Glucose regulation of adipose tissue browning by CBP/p300- and HDAC3-mediated reversible acetylation of CREBZF

Significance

Upon cold exposure, glucose is taken up by brown and beige adipose and serves as an energy source for thermogenesis. However, how glucose signals are transduced into cells and to what extent the thermogenic capacities are affected remain unknown. Here, we identified a metabolic regulator CREBZF as an important mediator of glucose effects on browning and thermogenic capacity in beige adipocytes. Mechanistically, CREBZF is modulated by glucose signals via CREB-binding protein (CBP)/p300- and HDAC3-mediated reversible acetylation and dynamic changes of protein stability. Deficiency of CREBZF potentiates the glucose-induced thermogenesis and energy expenditure. CREBZF-mediated inhibition of energy dissipation may represent a mechanism of glucose in regulating energy balance, which is likely required for repressing hyperactivation of browning and thermogenesis.

Abstract

Glucose is required for generating heat during cold-induced nonshivering thermogenesis in adipose tissue, but the regulatory mechanism is largely unknown. CREBZF has emerged as a critical mechanism for metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as nonalcoholic fatty liver disease (NAFLD). We investigated the roles of CREBZF in the control of thermogenesis and energy metabolism. Glucose induces CREBZF in human white adipose tissue (WAT) and inguinal WAT (iWAT) in mice. Lys208 acetylation modulated by transacetylase CREB-binding protein/p300 and deacetylase HDAC3 is required for glucose-induced reduction of proteasomal degradation and augmentation of protein stability of CREBZF. Glucose induces rectal temperature and thermogenesis in white adipose of control mice, which is further potentiated in adipose-specific CREBZF knockout (CREBZF FKO) mice. During cold exposure, CREBZF FKO mice display enhanced thermogenic gene expression, browning of iWAT, and adaptive thermogenesis. CREBZF associates with PGC-1α to repress thermogenic gene expression. Expression levels of CREBZF are negatively correlated with UCP1 in human adipose tissues and increased in WAT of obese ob/ob mice, which may underscore the potential role of CREBZF in the development of compromised thermogenic capability under hyperglycemic conditions. Our results reveal an important mechanism of glucose sensing and thermogenic inactivation through reversible acetylation.

Adipose tissue is critical for controlling nutrient and energy homeostasis. White adipose tissue (WAT) stores and releases energy and secretes adipokines to communicate with other metabolic organs (1, 2). Brown adipose tissue (BAT) dissipates energy to produce heat in response to cold exposure, which is called adaptive nonshivering thermogenesis. A distinct population of adipocytes named brown-like (beige) adipocytes is emerged in WAT by cold exposure or other environmental cues in a process called browning or beiging (3). Once stimulated, thermogenic capacity in beige adipocytes is comparable to brown adipocytes (4). Stimulating browning of WAT may be applicable for the treatment of obesity and metabolic diseases.

Glucose is the important fuel source utilized by brown and beige adipocytes for thermogenesis. In response to cold exposure, glucose uptake is increased in BAT of humans and rodents, which is correlated with an increased systemic energy expenditure (5, 6). Although elevated glucose is taken up under cold exposure as a fuel source for generating heat, how glucose regulates thermogenic capacity and energy expenditure is poorly understood.

The thermogenic program in brown and beige adipocytes is controlled by transcription factors or coregulators, such as PPARγ and PGC-1α (3, 7). In contrast to stimulatory regulators, various proteins have been identified as inhibitors for this process via inhibiting PGC-1α activity such as receptor-interacting protein 140 (RIP140) (8), liver X receptor (9), transcriptional intermediary factor-2 (TIF2), (10) and twist-related protein 1 (TWIST1) (11) or via attenuating adrenergic receptor signaling, such as potassium channel subfamily K member 3 (KCNK3) (12). The inhibitory effects of these regulators may represent negative feedback mechanisms to prevent hyperactivation of thermogenesis and maintain appropriate systemic energy homeostasis. However, it is unclear how nutrient signals are transduced into the cell and to what extent the thermogenic capacities are affected in brown and beige adipocytes.

