CHCHD10 Modulates Thermogenesis of Adipocytes by Regulating Lipolysis

The balance between energy expenditure and intake is essential for whole-body energy homeostasis. Excess energy is stored in white adipose tissue (WAT) in the form of triglycerides (TGs), which are linked to metabolic disorders (1). On the contrary, brown adipose tissue (BAT) dissipates energy through nonshivering thermogenesis. Brown adipocytes highly express UCP1, which mediates the leak of protons across the mitochondrial membrane and uncouples oxidative respiration from ATP synthesis (2). Beige adipocytes are inducible brown adipocytes residing in WAT and are functionally very similar to classical brown adipocytes (1). Brown/beige adipocytes are highly associated with metabolic health in human and mouse models (1,3). This feature of energy consumption makes brown or beige adipocytes a promising target for metabolic disease intervention.

External stimulation, such as chronic cold, β3-adrenergic receptor agonists, and exercise, activates brown adipocytes and beige adipocytes (4,5). In response to these stimuli, lipolysis in adipocytes is activated, hydrolyzing TGs to free fatty acids (FFAs) and glycerol, which is achieved by the sequential action of adipose TG lipase (ATGL), hormone-sensitive lipase (HSL), and monoglyceride lipase (6). FFAs are among the best-characterized regulators of UCP1. On the one hand, FFA serves as fuel for oxidation to establish a proton gradient through the electron transport chain in mitochondrial respiration. Moreover, FFAs exert long-term effects on thermogenic capacity by activating transcription factors, such as peroxisome proliferator–activated receptors (PPARs), that increase the expression of Ucp1 and other genes promoting oxidative phosphorylation (OXPHOS) for heat generation (7). On the other hand, long-chain FFAs liberated by lipolysis bind to UCP1 and activate UCP1-catalyzed proton leak (8). A recent study showed that FFAs generated through lipolysis in WAT but not BAT are essential for fat thermogenesis (9). Although numerous studies have identified several regulators of thermogenic adipocytes, the underlying molecular basis needs to be further clarified.

To identify novel activators of thermogenic adipocytes, we established models of cold-exposed mice and immortalized brown adipocytes and conducted an RNA sequencing (RNA-seq) analysis in these models. We found that coiled-coil-helix-coiled-coil-helix domain-containing 10 (Chchd10) is one of the top upregulated genes in both models. The Chchd10 gene encodes a mitochondrial intermembrane protein that is enriched at cristae junctions. CHCHD10 associates with mitofilin, CHCHD3, and CHCHD6 within the mitochondrial contact site and cristae organizing system (MICOS) complex, which is a key regulator to form cristae, structures that contain respiratory complexes (10). Higher cristae abundance is associated with augmented oxidative respiration, whereas defects in cristae architecture strongly compromise respiratory function (11,12). In BAT, cold stress promotes active respiratory cristae formation to fulfill thermogenic energy requirements (13). Many heterozygous mutations of Chchd10 have been identified in human patients, especially affecting the neuromuscular system, such as in amyotrophic lateral sclerosis (ALS) and mitochondrial myopathy (14–17). Clinical and epidemiological studies have indicated that dysregulated energy metabolism is closely related to ALS, which show a negative contribution to the overall pathogenic process of ALS (18). Mutation of Chchd10 leads to MICOS complex disassembly and loss of mitochondrial cristae (19,20). However, the role of CHCHD10 in thermogenic fat and metabolic regulation remains unknown.

In the current study, CHCHD10 was identified as an essential regulator of nonshivering thermogenesis. We found that CHCHD10 expression in adipose tissues was strongly induced by thermogenic stimuli through PPARγ and PGC1α. CHCHD10 deficiency in adipocytes led to a defect in lipolysis, thereby resulting in impaired thermogenesis and cold intolerance.

Two LoxP sites were flanked outside exon 2 and exon 3 of the Chchd10 gene by CRISPR/Cas9 technology. Single guide RNA was transcribed in vitro, and the donor vector was constructed. Cas9, single guide RNA, and the donor vector were microinjected into the fertilized eggs of C57BL/6JGpt mice. Fertilized eggs were transplanted to obtain positive F0 mice, which were confirmed by PCR and sequencing. A stable F1 generation mouse model was obtained by mating positive F0 generation mice with C57BL/6JGpt mice. The Chchd10 gene was knocked out in adipocytes after mating Chchd10 loxP mice with mice expressing adiponectin-Cre recombinase.

Male C57BL6/J mice aged 8–12 weeks were purchased from GemPharmatech. All mice were housed at room temperature (25°C), unless otherwise specified, with a 12 h light/dark cycle and ad libitum access to food and water. The high-fat diet (HFD) (D12492; Research Diets) contained 60% energy in the form of fat. All animal experiments were conducted in accordance with all relevant ethical regulations for animal studies. All studies involving animal experimentation were approved by the Fudan University Shanghai Medical College Animal Care and Use Committee and followed the National Institutes of Health guidelines on the care and use of laboratory animals.

Subcutaneous adipose tissue (SAT) and pericardial adipose tissue (PAT) were obtained from individuals who underwent thoracic surgery in Shanghai Jiaotong University Affiliated Xinhua Hospital. SAT was from the sternotomy incision and PAT from the inner midcourse. The fat tissues were dissected in TRIzol and stored at −80°C. We complied with all relevant ethical regulations. The human study was approved by the ethics committee of Fudan University. Human samples were collected in accordance with the ethical guidelines of the 1975 Declaration of Helsinki. Written informed consent was obtained from each human subject.

