Time-Resolved Effects of Short-term Overfeeding on Energy Balance in Mice

Obesity is a disease with multifactorial etiology that poses a significant risk for a series of severe comorbidities, including type 2 diabetes, cardiovascular diseases, and cancers (1). Despite considerable progress in developing effective weight loss therapies, long-term weight management strategies remain elusive, and the underlying causes of obesity are a subject of intense debate (2). Anti-obesity interventions typically result in rapid weight loss followed by a weight plateau and progressive regain. This weight regain is thought to be triggered by the body perceiving weight loss as a threat to energy homeostasis, prompting a response characterized by hypoleptinemia, increased appetite, and decreased energy expenditure (3). Mirroring this defensive response, the organism similarly perceives deliberate attempts to gain weight as a homeostatic insult and engages compensatory responses to restore energy balance (2,4). However, the mechanisms that counteract a positive energy balance and weight gain remain less understood (2,4–7).

For decades, it has been debated whether overfeeding-induced weight gain is countered by so-called luxuskonsumption, an adaptive increase in energy expenditure beyond what can be attributed to an increased body mass (3,4,8–10). Some rodent studies using intragastric overfeeding and indirect metabolic measures have observed subtle (11) or transient (6) increases in whole-body energy expenditure, linked to thyroid hormones (12) and norepinephrine (11,13). However, other studies have not detected changes in energy expenditure during (14–16) or after (17) overfeeding [for review, see the report by Bray and Bouchard (8)]. This inconsistency highlights the need for clearer evidence on whether adaptive thermogenesis helps counteract weight gain after overfeeding. Similarly, the role of fecal energy excretion in body weight regulation remains insufficiently characterized (18). In healthy humans, between 1% and 11% of ingested energy appears to be lost through stool (19) and large interindividual variability in fecal calorie loss has been observed in the context of overfeeding (ranging from 80 to 500 kcal/day) (19). This highlights a potential impact of fecal energy excretion on body weight, as losing such a substantial amount of energy through stool will reduce the amount of metabolizable energy (18).

In this study, we used a mouse model of intragastric overfeeding (20) combined with high-resolution indirect calorimetry in metabolic cages and bomb calorimetry of fecal outputs to comprehensively assess temporal changes in energy intake, whole-body energy expenditure, and fecal energy excretion in response to overfeeding. Additionally, to further explore the role of adipose thermogenesis in the homeostatic response to overfeeding, we subjected uncoupling protein 1 (UCP1) knockout (KO) mice to intragastric overfeeding.

All mouse studies were conducted at the University of Copenhagen, Denmark, and carried out in accordance with regulations regarding the care and use of experimental animals that were approved by the Danish Animal Experimentation Inspectorate (2018-15-0201–01457; 2023-15-0201–01442). Wild-type (WT) male C57BL/6J mice (Janvier) and germline UCP1 whole-body KO mice (21) (Ucp1^tm1Kz; maintained on a mixed C57BL/6JN background; The Jackson Laboratory, catalog no. 003124) were used. All experiments were done at 22°C with a 12:12–hour light-dark cycle (6:00 a.m.–6:00 p.m.). Mice had ad libitum access to water and chow diet (SAFE D30; Safe Diets), unless explicitly noted. All mice were single housed after surgery and during experimental overfeeding and recovery.

The surgical procedure to insert the gastric catheter and the material used for automated overfeeding were the same as previously described (20).

Following a modified surgical protocol (17,22) we previously described (20), 18-week-old male C57BL/6J (n = 16) were anesthetized with isoflurane and subjected to a midline abdominal incision to expose the gastric ventricle. A purse-string suture was placed above the gastric fundus, and a 22-gauge catheter (catalog no. C30PU-MGA2209; Instech Laboratories) was inserted through the center of the suture into the lumen of the stomach. A subcutaneous vascular access button (catalog no. VABM1B/22; Instech) was secured to the skin at the back of the mouse neck, and the catheter was tunneled under the skin to connect the button. After surgery, mice were single housed and allowed to recover for 5 weeks before proceeding with overfeeding interventions.

