Mitochondria play a vital role in white adipose tissue (WAT) homeostasis including adipogenesis, fatty acid synthesis, and lipolysis. We recently reported that the mitochondrial fusion protein optic atrophy 1 (OPA1) is required for induction of fatty acid oxidation and thermogenic activation in brown adipocytes. In the current study we investigated the role of OPA1 in WAT function in vivo. We generated mice with constitutive or inducible knockout of OPA1 selectively in adipocytes. Studies were conducted under baseline conditions, at thermoneutrality, following high-fat feeding or during cold exposure. OPA1 deficiency reduced mitochondrial respiratory capacity in white adipocytes, impaired lipolytic signaling, repressed expression of de novo lipogenesis and triglyceride synthesis pathways, and promoted adipose tissue senescence and inflammation. Reduced WAT mass was associated with hepatic triglycerides accumulation and glucose intolerance. Moreover, mice deficient for OPA1 in adipocytes had impaired adaptive thermogenesis and reduced cold-induced browning of subcutaneous WAT and were completely resistant to diet-induced obesity. In conclusion, OPA1 expression and function in adipocytes are essential for adipose tissue expansion, lipid biosynthesis, and fatty acid mobilization of WAT and brown adipocytes and for thermogenic activation of brown and beige adipocytes.

Obesity and its metabolic complications, such as type 2 diabetes and cardiovascular disease, are critical public health problems of modern society, with serious economic implications (1). White adipose tissue (WAT) plays a key homeostatic role in metabolic health by ensuring efficient energy storage and rapid mobilization of triglycerides to meet metabolic demands. Mitochondria are critical for maintaining metabolic homeostasis in white adipocytes via the regulation of processes such as adipogenesis, fatty acid synthesis and esterification, branched-chain amino acid catabolism, and lipolysis (2,3).

Recent studies suggest a role for mitochondrial dynamics in the regulation of adipose tissue function. Obesity in rodents and humans is associated with reduced expression of the fusion protein mitofusin 2 (MFN2) in white and brown adipocytes (4). Indeed, adipocyte-specific knockdown of Mfn2 in adult mice exacerbates adiposity, due in part to increased adipocyte proliferation and adipogenesis (4). Furthermore, the fission-controlling enzyme dynamin-related protein 1 (DRP1) is required for browning of human white adipocytes (5), and its deficiency in adipose tissue of mice impairs lipolysis and leads to reduced whole-body energy expenditure (6).

We recently demonstrated that selective deletion of the mitochondrial fusion protein optic atrophy 1 (OPA1) in brown adipose tissue (BAT) impairs fatty acid oxidation and thermogenic activation of BAT, inducing compensatory browning of WAT and improving thermoregulation and metabolism (7). To determine the role of OPA1 in WAT function, we constitutively and inducibly deleted OPA1 in adipocytes. Collectively, our data reveal a novel role for OPA1 in adipose tissue expansion and lipid synthesis and mobilization, defects in which promote adipocyte senescence and adipose tissue inflammation. They also identify a critical role for OPA1 in thermogenesis and browning of WAT.

Animals and Animal Care

Experiments were performed in male and/or female mice on a C57Bl/6J background. Animals were housed at 22°C with a 12-h light, 12-h dark cycle with free access to water and standard chow or special diets, unless otherwise noted. All mouse experiments were conducted in accord with the animal research guidelines from the National Institutes of Health and were approved by the University of Iowa Institutional Animal Care and Use Committee. Additional details on animal studies can be found in Supplementary Material.

Cell Culture and Treatments

WAT stromal vasculature fraction was isolated from 6-day-old OPA1 floxed mice harboring the tamoxifen-inducible cre recombinase, Cre-ERT2, under the control of the Adipoq gene promoter. Cells were grown and differentiated as previously described (8). Mature adipocytes were treated with 500 nmol/L 4-hydroxytamoxifen (Sigma-Aldrich, St. Louis, MO) for 72 h to induce recombination (knockout [KO] cells) or with vehicle solution (wild-type [WT] cells). Cells were then treated with the β3 receptor agonist CL 316,243 (CL) (5 μmol/L) for 6 h. Cells were washed with ice-cold PBS and harvested for subsequent analysis.

Histology

Fragments of gWAT and BAT were fixed in 10% Zinc-formalin (Thermo Fisher Scientific, Waltham, MA) and prepared for histology, as previously described (7). Additional details can be found in Supplementary Material.

Glucose and Insulin Tolerance Tests, Nuclear Magnetic Resonance, and Serum Analysis

Glucose and insulin tolerance tests were performed as previously described (7). Analysis of plasma insulin was measured after a 6-h fast with a commercially available kit (Ultra-Sensitive Mouse Insulin ELISA Kit; Chrystal Chem, Downers Grove, IL). Serum levels of FGF21 (ELISA kit; BioVendor, Asheville, NC), and free fatty acids (FFA) (LSBio, Seattle, WA) were measured with commercially available kits according to the manufacturers’ directions. Whole-body composition was measured by nuclear magnetic resonance in the Bruker minispec NF-50 instrument (Bruker, Billerica, MA).

Analysis of Triglyceride Levels

Liver triglycerides were extracted with use of a solution of isopropanol and Triton X-100 and measured with the EnzyChrom Triglyceride Assay Kit according to the manufacturer’s instructions (BioAssay Systems, Hayward, CA) (7).

RNA Extraction and Quantitative RT-PCR

Total RNA was extracted from tissues with TRIzol reagent (Invitrogen) and purified with the RNeasy kit (QIAGEN, Germantown, MD) as previously described (7). Additional details can be found in Supplementary Material.

Western Blot Analysis

Immunoblotting analysis was performed as previously described (9). Additional details can be found in Supplementary Material.

Mitochondrial Isolation and Respiratory Function

Mitochondrial fraction was isolated from adipose tissue as previously described (10), and the respiratory rates were measured with the Oroboros O2K Oxygraph system (Oroboros Instruments, Innsbruck, Austria). Additional details can be found in Supplementary Material.

Data Analysis

Unless otherwise noted, all data are reported as mean ± SEM. Student t test was performed for comparison of two groups, and two-way ANOVA followed by Tukey multiple comparisons test was used when more than three groups were compared. A probability value of P ≤ 0.05 was considered significantly different. Statistical calculations were performed with GraphPad Prism software (GraphPad, La Jolla, CA). The association between oxygen consumption and body mass was calculated with ANCOVA, with the CalR software (11).