Immortalized brown preadipocytes were generated as previously described and cultured with DMEM (11995065; Gibco) containing 10% FBS (12483020; Gibco) (21). Upon reaching 70% confluence (designated day −2), brown preadipocytes were induced to differentiate into brown adipocytes with differentiation medium (DMEM containing 10% FBS, 1 μg/mL insulin, and 1 nmol/L triiodothyronine [T3]) until day 0. Cells were then cultured in induction medium (DMEM containing 10% FBS, 1 μg/mL insulin, 1 nmol/L T3, 0.5 mmol/L isobutylmethylxanthine, 0.5 μmol/L dexamethasone, and 0.125 mmol/L indomethacin) for 2 days, after which cells were cultured with differentiation medium, which was changed every other day. Cells were fully differentiated and expressed high levels of UCP1 on day 6.

Adipose tissue was enzymatically digested with collagenase VIII (CVIII) (C2139; Sigma-Aldrich). The digested tissue was filtered using a 100 μm mesh filter and centrifuged. The adipocytes floated on top of the supernatant. The cell debris pellet containing the stromal vascular fraction (SVF) was resuspended in ammonium chloride lysis buffer to remove red blood cells.

Overexpression or knockdown of Chchd10 in fat pads was conducted using adenoviruses as previously described (22,23). For overexpression, recombinant adenovirus containing Chchd10 was generated using the ViraPower Adenoviral Expression System (Invitrogen, Carlsbad, CA), while LacZ recombinant adenovirus served as the negative control. For Chchd10 knockdown, recombinant adenovirus containing short hairpin Chchd10 was produced through the BLOCK-iT Adenoviral RNAi Expression System (Invitrogen). An adenoviral expression vector, pAd/BLOCK-iT, encoding shRNA against Chchd10 was constructed, while an shRNA against LacZ served as the control. The short hairpin Chchd10 sequence was as follows: 5′-CACCGGAGTAAGGAAGGATCGTTCGAAAACGATCCTTCCTTACTCC-3′.

Recombinant adenovirus was produced and amplified in HEK293A cells and purified using adenovirus purification kits (VS-AVPQ022; Sartorius, Göttingen, Germany). Purified adenovirus was subcutaneously injected twice a week into inguinal WAT (iWAT) or BAT of live mice.

A Comprehensive Lab Animal Monitoring System (Columbus Instruments) was used to record VO₂, VCO₂, energy expenditure, and food intake. Animals were acclimatized in the recording chambers for 24–48 h, and measurements were taken subsequently for 36 h under a 12 h light/dark cycle with free access to food and water. To measure fasting metabolic conditions, mice were fasted for 16 h before transfer to metabolic cages; food was removed during the assessment. To measure the metabolic conditions under cold stimulation, mice were exposed at 4°C in the metabolic cages.

High-purity mitochondria from adipose tissues or adipocytes were obtained using the Mitochondria Isolation Kit (C3601; Beyotime) according to the manufacturer’s instructions. Isolated mitochondria were resuspended with solubilization buffer (50 mmol/L imidazole, 2 mmol/L 6-aminohex-anoic acid, 50 mmol/L sodium chloride, 1 mmol/L EDTA, pH 7.0) (24). Digitonin (20%) was added to the sample and placed on ice for 20 min. Solubilized samples were centrifuged at maximum speed for 30 min at 4°C. The supernatant was mixed with 5% Coomassie G-250, loaded onto 4–15% Mini-PROTEAN TGX Stain-Free Gels (Bio-Rad), and electrophoresed in PowerPace Basic (Bio-Rad) at 4°C at 100 V for 60 min and 16 mA until the dye front exited the gel. The proteins were transferred to methanol-activated polyvinylidene fluoride membranes in PowerPace Basic at 4°C at 30 V overnight. OXPHOS complexes were measured by incubating the membranes with specific antibodies.

In cold exposure experiments, mice were housed in prechilled cages at 4°C with free access to food and water. Rectal temperature was measured using an animal electronic thermometer (Alcott Biotech, Shanghai, China) at the indicated times after cold stimulation. Surface temperature was measured using an infrared digital thermography camera T430SC (Teledyne FLIR) after cold exposure. The images were analyzed using FLIR Tools (Teledyne FLIR) to normalize the temperature ranges. To assess body temperature in the fasting state, mice were fasted 16 h before exposure to the cold environment.

ATP levels were measured using an Enhanced ATP Assay Kit (S0027; Beyotime) according to the manufacturer’s instructions. Using a chemiluminescence apparatus, a standard curve was drawn, and the ATP content was calculated. ADP/ATP Ratio Assay Kit (Bioluminescent) (ab65313; Abcam) and AMP Assay Kit (Colorimetric) (ab273275; Abcam) were used to measure the ADP/ATP and AMP/ATP ratios, respectively.