We have previously identified CREB/ATF bZIP transcription factor (CREBZF) as an important regulator of insulin-induced hepatic lipogenesis via inhibiting Insig and that hyperactivation of CREBZF leads to sustained lipogenesis and hepatic steatosis under the selective insulin resistance (13, 14). We also uncovered the function of hepatic CREBZF in liver regeneration and nonalcoholic steatohepatitis (NASH) progression (15, 16). However, the metabolic function of CREBZF in the adipocyte is not known. Here, we characterized the central role of CREBZF in glucose sensing and white adipose tissue (WAT) browning. These in vivo and in vitro studies demonstrate that 1) glucose stimulates CREBZF expression in WAT of mice and humans; 2) Lys208 acetylation is required for glucose-induced reduction of proteasomal degradation and augmentation of CREBZF protein stability; 3) glucose-induced browning and thermogenic capacity of adipose is potentiated by CREBZF deficiency; and 4) CREB-binding protein (CBP)/p300- and HDAC3-dependent CREBZF acetylation couples glucose sensing to adipose tissue browning.

Materials and Methods

Animal Model.

Adipose-specific CREBZF knockout (CREBZF FKO) mice for glucose treatment or cold exposure were produced by crossing floxed CREBZF mice (13) with Adipoq-Cre recombinase transgenic mice (Model Animal Research Center of Nanjing University); wild-type (WT) littermates (Adipoq-Cre negative, CREBZF^flox/flox) were used as the control. All mice were housed under a 12:12 h light/dark cycle at controlled temperature. All animal experimental protocols were approved by the Institutional Animal Care and Use Committee at Shanghai Institute of Nutrition and Health, Chinese Academy of Sciences.

Adipose Specimens from Humans.

Visceral white adipose samples were obtained from adult patients undergoing bariatric surgery. Patients gave written consent for their samples to be collected. The study protocol was approved by the Ethics Committee of Nanjing Drum Tower Hospital, The Affiliated Hospital of Nanjing University Medical School, and was conducted in accordance with the 1975 Declaration of Helsinki.

Cold-Induced Thermogenesis.

Mice were placed at 4 °C for 48 h. Control mice were placed at room temperature (RT, 22 °C). Rectal temperature was measured using a model BAT-12 thermometer (Physitemp Instruments).

Statistical Analysis.

Data are expressed as mean ± SEM. Statistical significance was evaluated using the unpaired two-tailed Student's t test for two-group comparison. For multiple group comparisons, one-way analysis of variance (ANOVA) or two-way ANOVA was applied. Differences were considered significant at a P value < 0.05.

Results

Glucose Promotes the Cold-Induced Thermogenic Program in WAT.

To investigate the role of glucose in thermogenesis in vivo, mice were fasted to deplete the glucose in circulation and then treated with acute cold exposure followed by a bolus of glucose injection. Interestingly, glucose induced a significant rise in blood glucose and rectal temperature (Fig. 1 A and B) and increased expression levels of thermogenic genes in inguinal WAT (iWAT) (Fig. 1C) but had little effects on the thermogenic program in BAT (Fig. 1D). Protein levels of UCP1 were also markedly induced by glucose in iWAT, rather than BAT (Fig. 1E). These results suggest that the augmentation of the thermogenic program in beige adipocytes in iWAT may directly mediate the glucose effects on thermogenesis. Consistently, in mouse primary adipocytes differentiated from stromal vascular fraction (SVF) isolated from iWAT and BAT, glucose supplementation with different duration and dosage remarkably increased the expression levels of thermogenic genes in both beige and brown adipocytes (Fig. 1 F and G and SI Appendix, Fig. S1). The protein levels of UCP1 in primary beige adipocytes were stimulated by glucose in a time- and dose-dependent manner (Fig. 1H). Remarkably, glucose administration further boosted β3-adrenergic agonist CL-316,243-induced UCP1 in beige adipocytes (Fig. 1 I and J). These findings suggest that glucose is sufficient to elevate the cold tolerance and activate the thermogenic program of beige adipocytes.

Fig. 1.

Glucose is sufficient to potentiate thermogenesis and browning of WAT. Ten-week-old male C57BL/6 mice were fasted for 7 h (4 h at RT and 3 h in cold) and injected with glucose intraperitoneally at 2 g/kg. (A) Blood glucose. (B) Rectal temperature. n = 11 to 14, *P < 0.05, vs. mice injected with vehicle. (C–E) mRNA levels of thermogenic genes in inguinal WAT (iWAT) (C) and BAT (D). n = 4 to 5, *P < 0.05, vs. vehicle. (E) Immunoblots of UCP1 in iWAT and BAT. (F–H) Glucose supplementation activates the adipocyte thermogenic program. Mouse primary adipocytes differentiated from iWAT-derived SVF were treated with glucose as indicated. (F and G) mRNA levels of thermogenic genes are stimulated by glucose in a timecourse-dependent (F) and dose-dependent manner (G). n = 3 to 4, *P < 0.05, vs. vehicle. (H) Protein levels of UCP1. (I and J) Mouse primary adipocytes were treated with vehicle or CL-316,243 for 12 h, followed by supplementation of glucose for 12 h. mRNA levels (I) and protein levels (J) of UCP1. n = 7, *P < 0.05, vs. cells treated with vehicle, #P < 0.05, vs. cells treated with CL-316,243.