After recovery from surgery, individual mouse body weight and food intake were tracked for 3 days before the mice were placed in metabolic cages. Mice were then divided into two groups: a control group (n = 6) and an overfed group (n = 10), matched by body weight at the day before study start. Five mice in the overfed group were excluded from the study because of tube clogging during the overfeeding period that impeded the completion of the diet infusions. Prior to their placement into an indirect calorimetry system, mice were connected to the automated infusion system. This system consists of infusion pumps (catalog nos. 704500, 704501, 703005, 703024; Harvard Apparatus) and syringes (10 mL syringes with a Luer Lock tip; catalog no. 302995; Becton and Dickinson) connected to the vascular access buttons using 22-gauge tethers with springs (catalog no. VABM1T/22; Instech Laboratories), polyethylene tubing (catalog no. BTPE-50; Instech Laboratories), multiaxis lever arms (catalog no. SMCLA; Instech Laboratories), and 22-gauge swivels (catalog no. 375/22PS; Instech Laboratories). Mice always had ad libitum access to chow and tap water during the entire study. Mice movement was not limited despite being connected to the infusion system.

The study was initiated after 5 days of acclimatization in metabolic cages (Promethion Core Metabolic System; Sable Systems International). All mice were continuously infused with sterile water (100 µL/h) during the last 3 days of this acclimatization period. Then, on day 0 (d0), the overfed group received 1 day of eucaloric liquid diet (100%) infusion and 6 days of 50% caloric surplus over baseline requirements (150%), whereas control mice were infused with the same volume of sterile water. Every day at 10 a.m. during the overfeeding period, the infusion system was stopped for 30 min, enabling replacement of the syringe with fresh liquid diet and flushing of the outer tubing with sterile water. After syringe replacement, mice received an injection of 0.5 mL sterile water through the vascular access button and were infused with 1,000 µL/h sterile water for 30 min (another 0.5 mL). Therefore, the amount of diet needed to achieve 150% overfeeding was calculated for 23 h of infusion per day. After 7 days of overfeeding, mice remained connected to the infusion system and were allowed to recover for 4 days while receiving 100 µL/h of water infusion. At the end of the experiment, mice were disconnected from the automated overfeeding system, euthanized, and body composition was immediately measured by magnetic resonance imaging (Bruker LF90II) after the removal of the magnetic access button.

The overfed group was infused with a commercial liquid diet (Nutridrink Vanilla; catalog no. 584421; Nutricia) that was supplemented with 12.5% (w/v) sucrose (catalog no. S0389; Sigma-Aldrich). The addition of sucrose increased the caloric density of the liquid diet from 1.5 kcal/mL to 2 kcal/mL (energy percentage from fats, 26 E%; carbohydrates, 62 E%; protein, 12 E%).

To calculate the flow rate of diet infusion, the daily food intake of the chow (estimated caloric density = 3.389 kcal/g) during the baseline period was averaged for all the mice and divided by the caloric density of the diet (2 kcal/mL). Specific flow rates, daily volume infused, and daily caloric infusion are indicated in Supplementary Table 1. Because of imprecisions in the body weight and food intake measurements with the automated system incorporated in the metabolic cages, we measured these parameters manually each day when syringes were replaced with fresh liquid diet. We also corrected for food spillage at the end of the overfeeding and recovery periods, as indicated in Fecal Energy Excretion, below. Metabolic and behavioral parameters (namely, water intake and locomotion) were measured every 15 min, and data were binned and averaged per hour before plotting them on graphs (Supplementary Table 2).

An independent cohort of 18 chow-fed male C57BL/6J mice (16 weeks of age) were subjected to the same surgical procedure as described above for study 1. Out of these, 16 mice were split into control (n = 8) and overfed (n = 8) groups and acclimatized to metabolic cages for 5 days before the start of the infusions. Five mice in the control group and three in the overfed group had clogging of the infusion tube and were discarded before the end of the experiment. Two mice were added to the control group on d1 of the experiment (mouse 7 and mouse 9; see Supplementary Table 4) and were infused with liquid diet for 6 days. In total, five control and five overfed mice reached the end of the overfeeding period and were studied in the recovery period.