Cultured brown adipocytes with altered expression of Chchd10 were collected by trypsinization and centrifuged. Cell pellets were resuspended in analysis buffer (Dulbecco PBS supplemented with 25 mmol/L glucose, 1 mmol/L pyruvate, and 2% BSA). For adipose tissue measurement, fat pads were cut into pieces in analysis buffer. Cellular respiration was measured with a Clark-type oxygen electrode (Oxygraph⁺ System; Hansatech Instruments). Data were normalized to total protein content or cell number.

For Seahorse analysis, immortalized brown preadipocytes were seeded in XF24 cell culture microplates (Seahorse Bioscience). After 6 day differentiation, cells were washed twice and incubated in XF medium (containing 25 mmol/L glucose, 2 mmol/L glutamine, and 1 mmol/L pyruvate) for 30 min at 37°C without CO₂. The oxygen consumption ratio (OCR) was measured by the XF24 Extracellular Flux Analyzer (Seahorse Bioscience). Oligomycin (3 μmol/L), carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) (3 μmol/L), and antimycin A (1 μmol/L) were added into cartridges and injected into XF24 wells in succession. OCR was calculated as a function of reaction time and normalized to the number of cells.

Tissues or cultured cells were lysed with cell lysis buffer containing 2% sodium dodecyl sulfate, 50 mmol/L Tris-HCl (pH 6.8), 10 mmol/L dithiothreitol, phenylmethylsulfonyl fluoride, and protease inhibitor cocktail (P8340; Sigma-Aldrich). The proteins were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes. The expression of proteins of interest was analyzed using the following antibodies: anti-PPARγ (81B8), anti-VDAC1 (D73D12), anti-ATGL (2138), anti-HSL (4107), and anti-p-HSL (45804) from Cell Signaling Technology; anti-UCP1 (ab10983) from Abcam; anti-HSP90 (13119) from Santa Cruz Biotechnology; and anti-CHCHD10 (25671-1-AP), anti-ATP5A1 (66037-1-Ig), anti-NDUFA9 (20312-1-AP), anti-SDHB (10620-1-AP), anti-COXIV (11242–1-AP), and anti-PGC1α (66369-1-Ig) from Proteintech.

Total RNA was extracted from cells or tissues with TRIzol. cDNA was synthesized with the PrimeScript RT Reagent Kit with gDNA Eraser (RR047B; Takara Bio, Shiga, Japan). mRNA levels were quantified by quantitative PCR (qPCR) using SYBR Green (A46112; Applied Biosystems, Carlsbad, CA). PCRs were run with a ViiA 7 Real-Time PCR System (Applied Biosystems) in technical triplicates, and mRNA levels were normalized against 18S rRNA in the same samples. The qPCR primers are listed in Supplementary Table 1.

Synthetic siRNA oligonucleotides specific for Chchd10 (5′-GGGCCUGCUCAUCGCCUAATT-3′) and (5′-GGAGUAAGGAAGGAUCGUUTT-3′) and negative control siRNA duplexes were synthesized by GenePharma. Immortalized brown preadipocytes were transfected with siRNA mixed with Lipofectamine RNAiMAX (13778150; Invitrogen) at 70% confluence. Four days later, cells were transfected for a second time. The silencing effects of oligonucleotides were detected by RT-qPCR and Western blot.

Fresh adipose tissues were isolated from mice, and tissues with minimal mechanical damage, such as pulling, contusion, and extrusion, were selected. Avoiding light, tissues were postfixed with 1% OsO₄ in 0.1 mol/L phosphate buffer (pH 7.4) for 2 h at room temperature. After three washes in 0.1 mol/L phosphate buffer (pH 7.4), tissues were dehydrated in graded ethanol, infiltrated in 50% propylene oxide and 50% resin for 1 h, treated with resin for 1 h, and embedded in a resin-based mold for polymerization overnight at 37°C. The resin blocks were cut into 60–80-nm–thick slices using an ultramicrotome and stained. The cuprum grids were placed in the grid board and dried overnight at room temperature and observed and photographed using a transmission electron microscope (HT7800/HT7700; Hitachi).

Fresh tissues fixed in 4% paraformaldehyde were embedded in paraffin followed by staining with hematoxylin-eosin at room temperature for 24 h. The histologic images were taken using a microscope. For immunohistochemistry, the slides were blocked with 2% horse serum, incubated with anti-UCP1 (1:150) overnight at 4°C, and washed. Slides were incubated with an appropriate peroxidase polymer–linked secondary antibody for 30 min. The slides stained with 3,3′-diaminobenzidine substrate were counterstained with hematoxylin.

In mitochondrial electron transport activity of adipose tissues was measured by staining with 2,3,5-triphenyltetrazolium chloride (TTC) (T8877; Sigma-Aldrich). ∼20 mg of adipose tissue was placed in 2% TTC for 15 min at 37°C and then fixed in 10% formalin for 30 min at room temperature. The tissues were transferred to 95% ethanol (400 μL/sample) to remove the formalin and incubated overnight at 4°C, and the solution was quantified by spectrophotometry at 485 nm and normalized by tissue weight.

