Selective Impairment of Glucose but Not Fatty Acid or Oxidative Metabolism in Brown Adipose Tissue of Subjects With Type 2 Diabetes

The stimulation of the sympathetic nervous system resulting from cold exposure activates responses to defend against the cold (1). This includes the stimulation of brown adipose tissue (BAT), whose function is to produce heat (2,3). The molecular signature of the largest and most prevalent BAT depot in adult humans, found in the supraclavicular region, also suggests that in contrast to the classical BAT found in rodents, human BAT consists of a mosaic of white and brown adipocytes (4). However, similar to rodents, it demonstrates remarkable plasticity based on chronic environmental conditions (5). In lean and healthy men, BAT can significantly contribute to cold-induced thermogenesis, fueled predominantly by fatty acids hydrolyzed from intracellular triglycerides (TG) (6). In contrast, the volume of BAT positive for ¹⁸F-fluoro-deoxyglucose (¹⁸FDG), a positron-emitting glucose tracer (i.e., metabolically active BAT volume), and the clearance of circulating glucose by BAT are lower in cold-exposed overweight or obese individuals compared with lean individuals (7–9). BAT glucose clearance was also significantly lower in overweight or obese insulin-resistant individuals (7). Older age, higher BMI, and the presence of type 2 diabetes are all independently associated with a lower prevalence of spontaneously activated BAT (10–12). BAT glucose uptake may not always reflect its thermogenic activity because glucose taken up by this tissue is used predominantly for de novo lipogenesis (13) and is highly dependent on intracellular TG content (5). Whether BAT thermogenesis is truly impaired in individuals with type 2 diabetes is therefore unknown. The current study was designed to address the hypothesis that BAT oxidative, nonesterified fatty acid (NEFA), and glucose metabolism are defective and are associated with abnormal whole-body thermogenic and metabolic responses upon well-controlled acute cold exposure in men with type 2 diabetes versus age-matched control subjects without diabetes and young healthy control subjects.

Nineteen of the 25 participants (6 with type 2 diabetes, 7 matched control participants, and 6 of the 12 young healthy control subjects) underwent two 4.5-h metabolic tests (protocols A and B; Supplementary Fig. 1) performed in random order within an average of 20 days of each other and designed to assess whole-body and BAT-specific energy substrate turnover and oxidation and energy expenditure at room temperature (∼23–25°C, time 0–120 min) and during an acute cold exposure (time 120–300 min) (6). Six of the 12 young control subjects only participated in a single 4.5-h metabolic test, which did not include the administration of ¹⁸FTHA (protocol C (5)). Note that original data are presented for 13 participants (6 with type 2 diabetes and 7 age-matched control participants). Experiments were conducted between 0730 and 1500 after a 12-h fast and 48 h without strenuous physical activity. Subjects wearing only shorts were weighed and instrumented with six autonomous wireless temperature sensors to measure mean skin temperature (14) and surface electromyography (EMG) electrodes (Delsys EMG System, Natick, MA) placed on the belly of eight muscles. Participants were fitted with a liquid-conditioned tube suit, ingested a telemetric thermometry capsule to measure core temperature, and performed a series of standardized muscle contractions to normalize EMG recordings to the maximal voluntary contraction (MVC) of each muscle being measured for shivering activity (6).

The same suit was used for all subjects to maintain consistent tubing density and water flow. After a 120-min baseline period at ambient temperature (∼23–25°C), the liquid-conditioned suit was perfused with 18°C water using a temperature- and flow-controlled circulation bath to minimize overt shivering while reducing skin temperature by ∼4.0 ± 0.2°C. Whole-body metabolic heat production was determined by indirect respiratory calorimetry corrected for protein oxidation (15), at room temperature and between 180–200 min and 280–300 min. Whole-body and muscle-specific shivering intensity and pattern as well as mean skin and core temperatures were measured continuously from time 90 to 300 min (16). Only the means of the final 30 min of the ambient period and final 120 min of the cold exposure are reported.

