Due to differential splicing, endoplasmic reticulum membrane complex subunit 10 has two isoforms: membrane EMC10 and secreted EMC10 (scEMC10) (45, 46). ScEMC10 serum contents in white and Chinese Han human populations were positively correlated with obesity and insulin resistance. After bariatric surgery or exercise and caloric restriction, serum scEMC10 levels decreased. The whole-body Emc10 knockout mice, displaying more adipocyte thermogenesis and higher whole-body energy expenditure, resisted obesity and insulin resistance induced by high-fat diets. Conversely, Emc10 overexpression mice were more easily induced into obesity and insulin resistance on high-fat diets, had less adipocyte thermogenesis, and lower whole-body energy expenditure. In addition, the circulating neutralizing antibody to EMC10 also helped mice to reduce body weight gain. Mechanically, secreted EMC10 can be taken up by adipocytes and interacted with the Protein Kinase A catalytic subunit α to inhibit it from phosphorylating CREB. So the decreased phosphorylated CREB proteins led to a reduction in adipocyte thermogenesis (47).

ADISSP (Adipose secreted signaling protein) is a new uncharacterized adipokine, highly and selectively expressed in brown adipose tissue in humans and mice. The expression of this protein is higher in adipose tissue in normal people than in obese people and is negatively correlated with body weight index. The overexpression of Adissp in adipose tissue by Ap2 promoter-driven promoted adipocyte thermogenesis, augmented oxygen consumption, increased glycolysis, elevated body temperature, and resisted body weight gain induced by high-fat diets. In contrast, adipose-specific Adissp knockout mice had lower body temperature, less adipocyte thermogenesis, and were more susceptible to high-fat diet-induced obesity and hyperglycemia. ADISSP promoted white adipocyte browning and brown adipocyte activity by paracrine signaling but not by endocrine signaling. A more sensitive assay is lacking to measure ADISSP endogenous circulating contents. The PKA signaling pathway can be activated independently of the beta-adrenergic signaling pathway by ADISSP (32). Though a surface receptor in adipocytes binds to ADISSP, the receptor gene is still not confirmed and unknown.

HIFα (Hypoxia-inducible factor α, HIFα) has two family members: HIF1α and HIF2α (48). These two genes were upregulated in white and brown adipose tissue after cold exposure. The adipose Hif1α, or Hif2α, or double-specific knockout mice had more beige adipocytes in the inguinal adipose tissue and higher energy expenditure compared to the wild-type mice upon cold exposure or CL316243 (beta-adrenergic agonist) stimulation. By RNA-Seq analysis, Prkaca, which encodes PKA catalytic subunit α (PKA Cα), was a highly ranked gene in adipose tissue of Hif2α knockout mice. Using in silico analysis, miR-3085-3p, directly regulated by HIF2α, was identified to target the evolutionarily conserved region of 3’ UTR of Prkaca. So, adipocyte HIF2α could suppress PKA Cα-mediated thermogenic properties by promoting miR-3085-3p expression (33).

Ovo-like zinc finger 2 (OVOL2) belongs to the Ovo family of zinc-finger transcription factors family and is highly conserved in invertebrates and vertebrates (49). The C57BL/6J mice with Oval2 mutations induced by Nethyl-N-nitrosourea (ENU), named Oval2 ^boh/boh, were characterized by increased body weight without affecting food intake under a standard chow diet. The boh mutation is a single nucleotide transition from G to A, causing the substitution of tyrosine for a conserved cysteine, and does not affect the OVOL2 protein’s stability. Ovol2 knockout is embryonically lethal. The heterozygotes with the boh allele and the null allele of Ovol2 (Ovol2 ^boh/-) exhibited overall stronger phenotypes: obesity, reduced energy expenditure, and adipocyte thermogenesis. Overexpression of Ovol2 in adipocytes reduced total body and liver fat and improved insulin sensitivity in mice fed on a high-fat diet. Ovol2 is highly expressed in white adipose stromal cells but not in mature white adipocytes. OVOL2 can block adipogenesis by interacting with the C-terminal portion of C/EBPα to inhibit its adipogenic function in both mouse and human adipocytes (34). The mechanism of OVOL2 in brown adipocytes remains elusive and needs more investigation.