Cultured brown adipocytes used to detect nascent protein synthetic rate were grown in 60-mm dishes until day 4 after differentiation. Before labeling, cells were incubated with methionine-free media for 1 h and then with added 50 μmol/L l-azidohomoalanine (AHA) (C10102; Life Technologies) for 4 h. Cells were washed with cold PBS and lysed in buffer containing Tris HCl (pH 8.0), 1% sodium dodecyl sulfate, and sonicated lysates. The completed lysates were subjected to Click-iT reaction to switch azido-modified nascent proteins to alkyne-biotin (B10185; Life Technologies) using the Click-iT Protein Reaction Buffer Kit (C10276; Life Technologies) according to the manufacturer’s protocol. Biotinylated nascent proteins were subjected to Western blot by using anti-ATGL or anti-VDAC1 antibodies.

Blood was collected from mice after overnight fasting, and serum was prepared for measurements. Serum glucose, TG, total cholesterol (TC), HDL, LDL, AST, and ALT levels were determined using an automatic biochemical analysis device (Roche). Fasting glycerol levels were measured using a kit (E1003; Applygen Technologies). Fasting FFA levels were measured using a nonesterified FFA assay kit (A042-1-1; Nanjing Jiancheng Bioengineering).

The MitoProbe JC-1 Assay Kit (MT09; Dojindo Molecular Technologies) was used to measure the mitochondrial membrane potential. Total DNA was extracted from tissues or cells by GenElute Mammalian Genomic DNA Miniprep Kits (G1N70; Sigma-Aldrich), and the relative mtDNA and nuclear DNA levels were determined by qPCR using Cebpα as the nuclear gene target and Cytob as the reference for mtDNA.

Differentiated cells were washed with PBS, fixed with 4% paraformaldehyde for 20 min at room temperature, and stained with BODIPY (LD02; Dojindo Molecular Technologies) and DAPI (BS04; Dojindo Molecular Technologies) for 10 min. The cells were photographed using a microscope and analyzed.

Statistical analyses were conducted using GraphPad Prism software. Results are shown as the mean ± SD. Two-tailed unpaired Student t test was used to compare two groups with normally distributed data. The two-tailed unpaired Student t test with Welch correction was used for two groups with unequal variances. The Mann-Whitney U test was used to determine significance between groups without a normal distribution. One-way ANOVA with Bonferroni multiple comparisons test was used to compare more than two groups; nonparametric statistical analysis was performed using the Kruskal-Wallis test with Dunn test for multiple comparisons. Two-way ANOVA with Bonferroni posthoc multiple comparison test was used to analyze multiple groups with two fixed factors. P < 0.05 was considered to indicate statistical significance.

The data sets and resource generated and/or analyzed during the current study are available from the corresponding authors upon reasonable request.

To identify putative activators of brown/beige adipocytes, we performed RNA-seq analysis in a differentiation model of brown adipocytes and iWAT of mice exposed to cold at 4°C. RNA-seq data showed that 5,876 genes were upregulated during brown adipocyte differentiation, and 139 genes were significantly upregulated in iWAT after cold stimulation, with 93 genes overlapping. Gene ontology analysis indicated that these 93 genes were strongly associated with metabolic processes, mitochondrial components, and oxidation reduction processes (Supplementary Fig. 1A and B). The data analysis highlighted Chchd10 as the top upregulated gene in both models (Fig. 1A and B).

Chchd10 was evidently increased in iWAT, epididymal WAT, and BAT of cold-induced mice (4°C) for 72 h, while other genes involved in mitochondrial cristae formation, like Chchd3, Chchd6, and Immt, showed low response to cold exposure (Fig. 1C and D). Consistently, CHCHD10 levels were profoundly increased in iWAT and BAT of mice after treatment with the β3-adrenoreceptor agonist CL316,243 (Fig. 1E). We found that CHCHD10 was rapidly increased within 1 h after cold induction, when the number of mitochondria remained unchanged, as indicated by the levels of VDAC1 (Supplementary Fig. 1C and D), which is a mitochondrial marker (25). We then inhibited thermogenic adipocytes by housing the mice at a thermoneutral temperature (30°C). The Chchd10 levels in iWAT and BAT in these mice were significantly decreased compared with mice housed at room temperature (Supplementary Fig. 1E).

Further experiments showed that CHCHD10 was abundantly expressed in mature adipocytes compared with the SVF (Fig. 1F), and the expression of CHCHD10 in adipocytes from iWAT was augmented by cold induction (Fig. 1G). Additionally, an in vitro adipogenic differentiation model showed that CHCHD10 expression was dramatically upregulated during the differentiation of brown and beige adipocytes (Fig. 1H and I).

