-utilising tissues of the mouse. FEBS Letters 188 257–261 (https://doi.org/10.1016/0014-5793(85)80383-5)
FeldmannHMGolozoubovaVCannonBNedergaardJ 2009 UCP1 ablation induces obesity and abolishes diet-induced thermogenesis in mice exempt from thermal stress by living at thermoneutrality. Cell Metabolism 9 203–209. (https://doi.org/10.1016/j.cmet.2008.12.014)
GilesDARamkhelawonBDonelanEMStankiewiczTEHutchisonSBMukherjeeRCappellettiMKarnsRKarpCLMooreKJ 2016 Modulation of ambient temperature promotes inflammation and initiates atherosclerosis in wild type C57BL/6 mice. Molecular Metabolism 5 1121–1130. (https://doi.org/10.1016/j.molmet.2016.09.008)
GilesDAMoreno-FernandezMEStankiewiczTEGraspeuntnerSCappellettiMWuDMukherjeeRChanCCLawsonMJKlarquistJ 2017 Thermoneutral housing exacerbates nonalcoholic fatty liver disease in mice and allows for sex-independent disease modeling. Nature Medicine 23 829–838. (https://doi.org/10.1038/nm.4346)
HoevenaarsFPMBekkenkamp-GrovensteinMJanssenRJHeilSGBunschotenAHoek-van den HilEFSnaas-AldersSTeerdsKvan SchothorstEMKeijerJ 2014 Thermoneutrality results in prominent diet-induced body weight differences in C57BL/6J mice, not paralleled by diet-induced metabolic differences. Molecular Nutrition and Food Research 58 799–807. (https://doi.org/10.1002/mnfr.201300285)
HoyAJBrandonAETurnerNWattMJBruceCRCooneyGJKraegenEW 2009 Lipid and insulin infusion-induced skeletal muscle insulin resistance is likely due to metabolic feedback and not changes in IRS-1, Akt, or AS160 phosphorylation. American Journal of Physiology. Endocrinology and Metabolism 297 E67–E75. (https://doi.org/10.1152/ajpendo.90945.2008)
KowalskiGMBruceCR 2014 The regulation of glucose metabolism: implications and considerations for the assessment of glucose homeostasis in rodents. American Journal of Physiology. Endocrinology and Metabolism 307 E859–E871. (https://doi.org/10.1152/ajpendo.00165.2014)
MasSMartínez-PinnaRMartín-VenturaJLPérezRGomez-GarreDOrtizAFernandez-CruzAVivancoFEgidoJ 2010 Local non-esterified fatty acids correlate with inflammation in atheroma plaques of patients with type 2 diabetes. Diabetes 59 1292–1301. (https://doi.org/10.2337/db09-0848)
OanaFTakedaHHayakawaKMatsuzawaAAkahaneSIsajiMAkahaneM 2005 Physiological difference between obese (fa/fa) Zucker rats and lean Zucker rats concerning adiponectin. Metabolism: Clinical and Experimental 54 995–1001. (https://doi.org/10.1016/j.metabol.2005.02.016)
RippeCBergerKBöiersCRicquierDErlanson-AlbertssonC 2000 Effect of high-fat diet, surrounding temperature, and enterostatin on uncoupling protein gene expression. American Journal of Physiology. Endocrinology and Metabolism 279 E293–E300. (https://doi.org/10.1152/ajpendo.2000.279.2.E293)
StemmerKKotzbeckPZaniFBauerMNeffCMüllerTDPflugerPTSeeleyRJDivanovicS 2015 Thermoneutral housing is a critical factor for immune function and diet-induced obesity in C57BL/6 nude mice. International Journal of Obesity 39 791–797. (https://doi.org/10.1038/ijo.2014.187)
SwoapSJLiCWessJParsonsADWilliamsTDOvertonJM 2008 Vagal tone dominates autonomic control of mouse heart rate at thermoneutrality. American Journal of Physiology. Heart and Circulatory Physiology 294 H1581–H1588. (https://doi.org/10.1152/ajpheart.01000.2007)
TianXYGaneshanKHongCNguyenKDQiuYKimJTangiralaRKTontonozPTonotonozPChawlaA 2016 Thermoneutral housing accelerates metabolic inflammation to potentiate atherosclerosis but not insulin resistance. Cell Metabolism 23 165–178. (https://doi.org/10.1016/j.cmet.2015.10.003)
TurnerNBruceCRBealeSMHoehnKLSoTRolphMSCooneyGJ 2007 Excess lipid availability increases mitochondrial fatty acid oxidative capacity in muscle: evidence against a role for reduced fatty acid oxidation in lipid-induced insulin resistance in rodents. Diabetes 56 2085–2092. (https://doi.org/10.2337/db07-0093)
TurnerNKowalskiGMLeslieSJRisisSYangCLee-YoungRSBabbJRMeiklePJLancasterGIHenstridgeDC 2013 Distinct patterns of tissue-specific lipid accumulation during the induction of insulin resistance in mice by high-fat feeding. Diabetologia 56 1638–1648. (https://doi.org/10.1007/s00125-013-2913-1)