Data and Resource Availability

All data generated during this study are included in the published article (and Supplementary Material). Resources related to this study, including mouse models, are available on reasonable request.

Constitutive Deletion of OPA1 in Adipocytes Impairs Mitochondrial Respiration and Induces Mild Lipodystrophy and Glucose Intolerance

OPA1 was selectively deleted in adipocytes (Ad-KO mice) through crossing of Opa1 floxed mice with mice harboring the Cre recombinase under the control of the adiponectin promoter (Adipoq Cre). OPA1 levels were significantly reduced in inguinal (iWAT) (Fig. 1A) and gonadal (gWAT) (Fig. 1B) WAT and in BAT of Ad-KO mice relative to WT littermate controls (Supplementary Fig. 1A). OPA1 deficiency impaired pyruvate/malate-supported state 3 mitochondrial respirations in mitochondria isolated from WAT (Fig. 1C and D) and BAT (Supplementary Fig. 1B). Similar to BAT-selective OPA1 KO mice (7), palmitoyl-carnitine–supported respirations were also attenuated in mitochondria isolated from BAT (Supplementary Fig. 1C). Morphologically, iWAT deficient for OPA1 had an increased number of enlarged adipocytes (Fig. 1E–G), while the number of adipocytes per area analyzed was reduced (Fig. 1H). Brown adipocytes had enlarged unilocular lipid droplets, suggesting whitening (Supplementary Fig. 1D). Body mass and body composition were unchanged in 8-week-old male mice (Fig. 1I and J). By 20 weeks of age, although body mass was not significantly different between genotypes, (Fig. 1I) fat mass (Fig. 1J) was significantly reduced in Ad-KO mice, while lean mass was unaffected (Fig. 1K). Although total fat mass was unchanged, gWAT (Fig. 1L) and iWAT (Fig. 1M) were significantly reduced in 8-week-old Ad-KO mice, while BAT mass was increased (Supplementary Fig. 1E). Finally, age-associated increases in gWAT (Fig. 1L) and iWAT (Fig. 1M) mass were prevented in 20-week-old Ad-KO mice. By 40 weeks of age, Ad-KO mice had reduced body mass (Supplementary Fig. 1F) and diminished gWAT (Supplementary Fig. 1G), iWAT (Supplementary Fig. 1H), and BAT (Supplementary Fig. 1I) mass, while liver mass (Supplementary Fig. 1J) was increased relative to body mass.

Figure 1

Constitutive deletion of OPA1 in adipocytes impairs mitochondrial function and induces mild lipodystrophy and glucose intolerance. A and B: OPA1 protein levels in 10-week-old WT and OPA1 Ad-KO mice normalized to GAPDH and the respective densitometric quantifications. A: iWAT. B: gWAT. C and D: Oxygen consumption rates in WAT from 20-week-old WT and Ad-KO mice. C: ADP-stimulated (state 3) pyruvate-malate (PM)-supported oxygen consumption rates in mitochondria isolated from iWAT. D: ADP-stimulated (state 3) pyruvate-malate–supported oxygen consumption rates in mitochondria isolated from gWAT. E: Representative images of H-E–stained histological sections of iWAT from 10-week-old WT and Ad-KO mice (n = 3). F: Calculation of mean adipocyte area. G: Frequency distribution for adipocyte area. H: Calculation of adipocyte number per area (data are represented as the average number of adipocytes per area, per mouse). IK: Body mass and body composition in 8 and 20-week-old Ad-KO mice and their age-matched WT controls. I: Body mass. J: Total body fat. K: Total lean mass. L: gWAT mass normalized to body mass (8 and 20 weeks of age). M: iWAT mass normalized to body mass (8 and 20 weeks of age). NP: Data collected in 10- to 12-week-old mice using the OxyMax Comprehensive Lab Animal Monitoring System at thermoneutrality (30°C). N: Oxygen consumption. O: Food intake. P: Locomotor activity. QT: Data collected in 20-week-old mice. Q: Liver mass normalized to body mass. R: Liver triglycerides levels. S: Glucose tolerance test. T: Area under the curve for the glucose tolerance test. U: Fasting glucose levels (after 6-h fast). Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or mice 8 weeks of age; #significantly different from WT at 20 weeks of age. hr, hour; OD, optical density.

Figure 1

Constitutive deletion of OPA1 in adipocytes impairs mitochondrial function and induces mild lipodystrophy and glucose intolerance. A and B: OPA1 protein levels in 10-week-old WT and OPA1 Ad-KO mice normalized to GAPDH and the respective densitometric quantifications. A: iWAT. B: gWAT. C and D: Oxygen consumption rates in WAT from 20-week-old WT and Ad-KO mice. C: ADP-stimulated (state 3) pyruvate-malate (PM)-supported oxygen consumption rates in mitochondria isolated from iWAT. D: ADP-stimulated (state 3) pyruvate-malate–supported oxygen consumption rates in mitochondria isolated from gWAT. E: Representative images of H-E–stained histological sections of iWAT from 10-week-old WT and Ad-KO mice (n = 3). F: Calculation of mean adipocyte area. G: Frequency distribution for adipocyte area. H: Calculation of adipocyte number per area (data are represented as the average number of adipocytes per area, per mouse). IK: Body mass and body composition in 8 and 20-week-old Ad-KO mice and their age-matched WT controls. I: Body mass. J: Total body fat. K: Total lean mass. L: gWAT mass normalized to body mass (8 and 20 weeks of age). M: iWAT mass normalized to body mass (8 and 20 weeks of age). NP: Data collected in 10- to 12-week-old mice using the OxyMax Comprehensive Lab Animal Monitoring System at thermoneutrality (30°C). N: Oxygen consumption. O: Food intake. P: Locomotor activity. QT: Data collected in 20-week-old mice. Q: Liver mass normalized to body mass. R: Liver triglycerides levels. S: Glucose tolerance test. T: Area under the curve for the glucose tolerance test. U: Fasting glucose levels (after 6-h fast). Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or mice 8 weeks of age; #significantly different from WT at 20 weeks of age. hr, hour; OD, optical density.

