Loss of tubby in MC4R– and VGLUT2–expressing neurons impairs energy homeostasis

Abstract

Obesity is a major global health crisis, yet the molecular mechanisms underlying adult-onset metabolic dysfunction remain incompletely understood. The tubby mouse is a foundational genetic model of maturity-onset obesity; however, the specific tissues and cell populations responsible for its metabolic phenotype have long remained elusive. Here, we demonstrate that the loss of tubby disrupts the coordinated regulation of energy intake and expenditure, leading to a sustained positive energy balance. Using cell-type-specific genetic tools, we identified MC4R-expressing and VGLUT2-expressing neurons as essential sites of tubby function. We found that tubby acts through the combined contribution of these neuronal populations, as selective deletion in either MC4R or VGLUT2 neurons is sufficient to phenocopy key features of the global Tub mutant. Together, these findings establish tubby as a central neuronal regulator of systemic energy homeostasis and define an excitatory MC4R–VGLUT2 circuit that governs feeding behavior and metabolic output.

INTRODUCTION

Obesity represents a growing global health challenge and a major risk factor for metabolic syndrome, cardiovascular dysfunction, and reduced lifespan (1). Pathological weight gain arises from the chronic disruption of energy homeostasis—the coordinated regulation of food intake and energy expenditure by central and peripheral systems. Although both peripheral metabolic pathways and central neuronal mechanisms have been extensively studied, how these systems are integrated to maintain long-term energy balance remains incompletely understood, particularly at the level of defined cell types.

Genetic models have provided critical insights into the biological processes governing energy homeostasis. For example, the discovery of the leptin-melanocortin pathway established the hypothalamus as a central regulator of appetite and metabolic rate (2). Among these models, the tubby mouse has long served as a canonical model of late-onset obesity (3). Tubby is the founding member of the tubby-like protein (TULP) family (4). A spontaneous mutation in the Tub gene causes maturity-onset obesity, retinal degeneration, and hearing loss in mice (5, 6). These phenotypes were later confirmed to result from loss of tubby function (7), highlighting its essential role in multiple physiological processes.

While the molecular bases of retinal degeneration and hearing loss in the tubby mouse have been extensively characterized (8, 9), the mechanisms by which tubby regulates body weight remain unclear. Notably, tubby mice consume less food than control animals in some reports (10, 11), suggesting that excessive caloric intake is unlikely to be the sole driver of obesity in this model. Instead, obesity has been associated with reduced energy expenditure, decreased physical activity (10, 11), and altered carbohydrate metabolism (10). Nevertheless, the number of studies systematically validating these observations remains limited. Defining the specific tissues and cell types responsible for these effects will be critical for understanding the mechanistic basis of tubby-dependent energy homeostasis.

In this study, we investigate the cellular basis of tubby-mediated control of energy homeostasis. We show that loss of Tub results in a dual disruption of energy balance, characterized by increased food intake and reduced energy expenditure. Using targeted deletions, we identify melanocortin-4 receptor (MC4R)–expressing neurons and VGLUT2-expressing neurons as key mediators of tubby function. While deletion of Tub in either neuronal population leads to significant weight gain, combined loss in both populations produces an additive exacerbation of obesity that more closely phenocopies the global Tub mutant. These findings define the neuronal substrates through which tubby regulates energy balance and provide mechanistic insight into the cellular basis of adult-onset obesity.

RESULTS

Tubby mice show adult-onset obesity accompanied by increased adiposity

To examine metabolic alterations associated with Tub deficiency, we longitudinally monitored body weight from juvenile stages (5 weeks of age) through adulthood (up to 20 weeks) in both male and female mice. Body weights of Tubby mice (Tub^tub/tub) did not differ from those of control littermates during early postnatal development. However, beginning at 8-9 weeks of age, Tubby mice exhibited a progressive increase in body weight in both males and females (Fig. 1A, B). By 20 weeks of age, body weight was significantly higher in Tubby mice than in controls in both sexes.

Fig. 1.