Plasma glucose appearance rate (Ra_glucose) was determined during protocol A (17). Ra_NEFA and Ra_glycerol were measured during protocol B, as previously described (18). Tissue oxidative metabolism was determined with ¹¹C-acetate dynamic positron emission tomography (PET) acquisition (19). Tissue oxidative metabolism index (the rapid fractional tissue clearance of ¹¹C-acetate, k, in sec⁻¹) was estimated from tissue ¹¹C activity over time using a monoexponential fit from the time of peak tissue activity (20) based on well-described assumptions (21).

In protocol A, tissue-specific glucose uptake was determined by ¹⁸FDG dynamic PET acquisition using the Patlak linearization method (22), with the image-derived arterial input function taken from the aortic arch (23). The slope of the plot in the graphical analysis is equal to the tissue glucose extraction constant of ¹⁸FDG (K_i in min^–1 of ¹⁸FDG). Tissue net glucose uptake (K_m) was then calculated by multiplying K_i by plasma glucose concentration corrected for changes in plasma glucose levels (24). After cold exposure (at time 300 min), a whole-body computed tomography (CT) scan (16 mAs), followed by a static whole-body PET acquisition, was performed to determine whole-body ¹⁸FDG organ distribution and tissue standard uptake value (SUV). In protocol B, the same procedure was performed but with intravenous bolus administration of ¹⁸FTHA (∼185 MBq) instead of ¹⁸FDG to determine tissue NEFA uptake after correction for plasma metabolites using Patlak linearization (19,22,25).

For dynamic PET acquisitions, the mean value of pixels (mean SUV) for each frame was recorded. Regions of interest (ROIs) were drawn on the aortic arch for blood activity (input functions), the larger skeletal muscles in the field of view (m. trapezius, m. pectoralis, m. deltoideus), posterior cervical subcutaneous adipose tissue, and on supraclavicular BAT according to the following criteria: a tissue radiodensity between –30 and –150 Hounsfield units (HU) and ¹⁸FDG uptake during cold exposure of more than 1.5 SUV units. Total ¹⁸FDG-positive BAT volume on whole-body scans was also quantified according to the latter criteria. ROIs were first defined from the transaxial CT slices and then copied to the respective ¹⁸FDG, ¹¹C-acetate, or ¹⁸FTHA image sequences. The supraclavicular BAT ROI was at least 15 mm distant from surrounding muscle or vascular structures (Supplementary Fig. 2) to minimize spill-in activity (23).

Plasma metabolites and hormones were determined as previously described (6,26).

Data are expressed as mean ± 95% CI or median and interquartile range (IQR) in case of nonnormal data distribution. Appropriate transformations of variables were performed when normal distribution was not observed for parametric statistical testing. The paired Student t test was used to compare between room temperature and cold exposure. One-way ANOVA was used to determine group-dependent differences in baseline characteristics, shivering intensity, ¹⁸FDG-positive BAT volume, and tissue metabolic activities. Two-way ANOVA for repeated measures, with group and temperature and their interaction as the independent variables, was used to analyze group- and temperature-dependent differences in averaged steady-state hormone and metabolite levels, thermal responses, and blood and tissue ¹¹C-acetate PET-acquired metabolic activities throughout the protocols. The Tukey multiple comparisons post hoc test was used, where applicable. Pearson correlation coefficients were used to determine zero-order and partial correlations between variables. A two-tailed P value of <0.05 was considered significant. Analyses were performed using GraphPad Prism 6.00 (GraphPad Inc., San Diego, CA) and IBM SPSS Statistics 21 software.

Participants were fully informed of the risks and methodologies applied and provided their written consent to participate in this study, in accordance with the Declaration of Helsinki. This study received ethics approval from the University of Ottawa Office of Research Ethics and Integrity and the Centre de Recherche du Centre hospitalier universitaire de Sherbrooke Human Ethics Committee.