Close modal

OPA1 deletion exclusively in BAT reduced WAT mass due to increased energy expenditure (7). We therefore performed indirect calorimetry in 12- to 14-week-old OPA1 Ad-KO mice, a time point at which KO and control mice had similar body mass. ANCOVA analysis indicated a slight leftward shift in the relationship between body weight and oxygen consumption in Ad-KO mice, however the group effect was not significantly different. Thus, relative to total body mass, Ad-KO mice have similar oxygen consumption rates versus WT animals. Food intake (Fig. 1O) and locomotor activity (Fig. 1P) were also unchanged between genotypes, indicating that changes in energy balance likely did not play a role in the reduced adiposity of Ad-KO mice.

Knockout of OPA1 in BAT induces ER stress that increases fibroblast growth factor 21 (FGF21), which promoted leanness in isocaloric feeding conditions (7). ER stress was also induced in BAT of Ad-KO mice, as suggested by increased BiP protein levels (Supplementary Fig. 1K), and was accompanied by increased FGF21 protein levels in BAT (Supplementary Fig. 1L) and serum (Supplementary Fig. 1M) of Ad-KO mice. OPA1 and FGF21 were simultaneously deleted in adipocytes (Ad-DKO mice) (Supplementary Fig. 1N), to test the role of FGF21 in the phenotype observed. FGF21 circulating levels were completely normalized in Ad-DKO mice (Supplementary Fig. 1O). However, similar to Ad-KO mice, Ad-DKO mice had unchanged body mass at 8 weeks of age (Supplementary Fig. 1P) but had reduced gWAT (Supplementary Fig. 1Q) and iWAT mass (Supplementary Fig. 1R) compared with WT mice.

Liver mass (Fig. 1Q) and liver triglycerides levels (Fig. 1R) were similarly increased in OPA1 Ad-KO and in Ad-DKO mice at 20 weeks of age (Supplementary Fig. 1S). Impaired WAT expansion and increased ectopic deposition of triglycerides in the liver correlated with impaired glucose tolerance and fasting hyperglycemia in KO mice (Fig. 1S–U). This impairment in glucose metabolism was also observed in Ad-DKO mice (Supplementary Fig. 1TV). Thus, adipose tissue–derived FGF21 does not mediate the reduction in WAT mass or glucose intolerance in OPA1 Ad-KO mice.

Lipid Metabolism Is Impaired in Adipocytes Lacking OPA1

The reduced adipose tissue mass in Ad-KO mice, in concert with hepatic steatosis and impaired systemic glucose homeostasis, suggested that OPA1 deletion directly impaired adipocyte function, resulting in a lipodystrophic phenotype. Genes involved in lipolysis and triglyceride synthesis were repressed in iWAT (Fig. 2A), gWAT (Fig. 2B), and BAT (Supplementary Fig. 2A) of Ad-KO mice. Moreover, protein levels of the de novo lipogenesis enzymes fatty acid synthase (FASN) and stearoyl-CoA desaturase 1 (SCD1) were reduced in iWAT (Fig. 2C and D) and gWAT (Fig. 2E and F) of Ad-KO mice. Because adipocyte number was reduced in KO mice, we investigated markers of adipogenesis and adipocyte senescence (12). Adipogenesis genes were unchanged in gWAT of 8-week-old (Supplementary Fig. 2D) and 40-week-old (Supplementary Fig. 2E) mice and in iWAT of 20-week-old mice (Supplementary Fig. 2F). However, the senescence marker P16 was significantly induced in gWAT regardless of age (Supplementary Fig. 2D and E), suggesting increased adipocyte death. Interestingly, adipogenesis markers were markedly induced in BAT of 20-week-old Ad-KO mice, suggesting a depot-specific role for OPA1 in adipogenesis (Supplementary Fig. 2F). Inflammation often accompanies adipose tissue dysfunction (13). Although in 8-week-old mice there were no changes in inflammatory markers (Supplementary Fig. 2D), OPA1 deletion in adipocytes was associated with increased inflammation in the gWAT of 40-week-old mice (Supplementary Fig. 2E).

Figure 2

Lipid metabolism is impaired in adipocytes lacking OPA1. AB: Relative mRNA expression of lipid metabolism genes (Pnpla, Gpat, and Dgat) normalized to Gapdh in 8- to 10-week-old WT and Ad-KO mice. A: iWAT. B: gWAT. C: Protein levels of FASN in iWAT normalized to GAPDH. D: Protein levels of SCD1 in iWAT normalized to GAPDH. E: Protein levels of FASN in gWAT normalized to GAPDH. F: Protein levels of SCD1 in gWAT normalized to GAPDH. G: Protein levels of phosphorylated (p)HSL in iWAT normalized to total HSL. H: Protein levels of ATGL in iWAT normalized to tubulin. I: Protein levels of phosphorylated HSL in gWAT normalized to tubulin. J: Protein levels of ATGL in gWAT normalized to tubulin. K: Serum FFA levels in the fed and fasted states (12-h fast). LN: Immunoblot data in WT and OPA1-deficient primary white adipocytes treated with CL for 6 h. L: Protein levels of OPA1 normalized to total tubulin. M: Protein levels of phosphorylated HSL normalized to total HSL. N: Protein levels of phosphorylated perilipin normalized to total perilipin. Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or WT vehicle; #significantly different from WT treated with CL. OD, optical density.

Figure 2

Lipid metabolism is impaired in adipocytes lacking OPA1. AB: Relative mRNA expression of lipid metabolism genes (Pnpla, Gpat, and Dgat) normalized to Gapdh in 8- to 10-week-old WT and Ad-KO mice. A: iWAT. B: gWAT. C: Protein levels of FASN in iWAT normalized to GAPDH. D: Protein levels of SCD1 in iWAT normalized to GAPDH. E: Protein levels of FASN in gWAT normalized to GAPDH. F: Protein levels of SCD1 in gWAT normalized to GAPDH. G: Protein levels of phosphorylated (p)HSL in iWAT normalized to total HSL. H: Protein levels of ATGL in iWAT normalized to tubulin. I: Protein levels of phosphorylated HSL in gWAT normalized to tubulin. J: Protein levels of ATGL in gWAT normalized to tubulin. K: Serum FFA levels in the fed and fasted states (12-h fast). LN: Immunoblot data in WT and OPA1-deficient primary white adipocytes treated with CL for 6 h. L: Protein levels of OPA1 normalized to total tubulin. M: Protein levels of phosphorylated HSL normalized to total HSL. N: Protein levels of phosphorylated perilipin normalized to total perilipin. Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or WT vehicle; #significantly different from WT treated with CL. OD, optical density.