Glucose and Cold Exposure Increase Expression Levels of CREBZF.

To further investigate whether the metabolic regulator CREBZF mediates the glucose signals in regulating thermogenic programs in adipose tissue, expression levels of CREBZF were measured in mice treated with glucose under cold exposure. CREBZF protein was significantly induced in iWAT by glucose (Fig. 2A). Consistently, the protein levels of CREBZF were also stimulated by cold exposure in iWAT, which was associated with an elevated expression of glucose transporter genes (Fig. 2 B and C). However, little changes were observed in transcriptional levels (SI Appendix, Fig. S2A), and these effects were not observed in BAT and differentiated primary brown adipocytes (SI Appendix, Fig. S2 B and C). Notably, CREBZF was also stimulated by glucose in the cultured human WATs (Fig. 2D). Compared with the epididymal WAT (eWAT), CREBZF was highly expressed in the iWAT and BAT, whose thermogenic programs would be activated by the external stimulus (SI Appendix, Fig. S2D). It is supposed that glucose-induced CREBZF may couple the glucose signals with thermogenesis in beige adipocytes.

Fig. 2.

CREBZF is induced by glucose signals in adipocytes. (A) Ten-week-old male C57BL/6 mice were fasted under cold exposure followed by glucose injection. Immunoblot of CREBZF in iWAT. The band intensity was quantified by densitometry. n = 3, *P < 0.05, vs. vehicle. (B and C) Ten-week-old male C57BL/6 mice were placed at RT or 4 °C (cold) for 48 h. (B) Immunoblots of CREBZF in iWAT. The band intensity was quantified by densitometry. n = 3, *P < 0.05, vs. RT. (C) mRNA levels of glucose transporters in iWAT. n = 5 to 7, *P < 0.05, vs. RT. (D) Glucose induces CREBZF in ex vivo human WAT. Human WATs were collected and minced and cultured ex vivo for 24 h treated with or without glucose. Protein levels of CREBZF were analyzed. (E–G) Glucose increases protein levels of CREBZF. HEK293T cells were transfected with (E and F) or without (G) FLAG-CREBZF and then treated with glucose as indicated. Immunoblots were performed. (H–K) Endogenous CREBZF is induced by glucose in adipocytes. 3T3-L1 adipocytes (H and I) and mouse primary adipocytes (J and K) were treated with glucose as indicated. Immunoblots and real-time PCR were performed. n = 4 to 6. (L and M) Expression levels of CREBZF are increased in WAT of ob/ob mice. CREBZF protein levels in WAT (L) and plasma glucose levels (M) of ob/ob mice or C57BL/6 mice fed with chow diet (lean). n = 4 to 6, *P < 0.05, vs. lean.

Next, the effects of glucose on CREBZF were evaluated in HEK293T cells. Interestingly, FLAG-tagged exogenous (Fig. 2 E and F) and endogenous CREBZF (Fig. 2G) were both induced by glucose in a dose- or timecourse-dependent manner. Treatment of glucose caused a potent induction of endogenous CREBZF protein levels but not mRNA levels in 3T3-L1 adipocytes or mouse primary adipocytes (Fig. 2 H–K). This notion is also supported by the fact that protein levels of CREBZF were potently higher in the WAT of hyperglycemia ob/ob mice (Fig. 2 L and M). These data suggest that CREBZF is likely induced by glucose signals in beiges at the posttranslational level after cold exposure.

Glucose-Induced Thermogenesis Is Potentiated by CREBZF Deficiency In Vivo.

To investigate the relative contribution of CREBZF loss of function to glucose-induced thermogenesis, adipose-specific CREBZF knockout (CREBZF FKO) mice were generated. CREBZF FKO and their WT littermates were fasted prior to cold exposure, followed by glucose injection, and then, rectal temperature and blood glucose levels were measured. In WT mice, rectal temperature was reduced following cold exposure, whereas an increased temperature was observed in 30 min after glucose injection, suggesting an induction of glucose-induced thermogenesis (Fig. 3 A and B). Surprisingly, compared with WT mice, CREBZF FKO mice showed a significant elevation in rectal temperature, as well as attenuated response in blood glucose elevation after glucose administration. Glucose-induced thermogenic genes were further elevated in iWAT of CREBZF FKO mice (Fig. 3C). These effects were not observed in BAT (Fig. 3D). These unexpected results suggest that CREBZF plays important roles in the regulation of the thermogenic program by glucose in beiges.