Loss of Tubby function causes obesity and progressive impaired glucose homeostasis. (A, B) Longitudinal body weight measurements of male (A) and female (B) control and tubby mice (tub/tub) (Male: control, n = 8; tub/tub, n = 7, Female: control, n = 12; tub/tub, n = 11). (C-E) Body-composition analysis in male mice at 8 and 16 weeks of age showing (C) total body mass, (D) fat mass, and (E) lean mass (control, n = 5; tub/tub, n = 7). (F) Fed and fasted blood glucose levels in male mice at 8 weeks of age (Fasted: control, n = 6; tub/tub, n = 6, Fed: control, n = 6; tub/tub, n = 7). (G, H) Glucose tolerance test (GTT; G) and insulin tolerance test (ITT; H) with corresponding area under the curve (AUC) analyses in male mice at 8 weeks of age (GTT: control, n = 6; tub/tub, n = 6, ITT: control, n = 6; tub/tub, n = 7). (I) Fed and fasted blood glucose levels in male mice at 16 weeks of age (Fasted: control, n = 8; tub/tub, n = 8, Fed: control, n = 6; tub/tub, n = 7). (J, K) GTT (J) and ITT (K) with corresponding AUC analyses in male mice at 16 weeks of age (GTT: control, n = 8; tub/tub, n = 6, ITT: control, n = 6; tub/tub, n = 7). Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with control. # indicates a significant genotype effect across the indicated age range (P < 0.05).

We next assessed body composition using dual-energy X-ray absorptiometry (DEXA) (Fig. 1C-E). At 8 weeks of age, total body mass, fat mass, and lean mass were comparable between Tubby and control mice. In contrast, at 16 weeks of age, when Tubby mice displayed a clear obesity phenotype, both total body mass and fat mass were significantly increased in Tubby mice compared with controls (Fig. 1C, D). Lean mass did not differ significantly between genotypes at either age (Fig. 1E). These results indicate that Tub deficiency leads to adult-onset obesity driven predominantly by increased adiposity.

Tubby mice exhibit progressive impairment of glucose homeostasis and insulin sensitivity

To evaluate the metabolic consequences of obesity in Tubby mice, we examined glucose homeostasis at 8 and 16 weeks of age, corresponding to pre-obese and obese stages, respectively. At both ages, fasting and fed blood glucose were comparable between Tubby mice and controls (Fig. 1F, I), indicating preserved basal glycemic control.

Despite similar baseline glucose levels, glucose tolerance tests revealed impaired glucose clearance in Tubby mice. At 8 weeks of age, Tubby mice exhibited a modest but significant delay in glucose clearance following glucose administration, as reflected by slightly increased glucose levels at middle time points of test and elevated glucose AUC values (Fig. 1G). At 16 weeks of age, this defect was markedly exacerbated with sustained hyperglycemia throughout the test and a pronounced increase in glucose AUC values (Fig. 1J).

Insulin tolerance tests (ITTs) further demonstrated an age-dependent decline in insulin sensitivity. While insulin-induced glucose lowering was comparable between genotypes at 8 weeks of age (Fig. 1H), Tubby mice at 16 weeks exhibited a blunted glycemic response to insulin, resulting in persistently elevated blood glucose levels and significantly increased ITT AUC values (Fig. 1H, K). Together, these data indicate that tubby-associated obesity is accompanied by progressive glucose intolerance, with insulin resistance emerging at later stages of Tubby mice.

A newly generated Tub deletion allele recapitulates the obesity phenotype

To enable tissue– and cell–type–specific analysis of tubby function, we generated a conditional Tub knockout allele (12) and first characterized the phenotype resulting from whole-body deletion. Germline deletion of the Tub coding sequence using E2A-Cre produced viable Tub-null mice (Tub^−/−). Both male and female Tub^−/− mice exhibited progressive body weight gain beginning at 8-9 weeks of age, reaching significantly higher body weights than control littermates by 20 weeks of age (Fig. 2A, B). Body composition analysis revealed that Tub^−/− mice displayed increased total body mass and fat mass at both 8 and 16 weeks of age, with only modest changes in lean mass (Fig. 2C-E). Thus, whole-body deletion of Tub phenocopies the adult-onset obesity observed in the original Tubby mouse model, demonstrating that this phenotype is robust and reproducible across independent loss-of-function alleles.

Fig. 2.