Transmembrane protein 86A (TMEM86A) is a putative lysoplasmalogenase, a close homolog of TMEM86B, and a YhhN family protein member (50, 51). Tmem86a expression was significantly upregulated in adipose tissue on high-fat diets compared to the standard chow diet and heavily enriched in mature adipocytes. Transcriptome-profiling (GEO: GSE94753) also shows that TMEM86A expression is upregulated in abdominal subcutaneous WAT from female patients with obesity compared to individuals without obesity. The adipose tissue-specific Tmem86a knockout mice had increased mitochondrial metabolism, adipocyte thermogenesis, and energy expenditure and exhibited significantly lower body weight gain. Mechanistically, the untargeted lipidomics analysis suggested that Tmem86a overexpression downregulated the content of lysoplasmalogens, including plasmatic lysophosphatidylethanolamine 18:0 (LPE P-18:0) and adipocyte-specific Tmem86a knockout (AKO) increases LPE P-18:0 content in adipose tissue. LPE P-18:0 could reduce the activity of PDE3B (phosphodiesterase 3b), which is abundantly expressed in adipocytes and can degrade cAMP. Furthermore, LPE P-18:0 treatment reduced body weight and fat mass, increased energy expenditure, and strongly activated the PKA signaling pathway in adipose tissue (35).

SSU72 is a dual-specific protein phosphatase and is expressed in a tissue-specific manner. Recent studies have demonstrated that SSU72 plays an important role in controlling the carboxyl-terminal domain (CTD) function of RNA polymerase II (RNAPII) and monitoring the liver (52–54). In adipose tissue, SSU72 phosphatase was highly enriched in brown adipose tissue relative to white adipose tissue. Ssu72 mRNA and protein levels were increased significantly in BAT and WAT after cold exposure. BAT from adipose-specific Ssu72 knockout mice appeared pale and showed enlarged mitochondria and disorganized cristae structures compared to WT mice. The knockout mice were much more cold-sensitive and intolerant after cold exposure. The expression of thermogenic genes and fatty acid β-oxidation genes was significantly attenuated in the brown adipose tissue of AKO mice.

When endoplasmic reticulum (ER) homeostasis is disrupted, PKR-like ER-regulated kinase (PERK) is activated to phosphorylate the α subunit of eukaryotic initiation factor 2 (eIF2α) at serine 51 (Ser51). EIF2α phosphorylation represses most of the proteins’ translation to reduce ER stress. Dephosphorylation of eIF2α by its phosphatase GADD34 in the liver can improve insulin sensitivity and reduce hepatosteatosis in mice fed high-fat diets. By RNA-Seq analysis, most PERK-eIF2α target genes were significantly upregulated in AKO BAT. Mechanistically, SSU72 can directly interact with eIF2α to inhibit its phosphorylation and increase the protein translation of mitochondrial oxidative phosphorylation and adipocyte thermogenesis in BAT. Furthermore, metabolic dysfunction in Ssu72-abated BAT could return to almost normal after restoring Ssu72 expression (36).

PR domain-containing 16 (PRDM16) is a master gene that controls the biogenesis of brown and beige adipocytes by forming a complex with transcriptional and epigenetic factors (12, 13, 15). PRDM16 is dynamically regulated at the post-translational level. Overexpression of euchromatic histone-lysine N-methyltransferase 1 (EHMT1) or chronic treatment with synthetic ligands of peroxisome proliferator-activated receptor-γ (PPARγ) prolongs PRDM16 protein’s half-life (55, 56). CUL2–APPBP2 complex as the ubiquitin E3 ligase was identified to determine PRDM16 protein stability by catalyzing its polyubiquitination. CUL2 functions as a scaffold protein by interacting with an E2 enzyme, elongation B (ELOB), elongation C (ELOC), and APPBP2 substrate receptor, also found in RING E3 ligase complexes (57, 58). Cul2 depletion in white adipocytes extended the half-life of PRDM16 protein and significantly increased uncoupled cellular respiration. Overexpression of Cul2 in adipocytes reduced brown/beige-fat-selective genes’ expression. Consistent with the results of Cul2 deletion, deletion of Appbp2 also led to higher PRDM16 protein levels and increased expression of brown/beige-fat-selective genes compared with the control cells. APPBP2 (S561N) variant, associated with lower levels of 2 h postprandial serum glucose and insulin, weakly interacted with PRDM16 protein relative to WT APPBP2. Differentiated primary adipocytes from Appbp2 mutant mice expressed higher thermogenic gene levels than WT adipocytes. Besides, adipose-specific Cul2 or Appbp2 knockout mice expressed higher levels of PRDM16 protein and adipose thermogenesis, displayed significantly higher whole-body energy expenditure, and gained less body weight than controls (37).