We then sought to identify the driver for CHCHD10 upregulation. We found that Chchd10 was significantly upregulated by two agents used for brown adipocyte differentiation: rosiglitazone and indomethacin (Supplementary Fig. 2A and B). Rosiglitazone is an agonist of PPARγ. Indomethacin activates PPARγ and inhibits Cox (26). Several predicted PPAR response elements were identified in the promoter of Chchd10, indicating that PPARγ might participate in the regulation of Chchd10 expression. To test this possibility, a luciferase reporter assay was conducted by subcloning the promoter of Chchd10 upstream of the luciferase-encoding gene. Luciferase activity was significantly increased by PPARγ and further augmented by its cofactor PGC1α or by rosiglitazone treatment (Supplementary Fig. 2C and D). Deletion of the predicted PPARγ binding site (−896 to −408 upstream of the Chchd10 transcription start site) almost abolished luciferase activity (Supplementary Fig. 2E). Chromatin immunoprecipitation coupled with qPCR confirmed the binding of PPARγ to the promoter of Chchd10 at the same site (Supplementary Fig. 2F). We then overexpressed PPARγ in cultured brown adipocytes and found that PPARγ promoted CHCHD10 expression (Supplementary Fig. 2G); in PGC1α-deficient cells, PPARγ could hardly augment the CHCHD10 level (Supplementary Fig. 2H), suggesting that PGC1α was required for PPARγ to regulate CHCHD10. Rosiglitazone treatment in vivo strongly induced Chchd10 expression in both iWAT and BAT (Supplementary Fig. 2I), while the PPARγ antagonist T0070907 repressed it (Supplementary Fig. 2J). Meanwhile, we found that PPARγ expression in adipose tissues was significantly increased upon cold stimulation and declined after mice were acclimated from cold to ambient temperature (25°C), showing a positive correlation with Chchd10 expression (Supplementary Fig. 2K). Disruption of PPARγ blocked the elevated CHCHD10 induced by cold exposure (Supplementary Fig. 2L and M) and inhibited thermogenic adipocyte activation (Supplementary Fig. 2N). These results together suggest that PPARγ transcriptionally activates Chchd10 expression during thermogenic adipocyte activation.

To investigate the function of Chchd10, we generated Chchd10-AKO mice by crossing Chchd10 flox/flox mice with adiponectin-Cre mice (Supplementary Fig. 3A and B). qPCR analysis showed that Chchd10 expression was specifically disrupted in adipose tissues but not in other tissues like liver and muscle (Supplementary Fig. 3C). Ablation of Chchd10 had no effect on other MICOS protein expressions like CHCHD3, CHCHD6, and IMMT (Supplementary Fig. 3D). Chchd10-AKO mice developed normally and had body weights similar to those of their wild-type (WT) littermates when fed a normal chow diet (Supplementary Fig. 3E–G). We challenged control and Chchd10-AKO mice in a 4°C environment and measured the rectal temperature every hour; no difference was observed between the two groups (Fig. 2A). To our surprise, when fasted overnight before cold stimulation, Chchd10-AKO mice exhibited severe hypothermia, with rectal temperatures <20°C, while WT mice did not (Fig. 2A). Another cohort of Chchd10-AKO mice showed similar cold intolerance upon fasting (Supplementary Fig. 3H). Consistently, infrared digital thermography showed that the surface temperature of Chchd10-AKO mice, especially in the interscapular BAT region, was much lower than that of control mice under fasting conditions (Fig. 2B). After cold exposure for 10 h, ∼90% of Chchd10-AKO mice died as a result of low body temperature (Fig. 2C). Indirect calorimetry study showed that there was no difference in VO₂, VCO₂, or energy expenditure between WT and Chchd10-AKO mice in the fed state at room temperature, while upon food deprivation, Chchd10-AKO mice had significantly lower rates of VO₂ consumption, VCO₂ production, and energy expenditure relative to WT (Supplementary Fig. 3I–K). Upon cold stimulation, compared with WT, VO₂, VCO₂, and energy expenditure were markedly decreased in Chchd10-AKO mice (Fig. 2D–F).

WAT and BAT from Chchd10-AKO mice developed to normal volumes but appeared pale in color and had significantly increased lipid deposition and enlarged lipid droplets compared with WT, as indicated by hematoxylin-eosin (H-E) staining; this often reflects thermogenic fat dysfunction (Fig. 2G and H). Consistently, UCP1 expression, at both mRNA and protein levels, was strongly downregulated in iWAT and BAT from Chchd10-AKO mice, while the expression of other genes responsible for mitochondrial formation and thermogenesis, like Prdm16 and Pgc1a, remained unchanged (Fig. 2I and J).

Because CHCHD10 is localized to the intermembrane space of mitochondria and associated with the MICOS, we then assessed mitochondrial morphology. We found that CHCHD10 deficiency led to mitochondrial swelling and ultrastructural major abnormalities, including loss, disorganization, and dilatation of cristae in BAT (Fig. 3A). However, the number of mitochondria was not affected, as indicated by the mtDNA copy number (Fig. 3B). In situ staining with the redox dye TTC showed that CHCHD10 deficiency decreased mitochondrial electron transport in iWAT and BAT (27) (Fig. 3C). Blue native PAGE analysis of isolated mitochondria revealed an impairment of OXPHOS CIV and CV ATP synthase assembly in BAT and iWAT of Chchd10-AKO mice, without changes in CI and CII (Fig. 3D). As a consequence of impaired OXPHOS CV, the ATP levels were significantly reduced in both WAT and BAT from Chchd10-AKO mice (Fig. 3E). In immortalized brown adipocytes, ATP levels were also decreased upon Chchd10 ablation (Fig. 3F) and increased by Chchd10 overexpression (Fig. 3G). Additionally, AMP/ATP and ADP/ATP ratios were elevated in CHCHD10-deficient adipocytes (Fig. 3H and I). These results indicate that CHCHD10 controls OXPHOS complex assembly, regulating ATP synthesis by affecting mitochondrial cristae formation.

