Brown adipose tissue is the key depot for glucose clearance in microbiota depleted mice

Brown adipose tissue (BAT) mediates nonshivering thermogenesis via UCP1 and contributes to glucose homeostasis. Gut microbiota have also been implicated in glucose regulation: germ-free and antibiotic-treated mice often show improved glucose tolerance. However, which host tissues drive enhanced glucose disposal after microbiota depletion is debated, with prior studies implicating browning of white adipose tissue (WAT) or reduced hepatic gluconeogenesis. Conflicting data also exist on whether microbiota depletion impairs or spares adaptive thermogenesis. This study asks: (1) does depleting gut microbiota consistently improve glucose clearance; (2) is any improvement secondary to adaptive thermogenesis; and (3) which tissues, particularly BAT versus WAT or liver, account for increased glucose uptake when microbiota are depleted.

Foundational work established BAT’s thermogenic role via UCP1 and its contribution to systemic glucose handling; human BAT is detectable and activation improves insulin sensitivity. Microbiota depletion (germ-free or antibiotics) has been reported to improve glucose tolerance in lean and obese mice. Some studies attributed improved glucose control to WAT browning or altered hepatic gluconeogenesis, while others suggested minimal changes when body mass differences are accounted for. Recent work indicated microbiota are required for optimal UCP1-dependent thermogenesis, but this has been challenged. Thus, the field shows conflicting reports on thermogenesis and on which tissues mediate glucose improvements after microbiota depletion, with little direct assessment of BAT uptake alongside WAT and liver across conditions.

Study design used male C57BL/6J background mice housed SPF. Microbiota depletion was induced with an antibiotic cocktail (ampicillin 300 mg/kg, metronidazole 500 mg/kg, vancomycin 250 mg/kg, neomycin 50 mg/kg) in sterile drinking water for ~3–5 weeks. Metabolic assessments included: (a) glucose tolerance tests (2 g/kg glucose i.p./i.v. with tail vein blood sampling at 0–120 min); (b) isotope tracing using [U-13C]-glucose and 13C-2-deoxyglucose (2DG) to quantify whole-body oxidation via 13CO2 breath sampling and tissue-specific uptake (22 tissues sampled 25–30 min post-injection) under room temperature (22 °C) and acute cold (4 °C) exposure; (c) indirect calorimetry (TSE) and respiratory exchange ratio following β3-adrenergic agonist (CL-316,243) to interrogate adaptive thermogenesis; (d) daily energy expenditure using indirect calorimetry and doubly labeled glucose/water approaches in step-down cooling protocols. Mouse models included wild-type, Ucp1 knockout (Ucp1-KO), and Ucp1-DTR mice enabling ablation of Ucp1+ adipocytes by diphtheria toxin (DT) injections. Diet interventions included low-fat diet (LFD) and high-fat diet (HFD) cohorts. Tissue uptake was quantified by mass spectrometry-based 13C enrichment in breath and oven-dried tissue samples. Western blotting assessed UCP1 protein where relevant. Statistical analyses employed Student’s t-tests, one- and two-way ANOVA with Bonferroni correction, and ANCOVA with body mass adjustments; significance thresholds P < 0.05.