Close modal

In addition to reduced lipid synthesis, lipolysis signaling was also attenuated in Ad-KO mice, as demonstrated by reduced phosphorylation of hormone sensitive lipase (HSL) and reduced adipose tissue triglyceride lipase (ATGL) in iWAT (Fig. 2G and H), gWAT (Fig. 2I and J), and BAT (Supplementary Fig. 2B and C) in ad libitum–fed mice. To assess lipolytic reserve in vivo, we measured FFA levels in the serum of WT and Ad-KO mice after a 12-h fast. FFA levels were unchanged between genotypes when mice were fed ad libitum, but the fasting-induced increase in serum FFA observed in WT mice was prevented in Ad-KO mice (Fig. 2K). OPA1 may regulate norepinephrine-induced lipolysis by facilitating perilipin 1 phosphorylation in white adipocyte cultures (14). After a 12-h fast, perilipin phosphorylation levels were significantly higher in OPA1 Ad-KO mice (Supplementary Fig. 2G), suggesting that alternative mechanisms exist that increase perilipin phosphorylation in response to a physiological stimulus in vivo when OPA1 is absent. To determine the effect of Opa1 downregulation on lipolysis signaling in vitro, we deleted Opa1 in fully differentiated primary adipocytes and stimulated lipolysis with the β3 receptor agonist CL. Opa1 downregulation in mature adipocytes (Fig. 2L) attenuated CL-induced HSL phosphorylation (Fig. 2M), but perilipin phosphorylation was not significantly changed (Fig. 2N).

Mice Lacking OPA1 in Adipocytes Are Cold Intolerant

OPA1 deficiency in brown adipocytes impairs cold-induced activation of the thermogenic gene program in BAT while paradoxically inducing browning of WAT and improving thermoregulation (7). We therefore investigated the ability of OPA1 Ad-KO to adapt to cold stress. At thermoneutrality (∼30°C for mice), core body temperature was unchanged between WT and Ad-KO mice (Fig. 3A). However, on cold exposure (4°C for 2 h, after a 12-h fast), core body temperatures dropped faster in Ad-KO mice relative to WT littermate controls (Fig. 3B). Moreover, in response to 3 days of cold exposure (4°C), in ad libitum–fed mice, the average core body temperature recorded during the dark cycle was significantly reduced in Ad-KO mice, leading to severe hypothermia (Fig. 3C), as evidenced by the last temperature recorded per mouse (Supplementary Fig. 3A). Body mass (Supplementary Fig. 3B) was unchanged between cold-exposed mice, and no differences were detected in food intake (Supplementary Fig. 3C) or locomotor activity (Supplementary Fig. 3D) between genotypes during the 3-day cold exposure experiments.

Figure 3

Mice lacking OPA1 in adipocytes are cold intolerant. A: Averaged core body temperature measured by telemetry in WT and Ad-KO mice at thermoneutrality. B: Rectal temperature in WT and Ad-KO mice exposed to acute cold stress (4°C) over a period of 2 h. C: Averaged core body temperature measured by telemetry in WT and Ad-KO mice exposed to 4°C over the course of 3 days. D: Relative mRNA expression of thermogenic genes normalized to Gapdh in BAT of WT and Ad-KO mice after 3 days of cold exposure. E: UCP1 protein levels normalized to GAPDH in BAT of WT and Ad-KO mice kept either at 30°C or 4°C for 3 days. F: Relative Opa1 mRNA levels normalized to Gapdh in iWAT of WT mice kept either at 30°C or 4°C for 3 days. G: Relative mRNA expression of thermogenic genes normalized to Gapdh in iWAT of WT and Ad-KO mice after 3 days of cold exposure. H: UCP1 protein levels normalized to GAPDH in iWAT of WT and Ad-KO mice kept either at 30°C or 4°C for 3 days. Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or mice at 30°C; #significantly different from WT at 4°C. OD, optical density.

Figure 3

Mice lacking OPA1 in adipocytes are cold intolerant. A: Averaged core body temperature measured by telemetry in WT and Ad-KO mice at thermoneutrality. B: Rectal temperature in WT and Ad-KO mice exposed to acute cold stress (4°C) over a period of 2 h. C: Averaged core body temperature measured by telemetry in WT and Ad-KO mice exposed to 4°C over the course of 3 days. D: Relative mRNA expression of thermogenic genes normalized to Gapdh in BAT of WT and Ad-KO mice after 3 days of cold exposure. E: UCP1 protein levels normalized to GAPDH in BAT of WT and Ad-KO mice kept either at 30°C or 4°C for 3 days. F: Relative Opa1 mRNA levels normalized to Gapdh in iWAT of WT mice kept either at 30°C or 4°C for 3 days. G: Relative mRNA expression of thermogenic genes normalized to Gapdh in iWAT of WT and Ad-KO mice after 3 days of cold exposure. H: UCP1 protein levels normalized to GAPDH in iWAT of WT and Ad-KO mice kept either at 30°C or 4°C for 3 days. Data are expressed as means ± SEM. Significant differences were determined with Student t test or two-way ANOVA with a significance level of P < 0.05. *Significantly different vs. WT mice or mice at 30°C; #significantly different from WT at 4°C. OD, optical density.

Close modal

Similar to mice lacking OPA1 exclusively in BAT (7), activation of the thermogenic gene program (Fig. 3D) and protein levels of UCP1 (Fig. 3E) were significantly reduced in BAT of cold-exposed Ad-KO mice relative to WT controls. Opa1 levels were induced in iWAT of WT mice in response to cold exposure (Fig. 3F), suggesting that OPA1 may play a role in adipose tissue browning. Indeed, cold-induced browning of iWAT was significantly attenuated in Ad-KO mice, as demonstrated by reduced induction of thermogenic genes (Fig. 3G) and decreased UCP1 protein levels (Fig. 3H). These data confirm a role for OPA1 in cold-induced browning of iWAT and suggest that, in the absence of normal BAT thermogenic function, browning of iWAT is essential for maintaining core body temperature in mice.