The transforming growth factor β (TGFβ) signaling pathway is complex and associated with various human pathologies (59). By comparing the transcriptome of human supraclavicular BAT (scBAT) and subcutaneous WAT and analyzing the transcriptome of the human multipotent adipose-derived stem (hMADS) cells differentiated into beige and white adipocytes, Gpr180 was found to be upregulated in brown fat on both tissue and cellular level. The knockdown of Gpr180 shifted brown and beige adipocytes towards a white-like phenotype. The whole body or inducible adipose tissue knockout of Gpr180 diminished brown and beige adipocyte function and impaired insulin resistance.

The knockdown of Gpr180 did not affect cAMP levels and phosphorylation of PKA substrates. However, it reduced the phosphorylation of SMAD3 protein at serine 423 in the matured adipocyte. Besides, TGFβ1-induced phosphorylation of SMAD3 and upregulation of Ucp1 were attenuated in beige adipocytes without GPR180, which indicates that GPR180 is required for full activation of the TGFβ signaling machinery. Collagen triple helix repeat containing 1 (CTHRC1) was identified as GRP180’s potential ligand. Overexpression of Cthrc1 in male mice prevented body weight gain and increased energy expenditure during the HFD-induced weight gain. In sum, GPR180 and CTHRC1 are the components of the TGFβ signaling pathway for adipocyte thermogenesis. These components regulate low-grade SMAD3 phosphorylation and control thermogenic adipocyte function, whole-body energy, and glucose homeostasis (38).

Circulating IL18 level is associated with body weight, insulin resistance, and metabolic syndrome in humans and mice (60). Il18 ablation in mice led to hyperphagia, obesity, insulin resistance, and decreased energy expenditure. Whole-body deletion of Il18r increased body weight and decreased energy expenditure, but white adipocyte browning was enhanced in mice on a chow diet (61, 62). This can be explained by the NaCl co-transporter (NCC) acting as an alternative receptor for IL18’s differential function in adipocyte thermogenesis. A single knockout of Ncc or a combined knockout of Il18r and Ncc, but not a single knockout of Il18r, blocked adipocyte thermogenesis. Consistent with this, brown adipocytes from Ncc ^−/− and Il18r ^−/− Ncc ^−/− mice, but not those from Il18r ^−/− mice, showed decreased levels of IL18 and induced uncoupled respiration.

Furthermore, Ncc ^fi/fi Ucp1 ^Cre mice gained more body weight, had lower adipocyte thermogenesis, and showed worse glucose intolerance and insulin resistance on high-fat diets. There is no difference in UCP1 positive areas and thermogenic genes’ expression in adipose tissue between Il18r ^fi/fi Ucp1 ^Cre mice and control mice. However, the concise mechanism of IL18 for Ucp1 promotion is still unclear. It is only known that it is not dependent on the cAMP signaling pathway in brown adipocytes. Overall, IL18 uses NCC to promote thermogenesis in BAT but uses IL18R to enhance glucose sensitivity in WAT (39).

The mitochondrial cristae biogenesis protein optic atrophy 1 (OPA1) was significantly downregulated in the human subcutaneous adipose tissue (SAT) of heavy co-twins by gene expression analysis and correlated with mitochondrial gene expression. Opa1 ^tg mice with ubiquitous mild Opa1 overexpression were slightly resistant to high-fat diet-induced obesity, glucose intolerance, and hepatic steatosis compared to their littermate controls. Similarly, resistance to obesity was observed in a mouse model of Opa1 haploinsufficiency (63). Opa1 ^tg mice produce more heat at room temperature by promoting BAT function and WAT browning. Opa1 facilitates cell-autonomous adipocyte browning by upregulating Kdm3a, a member of the Jumanji demethylase. This increases brown and beige thermogenic activity by controlling the H3K9 methylation status of Adrb1 and Ucp1 (64, 65). Metabolomic profiling further revealed that fumarate levels produced from the urea cycle derived Kdm3a-dependent Ucp1 induction in adipocytes. Moreover, overexpression of Opa1 can increase cAMP contents to activate CREB to upregulate the rate-limiting urea cycle enzyme Cps1 (carbamoyl phosphate synthetase-1) in adipose tissue (40). Although OPA1 may not depend on its pro-mitochondrial fusion role, more research is needed to determine the mechanism by which it activates CREB.