- Microbiota depletion by ABX improved glucose tolerance: in LFD-fed mice at 22 °C, ABX reduced GTT AUC versus controls (P < 0.001). Cold exposure improved GTT in controls but did not further augment clearance in ABX mice; both conditions showed lower rectal temperature during cold challenge. - Whole-body glucose oxidation: After [U-13C]-glucose, ABX mice at 4 °C exhibited ~1.2-fold higher peak 13CO2 output than controls (two-way ANOVA F3,42 = 9.243, P = 0.001); no difference at 22 °C. - Tissue-specific uptake: At 4 °C, ABX significantly increased 13C glucose/2DG in BAT (two-way ANOVA F5,7 = 106.8, P < 0.001), with increases also in heart (F5,7 = 6.589, P = 0.017) and liver in some analyses; cecum showed consistently elevated 13C content. At 22 °C, ABX increased uptake in the cecum. No significant differences were detected across WAT depots (inguinal, epididymal, mesenteric, retroperitoneal) or skeletal muscles. - Thermogenesis: ABX impaired sustained β3-adrenergic-induced oxygen consumption in wild-type mice (drop ~10% after 6 h of CL-316,243; ANCOVA F1,1 = 3.36, P = 0.033) and reduced UCP1 expression, indicating compromised UCP1-dependent thermogenesis. However, ABX had negligible impact on UCP1-independent thermogenesis: in Ucp1-KO mice, daily energy expenditure at 4 °C did not differ between ABX and controls (one-way ANOVA F3 = 0.5414, P > 0.999). - UCP1 is not required for ABX-induced glucose improvement: In HFD-fed cohorts, both wild-type and Ucp1-KO mice showed accelerated glucose clearance after ABX with reduced AUC (one-way ANOVA F3,20 = 4.831, P = 0.016), despite no changes in body weight, fat mass, or daily energy expenditure. - BAT is necessary: Genetic ablation of Ucp1+ cells (Ucp1-DTR + DT) blunted the ABX-induced improvement in ipGTT and reduced 13C-2DG uptake in BAT, while WAT and liver uptake remained unchanged. This indicates BAT, rather than WAT or liver, is the principal depot for enhanced glucose disposal when microbiota are depleted. - Cecum consistently showed increased glucose uptake in ABX mice (lean and obese), supporting a microbiota depletion-driven shift in intestinal substrate utilization from SCFAs toward glucose.

The study resolves a key question by demonstrating that improved glucose clearance after microbiota depletion is mediated predominantly by BAT and the cecum, not by WAT browning or hepatic gluconeogenesis. Isotope tracing pinpointed BAT as a major glucose sink under ABX, particularly with cold exposure, while WAT and skeletal muscle showed negligible changes. Although ABX impairs UCP1-dependent thermogenesis and reduces UCP1 expression, the glucose-lowering effect persists in Ucp1-KO mice and is abrogated by depleting Ucp1+ cells, indicating that BAT’s glucose uptake function is dissociable from UCP1-driven adaptive thermogenesis. Energy expenditure analyses further show that ABX does not alter UCP1-independent thermogenesis in Ucp1-KO mice, countering arguments that microbiota depletion broadly affects thermogenesis when mass is adjusted. Together, these findings establish a gut-BAT axis in glucose regulation, with potential translational implications: targeting BAT glucose utilization independent of thermogenic activation may improve glycemic control without cardiovascular risks associated with strong β3-adrenergic stimulation.

Antibiotic-mediated depletion of the gut microbiota enhances glucose clearance in mice by increasing glucose uptake predominantly in BAT and the cecum, with minimal contribution from WAT or liver. This effect is independent of UCP1 expression and UCP1-dependent thermogenesis but requires the presence of Ucp1-expressing adipocytes. Microbiota depletion does not impact UCP1-independent thermogenesis. The results highlight BAT as a key depot for glucose disposal in the absence of gut microbiota and delineate thermogenesis from glucose-handling functions in BAT. Future work should elucidate the molecular mechanisms of the gut-BAT axis that augment glucose uptake in BAT, identify the specific brown adipocyte subpopulations involved, characterize the role of BAT-like depots, and determine gut-derived signals and hormones mediating this crosstalk for therapeutic exploitation in diabetes.

Only a single post-injection time point (25–30 min) for tissue 13C uptake was analyzed, potentially missing dynamics in other tissues. The subset of brown adipocytes responsible for enhanced uptake under ABX was not identified; contributions from BAT-like/beige depots remain unclear. 13C-2DG is a surrogate for glucose uptake and can perturb glucose sensing, possibly biasing distribution estimates; although findings were consistent with 13C-glucose, limitations remain. Experiments were conducted in C57BL/6J background; strain, housing, temperature acclimation, antibiotic administration route, and diet could influence microbiota composition and responses. Some statistical reanalyses using adjusted body mass rely on cecum-content estimates from separate cohorts, introducing uncertainty.