Few modifications were made in comparison with study 1. Both control and overfed mice received liquid diet infusion. Control mice received 100% of energy requirements to maintain weight stability, whereas overfed mice received 140% of energy requirements. Mice in both groups received liquid diet infusion from day −1 at 100% of energy requirements and were then switched to either 100% or 140% of energy requirements from d0 until d7. Infusion flow rate was modified every day in the control group to maintain weight stability. None of the mice had access to solid chow diet during the overfeeding period (d0–d7). The presence of chow from day −1 to d0, together with the liquid diet infusion at 100% of energy requirements, explains the increased energy intake and transient overfeeding in both groups from day −1 to d0. Chow was reintroduced at the start of the recovery period. Mice were kept in metabolic cages for 14 days after the end of the overfeeding period. As in study 1, metabolic and behavioral parameters (namely, water intake and locomotion) were measured every 15 min, and data were binned and averaged per hour before plotting them on graphs (Supplementary Table 4).

Chow-fed male WT (n = 4) and UCP1 KO (n = 7) mice on a C57BL/6NJ mixed background at 22 to 30 weeks of age were overfed for 10 days, as previously described (20), with the minor modification that the infusion flow rate was increased gradually until reaching 150% of energy infusion on d9 (Supplementary Table 1). Mice were matched for baseline body weight (WT = 30.6 ± 2.2 g; UCP1 KO = 31.3 ± 1.4 g) and daily energy intake (WT = 13.1 ± 1.8 kcal; UCP1 KO = 13.2 ± 0.7 kcal). Individual trajectories of body weight and food intake are shown in Fig. 5B and D because mice in this study had tube clogging at different times during the overfeeding period. Overfeeding was ended on d8, d10, or d11, after we observed tube clogging, and this is indicated by vertical dashed lines in Fig. 5B and D. Mice were observed after overfeeding to evaluate the changes in body weight and food intake until they stabilized.

For assessment of excreted energy via feces (23), fecal pellets were collected after the overfeeding (7 days for both studies) and recovery (4 days in study 1; 11 days in study 2) periods. The bedding and nesting materials from each cage were dried at room temperature for 1 week. Subsequently, the cage bedding was separated by size using kitchen sieves with varying mesh sizes (Veras Verden). The flow-through material was collected in a tray and manually examined for food remnants and fecal pellets. Food remnants were weighed to determine total food spillage, and this was adjusted for in the calculation of food intake for energy balance. Feces were carefully collected using tweezers and then weighed and desiccated in a drying oven (50°C) before bomb calorimetric combustion (catalog no. IKA C5003; IKA Werke). For estimation or calculation of the absorbed energy in study 1, the energy density of the chow diet was determined by combusting a representative sample of chow pellet. The digestive efficiency was calculated by dividing the absorbed energy (energy intake minus energy lost with feces) by the ingested energy.

Gene expression profiling in interscapular brown adipose tissue (iBAT), inguinal white adipose tissue (iWAT), and quadriceps femoris muscle was performed on tissues obtained from control and overfed mice at d14 and d14 + 3 from a previous study (20). After obtaining cDNA, as previously described (20), quantitative PCR (qPCR) was performed using Precision Plus qPCR Mastermix containing SYBR green (Primer Design). For primer sequences, see Supplementary Table 3. Quantification of mRNA expression was performed according to the comparative cycle-threshold method. All results were normalized to housekeeping genes Rplp0 (iBAT) and Hprt (iWAT and muscle).

Data were analyzed and plots were generated using GraphPad Prism software, version 10.3. All data are presented as mean ± SEM, unless indicated otherwise. Findings with P ≤ 0.05 were considered statistically significant. All statistical analyses are indicated in figures and/or figure legends. Indirect calorimetry data were exported with Macro Interpreter, macro 13 (Sable Systems) prior to analysis in CalR, version 1.3 (24). Regression-based ANCOVA of energy expenditure using body weight as a covariate were performed using absolute body weight at the end of overfeeding (d7 for study 1 and study 2) and recovery period (d7 + 4 for study 1, d7 + 7 for study 2), respectively.