OPA1 Ad-KO Mice Are Resistant to Diet-Induced Obesity

Reduced expression of TG synthesis genes and enzymes that promote de novo lipogenesis in OPA1 Ad-KO mice led us to investigate whether OPA1 deficiency would limit adipose tissue expansion in response to an obesogenic diet. High-fat diet–induced weight gain (Fig. 4A and B) and fat mass expansion (Fig. 4C) were completely prevented in Ad-KO mice after 12 weeks of high-fat feeding, while lean mass was unchanged between genotypes (Fig. 4D). BAT mass was unchanged regardless of diet or genotype (Fig. 4E), but the diet-induced increase in gWAT (Fig. 4F) and iWAT (Fig. 4G) mass observed in WT mice was absent in Ad-KO mice. Indirect calorimetry revealed no change in oxygen consumption between groups (Fig. 4H). ANCOVA analysis for VO2 as a function of body weight showed no significant differences in the group effect, despite an accentuated leftward shift in the Ad-KO mice group relative to WT mice fed HFD (Fig. 4I). Locomotor activity was unchanged between groups (Fig. 4J), while food intake was only significantly reduced in WT mice fed HFD relative to WT mice fed control diet (Fig. 4K).

Figure 4

OPA1 Ad-KO mice are resistant to diet-induced obesity. AP: Data from WT and Ad-KO mice fed either a control diet (Cont) (10% fat) or a high-fat diet (HFD) (60% fat) for 12 weeks (animals started the diet at ∼6 weeks of age). A: Growth curve. B: Total body mass. C: Total fat mass. D: Total lean mass. E: BAT mass normalized to body mass. F: gWAT mass normalized to body mass. G: iWAT mass normalized to body mass. H: Oxygen consumption. I: Oxygen consumption expressed as a function of body mass. J: Locomotor activity. K: Food intake. L: Liver triglycerides levels. M: Glucose tolerance test. N: Area under the curve for the glucose tolerance test. O: Insulin tolerance test. P: Area under the curve for the insulin tolerance test. Q: Fasting insulin levels. Data are expressed as means ± SEM. Significant differences were determined with two-way ANOVA, with a significance level of P < 0.05. *Significantly different vs. WT control; #significantly different vs. WT HFD. GTT, glucose tolerance test; hr, hour.

Figure 4

OPA1 Ad-KO mice are resistant to diet-induced obesity. AP: Data from WT and Ad-KO mice fed either a control diet (Cont) (10% fat) or a high-fat diet (HFD) (60% fat) for 12 weeks (animals started the diet at ∼6 weeks of age). A: Growth curve. B: Total body mass. C: Total fat mass. D: Total lean mass. E: BAT mass normalized to body mass. F: gWAT mass normalized to body mass. G: iWAT mass normalized to body mass. H: Oxygen consumption. I: Oxygen consumption expressed as a function of body mass. J: Locomotor activity. K: Food intake. L: Liver triglycerides levels. M: Glucose tolerance test. N: Area under the curve for the glucose tolerance test. O: Insulin tolerance test. P: Area under the curve for the insulin tolerance test. Q: Fasting insulin levels. Data are expressed as means ± SEM. Significant differences were determined with two-way ANOVA, with a significance level of P < 0.05. *Significantly different vs. WT control; #significantly different vs. WT HFD. GTT, glucose tolerance test; hr, hour.

Close modal

Despite baseline differences in chow diet, hepatic triglyceride accumulation following high-fat feeding was similar between WT and Ad-KO mice (Fig. 4L). Glucose homeostasis, measured by glucose tolerance testing (Fig. 4M and N), was significantly impaired in mice fed a high-fat diet but slightly improved in Ad-KO mice. Similarly, impaired insulin tolerance was mildly attenuated in Ad-KO mice, (Fig. 4O and P), while fasting insulin levels were similarly elevated in Ad-KO and WT mice (Fig. 4Q). To determine whether housing at room temperature contributed to the lean phenotype in these mice, we reared a separate cohort of mice at thermoneutral conditions (30°C), while simultaneously feeding them HFD or control diet for 8 weeks. Ad-KO mice raised at thermoneutrality remained resistant to DIO, as shown by reduced body mass (Supplementary Fig. 4A) and total fat mass (Supplementary Fig. 4B) relative to WT mice, whereas lean mass was unaffected (Supplementary Fig. 4C).

While Opa1 deletion exclusively in BAT promoted resistance to DIO that was independent of FGF21 induction (7), downregulation of Opa1 in skeletal muscle induced FGF21, which played a central role in mediating the resistance to DIO in that model (15). We therefore tested the requirement of adipose tissue FGF21 for the resistance to DIO in Ad-KO mice. Ad-DKO mice fed HFD for 12 weeks remained lean (Supplementary Fig. 4D) and had reduced total fat mass (Supplementary Fig. 4E) relative to WT littermate controls, while lean mass was unaltered (Supplementary Fig. 4F). Chow-fed Ad-DKO mice were glucose intolerant relative to WT controls and were equivalently glucose intolerant relative to WT mice following HFD (Supplementary Fig. 4G and H), suggesting a possible role for FGF21 in regulating glucose homeostasis in Ad-KO mice. Together these data suggest that Opa1 deletion in adipocytes induced resistance to DIO by reducing adipose tissue expansion via the repression of lipid synthesis pathways. These effects are independent of FGF21 or impaired thermogenesis.