Fig. 3.

Ablation of CREBZF in adipocytes enhances glucose-induced thermogenesis in vivo. Ten-week-old male CREBZF FKO and WT mice were fasted with cold exposure and then injected with glucose intraperitoneally. (A) Blood glucose levels. (B) Rectal temperature and calculated average rectal temperatures after glucose injection. n = 12 to 17, *P < 0.05 vs. WT. (C) Glucose-induced browning of beige adipocytes was augmented by CREBZF deficiency. Protein and mRNA levels of thermogenesis-related genes in iWAT. n = 5 to 6, *P < 0.05 vs. WT. (D) Protein and mRNA levels of thermogenic genes in BAT. n = 5 to 6. (E) Oil Red O staining of CREBZF+/+ and CREBZF−/− adipocytes. (F) Primary adipocytes were treated without or with glucose for 12 h as indicated. mRNA levels of UCP1 and PGC-1α. n = 6 to 8, *P < 0.05, vs. CREBZF+/+. (G and H) CREBZF overexpression suppresses CL-316,243-induced thermogenic genes. Primary adipocytes were infected with adenovirus encoding CREBZF (Ad-CREBZF) or control virus (Ad-GFP) and treated with CL-316,243. Protein (G) and mRNA (H) levels of thermogenic genes. n = 5, *P < 0.05 vs. Ad-GFP and vehicle, #P < 0.05 vs. Ad-GFP and CL-316,243. (I) The OCRs of matured adipocytes. The uncoupled OCRs were calculated as average. n = 5, *P < 0.05, vs. CREBZF+/+ + glucose. (J) The expression levels of CREBZF are negatively correlated with UCP1 in WAT from human patients. The protein levels of CREBZF and UCP1 in WAT of human obese patients were measured by immunoblots. The quantification and correlation of the band intensity were analyzed. n = 9.

Next, CREBZF gain and loss of function on glucose-induced thermogenic gene expression were examined in primary adipocytes. CREBZF knockout did not affect adipocyte differentiation as evidenced by Oil Red O staining showing similar lipid accumulation in both CREBZF+/+ and CREBZF−/− adipocytes (Fig. 3E and SI Appendix, Fig. S2E). Strikingly, ablation of CREBZF caused a profound induction of thermogenic gene expression in the presence of glucose (Fig. 3F). Adenovirus-mediated overexpression of CREBZF inhibited the expression levels of UCP1 and PGC-1α under adrenergic stimulation (Fig. 3 G and H). The bioenergetics profiles of CREBZF−/− adipocytes showed a potent induction of oxygen consumption rate (OCR) in the presence of glucose, and uncoupled OCR was significantly increased (Fig. 3I). Importantly, the protein levels of CREBZF were negatively correlated with the UCP1 in human WATs (Fig. 3J). Taken together, these data suggest that CREBZF may function as an inhibitor of the thermogenic program, which mediates the effects of glucose signals on thermogenesis and energy expenditure in beige adipocytes.

Loss of CREBZF in Adipocytes Causes Enhanced Browning and Thermogenesis of WAT in Response to Cold and Adrenergic Stimulation.

Next, to further validate the loss-of-function roles for CREBZF in adipose tissue browning and thermogenesis, CREBZF FKO and WT mice were placed at RT or exposed to 4 °C for 48 h. CREBZF FKO mice did not show significant histological changes in iWAT compared with WT mice under RT. However, CREBZF deficiency enhanced cold-induced browning phenotypes as evidenced by increased emergence of adipocytes containing smaller, paucilocular or multilocular lipid droplets, and increased immunoreactivity for UCP1 (Fig. 4A). Moreover, expression levels of genes related to thermogenesis, mitochondria oxidative programs, and lipolysis in iWAT were significantly increased in CREBZF FKO mice (Fig. 4 B–D). Consistently, CREBZF FKO mice showed increased body temperature and enhanced cold tolerance, suggesting improved capability to defend the body temperature during cold exposure (Fig. 4E). However, little changes were observed in BAT as evidenced by histological analysis and immunoblots (Fig. 4 F and G), suggesting a BAT-independent role of CREBZF during cold exposure. CREBZF knockout efficiency and specificity were verified by immunoblots (Fig. 4H). The suppressive roles of CREBZF on browning and thermogenesis were further verified in mice under adrenergic stimulation (SI Appendix, Fig. S3). Consistently, CREBZF deficiency enhanced browning phenotypes of iWAT, expression levels of thermogenesis-related genes, and body temperature induced by CL-316,243 treatment. These changes were not observed in BAT. Importantly, CREBZF deficiency also protects mice from high-fat, high-sucrose diet–induced obesity and systemic metabolic disorders (SI Appendix, Fig. S4), as shown by the smaller adipocytes in iWAT, improved dyslipidemia, reduced body weight, and improved systemic energy metabolism. Taken together, these data demonstrate that deletion of CREBZF results in improved thermogenic gene expression, browning of iWAT, cold tolerance, and systemic metabolic disorders in obesity.