Tub knockout mice display heterogeneous food intake with reduced energy expenditure. (A, B) Longitudinal body weight measurements of male (A) and female (B) control and Tub^−/− mice (Male: control, n = 13; Tub^−/−, n = 11, Female: control, n = 10; Tub^−/−, n = 5). (C-E) Body-composition analysis in male mice showing (C) total body mass, (D) fat mass, and (E) lean mass (control, n = 4; Tub^−/−, n = 5). (F, H) Weekly food intake in male (F) and female (H) mice, shown as the average across an 8-week period (Male: control, n = 14; Tub⁻/⁻, n = 12, Female: control, n = 11; Tub⁻/⁻, n = 10). (G, I) Weekly food-intake trajectories of Tub⁻/⁻ mice over an 8-week period after stratification into low–food-intake (Lo) and high–food-intake (Hi) subgroups in males (G) and females (I). (J, L) Body weight at 18 weeks of age in male (J) and female (L) mice, shown for control and for Tub⁻/⁻ mice stratified into Lo and Hi subgroups. (K, M) Energy expenditure plotted against body weight in male (K) and female (M) mice at 8 weeks of age (Male: control, n = 6; Tub⁻/⁻, n = 5, Female: control, n = 5; Tub⁻/⁻, n = 5). (N, O) Locomotor activity measured across light and dark phases in male (N) and female (O) mice at 8 weeks of age (Male: control, n = 6; Tub⁻/⁻, n = 5, Female: control, n = 5; Tub⁻/⁻, n = 5). Data are presented as mean ± SEM. Two-sample t-tests were used unless otherwise indicated. Energy-expenditure analyses (K, M) were performed using ANCOVA with body weight as a covariate. *P < 0.05, **P < 0.01, ***P < 0.001 compared with control. ^#indicates a significant genotype effect across the indicated age range (P < 0.05).

Increased food intake and reduced energy expenditure jointly drive obesity in Tub-deficient mice

To define the physiological basis of obesity in Tub-deficient mice, we quantified food intake in both male and female mice. Average weekly food intake was significantly higher in Tub⁻/⁻ mice than in control littermates over an 8-week period in males, but not in females (10-18 weeks of age) (Fig. 2F, H). Food intake among Tub⁻/⁻ mice showed marked inter-individual variability and could be divided into higher- and lower-intake subgroups relative to the group average (Fig. 2F-I). In males, the lower-intake Tub⁻/⁻ subgroup consumed amounts comparable to controls (Fig. 2G), whereas in females, the lower-intake Tub⁻/⁻ subgroup consumed less food than controls (Fig. 2I).

To assess the contribution of food intake to obesity, we analyzed the relationship between food intake and body weight. Higher-intake Tub⁻/⁻ mice exhibited markedly increased body weight compared with controls, while lower-intake Tub⁻/⁻ mice also displayed increased body weight despite normal or reduced food intake in male and females, respectively (Fig. 2J, L). These findings indicate that hyperphagia contributes to obesity in a subset of Tub-deficient mice but is not sufficient to fully explain the obese phenotype.

We therefore examined energy expenditure using indirect calorimetry. In male mice, the positive relationship between body weight and energy expenditure observed in controls was attenuated in Tub⁻/⁻ mice, indicating reduced mass-adjusted metabolic output (Fig. 2K). In females, energy expenditure correlated with body weight similarly to controls, although the relationship was shifted downward, resulting in lower absolute energy expenditure at matched body weights (Fig. 2M). Locomotor activity did not differ between genotypes in either sex (Fig. 2N, O). Together, these data demonstrate that tubby-associated obesity results from the combined effects of increased caloric intake and impaired energy expenditure.

Tubby-associated obesity is independent of developmental defects

To distinguish developmental defects from adult tubby function, we induced Tub deletion after development by tamoxifen-mediated recombination at 5 weeks of age. Despite post-developmental loss of Tub, R26-CreERT; Tub^fl/fl mice exhibited a progressive increase in body weight in both males and females compared with controls (Fig. 3). These results indicate that tubby is required in adulthood to maintain energy balance and that tubby-associated obesity is not a consequence of developmental defects.

Fig. 3.