Raw data collected from indirect calorimetry cages are available as supplementary tables. Other data are available upon request. Resources are available upon request.

We measured changes in body weight and voluntary food intake in mice subjected to 7 days of intragastric overfeeding (150%; i.e., 50% excess calories) followed by 4 days of recovery (Fig. 1A). Overfed mice gained 10.7% of their baseline weight on average (Fig. 1B), corresponding to 3.4 g (Fig. 1C), whereas control mice remained weight stable (Fig. 1B and C). After overfeeding, mice returned to their original body weight within 4 days (Fig. 1B and C). Despite similar body weights in control and overfed mice after 4 days of recovery, overfed mice had a slightly higher percentage of fat mass and a slightly lower percentage of lean mass compared with controls (Fig. 1D). Consistent with our previous results as well as those in other reports (5,6,11,12,14,17,20,25), weight gain during overfeeding occurred together with a marked reduction in voluntary food intake, which gradually returned to baseline levels after calorie infusion was stopped (Fig. 1E). Overfeeding was also associated with a pronounced reduction in water intake that persisted throughout the recovery period (Fig. 1F).

We found no difference in average daily energy expenditure between the groups, either during overfeeding or during the recovery period (Fig. 2A and B). Although the mice displayed the expected circadian variations, with a higher metabolic rate during their active phase, overfeeding did not induce an increase in total energy expenditure, and the subtle increase in energy expenditure observed during the light phases of overfeeding might be attributed to the increased body weight (Fig. 2A). This finding was confirmed by regression-based ANCOVA controlling for body weight (Fig. 2B). In contrast, overfeeding triggered significant shifts in metabolic fuel utilization, revealed by changes in the respiratory exchange ratio (RER). During overfeeding, mice relied more on carbohydrate oxidation, indicated by an elevated RER (Fig. 2C and D). This shifted rapidly after infusion stopped, with RER dropping sharply in overfed mice (Fig. 2C and D) and gradually returning to control levels by d4 of recovery (Fig. 2C), reflecting a swift change to fat oxidation after overfeeding.

Although previous studies have linked increased non-exercise activity thermogenesis to weight gain resistance in humans (26), we did not observe any significant differences in locomotor activity during the overfeeding or recovery periods (Fig. 2E and F). However, a trend toward decreased locomotor activity was observed in the light phase during the overfeeding period (Fig. 2F).

To measure fecal energy excretion, we used bomb calorimetry. Although fecal energy density remained unchanged by overfeeding (Fig. 2G), overfed mice excreted approximately three times less feces during both the overfeeding and recovery periods (Fig. 2H), leading to lower fecal energy excretion (Fig. 2I). This reduction possibly reflects increased absorption of the liquid overfeeding diet compared with solid fiber-rich chow, as suggested by the increased digestive efficiency observed in overfed mice (95%) compared with control mice (81%) (Fig. 2J). Energy balance calculations revealed a positive energy balance in overfed mice during overfeeding, followed by a negative energy balance during recovery, as expected (Fig. 2K), whereas control mice had a slightly positive energy balance in both phases (Fig. 2K). Overfed mice had a higher positive energy balance across the entire study compared with controls, despite similar body weights (Fig. 2K). These findings suggest that fecal energy excretion plays a minor role in the homeostatic response to overfeeding.

To minimize potential confounding effects from diet differences between the control group (water infusion and chow ad libitum) and the intervention group that was overfed with a liquid diet, we conducted an experiment in metabolic cages in which the control group received the same liquid diet used for overfeeding but adjusted in infusion volume to meet their weight-stable energy requirements. Importantly, none of the groups had access to solid chow diet during the overfeeding period. For this experiment, we extended the measurements in the recovery period up to 14 days after overfeeding (Fig. 3A).