Fig. 4.

Adipocyte-specific deletion of CREBZF promotes browning of subcutaneous WAT and adaptive thermogenesis during cold exposure. Male CREBZF FKO and WT mice at 10 wk old were kept at RT or 4 °C for 48 h. (A) Representative H&E and UCP1 immunohistochemical staining of iWAT sections (Scale bar: 50 μm). (B) Protein levels of UCP1 in iWAT. (C) mRNA levels of thermogenic genes. n = 4 to 5, *P < 0.05 vs. RT and WT; #P < 0.05 vs. cold and WT. (D) mRNA levels of mitochondria oxidative and lipolytic genes. n = 5 to 6, *P < 0.05 vs. cold and WT. (E) Rectal temperature. n = 6 to 8, *P < 0.05 vs. WT. (F) Representative H&E and UCP1 immunohistochemical staining of BAT sections (Scale bar: 50 μm). (G) Expression levels of UCP1 in BAT. (H) The verification of adipose-specific CREBZF deficiency. Expression levels of CREBZF in iWAT, eWAT, liver, and muscle were measured.

CBP/p300- and HDAC3-Mediated Reversible Acetylation of CREBZF.

Given that mRNA levels of CREBZF were not significantly changed by glucose, glucose may regulate CREBZF protein stability through posttranslational modification. The acetylation and phosphorylation of CREBZF were rigorously assessed in CREBZF-transfected HEK293T cells under glucose treatment. As shown in Fig. 5A and SI Appendix, Fig. S5A, CREBZF acetylation, not phosphorylation, was potently increased after glucose treatment, suggesting that glucose may regulate CREBZF protein stability via acetylation. Consistently, the CREBZF acetylation and global acetylation of cellular proteins were up-regulated by glucose and CL-316,243 in primary adipocytes (Fig. 5 B–D and SI Appendix, Fig. S5B). The effects of acetylation on CREBZF stability are verified by treatment with trichostatin A (TSA) and nicotinamide (NAM) that inhibit all four classes of known deacetylases (17). CREBZF protein was potently induced by TSA and NAM, although only TSA caused a robust induction of CREBZF acetylation (Fig. 5E and SI Appendix, Fig. S5C). Similar results were observed in primary adipocytes (Fig. 5F). These data suggest that glucose may increase CREBZF protein levels via inducing acetylation of CREBZF in adipocytes.

Fig. 5.

Reversible acetylation of CREBZF through CBP/p300 and HDAC3. (A and B) Glucose increases CREBZF acetylation. HEK293T cells were transfected with FLAG-CREBZF (A), and primary adipocytes were infected with Ad-CREBZF (B). Cells were treated with glucose as indicated. The cell lysates were incubated with FLAG or CREBZF antibody and purified with protein A/G-Sepharose beads. The precipitates were immunoblotted with acetylated-lysine antibody. (C and D) CREBZF acetylation was induced under adrenergic stimulation. Primary adipocytes infected with Ad-CREBZF were treated with CL-316,243. Cell lysates were incubated with CREBZF (C) or acetylated-lysine (D) antibody and purified with protein A/G-Sepharose beads and precipitates were immunoblotted with acetylated-lysine or CREBZF antibody. (E) HEK293T cells were transfected with FLAG-CREBZF or pcDNA and then treated with deacetylase inhibitors TSA or NAM. Acetylation levels of CREBZF were analyzed. (F) Endogenous levels of CREBZF are increased by TSA in primary adipocytes. (G) HEK293T cells were transfected with the plasmids as indicated, and CREBZF acetylation was measured. (H) HEK293T cells were transfected with FLAG-CREBZF and HA-CBP. Coimmunoprecipitation analysis was performed. (I) HEK293T cells were transfected with the plasmids as indicated, and CREBZF acetylation was detected. (J and K) HEK293T cells were transfected with GST-CREBZF, FLAG-HDAC1/2/3/6, or empty vector pEBG (GST) (J) or with GST-CREBZF and FLAG-HDAC3 (K) as indicated. The lysates were purified with GSH Sepharose beads and analyzed with antibodies against FLAG, GST, or HDAC3. (L) Primary adipocytes were infected with Ad-CREBZF, treated without or with RGFP966 or TSA, followed by measurement of CREBZF acetylation. (M) Endogenous levels of CREBZF are induced by pharmacological inhibition of HDAC3 in adipocytes.