To test whether CHCHD10 functions in a cell autonomous manner, we disrupted Chchd10 expression by two sets of siRNAs in immortalized brown preadipocytes and then induced them to differentiate into mature adipocytes. Ablation of Chchd10 had no effect on adipogenic differentiation (Supplementary Fig. 4A) or the expression of other genes related to MICOS (Supplementary Fig. 4B). Chchd10 disruption resulted in enlarged lipid droplets in adipocytes (Fig. 4A) and inhibited the OCR of adipocytes (Fig. 4B). Both mRNA and protein levels of UCP1 were significantly downregulated by Chchd10 deficiency (Fig. 4C and D). We also assessed the mitochondrial membrane potential, which is critical for ATP generation by OXPHOS and serves as an indicator for cell health, by JC-1 staining. Depletion of Chchd10 diminished the overall mitochondrial membrane potential (Fig. 4E), with no effect on the number of mitochondria as indicated by MitoTracker staining and the unchanged mtDNA copy number (Supplementary Fig. 4C and D). We then knocked down Chchd10 expression in fat pads by adenovirus injection. We found that 1) BAT-specific Chchd10 deficiency increased the size of lipid droplets and reduced UCP1 levels in BAT (Supplementary Fig. 5A), 2) Chchd10 deficiency in iWAT resulted in enlarged lipid droplets both at room temperature and under cold conditions (Supplementary Fig. 5B), and 3) Chchd10 deficiency repressed UCP1 expression and VO₂ in iWAT (Supplementary Fig. 5C and D), all of which were consistent with the phenotype of Chchd10-AKO mice.

To complement the loss-of-function model of Chchd10, we overexpressed Chchd10 in immortalized brown adipocytes. Chchd10 overexpression led to smaller lipid droplets compared with control (Fig. 4F) and augmented UCP1 expression (Fig. 4G) and VO₂ (Fig. 4H), indicative of activated brown adipocytes. We then ectopically expressed Chchd10 in BAT and iWAT by adenovirus harboring the Chchd10 coding sequence. Overexpression of Chchd10 in BAT dramatically inhibited lipid deposition and upregulated UCP1 expression (Supplementary Fig. 5E). Overexpression of Chchd10 in iWAT activated beige adipocytes and retained beige adipocyte morphology, even after mice were transferred from cold conditions to room temperature (Supplementary Fig. 5F and G). These results collectively suggest that CHCHD10 promotes the activation and maintenance of brown and beige adipocytes.

We then measured the regulation of VO₂ by CHCHD10 in Ucp1 knockout (Ucp1-KO) adipocytes to clarify whether function of CHCHD10 was UCP1 dependent. In WT adipocytes, knockdown of Chchd10 decreased the OCR, while in Ucp1-KO adipocytes, CHCHD10 had no effect (Fig. 4I and Supplementary Fig. 5H). Additionally, UCP1 ablation totally abolished the promoted effect on VO₂ induced by Chchd10 overexpression (Fig. 4J). These data together demonstrate that UCP1 mediates function of CHCHD10 in energy expenditure.

Having observed that CHCHD10 affected the activation of thermogenic fat, we then sought to explore the underlying mechanisms. Since Chchd10 deficiency in adipocytes impaired thermogenesis in the fasting state but not the fed state, we speculated that lipolysis is involved in the regulatory effects of CHCHD10 on thermogenesis because lipolysis, generating FFAs for thermogenesis, is highly responsive to fasting (6). Meanwhile, it has been reported that mice defective in adipose lipolysis are not cold sensitive when food is present but cold intolerant upon fasting, which is similar to the phenotype of Chchd10-AKO mice (9). To test our hypothesis, we assessed lipolytic capacity by measuring the serum levels of glycerol and FFAs after fasting, which are lipolysis products. We found that serum levels of glycerol and FFAs were significantly reduced in Chchd10-AKO mice compared with control (Fig. 5A and B) and that glycerol release in immortalized brown adipocytes was repressed upon knockdown of Chchd10 (Fig. 5C), reflecting an impaired lipolytic capacity under Chchd10 deficiency. The levels of ATGL and HSL, which catalyze the first step and the rate-limiting step in the hydrolysis of TGs, respectively, as well as the active form p-HSL, were dramatically decreased in all three fat tissue types from Chchd10-AKO mice compared with the same tissues from WT mice (Fig. 5D).

We then aimed to elucidate whether lipolysis mediates the effects of CHCHD10 on brown/beige adipocyte activation. To this end, we replenished FFAs in the adipocytes and found that FFA addition reversed the decreased OCR in Chchd10-deficient cells (Fig. 5E). We next overexpressed Chchd10 in immortalized adipocytes by adenovirus infection and then inhibited lipolysis using the ATGL inhibitor Atglistatin (28). We found that Chchd10 overexpression increased VO₂, while Atglistatin reversed this effect (Fig. 5F). We repeated similar experiments in iWAT and found that Chchd10 induced beige adipocyte formation, as indicated by increased Ucp1 mRNA levels, and multilocular lipid droplets morphology, was diminished by Atglistatin (Fig. 5G and H). We also overexpressed ATGL in iWAT of Chchd10-AKO mice and found that replenishment of ATGL increased UCP1 level and decreased the size of lipid droplets, improving thermogenic adipocyte activation in Chchd10-AKO mice (Fig. 5I–K). In addition, we found that the CHCHD10, ATGL, HSL, and p-HSL protein levels were significantly increased in WAT and BAT from fasting mice (Fig. 5L), thus supporting the notion that CHCHD10 regulates lipolysis in adipose tissues. These results indicate that Chchd10 modulates thermogenic fat activation through regulating lipolysis.