Overfed mice had an average weight gain of 15.9% relative to their baseline weight (Fig. 3B), equivalent to 5 g (Fig. 3C), and returned gradually to baseline after the 7-day overfeeding period. Control mice maintained a stable body weight throughout the experiment (Fig. 3B and C). Voluntary food intake by overfed mice gradually returned to baseline levels after the cessation of liquid diet infusion (Fig. 3D). Control mice also exhibited a transient decrease in voluntary food intake, though to a lesser extent, at the beginning of the recovery period (Fig. 3D). Water intake in the overfed group was decreased, both during overfeeding and during the recovery period (Fig. 3E).

Controlling for the liquid diet, we observed a gradual increase in energy expenditure in response to overfeeding (Fig. 4A). Regression-based ANCOVA analysis, adjusted for body weight, indicated that this increase exceeds what would be expected from the higher body weight alone, although only a statistical trend was observed. The absence of statistical significance may be attributed to the short duration of overfeeding and/or limited statistical power due to the relatively small sample size. The changes in RER and locomotor activity (Fig. 3C–F) mirrored those observed in the experiment in which chow diet and water infusion was used for the control group.

In contrast to the experiment using a chow diet combined with water infusion as the control condition (Fig. 2G–J), the experiment using an infused liquid diet and no access to chow diet as the control showed that fecal energy density, fecal output, and fecal energy excretion were similar between the overfed and control groups in the period when the liquid diet was infused (Fig. 4G–I). This suggests that the liquid diet, when administered at energy levels matching the weight-stable requirements of the control group, does not significantly alter fecal energy excretion or output compared with the overfed mice. Energy balance calculations showed a positive balance in overfed mice during overfeeding, followed by a negative balance during recovery (Fig. 4K). Unlike the experiment with chow-fed controls (Fig. 2K), the experiment with liquid-fed groups exhibited similar overall energy balances and body weights between overfed and control mice by the end of the study (Fig. 4K).

To evaluate the potential impact of overfeeding on adipose thermogenesis, we subjected WT and germline UCP1 KO mice to overfeeding (Fig. 5A) and found similar relative (+17.7% in WT vs. +19% in UCP1 KO mice) and absolute weight changes (+5.4 g in WT vs. +6.0 g in UCP1 KO mice) (Fig. 5B and C). Weight loss after overfeeding was rapid and genotype-independent (Fig. 5B and C). Additionally, WT and UCP1 KO mice exhibited similar suppression of voluntary food intake during overfeeding, gradually returning to baseline levels after overfeeding (Fig. 5D). These findings suggest that UCP1-mediated adipose thermogenesis is not essential for protection against forced weight gain in mice housed at standard room temperature (22°C).

Despite UCP1 not being necessary for the defense against overfeeding-induced weight gain, our previous results in WT mice demonstrated an induction of adaptive thermogenesis gene programs in iBAT, including Ucp1, after 14 days of 150% overfeeding (20). Using tissues from that study (Fig. 5E) (20), we now further assessed the expression of genes involved in futile cycles (27–31) (i.e., the creatine, calcium, and lipid futile cycles; Fig. 5F–H). For the creatine futile cycle, we detected modest increases in Alpl/Tnap and Gamt expression in iBAT, and Gamt and Slc6a8 expression in iWAT in response to overfeeding. These effects were transient for Gamt, whereas Tnap and Slc6a8 expression remained increased during recovery. Ckb expression increased during recovery in iBAT and slightly decreased in iWAT (Fig. 5F). In the calcium futile cycle, alterations were observed in Serca2b and Ryr2 expression in iWAT, but no changes in gene expression were found in iBAT or skeletal muscle (Fig. 5G). The lipid futile cycle showed more pronounced changes in response to overfeeding. Although expression of some genes decreased (e.g., Atgl in iBAT, Dgat1 in iWAT), that of others increased (Gk in iBAT). Many transcripts (namely, Hsl, Mgl, Agpat, and Dgat1 in iBAT; and Atgl, Gk, Gpat, Agpat, and Dgat2 in iWAT) transiently increased during overfeeding (Fig. 5H).