Inducible Deletion of OPA1 in Adipose Tissue of Adult Mice Largely Recapitulates the Phenotype of Ad-KO Mice

To exclude potentially confounding developmental effects, we generated mice with inducible knockout of OPA1 in adipose tissue of young adults by generating a tamoxifen-inducible mouse model (Ind-KO). OPA1 levels were significantly reduced in WAT (Fig. 5A) and in BAT (Supplementary Fig. 5A) of Ind-KO mice 4 weeks after tamoxifen injections (mice were injected at ∼12 weeks of age). At this time point, body mass (Supplementary Fig. 5B), total fat mass (Supplementary Fig. 5C), and total lean mass (Supplementary Fig. 5D) were unchanged between Ind-KO mice and their WT littermate controls. However, gWAT mass was significantly reduced (Supplementary Fig. 5E), while iWAT was unchanged (Supplementary Fig. 5F), and BAT mass was significantly increased (Supplementary Fig. 5G) in Ind-KO mice. At 10–12 weeks post–tamoxifen injections, body mass (Fig. 5B) and total fat mass (Fig. 5C) were significantly decreased in Ind-KO mice, whereas lean mass (Fig. 5D) remained unchanged. The weights of various adipose depots normalized to body weight, namely, gWAT (Fig. 5E), iWAT (Fig. 5F), and BAT (Fig. 5G), were all reduced in Ind-KO mice. These changes in body composition occurred independently of changes in oxygen consumption (Supplementary Fig. 5H), food intake (Supplementary Fig. 5I), and locomotor activity (Supplementary Fig. 5J) between WT and Ind-KO mice. Liver mass (Fig. 5H) and triglyceride levels (Fig. 5I) were elevated in Ind-KO animals, which correlated with impaired glucose tolerance (Fig. 5J and K) and higher fasting glucose levels (Fig. 5L). At baseline conditions, genes involved in triglycerides and fatty acids synthesis were uniformly repressed in WAT (Fig. 5M) and to a variable degree in BAT (Supplementary Fig. 5K). Similarly, thermogenic genes, such as Ucp1, were also repressed in iWAT (Fig. 5N) and in BAT (Supplementary Fig. 5L). Core body temperature was unchanged between WT and Ind-KO mice at thermoneutrality but was significantly reduced in Ind-KO mice after 3 days of cold exposure. Cold-induced activation of the thermogenic gene program was attenuated in BAT (Fig. 5Q) and iWAT (Fig. 5R) of Ind-KO mice. Moreover, UCP1 protein levels were reduced in iWAT of Ind-KO mice after cold exposure (Fig. 5S). Finally, after 12 weeks of high-fat feeding, Ind-KO mice had significantly reduced body mass (Fig. 5T and U) and total fat mass (Fig. 5V) relative to WT mice, while lean mass remained unchanged (Fig. 5X). Together, our data demonstrate that deleting Opa1 in young adult mice essentially recapitulates the changes in lipid metabolism, browning, and resistance to DIO observed when OPA1 is deleted during embryonic development.

Figure 5

Inducible deletion of OPA1 in adipose tissue of adult mice recapitulates the phenotype of Ad-KO mice. A: OPA1 protein levels in iWAT normalized to GAPDH and the respective densitometric quantifications. BN: Data collected from WT and Ind-KO mice 10–12 weeks after tamoxifen injections (tamoxifen treatment started at ∼8 weeks of age). B: Body mass. C: Total fat mass. D: Total lean mass. E: gWAT mass normalized to body mass. F: iWAT mass normalized to body mass. G: BAT mass normalized to body mass. H: Liver mass normalized to body mass. I: Liver triglycerides levels. J: Glucose tolerance test (GTT). K: Area under the curve for the GTT. L: Fasting glucose levels (after a 6-h fast). M: Relative mRNA expression of lipid metabolism genes in iWAT normalized to Gapdh levels. N: Relative mRNA expression of thermogenic genes in iWAT normalized to Gapdh levels. OS: Data collected in WT and Ind-KO mice after 3 days at either thermoneutrality (30°C) or at 4°C. O: Averaged core body temperature at thermoneutrality. P: Averaged core body temperature at 4°C. Q: Relative mRNA expression of thermogenic genes in BAT after 3 days of cold exposure normalized to Gapdh levels. R: Relative mRNA expression of thermogenic genes in iWAT after 3 days of cold exposure normalized to Gapdh levels. S: UCP1 protein levels in iWAT after 3 days of cold exposure normalized to tubulin and the respective densitometric quantification. TV: Data from WT and Ind-KO mice fed a high-fat diet (HFD) (60% fat) for 12 weeks (animals started the diet at ∼2 weeks after tamoxifen injections, ∼10 weeks of age). T: Growth curve. U: Final body mass. V: Total fat mass. W: Total lean mass. Data are expressed as means ± SEM. Significant differences were determined with Student t test, with a significance level of P < 0.05. *Significantly different vs. WT mice. OD, optical density.

Figure 5

Inducible deletion of OPA1 in adipose tissue of adult mice recapitulates the phenotype of Ad-KO mice. A: OPA1 protein levels in iWAT normalized to GAPDH and the respective densitometric quantifications. BN: Data collected from WT and Ind-KO mice 10–12 weeks after tamoxifen injections (tamoxifen treatment started at ∼8 weeks of age). B: Body mass. C: Total fat mass. D: Total lean mass. E: gWAT mass normalized to body mass. F: iWAT mass normalized to body mass. G: BAT mass normalized to body mass. H: Liver mass normalized to body mass. I: Liver triglycerides levels. J: Glucose tolerance test (GTT). K: Area under the curve for the GTT. L: Fasting glucose levels (after a 6-h fast). M: Relative mRNA expression of lipid metabolism genes in iWAT normalized to Gapdh levels. N: Relative mRNA expression of thermogenic genes in iWAT normalized to Gapdh levels. OS: Data collected in WT and Ind-KO mice after 3 days at either thermoneutrality (30°C) or at 4°C. O: Averaged core body temperature at thermoneutrality. P: Averaged core body temperature at 4°C. Q: Relative mRNA expression of thermogenic genes in BAT after 3 days of cold exposure normalized to Gapdh levels. R: Relative mRNA expression of thermogenic genes in iWAT after 3 days of cold exposure normalized to Gapdh levels. S: UCP1 protein levels in iWAT after 3 days of cold exposure normalized to tubulin and the respective densitometric quantification. TV: Data from WT and Ind-KO mice fed a high-fat diet (HFD) (60% fat) for 12 weeks (animals started the diet at ∼2 weeks after tamoxifen injections, ∼10 weeks of age). T: Growth curve. U: Final body mass. V: Total fat mass. W: Total lean mass. Data are expressed as means ± SEM. Significant differences were determined with Student t test, with a significance level of P < 0.05. *Significantly different vs. WT mice. OD, optical density.