Six men with well-controlled type 2 diabetes (HbA_1c <7.5%) on diet or oral hypoglycemic agents (type 2 diabetes group) were compared with 7 age-matched men of similar weight and BMI (matched control subjects) and with 12 young healthy men (healthy young control subjects) (Table 1) who participated in previous studies (5,6). Insulin, thyroid-stimulating hormone, and leptin declined, whereas plasma NEFA levels and appearance rate, glycerol appearance rate, thyroxine (T₄), and cortisol increased significantly across the three groups during cold exposure (Table 2). Plasma glucose and TG concentrations were greater in individuals with type 2 diabetes than in the healthy young control subjects and age-matched control subjects, with the plasma glucose decreasing during cold exposure in individuals with type 2 diabetes. Adiponectin increased during cold exposure across the three groups, but the increase was greater in the age-matched control subjects. During cold exposure, the mean skin temperature fell by 4.5°C (95% CI 3.7–5.2), 3.2°C (95% CI 2.0–4.3), and 4.1°C (95% CI 3.1–5.0) in the healthy young control subjects, age-matched control subjects, and individuals with type 2 diabetes, respectively (P = 0.62) (Fig. 1A and B). The core temperature fell slightly and similarly in all three groups (P = 0.33) (Fig. 1C and D), and the difference in water temperature entering and leaving the cooling suit was the same in all three groups (Fig. 1E). Cold exposure evoked a ∼1.8-fold increase in energy expenditure in the three groups of participants (Δ energy expenditure of 1.3 [95% CI 1.0–1.6], 0.8 [95% CI 0.4–1.3], and 1.1 [95% CI 0.7–1.4] kcal/min in the healthy young control subjects, age-matched control subjects, and individuals with type 2 diabetes, respectively; Fig. 1F and G and Table 2). By design, shivering was controlled and minimized, reaching 1.9% of a MVC (95% CI 1.1–2.7) in the young healthy group but was an average of twofold greater in subjects with type 2 diabetes (4.0% MVC [95% CI 1.6–6.5], P = 0.06 vs. young healthy control subjects) for the same environmental stress (Fig. 1H). The median BAT volume of activity, based on the volume of ¹⁸FDG uptake, was 48 mL (IQR 26–70) in the young healthy individuals (Fig. 1I), significantly greater than the median volumes observed in the age-matched control subjects (13 mL [IQR 9–20]; P = 0.002) or participants with type 2 diabetes (4 mL [IQR 3–14], P = 0.0001).

During cold exposure, fractional glucose uptake (K_i) (Fig. 2A) in supraclavicular BAT and subcutaneous white adipose tissue (scWAT) was higher in the young individuals compared with age-matched control subjects or participants with type 2 diabetes, whereas fractional glucose uptake in skeletal muscle was not different among the three groups. Net tissue glucose uptake (K_m) (Fig. 2B) was also higher in BAT and scWAT of the young control subjects compared with matched control subjects or participants with type 2 diabetes, whereas muscle net glucose uptake was higher in participants with type 2 diabetes than in matched control subjects. Glucose uptake by BAT summed up to a median of 3.2 µmol/min (IQR 0.6–91.2) in young participants but was limited to 0.1 µmol/min (IQR 0.0–0.4) in the matched control subjects and 0.1 µmol/min (IQR 0.0–0.7) in the participants with type 2 diabetes. Fractional NEFA uptake (K_i) (Fig. 2C) and net NEFA uptake (K_m) (Fig. 2D) were not different among the three groups. Based on ¹⁸FDG BAT volume of activity, median total BAT NEFA uptake was 0.7 µmol/min (IQR 0.1–1.2) in young participants but was limited to 0.1 µmol/min (IQR 0.0–0.2) in the matched control subjects and 0.1 µmol/min (IQR 0.0–0.1) in the participants with type 2 diabetes. There was a direct correlation between BAT fractional glucose and NEFA uptake in younger healthy individuals (r = 0.85, P = 0.03) that was not significant in those with type 2 diabetes (r = −0.40, P = 0.44) or in the age-matched control subjects (r = −0.43, P = 0.34).