By dissipating energy as heat, UCP1 mediates adaptive thermogenesis. Long-chain fatty acids bind on UCP1 to drive proton leak, while purine nucleotides bind on UCP1 to block this uncoupling process (66, 67). Sulfenylation of UCP1 on Cys253 is essential for acute cold-induced uncoupled respiration, while the lysine succinylation of UCP1 reduces its activity and stability (68). Cytosolic calcium can directly stimulate adenylyl cyclase activity to increase cAMP production and PKA activation to induce thermogenesis (69). Two endoplasmic reticulum-located calcium channels, sarco/endoplasmic reticulum Ca²⁺-ATPase 2b (SERCA2b) and ryanodine receptor 2 (RyR2), are involved in ATP-dependent and UCP1-independent calcium cycling machinery to dissipate energy as heat in a beige adipocyte (70).

The mitochondrial calcium uniporter (MCU) complex, which consists of a pore-forming subunit (MCU) and several regulatory subunits, including essential MCU regulator (EMRE) and mitochondrial calcium uptake 1 (MICU1) et al., is a key regulator of mitochondrial calcium (71). Mcu BKO (Mcu ^f/f with Ucp1 ^Cre mice) and Emre BKO (AAV-Emre gRNA into BAT of Rosa26-LSL-Cas9 with Adipoq^Cre) mice are hypothermic. They could not maintain their core body temperatures when challenged with cold exposure. Upon adrenergic stimulation, MCU recruits UCP1 through EMRE to form an MCU-EMRE-UCP1 complex (thermo porter), which increases mitochondrial calcium uptake to accelerate the tricarboxylic acid cycle and supply more protons that promote uncoupled respiration. A mutant EMCU (unable to conduct Ca²⁺) could interact with UCP1 at a similar level as the WT EMCU, decreasing the level of UCP1-dependent respiration. MICU1 is the gatekeeper to prevent Ca²⁺ overload in mitochondria by interacting with MCU. Their interaction markedly decreased upon cold exposure or NE/CL-316,243 treatment. AAV-Micu1 BKO (AAV-Micu1 gRNA into Rosa26-LSL-Cas9 with Adipoq-Cre) enhanced brown adipocyte uncoupled respiration and energy expenditure when activated by NE. The mice carrying the enforcedly assembled thermo porter gained less body weight, more glucose and insulin tolerance, and increased animal energy expenditure; the opposite metabolic phenotypes were observed in Mcu BKO and Emre BKO mice (41).

Because UCP1 loss causes the depletion of most components of the mitochondrial electron transport chain (ETC) in BAT, the interpretation of phenotypes of Ucp1 KO mice is bewildering and confusing (72). Thermogenic reactive oxygen species (ROS) could reversibly modify a regulatory site (C253) on UCP1 to elevate UCP1-dependent respiration (67). Based on these findings, a mouse model in which the regulatory C253 site on UCP1 is mutated to an alanine (UCP1 C253A mouse) was generated to examine its role in regulating energy homeostasis and metabolic disease. Quantitative proteomics of BAT from WT, UCP1 KO, and UCP1 C253A mice demonstrated that BAT from UCP1 C253A mice maintained expression of the full mitochondrial metabolic proteins. Both male and female UCP1 C253A mice exhibited significantly lower VO2 consumption, energy expenditure, and VCO2 production in response to cold exposure. Fed on a high-fat, high-sucrose (HFHS) diet, male, and female WT and UCP1 C253A mice gained indistinguishable body weight, and their food intake was identical. Remarkably, male but not female C253A mice exhibited more glucose intolerance than WT mice. The proteomic analysis of adipose tissues from both male and female mice on HFHS diets suggested that UCP1 C253A strongly agonizes WAT inflammation in male but not female mice. The mutation of UCP1 C253A increased mitochondrial protein oxidation and systemic inflammation in male mice since the inflammatory cytokine expression was attenuated upon the supplementation of MitoQ (73), a mitochondria-targeted antioxidant. UCP1 C253A male mice treated with β-estradiol showed the decreased expression of inflammatory cytokines, which were significantly increased in BAT of untreated C253A mice (42).