Adult-onset loss of Tub is sufficient to drive obesity. (A, B) Body-weight trajectories of male (A) and female (B) mice following tamoxifen-induced, adult-stage deletion of Tub. Control mice were Tub^fl/fl (Male: control, n = 3; R26-CreERT; Tub^fl/fl, n = 3, Female: control, n = 7; R26-CreERT; Tub^fl/fl, n = 6). Data are presented as mean ± SEM. *P < 0.05 compared with control. Arrows indicate the timing of tamoxifen administration.

MC4R- and VGLUT2-expressing neurons jointly mediate tubby-dependent control of body weight

To identify the cellular substrates underlying tubby-dependent regulation of energy homeostasis, we performed a systematic Cre-based deletion screen across peripheral and central tissues using Tub^fl/fl mice. Deletion of Tub in peripheral metabolic tissues, including adipocytes (Adipoq-Cre) and brown adipocytes (Ucp1-Cre), did not affect body weight, indicating that tubby function in adipose tissues is dispensable for energy balance (Supplementary Fig. 1A, B).

Within the nervous system, deletion of Tub in VGLUT2-expressing glutamatergic neurons under high fat diet conditions resulted in a clear obesity phenotype, whereas deletion in VGAT-expressing GABAergic neurons had no obvious effect (Supplementary Fig. 1C, D). We then examined Vglut2-Cre; Tub^fl/fl mice under normal diet conditions to exclude potential confounding effects of nutrient excess, which confirmed a selective requirement for tubby in excitatory neuronal populations (Fig. 4A, B). Given tubby’s established role in ciliary GPCR trafficking (13-16), we next examined MC4R-expressing neurons and found that Tub deletion using Mc4r-Cre also led to increased body weight (Fig. 4C, D).

Fig. 4.

Deletion of Tub in VGLUT2- and MC4R-expressing neurons recapitulates the obesity phenotype. (A-F) Body-weight trajectories of male (A, C, E) and female (B, D, F) mice with conditional deletion of Tub (cKO) in VGLUT2-expressing neurons (A, B; Vglut2-Cre; Tub^fl/fl), MC4R-expressing neurons (C, D; Mc4r-Cre; Tub^fl/fl), or both neuronal populations (E, F; Vglut2-Cre; Mc4r-Cre; Tub^fl/fl), compared with corresponding littermate controls. Vglut2-Cre cohort (A, B): Male—control, n = 9; cKO, n = 7. Female–control, n = 7; cKO, n = 3. Mc4r-Cre cohort (C, D): Male—control, n = 16; cKO, n = 20. Female—control, n = 7; cKO, n = 3. Vglut2-Cre; Mc4r-Cre cohort (E, F): Male—control, n = 8; cKO, n = 14. Female—control, n = 2; cKO, n = 6. Control groups include Tub_fl/+, Tub^fl/fl, or Cre^+/−; Tub_fl/+. (G, H) Body weight at 18 weeks of age for male (G) and female (H) mice corresponding to the cohorts shown in (A-F). All comparisons were performed using littermate controls. Data are presented as mean ± SEM. Two-sample t-tests were used for statistical analysis. *P < 0.05, **P < 0.01 compared with control. ^#indicates a significant genotype effect across the indicated age range (P < 0.05).

Because MC4R regulates energy homeostasis through multiple brain regions (17), we further tested several MC4R-expressing neuronal populations implicated in metabolic control. However, Tub deletion using Sim1-Cre (paraventricular nucleus of the hypothalamus), ChAT-Cre (cholinergic preganglionic parasympathetic neurons in the dorsal motor nucleus of the vagus and sympathetic neurons in the intermediolateral column), Dbh-Cre (noradrenergic neurons), or Phox2b-Cre (cholinergic preganglionic parasympathetic neurons) failed to produce a metabolic phenotype. Similarly, deletion of Tub in canonical hypothalamic feeding circuits––including Agrp-Cre and Pomc-Cre (arcuate nucleus), and Sf1-Cre (ventromedial hypothalamus)––did not recapitulate the obesity phenotype (Supplementary Fig. 1D-J). These findings argue against a primary role for canonical hypothalamic feeding neurons or autonomic efferent pathways in tubby-associated obesity.

Notably, combined deletion of Tub using Vglut2-Cre and Mc4r-Cre produced a more severe obesity phenotype than either manipulation alone, indicating additive effects and suggesting that tubby functions in partially distinct, rather than fully overlapping, VGLUT2^＋ and MC4R^＋ neuronal populations to regulate body weight. Together, these results define a restricted neuronal subset in which Tubby is required for the maintenance of energy homeostasis. 