The gene expression data indicate that both the creatine futile cycle and the lipid futile cycle may contribute to the increased energy expenditure associated with overfeeding. However, the use of water-infused, chow-fed control mice from previous studies may have obscured some of the observed effects. To gain a clearer understanding of the contributions of various thermogenic gene programs to the metabolic phenotype, additional studies using liquid-diet control conditions are warranted.

Rodent studies investigating changes in energy expenditure during overfeeding present contrasting findings. Two rat studies observed a transient increase in energy expenditure at the end of overfeeding periods, with the effects disappearing by either d1 (6) or d3 (11) into the recovery phase. Conversely, other studies in rats (14–16) found no changes in energy expenditure beyond those attributable to increased body size during overfeeding, even at high calorie surpluses (200%). Similarly, the only mouse study, to our knowledge, that measured energy expenditure in response to overfeeding reported no increase after 2 weeks of intragastric overfeeding (17). Ravussin et al. (17) suggested that a decrease in energy efficiency (weight gain per calorie ingested) after overfeeding in mice could indicate increased energy expenditure or excretion. Here, we show that during overfeeding, energy expenditure increases slightly beyond what can be explained by the increased body weight. However, this change did not reach statistical significance, possibly due to the duration of the overfeeding period and/or limited sample size. The pronounced negative energy balance in the recovery period appears to be driven by a strong suppression of energy intake and potentially a sustained increase in energy expenditure. It remains to be determined if longer-term intragastric overfeeding increases energy expenditure and/or if other variables, such as varying levels of caloric excess and differences in animal species and strains, affect energy expenditure differently.

Assessment of energy expenditure in human overfeeding trials is challenging and has yielded variable and inconsistent results (4,8). Our findings in mice show a trend toward increased energy expenditure, with similar weight gain and recovery patterns in UCP1 KO mice compared with WT littermates in response to overfeeding. Although our data from UCP1 KO mice provide valuable insights, it is crucial to consider that germline UCP1 deficiency may induce compensatory mechanisms and secondary metabolic defects (32,33), which could potentially confound our findings. Moreover, the UCP1 KO study was conducted with suboptimal control conditions, using chow-fed, water-infused controls instead of liquid-diet-infused controls, which may have masked subtle effects in this experiment. Futile metabolic cycles in thermogenic fat and skeletal muscle have been linked to increased energy expenditure (28–31). Here, we observed subtle changes in the expression of genes encoding components of futile cycles in adipose tissue, suggesting that these futile metabolic cycles may play a minor role in counteracting weight gain in response to overfeeding. However, it is important to note that these experiments were conducted with suboptimal controls, which may have led to an underestimation of these effects. More studies are needed to better elucidate the potential contributions of both UCP1-dependent heat production and other thermogenic futile cycles.

Irrespective of these limitations, our data largely align with those of previous human studies suggesting minimal BAT activation and limited changes in energy expenditure in response to overfeeding (8–10,34). This supports the notion that luxuskonsumption might not be a major contributor to the defense against weight gain in humans. However, in some human studies, researchers found an increase in total energy expenditure in response to overfeeding (3,35). These contrasting findings and methodological differences between human studies and rodent intragastric overfeeding limit the direct translation of our mouse data to humans. Although our findings suggest a limited impact of UCP1-mediated thermogenesis in mitigating overfeeding-induced weight gain in mice, this does not preclude the possibility that more profound adaptive increases in energy expenditure and/or energy excretion could offer protection against weight gain in other scenarios. For instance, some mouse strains and humans with higher resistance to obesity may exhibit such adaptive responses (4,18). In addition to factors such as overfeeding duration, degree of calorie excess, and interindividual variability, the potential influence of diet composition on the response to overfeeding should also be considered. Unlike our study, which used a liquid diet lacking fiber, human overfeeding studies typically involve solid foods that contain fiber. Fiber influences fecal energy excretion and energy balance (36). Human studies have reported an increase in total fecal energy output in response to overfeeding (3,19), potentially explained by the increased food intake and presence of fiber in the diet, which promotes fecal bulking and potentially higher energy excretion compared with a fiber-deficient liquid diet.