Supraclavicular BAT ¹¹C radioactivity over time at room temperature and during cold exposure are shown in Fig. 3A and B, respectively. Peak BAT ¹¹C radioactivity, an index of tissue perfusion (27), increased in all groups during cold exposure, but was lower in participants with type 2 diabetes than in young control participants (Fig. 3C). The cold-induced change in tissue perfusion was the same among the three groups. There was a significant correlation between peak BAT ¹¹C radioactivity and the fractional (Fig. 3D) and net glucose uptake by BAT (Pearson r = 0.58, P = 0.003; Pearson r = 0.58, P = 0.003, respectively). Results from the moderation regression analysis showed that age blunted the relationship between peak BAT ¹¹C radioactivity and fractional (β² = −0.28, P = 0.04) and net glucose uptake by BAT (β² = −0.31, P = 0.04) (Supplementary Tables 1 and 2). The monoexponential decay slope from tissue peak ¹¹C activity (¹¹C-acetate k), an index of tissue oxidative metabolism (28,29), increased during cold exposure in all participants (Fig. 3E). However, the cold-induced change in tissue oxidative metabolism was the same among the three groups. BAT CT radiodensity (in HUs), which is inversely related to the tissue lipid content, increased in all participants during cold exposure (Fig. 3F) but was significantly greater in age-matched control subjects and participants with type 2 diabetes than in the young control subjects. However, the cold-induced change in BAT CT radiodensity was the same among all 3 groups, increasing by 16 HU (95% CI 8–24) in the healthy young control subjects, 7 HU (95% CI 2–13) in the age-matched control subjects, and 10 HU (95% CI 3–17) in the individuals with type 2 diabetes (P = 0.17).

Although BAT oxidative metabolism and shivering intensity were not independently associated with cold-induced energy expenditure (Fig. 4A and B), nor were shivering intensity and BAT oxidative metabolism associated (Fig. 4C), a trend toward an inverse correlation between ¹⁸FDG-positive BAT volume and shivering was evident (Pearson r = −0.36, P = 0.08) (Fig. 4D). Interestingly, BAT fractional glucose uptake and ¹⁸FDG-positive BAT volume were both inversely related to age and diabetes status, whereas shivering intensity and net muscle glucose uptake was directly associated with age and diabetes (Supplementary Table 3). When correcting for age, the effect of diabetes on BAT fractional glucose uptake and shivering intensity was significantly blunted, whereas the association between the presence of diabetes and skeletal muscle fractional glucose uptake improved (Supplementary Table 3).

The NEFA and glycerol appearance rate increased significantly during cold exposure, but there was no difference among the three groups (Table 2). A significant correlation was found between the appearance rate of NEFA during cold exposure and ¹⁸FDG-positive BAT volume (Pearson r = 0.57, P = 0.004) (Fig. 5A) but not with cold-induced BAT oxidative index (ΔBAT oxidative index; Pearson r = −0.32, P = 0.14) (Fig. 5B). The NEFA appearance rate during cold exposure was also significantly correlated with the cold-induced change in BAT radiodensity (ΔBAT radiodensity; Pearson r = 0.44, P = 0.03) (Fig. 5C).

BAT glucose uptake is lower in overweight or obese individuals compared with lean individuals, which is associated with lower ¹⁸FDG-positive BAT volume in some (7) but not all studies (8). Here we show that glucose uptake in BAT and ¹⁸FDG-positive BAT volume are also diminished in older overweight men with or without type 2 diabetes. Although previous investigations have interpreted these findings as a reflection of impaired BAT thermogenesis, here we show that this is not the case. We show that BAT oxidative metabolism and NEFA uptake do not differ among young healthy control subjects, age-matched control subjects, or participants with type 2 diabetes. This demonstrates that BAT activation of oxidative metabolism is not impaired in older men with and without type 2 diabetes compared with young healthy men. This conclusion is further supported by the similar degree of cold-induced increase in BAT perfusion index and by a similar cold-induced increase in BAT radiodensity. The similar reduction in ¹⁸FDG-positive BAT volume combined with lower BAT fractional glucose uptake in age-matched control subjects and in individuals with type 2 diabetes compared with young healthy men is suggestive of an age-dependent decrease in BAT recruitment.