Next, we sought to clarify the mechanism underlying the regulation of lipolysis by CHCHD10. It has been established that lipolysis is an energy-consuming process. Considering that ATP was reduced in Chchd10-deficient cells (Fig. 3E and F), we wondered whether decreased ATP levels mediated the inhibitory effect of Chchd10 deficiency on lipolysis. To test this possibility, we disrupted Chchd10 expression in immortalized brown adipocytes and then treated the cells with ATP, followed by detection of lipolysis. We used liposomes to deliver ATP into adipocytes and detected a significant increased ATP level in immortalized brown adipocytes treated with ATP-liposome (Supplementary Fig. 6A). Interestingly, we found that glycerol release was inhibited by Chchd10 ablation, while this was totally reversed after replenishment of ATP (Fig. 6A). The decreased ATGL protein levels in Chchd10-deficient cells were totally restored upon ATP treatment (Fig. 6B), implying that ATP could rescue the defective lipolysis upon loss of Chchd10. We then treated cells with FCCP to inhibit mitochondrial ATP production (Fig. 6C) and found that FCCP treatment markedly downregulated ATGL level, while ATP replenishment counteracted the decreased ATGL (Fig. 6D).

We then explored how ATP regulates ATGL expression. We found that ATP upregulated the ATGL protein level without affecting its mRNA level, indicating a posttranscriptional regulatory mechanism (Fig. 6E and F). Translation regulation and stability are pivotal to the protein abundance. Further experiments showed that ATP treatment had little effect on the half-life of ATGL (Fig. 6G), which prompted us to postulate that ATP might regulate translation of ATGL. To test this possibility, we pretreated immortalized brown adipocytes with cycloheximide (CHX) to inhibit protein translation before ATP treatment. We found that ATP augmented ATGL protein levels, while pretreatment with CHX blocked the ATP-induced upregulation of ATGL (Fig. 6H), indicating that ATP might regulate translation of ATGL. To verify this possibility, we used AHA to label newly synthesized protein and found that knockdown of Chchd10 dramatically inhibited newly synthesized ATGL protein, while replenishment of ATP reversed it (Fig. 6I). Consistently, Chchd10 overexpression promoted nascent protein synthesis of ATGL (Fig. 6J). These results together reveal that Chchd10 deficiency leads to decreased ATP levels, which in turn downregulated lipolysis by inhibiting Atgl translation.

We then challenged Chchd10-AKO mice and WT littermates with HFD. No difference in body weight was observed between the two groups (Fig. 7A and Supplementary Fig. 7A–C). Glucose tolerance and insulin sensitivity were not significantly different (Supplementary Fig. 7D and E). However, serum TG levels were significantly higher in Chchd10-AKO mice compared with WT mice under both control diet and HFD conditions; serum TC and glucose levels were not significantly different (Fig. 7B). In addition, lipid droplets were larger in adipose tissues of Chchd10-AKO mice compared with those of WT mice (Fig. 7C). Livers from Chchd10-AKO mice exhibited increased lipid deposition compared with WT livers (Fig. 7C). Consistent with hepatic steatosis, AST levels were significantly increased in Chchd10-AKO mice, reflecting metabolic damage (Fig. 7D). These data indicate that Chchd10-AKO mice are prone to develop dyslipidemia.

The core temperature of Chchd10-AKO mice was obviously reduced in the fasting state (Fig. 7E). Examination of surface temperature showed similar results (Fig. 7F). OCR and heat expenditure were reduced in Chchd10-AKO mice in the fasting state (Supplementary Fig. 7F). In support of thermogenesis defects, VO₂ in iWAT from Chchd10-AKO mice was dramatically repressed (Fig. 7G). ATP levels in adipose tissues were significantly lower (Fig. 7H). Serum glycerol and FFA levels were reduced, reflecting decreased lipolysis in Chchd10-AKO mice (Fig. 7I and J). Consistently, both mRNA and protein levels of UCP1 were downregulated in adipose tissues of Chchd10-AKO mice; the protein level of ATGL was decreased in adipose tissues from Chchd10-AKO mice (Fig. 7K and L).

We next investigated Chchd10 expression in adipose tissues under different physiological statuses. Thermogenesis in BAT declines with age (29); therefore, we also assessed Chchd10 and Ucp1 expression in an aging mouse model. We found that Chchd10 expression was dramatically decreased in adipose tissues of aged mice accompanied by reduced expression of ATGL and UCP1 (Fig. 8A and B). We also collected human adipose tissues, including SAT and PAT. PAT was recognized as brown-like adipose tissues (30,31). Expression of both Chchd10 and Ucp1 were higher in PAT than SAT (Fig. 8C). In PAT and SAT, Chchd10 mRNA level was highly positively correlated with Ucp1 (Fig. 8D and E). These data overall suggest that CHCHD10 expression is positively associated with UCP1 and ATGL expression.