Close modal

Mitochondria are crucial for WAT function and homeostasis (3). Defects in mitochondrial dynamics contribute to adipose tissue dysfunction in obesity and impair thermogenesis in brown and beige adipocytes (4,16). We recently demonstrated an important role for OPA1 in the thermogenic activation of brown adipocytes and in the regulation of systemic metabolic homeostasis (7). However, the role of OPA1 in WAT in vivo remained largely unexplored. In the current study, we show that OPA1 is required to maintain mitochondrial function and lipid homeostasis, with its deficiency leading to impaired lipid storage and mobilization, accelerated adipocyte senescence, and adipose tissue inflammation, which promotes systemic metabolic dysregulation. Furthermore, OPA1 is required for cold-induced browning of subcutaneous WAT; thus, combined deletion of OPA1 in both white and brown adipocytes rendered mice cold-intolerant.

Studies in vitro and in vivo demonstrated that OPA1 levels increase during lipid accumulation (14,17), whereas in humans, Opa1 transcript levels are significantly reduced in visceral adipose tissue of obese versus lean individuals (18). These studies suggest the potential involvement of OPA1 in adipose tissue homeostasis. Indeed, studies in 3T3-L1 white adipocytes and in human adipocytes in vitro suggested a role for OPA1 in lipolysis. OPA1 was found to be localized in lipid droplets, where it functions as a previously unidentified A-kinase anchoring protein (AKAP) responsible for perilipin phosphorylation in response to adrenergic stimuli (14,19). Our data in Ad-KO mice confirm a role for OPA1 in lipolysis. However, defective lipolytic signaling in Ad-KO mice seems to be independent of perilipin phosphorylation, suggesting that OPA1 likely plays a fundamentally different regulatory role in lipolysis in vivo under physiological conditions.

A novel role of OPA1 revealed in this study is the regulation of lipid synthesis. Mitochondria play an important role in lipogenesis and triglyceride synthesis. Oxidative phosphorylation provides ATP to support lipogenesis (20), and mitochondrial metabolism of glucose provides acetyl-CoA, a substrate for fatty acid (18) synthesis. Mitochondria generate glycerol-3-phosphate, a precursor substrate for fatty acid esterification and triglyceride synthesis (21). Thus, reduced mitochondrial oxidative capacity in adipocytes lacking OPA1 could decrease the availability of critical precursors required for fatty acid synthesis and esterification. These data support the hypothesis that defective lipid synthesis capacity prevents age- and diet-induced WAT expansion in Ad-KO mice. Impaired adipose tissue expansion also occurred in part because of transcriptional repression of critical enzymes required for triglyceride synthesis and lipogenesis. Finally, augmented adipocyte senescence and inflammation may prevent age- and diet-induced adipose tissue expansion in Ad-KO mice. In agreement with our data, mild induction of Opa1 increased adipose tissue expandability in response to HFD (18). Altered Opa1 expression in adipocytes induces metabolic pathways that regulate epigenetic modifications, thereby changing thermogenic gene expression (18). Although not directly tested in our study, it is conceivable that Opa1 deletion in adipocytes may reduce certain metabolites, driving epigenetic changes that repress gene programs, including lipogenic genes. Together, our data suggest that altered mitochondrial function and dynamics in Ad-KO mice impair lipid metabolism through transcriptional repression of lipogenesis and reduced fatty acid biosynthesis.

Studies in mice lacking the mitochondrial fusion protein MFN2 in adipocytes also linked mitochondrial dynamics with adipose tissue function and lipid metabolism. However, in contrast to our data, constitutive (22) or inducible (4) deletion of Mfn2 increased adipose tissue expandability in response to an obesogenic diet. Transcriptionally, this correlated with increased expression of genes regulating adipocyte proliferation and lipogenesis (4). Increased adipocyte expandability in Mfn2-deficient mice was associated with improved glucose homeostasis in obesity, while Opa1 deletion impaired glucose homeostasis by preventing adipocyte expandability. Finally, Mfn2 deletion in Ucp1-expressing adipocytes prevented diet-induced obesity and insulin resistance; however, the proposed mechanisms were different than in our OPA1 model (7) and involved a sex-specific remodeling of BAT mitochondrial function (16).

Mitochondrial dynamics have been shown to contribute to WAT browning in part via DRP1 activation in human white adipocytes (5). We showed that Opa1 deletion in BAT promotes compensatory browning of WAT, which correlates with increased protein levels of OPA1 (7). Here, we show that cold exposure induces Opa1 levels in iWAT of WT mice, suggesting a role for OPA1 in regulating cold-induced browning. Indeed, cold-induced browning of iWAT was significantly attenuated in Ad-KO mice. In concert with reduced thermogenic activation of BAT, impaired iWAT browning rendered Ad-KO mice cold intolerant. These data support our earlier conclusion that improved thermoregulation in mice lacking OPA1 in BAT is mediated by compensatory browning of WAT (7), which is absent in Ad-KO mice. A recent report linked fumarate induction of the demethylase KDM3a that correlated with increased Ucp1 and browning of WAT in response to OPA1 overexpression (18). In contrast, Opa1 deletion limits urea cycle and generation of fumarate, which was linked to reduced browning in vitro (18). Whether alternative mechanisms limit cold-induced browning when Opa1 is deleted in vivo remains to be determined, but our data confirm repression of transcriptional programs, including cold-induced activation of thermogenic genes, in OPA1-deficient adipocytes. Additional potential mechanisms attenuating browning in our model include reduced lipolysis and impaired mitochondrial respiratory function.

In conclusion, OPA1 deficiency disrupts fundamental aspects of adipocyte biology, including lipid storage and mobilization, and cold-induced browning of WAT, in a manner that is independent of developmental changes in mice. We provide novel information linking Opa1 deletion with transcriptional repression of lipogenic genes and lipolytic signaling, which results in reduced adiposity. Furthermore, OPA1 deletion in WAT accelerated adipocyte senescence and induced adipose tissue inflammation, which, in conjunction with impaired lipogenesis, contributed to systemic metabolic dysregulation. Our study reveals a link between OPA1 expression and mitochondrial respiratory capacity, with lipolytic signaling, and the transcriptional regulation of lipogenic and thermogenic genes in white adipocytes. These findings provide important perspective that informs the development of mitochondria-based therapies to treat obesity, insulin resistance, and other metabolic disorders.

E.D.A. is currently affiliated with Department of Medicine, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA.

This article contains supplementary material online at https://doi.org/10.2337/figshare.21171184.