BAT glucose uptake is reduced in overweight or obese individuals (7,8). This reduced cold-induced net glucose uptake in BAT was previously attributed to low BAT blood flow because BAT fractional glucose extraction was similar to lean individuals (7). Here we show that fractional and net glucose uptake by BAT are both reduced in overweight age-matched control subjects and in participants with type 2 diabetes. Although lower blood flow and more advanced age were both independently associated with lower fractional and net glucose uptake, age appeared to have a predominant effect. WAT blood flow is reduced in obese individuals as well as at all stages of impaired glucose tolerance (30). This is explained, at least in part, by capillary rarefaction in WAT (31). A similar obesity-induced capillary rarefaction in BAT, leading to the development of a WAT-like phenotype with the BAT shifting functions toward lipid storage rather than thermogenesis, was observed in high fat–fed mice (32). The lower radiodensity of the BAT in the overweight age-matched control subjects and in the participants with type 2 diabetes could suggest a similar shift toward lipid storage or a lipolytic dysfunction. BAT glucose uptake appears to be significantly influenced by BAT TG content (5,33). Despite these observations, our results do not suggest impairment in intracellular TG lipolysis in BAT or WAT in individuals with type 2 diabetes because the cold-induced reduction in BAT radiodensity and systemic NEFA turnover were the same compared with young healthy individuals. Interestingly, the lower BAT glucose uptake was offset by the improved clearance of shivering skeletal muscle in individuals with type 2 diabetes, which likely accounts for the cold-induced reduction of glycemia in this group. These results are consistent with a predominant effect of muscle versus BAT on glucose clearance during acute cold exposure in humans (34). Although our study (34) and that of Muzik et al. (35) suggest that BAT in adults may not contribute much to energy metabolism, acute cold exposure may not maximally stimulate BAT oxidative metabolism because we demonstrated that short-term mild daily cold exposure sustained over 4 weeks further increased acute cold-stimulated BAT oxidative metabolism and BAT volume of metabolic activity (5).

Although ¹⁸FDG-positive BAT volume and glucose uptake per BAT volume was smaller in age-matched control subjects and in participants with type 2 diabetes versus younger healthy individuals, all three groups exhibited 2.4- to 3.6-fold increases in oxidative metabolism upon cold exposure and similar NEFA uptake per supraclavicular BAT volume. Previous investigations have concluded that a decrease in BAT glucose uptake in overweight or obese individuals (7) was suggestive of blunted BAT oxidative metabolism. Human autopsy studies and necropsy studies in various mammalian models, ranging from rodents to primates, have demonstrated age-dependent decreases in BAT capacity (volume) based on histopathological examination (36,37). However, the functionality of this tissue as a result of aging, obesity, or type 2 diabetes has been far less clear. Here we provide the first in vivo evidence that BAT oxidative metabolism and NEFA uptake per ¹⁸FDG-positive BAT volume is maintained through aging or as a result of obesity and obesity-associated metabolic diseases despite a clear reduction in fractional and net glucose uptake. This demonstrates that ¹⁸FDG has its limitations in assessing BAT oxidative metabolism, as recently acknowledged (38).

Our findings also imply that ¹⁸FDG-positive BAT volume may not accurately reflect the true metabolically active BAT volume in humans. Although we calculated that total BAT NEFA uptake was reduced in subjects with type 2 diabetes and age-matched control subjects versus younger healthy subjects, this was fully attributable to the lower ¹⁸FDG BAT volume in the first two groups. Because ¹⁸FTHA uptake offers little contrast even in healthy individuals between ¹⁸FDG-positive BAT and WAT, performing a reliable measurement of BAT volume based on ¹⁸FTHA from the present data was not possible.

The cold-induced increase in total body energy expenditure was similar among the three groups of participants, but shivering tended to be increased in individuals with type 2 diabetes and was inversely associated with ¹⁸FDG-positive BAT volume. This suggests, as could be expected, that BAT activation and shivering are reciprocal mechanisms recruited to combat cold (39). Thus, our study does not rule out a significant effect of BAT on energy homeostasis in humans or a potential role of BAT in type 2 diabetes. However, BAT activity is imperfectly described by its glucose metabolism.

Consistent with our previous findings, the increase in BAT radiodensity during cold exposure suggests that intracellular TG utilization is the main fuel source for BAT oxidative metabolism and that plasma glucose and NEFA utilization rates by BAT are both trivial during acute cold exposure (5,6,34). Blunted lipolytic action of catecholamines in WAT is an early event in obesity in humans (40) and in BAT of obese Zucker rats (41) due to a reduction in β-adrenergic receptor density. Reduced β-adrenergic receptor density was also demonstrated in BAT of high fat–, high sucrose– fed mice (32). This may have an effect on total volume of BAT recruited but not necessarily on oxidative capacity of activated BAT regions.