Next, the effects of transacetylases, including p300, CBP, GCN5, and PCAF, on CREBZF were evaluated. CBP and p300 caused a remarkable induction of CREBZF acetylation, which is correlated with increased expression of CREBZF protein (Fig. 5G). The coimmunoprecipitation assay showed a potent association between CBP and CREBZF (Fig. 5H), although the magnitude of the association between p300 with CREBZF is relatively lower (SI Appendix, Fig. S5D). These results support that CBP and p300 are the transacetylases of CREBZF, leading to increased protein levels of CREBZF.

Given that treatment with class I/II/IV deacetylases inhibitor TSA induced CREBZF acetylation and protein levels, the effects of class I deacetylases HDAC1/2/3 and class II deacetylase HDAC6 on CREBZF were determined. As shown in Fig. 5I, treatment with HDAC1 and HDAC3 caused a potent reduction of CREBZF acetylation correlating with decreased protein levels of CREBZF. However, only HDAC3 interacts with CREBZF (Fig. 5J), which is further verified by the interaction between CREBZF and endogenous HDAC3 (Fig. 5K). Furthermore, treatment with specific HDAC3 inhibitor RGFP966 (18) increased acetylation and protein levels of CREBZF in adipocytes and HEK293T cells (Fig. 5 L and M and SI Appendix, Fig. S5E). These results indicate that HDAC3 is the major deacetylase that acts on CREBZF. Together, these data demonstrate that CREBZF is tightly regulated by CBP/p300- and HDAC3-mediated reversible acetylation.

Glucose Enhances Protein Stability of CREBZF via Lys208 Acetylation.

To further identify acetylation sites in CREBZF, liquid chromatography–tandem mass spectrometry (LC–MS/MS) analysis was performed. In HEK293T cells treated without or with TSA and NAM, an evolutionally conserved residue, Lys208, is acetylated by TSA and NAM treatment (Fig. 6 A and B). As shown in Fig. 6C, a nonacetylatable mutant of CREBZF (K208R) eliminated RGFP966- or TSA-induced acetylation and protein levels of CREBZF. Alternatively, an acetylation mimetic mutant of Lys208 to glutamine (K208Q) caused a potent induction of CREBZF protein at the basal condition and was resistant to effects of RGFP966 or TSA. Importantly, both K208R and K208Q mutants were resistant to glucose-induced acetylation and protein levels of CREBZF (Fig. 6 D and E). These data demonstrate that Lys208 is required and sufficient for the acetylation and stabilization of CREBZF in response to glucose signals or HDAC3 inhibition.

Fig. 6.

Lys208 acetylation is required for glucose-induced stabilization of CREBZF. (A) Acetylation of Lys208 was demonstrated by LC–MS/MS analysis. HEK293T cells were transfected with FLAG-CREBZF and treated with TSA and NAM. The cell lysates immunoprecipitated with anti-FLAG were prepared for proteomics analysis. (B) Sequence alignments. (C and D) HEK293T cells were transfected with WT or K208R, K208Q CREBZF mutants and then treated with RGFP966 or TSA (C) or with glucose (D). The acetylation of CREBZF was measured. (E) Densitometric quantification of band intensity. n = 4. (F) HEK293T cells were cotransfected with FLAG-CREBZF and HA-Ub and treated with TSA and NAM in the absence or presence of MG132. The ubiquitination levels of CREBZF were determined. (G) HEK293T cells were transfected with WT or K208R, K208Q mutants of CREBZF and HA-Ub and then treated with glucose. CREBZF ubiquitination was determined. (H) HEK293T cells were transfected with WT, K208R, or K208Q mutants of CREBZF and treated with 100 ng/mL cycloheximide for the indicated time. Immunoblots were performed. (I) Densitometric quantification for WT or CREBZF mutants. n = 7 to 10.