DISCUSSION

In this study, we provide a physiological and cellular characterization of the adult-onset obesity phenotype in the canonical Tubby mouse model. We show that Tub deficiency disrupts energy homeostasis through a dual mechanism, combining increased food intake in a subset of animals with a marked reduction in mass-adjusted energy expenditure. Our data identify VGLUT2-expressing excitatory neurons and MC4R–expressing neurons as key cellular substrates through which tubby regulates body weight.

A central finding of this work I[s the resolution of conflicting reports regarding feeding behavior in Tubby mice (6, 10, 11). By analyzing individual animals, we observed substantial inter-individual variability in food intake among Tub⁻/⁻ mice (Fig. 2G, I). While some mice exhibited pronounced hyperphagia, others consumed amounts comparable to—or lower than—controls yet still developed severe obesity (Fig. 2J, L). These results indicate that increased caloric intake can contribute to accelerated weight gain but is neither necessary nor sufficient to explain the phenotype. Instead, impaired energy expenditure emerges as the dominant and consistent driver of sustained positive energy balance (Fig. 2K, M). Notably, the lower-intake Tub⁻/⁻ subgroup also differed by sex. In males, lower-intake Tub⁻/⁻ mice consumed food in amounts similar to controls (Fig. 2G), whereas in females, lower-intake Tub⁻/⁻ mice consumed less food than controls. This observation is in line with previous reports that female tubby mice show significantly reduced food intake during the light period, whereas male tubby mice display only a similar non-significant tendency (10), raising the possibility of a sex-specific component in feeding behavior. Despite the development of obesity in both sexes, the reduced-intake phenotype was more evident in females. Further studies will be required to determine whether this sex difference reflects distinct underlying metabolic mechanisms.

Using a Cre-based deletion strategy, we identified the neuronal populations through which tubby regulates energy balance. Deletion of Tub in either VGLUT2-expressing excitatory neurons or MC4R-expressing neurons was sufficient to induce obesity, while combined deletion in both populations produced an additive phenotype that more closely resembled whole-body Tub deficiency (Fig. 4). These findings suggest that tubby functions within excitatory and melanocortin-related circuits that together regulate metabolic homeostasis. Whether these represent parallel pathways or reflect incomplete deletion within overlapping neuronal populations will require further investigation.

Importantly, inducible deletion of Tub in adulthood recapitulated the obesity phenotype, demonstrating that tubby-associated obesity is not a consequence of developmental abnormalities (Fig. 3). Rather, tubby is required for the maintenance of energy-balance circuits in the mature brain, highlighting its role in adult neuronal function.

Given tubby’s reported involvement in ciliary GPCR trafficking, tubby may modulate the localization or signaling of metabolic GPCRs within excitatory neurons (15). While it is reported that ciliary localization of MC4R is important for energy homeostasis (18), tubby mouse or deletion of Tub in MC4R-relevant brain regions did not phenocopy the full spectrum of MC4R dysfunction. These observations suggest that tubby is likely to influence other signaling pathways beyond MC4R. Identifying these mechanisms will be an important direction for future studies.

A limitation of the present study is that, although we identify VGLUT2- and MC4R-expressing neurons as critical sites of Tub action, metabolic phenotyping was not performed in the Vglut2-Cre; Tub^fl/fl, Mc4r-Cre; Tub^fl/fl, or Vglut2-Cre; Mc4r-Cre; Tub^fl/fl cohorts. Therefore, we cannot determine the relative contribution of altered food intake and reduced energy expenditure to the phenotype in each conditional knockout line. Moreover, the precise brain regions and circuit architectures, and neuronal activity dynamics involved remain unresolved. Further studies employing spatially restricted genetic and functional approaches will be necessary to define the relevant neural substrates.

In summary, our findings establish tubby as an essential regulator of adult energy homeostasis acting within excitatory and melanocortin-related neuronal populations. By reconciling variability in feeding behavior and identifying a primary defect in energy expenditure, this study reframes tubby-associated obesity as a disorder of metabolic regulation rather than appetite alone.