Second, the fatty acids released through the hydrolysis of intracellular TG serve to both fuel thermogenesis and to activate uncoupling protein 1 (UCP1), the source of BAT thermogenesis. Adrenergic UCP1 activation largely depends on adipose TG lipase (ATGL) rather than hormone-sensitive lipase (HSL) (42). ATGL and HSL protein expression in WAT are both decreased in an obese, insulin-resistant state, which is associated with the degree of insulin-resistance and hyperinsulinemia in obesity rather than fat mass and fat distribution (43). Whether this decreased expression in ATGL and HSL also extends to human BAT is unclear. However, mice lacking HSL in brown adipocytes are not cold sensitive because they exhibit normal BAT thermogenesis; however, they do show signs of greater TG and diacylglycerol accumulation but without a detectable effect on thermogenesis (44). In contrast, mice lacking ATGL showed severe thermoregulatory defects and lipid accumulation (45). To what extent and in what proportions these enzymes are decreased in human BAT is unclear. That cold-induced increase in BAT oxidative metabolism and radiodensity was normal in individuals with type 2 diabetes despite a more WAT-like appearance does not suggest a major defect in BAT lipolysis.

One important limitation of the current study is the small number of participants that limits the power to detect small to moderate differences in outcomes. However, this limitation did not prevent us from detecting significant metabolic differences between participants with type 2 diabetes and healthy subjects. Furthermore, BAT oxidative metabolism index during cold exposure tended to be higher—not lower—in subjects with type 2 diabetes versus healthy control subjects. Thus, it is unlikely that the small number of subjects with type 2 diabetes in the current study affected our main conclusion. Another limitation is the absence of BAT biopsy specimens to provide insight into the molecular mechanisms for the observed metabolic responses of BAT in type 2 diabetes. Future studies are needed in that respect.

Cold-induced oxidative metabolism and NEFA uptake are not impaired in men with type 2 diabetes despite a clear reduction in glucose uptake per volume of BAT and “whitening” of the BAT phenotype of these individuals based on lower CT radiodensity and reduced basal tissue perfusion. We observed a major reduction in BAT glucose uptake in older individuals, whether or not they had type 2 diabetes. Our results illustrate the limitations of ¹⁸FDG to assess oxidative metabolism in human BAT. Despite this limitation, ¹⁸FDG-positive BAT volume and muscle shivering tend to be activated reciprocally to combat cold exposure, suggesting that total BAT thermogenic capacity may nevertheless be reduced in older individuals with and without type 2 diabetes. Acute cold exposure reduces the glucose level in subjects with type 2 diabetes, but this effect is likely driven by increased glucose uptake in shivering skeletal muscles, whereas BAT plays a minimal role in systemic glucose clearance.

Acknowledgments. The authors acknowledge the excellent technical assistance provided by Diane Lessard, Caroll-Lynn Thibodeau, Maude Gérard, and Éric Lavallée from the Centre de recherche du Centre hospitalier universitaire de Sherbrooke, and also thank all of the subjects who participated in this study.

Funding. D.P.B. is the recipient of the NSERC Postgraduate Scholarship award. S.M.L. is the recipient of a CIHR Postdoctoral Fellowship award. C.N. is the recipient of a CIHR Postdoctoral Fellowship award. This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC Canada) to F.H. and was performed at the Centre de recherche du Centre hospitalier universitaire de Sherbrooke, a research center funded by the Fonds de la recherche du Québec–Santé (FRQS). D.R. is the recipient of the Research Chair on Obesity of Laval University. This work was supported by a grant from the Canadian Diabetes Association (OG-3-10-2970-AC) and Canadian Institutes of Health Research (CIHR) to A.C.C. A.C.C. is the CIHR-GlaxoSmithKline Chair in Diabetes.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. B.G., É.E.T., F.H., D.R., and A.C.C. conceived and designed the experiments. D.P.B., S.M.L., C.N., M.K., S.P., É.E.T., F.H., D.R., and A.C.C. collected, analyzed, and interpreted data. D.P.B., S.M.L., C.N., M.K., B.G., É.E.T, F.H., D.R., and A.C.C. drafted the article or revised it critically for important intellectual content. A.C.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.