To test whether acetylation regulates CREBZF activity via the ubiquitin–proteasome pathway of protein degradation, FLAG-CREBZF and HA-ubiquitin (HA-Ub) were transfected in HEK293T cells. CREBZF ubiquitination was observed in the presence of HA-Ub. Strikingly, both TSA and glucose treatment significantly reduced ubiquitination levels of CREBZF (Fig. 6 F and G). However, these effects were ablated in K208R or K208Q mutants, suggesting that Lys208 is required for glucose-repressed ubiquination of CREBZF. Consistently, K208R mutant degraded much more rapid, whereas degradation of K208Q was significantly reduced comparing with WT CREBZF (Fig. 6 H and I). These results demonstrate that CREBZF is regulated by the ubiquitin–proteasome pathway, and glucose prevents protein degradation of CREBZF via increasing acetylation of Lys208 in CREBZF.

CREBZF Binds to and Inhibits Transcriptional Activity of PGC-1α.

Previously, CREBZF was identified as a transcriptional coregulator for PGC-1α in the regulation of nuclear receptor expression (19, 20). Given that PGC-1α functions as the master regulator of energy dissipation in brown and beige fat (21, 22), we hypothesize that CREBZF may mediate the effects of glucose on thermogenesis through directly associating with PGC-1α. Therefore, the transcriptional activity of PGC-1α on thermogenic program was inhibited. As shown in Fig. 7 A and B, CREBZF physically interacted with PGC-1α. Furthermore, CREBZF overexpression significantly ablated PGC-1α-induced expression levels of UCP1 and its downstream target genes in adipocytes (Fig. 7 C and D). Chromatin immunoprecipitation assays showed that glucose-induced occupancy of PGC-1α over UCP1, PGC-1α, and COX8b promoters was significantly augmented in CREBZF-deficient mouse primary adipocytes (Fig. 7E). Taken together, these data demonstrate that CREBZF deficiency potentiates glucose-induced thermogenesis in vivo and in vitro and that glucose may impair iWAT thermogenic capacity through the CREBZF-PGC-1α pathway.

Fig. 7.

CREBZF coordinates with PGC-1α to inhibit the thermogenic gene program. (A and B) CREBZF associates with PGC-1α. GST-CREBZF and FLAG-PGC-1α (A) or FLAG- PGC-1α and myc-CREBZF (B) were transfected in HEK293T and then purified with GSH Sepharose beads or FLAG beads. The precipitates and lysates were analyzed. (C and D) Primary adipocytes were infected with adenovirus encoding PGC-1α (Ad-PGC-1α) or Ad-GFP, followed by transfection with CREBZF. mRNA levels of UCP1 (C) and PGC-1α target genes (D) were measured. n = 7 to 8, *P < 0.05, vs. Ad-GFP and pcDNA; #P < 0.05, vs. Ad-PGC-1α and pcDNA. (E) The quantitative chromatin immunoprecipitation assays were performed using PGC-1α antibody or control antibody against IgG in adipocytes treated with or without glucose. The occupancy of endogenous PGC-1α on the UCP1, PGC-1α, and COX8b promoters was analyzed via real-time PCR. n = 4 to 5. *P < 0.05 vs. vehicle and CREBZF-/- with PGC-1α immunoprecipitation; #P < 0.05 vs. glucose and CREBZF+/+ with PGC-1α immunoprecipitation. (F) The proposed model for the CREBZF-mediated glucose-sensing pathway in WAT. In white adipocytes, when the cellular glucose levels are high, such as cold exposure or β3-adrenergic receptor signaling activation, CBP/p300-mediated acetylation of CREBZF at Lys208 caused a potent stabilization of CREBZF and inhibition of PGC-1α’s transcriptional activity, leading to impaired mitochondrial function, thermogenic capacity, energy expenditure, and cold tolerance. In contrast, under low-glucose conditions, deacetylation of CREBZF by HDAC3 increases degradation of CREBZF via the ubiquitin–proteasome pathway, which leads to enhanced thermogenic capacity. The CREBZF-mediated glucose-sensing pathway may represent a central mechanism of glucose in regulating energy balance in adipocytes, which is required for repressing hyperactivation of browning and thermogenesis.

Discussion

The up-regulated glucose uptake and glucose oxidation are essential to support thermogenesis in brown and beige adipocytes upon cold exposure or environmental stimulus. This study uncovers a metabolic regulator CREBZF as an important mediator of glucose effects on browning and thermogenic capacity in white adipocytes. Mechanistically, CREBZF is modulated by glucose signals via CBP/p300- and HDAC3-mediated reversible acetylation modification and dynamic changes of protein stability. Deficiency of CREBZF potentiates the glucose-induced thermogenesis and energy expenditure. CREBZF-mediated inhibition of thermogenic gene expression and energy dissipation may represent a critical mechanism of glucose in regulating energy balance.