In this study, we integrated intragastric overfeeding in rodents with high-resolution assessments of whole-body energy expenditure, metabolic fuel utilization, and locomotor activity across both overfeeding and recovery phases. This comprehensive approach, complemented by measurements of fecal energy excretion, provides a detailed picture of the metabolic and energy balance changes associated with short-term overfeeding in mice. One of our key findings, the reduced food intake seen in response to intragastric overfeeding, aligns with prior research showing significant appetite suppression in animals subjected to overfeeding (5,20,25,37). Although we observed a trend toward an adaptive increase in energy expenditure during overfeeding, the immediate shift toward fat oxidation after overfeeding might constitute another important adaptive response contributing to resistance against excessive weight gain (38). Hence, although the forced increase in energy intake caused by intragastric infusion seems to drive weight gain during overfeeding, profound shifts in metabolic fuel oxidation may causally contribute to the hypophagic and anorectic responses seen when overfeeding stops (39,40). Further enhancing temporal data resolution could be valuable in elucidating the precise sequence of physiological changes underlying the homeostatic response to overfeeding. Additionally, longer overfeeding interventions, using larger animal models, exploring diverse diet compositions, and evaluating alternative routes of energy loss such as urine and skin (18,41), would offer a more comprehensive understanding of the physiological response to overfeeding.

Acknowledgments. The authors thank Dr. Vibeke Kruse and Dr. Morten Dall and members of the Rodent Metabolic Phenotyping Platform at the Novo Nordisk Foundation Center for Basic Metabolic Research (CBMR) for their help with intragastric surgeries, data export, and setup of the Sable System. We also thank members of the Clemmensen Group for stimulating discussions and Dr. Vaida Juozaityté for helping with collection of fecal pellets. The authors thank Dr. Johanna Bruder and Dr. Martin Klingenspor (Technical University of Munich) for kindly providing advice on fecal pellet collection. Dr. Tao Ma and Dr. Astrid Linde Basse (Gerhart-Hines Group at the CBMR) kindly provided founders to establish a UCP1 KO mouse colony for overfeeding experiments.

Funding. P.R.-R. is supported by the Consejería de Universidad, Investigación e Innovación (grant A-EXP-161-UGR23), and by the ERDF Andalusia Program 2021–2027. J.L. is supported by the BRIDGE – Translational Excellence Programme (www.bridge.ku.dk) at the Faculty of Health and Medical Sciences, University of Copenhagen, which is funded by the Novo Nordisk Foundation (NNF) (grant NNF20SA0064340). Financial support for C.C. was provided through Lundbeck Foundation (fellowship R238-2016-2859) and the NNF (grants NNF17OC0026114 and NNF22OC0073778). The Novo Nordisk Foundation Center for Basic Metabolic Research is an independent Research Center based at the University of Copenhagen, Denmark, and partially funded by an unconditional donation from the NNF (www.cbmr.ku.dk; grants NNF18CC0034900 and NNF23SA0084103). M.K. was supported by the Deutsche Forschungsgemeinschaft (grant KL 3285/5-1), the German Center for Diabetes Research (grants 82DZD03D03 and 82DZD03D1Y), and the NNF (grant NNF19OC0055192).

Duality of Interest. C.C. is a cofounder of Ousia Pharma ApS, a biotech company developing therapeutics for treatment of metabolic disease. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. P.R.-R., C.L., and C.C. conceived the study. P.R.-R., C.S., and C.L. performed animal surgeries and executed the mouse in vivo studies. M.K. analyzed the energy density of food and feces. C.G. performed gene expression analyses. P.R.-R. analyzed the data and wrote the draft of the manuscript. C.L. and J.L. reviewed and edited the draft of the manuscript. All authors contributed to data interpretation and provided input to the manuscript. C.C. supervised the study and reviewed and edited the manuscript. C.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes full responsibility for the integrity of data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at CPH BAT 2024, Copenhagen, Denmark, 27–31 May 2024.