Acknowledgments. Metabolic phenotyping was performed at the Metabolic Phenotyping Core at the Fraternal Order of Eagles Diabetes Research Center. Analysis of mRNA expression was performed with quantitative PCR in the Genomics Division of The Iowa Institute of Human Genetics. The authors thank Dr. Hiromi Sesaki, at Johns Hopkins University, for providing the OPA1 floxed mice.

Funding. This work was supported by National Heart, Lung, and Blood Institute, National Institutes of Health, grants HL127764 and HL112413 and the Teresa Benoit Diabetes Research Fund to E.D.A., who is an established investigator of the American Heart Association (AHA), and by AHA Scientist Development Grant 15SDG25710438 (to R.O.P.).

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

Author Contributions. R.O.P. and E.D.A conceived the project and coordinated all aspects of this work. R.O.P. conceptualized, designed, and conducted the experiments; analyzed data; and wrote the manuscript. A.C.O., A.M., S.F., J.R.W., and L.M.G.P. conducted the experiments and analyzed data. M.W., R.H., S.T.A., and J.K. conducted experiments. M.J.P. provided essential materials and critical expertise. E.D.A. edited the manuscript. R.O.P. and E.D.A. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Arroyo-Johnson
C
,
Mincey
KD
.
Obesity epidemiology worldwide
.
Gastroenterol Clin North Am
2016
;
45
:
571
579
2.
Kusminski
CM
,
Scherer
PE
.
Mitochondrial dysfunction in white adipose tissue
.
Trends Endocrinol Metab
2012
;
23
:
435
443
3.
Boudina
S
,
Graham
TE
.
Mitochondrial function/dysfunction in white adipose tissue
.
Exp Physiol
2014
;
99
:
1168
1178
4.
Mancini
G
,
Pirruccio
K
,
Yang
X
,
Blüher
M
,
Rodeheffer
M
,
Horvath
TL
.
Mitofusin 2 in mature adipocytes controls adiposity and body weight
.
Cell Rep
2019
;
26
:
2849
2858.e4
5.
Pisani
DF
,
Barquissau
V
,
Chambard
JC
, et al
.
Mitochondrial fission is associated with UCP1 activity in human brite/beige adipocytes
.
Mol Metab
2018
;
7
:
35
44
6.
Li
X
,
Yang
L
,
Mao
Z
, et al
.
Novel role of dynamin-related-protein 1 in dynamics of ER-lipid droplets in adipose tissue
.
FASEB J
2020
;
34
:
8265
8282
7.
Pereira
RO
,
Marti
A
,
Olvera
AC
, et al
.
OPA1 deletion in brown adipose tissue improves thermoregulation and systemic metabolism via FGF21
.
eLife
2021
;
10
:
e66519
8.
BonDurant
LD
,
Ameka
M
,
Naber
MC
, et al
.
FGF21 regulates metabolism through adipose-dependent and -independent mechanisms
.
Cell Metab
2017
;
25
:
935
944.e4
9.
Pereira
RO
,
Wende
AR
,
Olsen
C
, et al
.
Inducible overexpression of GLUT1 prevents mitochondrial dysfunction and attenuates structural remodeling in pressure overload but does not prevent left ventricular dysfunction
.
J Am Heart Assoc
2013
;
2
:
e000301
10.
Garcia-Cazarin
ML
,
Snider
NN
,
Andrade
FH
.
Mitochondrial isolation from skeletal muscle
.
J Vis Exp
2011
;
49
:
2452
11.
Mina
AI
,
LeClair
RA
,
LeClair
KB
,
Cohen
DE
,
Lantier
L
,
Banks
AS
.
CalR: a web-based analysis tool for indirect calorimetry experiments
.
Cell Metab
2018
;
28
:
656
666.e1
12.
Lee
G
,
Kim
YY
,
Jang
H
, et al
.
SREBP1c-PARP1 axis tunes anti-senescence activity of adipocytes and ameliorates metabolic imbalance in obesity
.
Cell Metab
2022
;
34
:
702
718.e5
13.
Kawai
T
,
Autieri
MV
,
Scalia
R
.
Adipose tissue inflammation and metabolic dysfunction in obesity
.
Am J Physiol Cell Physiol
2021
;
320
:
C375
C391
14.
Pidoux
G
,
Witczak
O
,
Jarnæss
E
, et al
.
Optic atrophy 1 is an A-kinase anchoring protein on lipid droplets that mediates adrenergic control of lipolysis
.
EMBO J
2011
;
30
:
4371
4386
15.
Pereira
RO
,
Tadinada
SM
,
Zasadny
FM
, et al
.
OPA1 deficiency promotes secretion of FGF21 from muscle that prevents obesity and insulin resistance
.
EMBO J
2017
;
36
:
2126
2145
16.
Mahdaviani
K
,
Benador
IY
,
Su
S
, et al
.
Mfn2 deletion in brown adipose tissue protects from insulin resistance and impairs thermogenesis
.
EMBO Rep
2017
;
18
:
1123
1138
17.
Grant
RW
,
Vester Boler
BM
,
Ridge
TK
,
Graves
TK
,
Swanson
KS
.
Adipose tissue transcriptome changes during obesity development in female dogs
.
Physiol Genomics
2011
;
43
:
295
307
18.
Bean
C
,
Audano
M
,
Varanita
T
, et al
.
The mitochondrial protein Opa1 promotes adipocyte browning that is dependent on urea cycle metabolites
.
Nat Metab
2021
;
3
:
1633
1647
19.
Greenberg
AS
,
Kraemer
FB
,
Soni
KG
, et al
.
Lipid droplet meets a mitochondrial protein to regulate adipocyte lipolysis
.
EMBO J
2011
;
30
:
4337
4339
20.
Lu
RH
,
Ji
H
,
Chang
ZG
,
Su
SS
,
Yang
GS
.
Mitochondrial development and the influence of its dysfunction during rat adipocyte differentiation
.
Mol Biol Rep
2010
;
37
:
2173
2182
21.
Nye
C
,
Kim
J
,
Kalhan
SC
,
Hanson
RW
.
Reassessing triglyceride synthesis in adipose tissue
.
Trends Endocrinol Metab
2008
;
19
:
356
361
22.
Boutant
M
,
Kulkarni
SS
,
Joffraud
M
, et al
.
Mfn2 is critical for brown adipose tissue thermogenic function
.
EMBO J
2017
;
36
:
1543
1558
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.