Glucose Plays Essential Roles in Maintenance of Energy Homeostasis in Adipose Tissue.

Glucose uptake increases in brown and beige adipose upon cold exposure. Previous studies have proposed that glucose supports the energy demand for thermogenesis as an energy source (23, 24). However, whether glucose regulates the thermogenic program in the transcriptional level is not clear. In our study, we show that glucose significantly increases transcriptional activity of the thermogenic program. This notion is supported by our observation that fasted mice with glucose injection have increased body temperature and expression levels of thermogenic genes in iWAT. The thermogenic program is also stimulated by glucose in a beige adipocyte-autonomous manner. Unexpectedly, we also uncovered a critical mechanism by which CREBZF-mediated glucose sensing may contribute to inhibition on brown features in white adipose.

Glucose treatment induces body temperature and thermogenesis as well as CREBZF in white adipose, which are further augmented by CREBZF deficiency. Furthermore, CREBZF deficiency protects the mice from diet-induced systemic metabolic disorders under hyperglycemia conditions. We hypothesize that in physiological conditions, glucose-induced CREBZF may protect beige adipocytes from hyperactivation and unnecessary energy dissipation. Given that the thermogenic capacity is attenuated in hyperglycemia and obesity (25, 26), glucose-dependent activation of CREBZF may represent the mechanism for reduced energy expenditure in these pathological conditions, and CREBZF activity is potently increased in WAT of hyperglycemia ob/ob mice and by glucose treatment in human adipose tissues. Therefore, it is likely that glucose may regulate thermogenesis through a feedback mechanism, although further investigations are needed.

CREBZF Couples Glucose Signals to Repress Adipose Tissue Browning.

Several lines of evidence support the critical roles of glucose-induced CREBZF in regulating browning capacity and thermogenesis in beiges. First, CREBZF is highly expressed in iWAT and BAT, and protein levels of CREBZF are potently increased by glucose treatment or cold-induced glucose uptake in mice iWAT and human adipose tissues, which are negatively correlated with UCP1 in human adipose tissues. Second, the induction of thermogenic capacity was only observed in iWAT of CREBZF FKO mice in response to cold exposure or CL-316,243 treatment. Third, in vitro assays show the increased capacity of thermogenesis in CREBZF-deficient beiges, suggesting an adipocyte-autonomous regulation. Moreover, CREBZF deficiency relieves the suppression of transcriptional activity of PGC-1α by glucose, whereas CREBZF gain of function is sufficient to inhibit PGC-1α-induced thermogenic genes in adipocytes. These findings demonstrate that CREBZF activation mediates glucose signals to regulate browning of WAT and maintenance of energy balance.

The present study demonstrates important effects of CREBZF on browning and thermogenesis selectively in beiges. Consistent with the effects of CREBZF, various regulators promote browning of iWAT with mild effects on BAT function, such as p38α, SIRT1, and FGF21 (27–29). Together, these findings suggest that compared with BAT activation, iWAT browning is regulated by distinct and complex mechanisms, which requires further investigation.

Glucose Induces CREBZF Activity by CBP/p300- and HDAC3-Mediated Reversible Acetylation.

The present study establishes CREBZF as a glucose sensor in the process of browning and thermogenesis. CREBZF activity is dynamically regulated by reversible acetylation in response to glucose signaling. CREBZF is deacetylated by HDAC3 and then degraded via the ubiquitin–proteasome pathway under low-glucose conditions. In contrast, upon cold exposure or β3-adrenergic receptor agonist treatment when glucose uptake is elevated, CREBZF is acetylated by CBP/p300 in adipocytes, which leads to an induction of CREBZF activity and inhibition of PGC-1α-mediated thermogenic gene expression. As acetyl-coenzyme A is the sole donor of acetyl groups for protein acetylation, it is likely that glucose oxidation–produced acetyl-coenzyme A (23, 30) may serve as a cellular signal and directly elevate the acetylation and activity of CREBZF.

Given that nutritional and environmental cues are sufficient to activate CREBZF-mediated regulation of browning in iWAT, lysine 208 acetylation of CREBZF may be exploited as readout to evaluate the browning process and develop therapeutic approaches for treating hyperglycemia and obesity.

In summary, these findings demonstrate that in addition to fuel source for thermogenesis, glucose may act as a chemical signal to control the thermogenic program and energy metabolism via reversible acetylation of the metabolic coregulator CREBZF. Targeting CREBZF may have therapeutic implications for treating obesity and related metabolic diseases, such as insulin resistance.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.