MATERIALS AND METHODS

Animals

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Yonsei University College of Medicine (19). Tub floxed mice (Tub^fl/fl) were crossed with E2a-Cre mice expressing Cre in germ cells to produce Tub general knockout mice (20). To generate Cre-dependent tub conditional knockout mice, Tub^fl/fl were crossed with cell type specific Cre-driver lines shown in Supplementary Table 1. All Cre-driver lines were backcrossed five times into a C57BL/6J background before being used for experiments.

Mice were fed either a normal chow diet (PicoLab^Ⓡ Rodent Diet 20, #5053, LabDiet, St. Louis, MO) or a high-fat diet (60 kcal% fat, D12492, Research Diets, New Brunswick, NJ) ad libitum. All mice were housed under controlled temperature and humidity conditions with a 12-h light/12-h dark cycle (lights off, 20:00-08:00) at the animal facility of Yonsei University College of Medicine.

All experiments were performed using littermate control and KO mice collected and analyzed together. Detailed information on cohort and litter composition for each dataset is provided in Supplementary Table 2.

Glucose and insulin tolerance tests

Glucose tolerance tests were performed in mice 8 or 16 weeks of age after an 18-h fast with water available ad libitum. Blood glucose was measured from tail nick samples using Ascensia Contour glucometer (Bayer HealthCare, Leverkusen, Germany). After baseline (fasting) glucose measurement, mice were injected intraperitoneally with glucose (1 g/kg), and blood glucose was measured at 0, 15, 30, 60, 90, and 120 min post-injection. For insulin tolerance tests, mice at 9 or 17 weeks of age were fasted for 2 hours with water available ad libitum. After baseline glucose measurement, insulin (0.9 U/kg) was injected intraperitoneally, and blood glucose was measured at 0, 15, 30, 60, 90, and 120 min post-injection.

Analysis of body composition by dual-energy x-ray absorptiometry (DEXA)

Body composition was assessed by DEXA using an InAlyzer system (MediKors, Seongnam-si, Korea). Mice were anesthetized by intraperitoneal injection of a Zoletil (tiletamine–zolazepam; Virbac, Carros, France)–xylazine mixture (Rompun; Elanco, Greenfield, IN) at doses of 30 mg/kg Zoletil and 10 mg/kg xylazine. Following induction of anesthesia, mice were positioned prone on the scanner bed and immobilized to minimize motion. The device was calibrated prior to scanning according to the manufacturer’s instructions using the manufacturer-supplied calibration phantom. One whole-body scan per mouse was acquired, and fat mass and lean mass were calculated using the built-in analysis program.

Metabolic analysis

The metabolic rate was measured in control (Tub^＋/＋ or Tub^＋/−) and Tub^−/− mice at 8 weeks of age fed a normal chow diet. The mice were individually housed for 2 days before the experiment to allow acclimation in CaloSys indirect calorimetry System (TSE Systems, Bad Homburg, Germany). After adaptation period, O₂ consumption, CO₂ production and heat generation were measured continuously for 5 days.

Tamoxifen administration

Mice at 5 weeks of age (Tub^fl/fl and R26-CreERT; Tub^fl/fl) were administered daily by oral gavage (150 mg/kg) for 5 days. Tamoxifen (T5648, Sigma-Aldrich, St. Louis, MO) was prepared at 30 mg/ml in corn oil. Briefly, corn oil was sterilized by filtration through a 0.2-μm filter and warmed to 50°C, after which tamoxifen was added. The mixture was shaken overnight at 50°C until fully dissolved. Tamoxifen was administered by oral gavage at a dose of 3 mg in a volume of 100 μl.

Statistics

All data are presented as the mean ± s.e.m. Normality of residuals was assessed with Q–Q plots and homoscedasticity (equal variance) was assessed using F-test for two-group comparisons. Statistical differences between two groups were analyzed using two-tailed unpaired t-tests. For experiments with three or more groups, one-way ANOVA followed by Tukey’s post hoc test was applied; homogeneity of variances was evaluated using the Brown–Forsythe test. Analyses were performed using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA). Indirect calorimetry data (energy expenditure) were analyzed using the CalR web tool (CalR; CalRapp.org) based on a general linear model/ANCOVA with body weight as a covariate.