On the basis of the present findings, we propose a potential mechanism by which CHCHD10 regulates lipolysis and thermogenesis in adipocytes. CHCHD10 is obviously increased during activation of thermogenic adipocytes and contributes to the organization of mitochondrial cristae. When CHCHD10 is disrupted, cristae are disorganized, leading to impaired OXPHOS complex assembly, thereby repressing ATP production. A decreased ATP level results in downregulation of lipolysis by reducing the ATGL protein level, which in turn inhibits thermogenesis and energy expenditure. Consequently, Chchd10-AKO mice are prone to develop dyslipidemia and NAFLD.

The current study highlights lipolysis as the key mediator of the regulatory effects of CHCHD10 on thermogenesis in adipose tissues, which is based on the following evidence. First, Chchd10-AKO mice showed impaired thermogenesis and energy expenditure in the fasting state but not in the fed state, indicating that the processes involved in thermogenesis during fasting might be regulated by CHCHD10. Consistent with this phenotype of Chchd10-AKO mice, mice defective in adipose lipolysis were not cold sensitive when food was present but exhibited cold intolerance under fasting (9). Second, lipolysis not only activated UCP1 protein but also increased mRNA expression of Ucp1. Both mRNA and protein levels of UCP1 were decreased in Chchd10-AKO mice. As a mitochondrial located protein, CHCHD10 might indirectly regulate Ucp1 transcription, while lipolysis could be the potential mediator. Finally, and most importantly, lipolysis was dramatically inhibited upon Chchd10 ablation.

When CHCHD10 is disrupted, cristae are disorganized, leading to impaired OXPHOS complex assembly, which is the primary mediator for the other phenotypes of CHCHD10 deficiency, including decreased ATP generation, impaired lipolysis, and thermogenesis. Decreased lipolysis in part might be a compensatory event for defective OXPHOS complexes caused by loss of CHCHD10. Nevertheless, decreased ATP was the driving factor for defective lipolysis. Lipolysis is an energy-consuming process. Inhibitors of mitochondrial OXPHOS, like rotenone and oligomycin, have been shown to repress hormone-induced lipolysis in adipose tissues, accompanied by low ATP levels (32,33), which linked ATP production with lipolysis regulation. Extracellular ATP could also induce lipolysis in adipocytes (34). These results indicated that lipolysis needs respiration coupled with phosphorylation, requiring a continuous supply of energy. ATP was required for cAMP-dependent protein kinase to phosphorylate HSL and stimulate lipolysis (34); however, little evidence regarding the direct regulatory effects of ATP on lipolysis have been reported. Actually, mitochondria physically and functionally interact with lipid droplets (35); thus, it is reasonable to hypothesize that local mitochondrial-generated ATP might induce lipolysis in lipid droplets. Indeed, our observation showed that Chchd10 deficiency led to decreased ATP levels, which in turn downregulated lipolysis by inhibiting Atgl translation. However, the molecular mechanism by which ATP regulates Atgl translation needs further clarification.

Ucp1-KO mice exhibit lower energy expenditure compared with WT under norepinephrine treatment but not under basal conditions (36). Chchd10 deficiency resulted in decreased energy expenditure under fasting, without changes under basally fed conditions. We also found that VO₂ and energy expenditure in Chchd10-AKO mice was obviously lower than WT mice at cold. These data together indicate that Chchd10-AKO mice show reduced energy expenditure compared with WT mice under stimulation like fasting or cold exposure but have a comparable energy expenditure level with WT under basal conditions, which was in line with the observation that energy expenditure is not lower basally in Ucp1-KO mice. As for the VO₂ in adipocytes, cells were incubated with adipogenic cocktail including thyroxine T3 and indomethacin to promote adipogenic differentiation; therefore, we observed that the basal VO₂, which was actually under thermogenic stimulation, was decreased by CHCHD10 or UCP1 ablation, which did not contradict the notion that UCP1 is not basally leaky (36).

Mutations of Chchd10 have been found in patients with ALS (16). These mutations may lead to CHCHD10 loss of function, thereby inducing abnormal cristae formation and mitochondrial dysfunction. It has been reported that patients with ALS often experience metabolic disorders. ALS is associated with several defects in energy metabolism, including weight loss, hypermetabolism, and hyperlipidemia, which to some extent support our findings that Chchd10-AKO mice are prone to develop hyperlipemia (18). Indeed, it will be interesting to assess the function of adipose tissues in patients with ALS with Chchd10 mutations. It is also meaningful to introduce mutations in the ALS gene in Chchd10-AKO mice to study its effects on lipolysis and thermogenesis. Overall, our findings reveal an essential role of CHCHD10 in the control of mitochondrial ATP generation, thereby regulating lipolysis and thermogenesis in adipocytes.

Funding. Ya.L. has received National Natural Science Foundation of China (NSFC) grants 82170884 and 81970744 and scientific research projects of Shanghai Municipal Health Commission grant 20204Y0116. Q.-Q.T. has received National Key R&D Program of China grant 2018YFA0800400 and NSFC grants 81730021 and 32070760.

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

Author Contributions. M.D., Y.M., R.D., W.Z., and X.D. performed the experiments and prepared the data. Q.Y. conducted the metabolic cage measurements. Y.T., S.Q., Yu.L., and D.P. participated in the experimental design and results discussion. Q.-Q.T. and Ya.L. conceived, devised, and supervised the project and wrote the manuscript. M.D., Q.-Q. T., and Ya.L. are the guarantors of this work and, as such, had full access to all the data and take responsibility for the integrity of data and the accuracy of data analysis.