Introduction

Melatonin is produced by the pineal gland and is involved in the regulation of diverse physiological processes, such as anti-inflammatory, antioxidant, lipid metabolic, and glucose metabolic processes1,2,3. Melatonin coordinates several molecular factors, including sirtuin 1 (SIRT1), peroxisome proliferator-activated receptor gamma-coactivator-1α, nuclear respiratory factor-1, transcription factor A, retinoid orphan receptor, and noncoding RNAs4,5,6,7,8. Moreover, the melatonin-AKT signaling pathway suppresses hepatic gluconeogenesis via melatonin receptors (MT1/MT2) in rodent models9. Melatonin prevents sepsis-induced liver injury and dysregulation of hepatic gluconeogenesis by targeting SIRT1-signal transducer and activator of transcription 3 (STAT3) activation in the liver10. However, the absence of melatonin prompts hepatic gluconeogenesis by stimulating the unfolded protein response in rats with pinealectomy11.

Cereblon (CRBN) is primarily expressed in diverse tissues and is involved in the regulation of metabolic disturbance by downregulating AMP-activated protein kinase (AMPK)12,13,14. Intriguingly, Crbn-deficient mice exhibit hyperphosphorylated AMPK and improved insulin sensitivity15. B-cell translocation gene 2 (BTG2) is a key regulator of various cellular processes, including cell growth, differentiation, and apoptosis. This molecule is induced by hypoxia, oxidative stress, and metabolic changes and inhibited by insulin and growth factors16,17. Previous studies have shown that BTG2 regulates hepatic gluconeogenic gene transcription by glucagon-CREB signaling and/or Nur77 interaction18,19. Cyclic adenosine monophosphate (cAMP)-responsive element-binding protein H (CREBH) is a basic leucine zipper transcription factor and is highly expressed in diverse tissues, such as the liver, skeletal muscle, and stomach20. CREBH is induced by fasting, glucocorticoid receptor, proinflammatory cytokines, and endoplasmic reticulum (ER) stress and is repressed by insulin. CREBH plays a crucial role in regulating glucose and lipid metabolism by interacting with several nuclear receptors and transcription factors21. However, the potential role of the hepatic CRBN-BTG2-CREBH axis during fasting- and diabetes-mediated glucose metabolism remains largely unexplored.

Sestrin2 (SESN2) is induced by pathophysiological events, such as oxidative stress, hypoxia, and nutritional stimuli. SESN2 has a beneficial effect on physiological processes, including obesity, insulin resistance, and inflammation22. Loss of SESN2 is known to increase oxidative stress and ER stress, which can lead to liver damage and fibrosis23. Small heterodimer partner (SHP; NR0B2) is an orphan member of the nuclear receptor superfamily and lacks the classical DNA-binding domain. SHP functions as a transcriptional corepressor by directly interacting with several nuclear receptors and transcription factors. SHP has a vital role in the maintenance of metabolic substances, including glucose, lipids, bile acid metabolism, and inflammation24,25. However, the potent role of the SESN2-SHP signaling pathway in hepatic gluconeogenesis and its subsequent role remain unknown.

Herein, we demonstrate that elevation of the CRBN-BTG2-CREBH signaling pathway is induced by fasting and diabetic conditions. CRBN and BTG2 significantly increase hepatic glucose metabolism via the induction of CREBH in diabetic mice and primary hepatocytes. Melatonin ameliorates fasting- and diabetes-mediated induction of hepatic gluconeogenesis via the stimulation of the SESN2-SHP cascade. Overall, our findings suggest that the melatonin-SESN2-SHP signaling network provides a strong basic molecular mechanism. This study will be preclinically and/or therapeutically important to target one or more of our identified members to alleviate hepatic metabolic dysfunction under fasting and diabetic conditions.

Materials and methods

Human subjects

Human liver samples were collected at Yonsei University College of Medicine (Republic of Korea) following the Institutional Review Board approved study protocol (No. 2019-0595-016, 2021-2931-001). Liver biopsy from healthy donors (NOR, n = 5) and patients diagnosed with diabetes mellitus (DM, n = 6) was performed with consent. All protocols were approved by the Medical Ethics and Clinical Research Committee following the guidelines and policies of the Declaration of Helsinki.

Animals

Male wild-type C57BL/6 J (Samtako BioKorea, Seoul, Republic of Korea) mice and CRBN null mice were used as described previously15. For fasting and feeding condition studies, the mice were fed ad libitum and fasted for 9 h. WT and streptozotocin (STZ, 80 mg/kg body weight, Sigma‒Aldrich, St. Louis, MO, USA)-induced diabetic mice were treated for 14 days19. WT and CRBN null mice were fed a chow diet (CD) and a high-fat diet (HFD) as mentioned previously15. WT and db/db mice were administered melatonin (Sigma‒Aldrich, 20 mg/kg of body weight) once daily for 14 days. For viral transduction, adenoviruses expressing Crbn, Btg2, Sesn2 (Se2), and Shp (1 × 109 plaque-forming units) were intravenously injected into WT and db/db mice. For the silencing of Btg2, lentivirus expressing shBtg2 (1 × 109 transducing units/mL) was intravenously injected into db/db mice. For silencing viral transduction, WT and db/db mice were injected with Ad-shCrbn, Ad-shSesn2 (shSe2), Ad-small interfering (si)RNA Crebh, and Ad-siShp (1 × 109 plaque-forming units) via tail vein injection. Total RNA was extracted from the livers of adenovirus- and lentivirus-infected mice for quantitative real-time polymerase chain reaction (qPCR) analysis. Blood glucose levels of all mice were measured with Glucostix Accu-Check (Roche Diagnostics, Mannheim, Germany). Animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Illinois at Chicago (#20-079) and Raon Bio (KPCP-140029) according to the policy of the National Institutes of Health.

Measurement of metabolic parameters

Plasma glucose was measured using Glucostix Accu-Check (Roche Diagnostics, Mannheim, Germany) as previously described19,26. For glucose tolerance tests, the mice were fasted for 14 h and then intraperitoneally injected with 1 g/kg glucose. For insulin tolerance tests, the mice were fasted for 4 h and intraperitoneally injected with 0.5 units/kg insulin. The blood glucose levels were measured under the indicated conditions after each injection.

RNA extraction and qPCR analysis

Total RNA was isolated using the TRIzol method (Invitrogen, Carlsbad, CA, USA), as mentioned previously19. mRNA expression was determined using the Power SYBR® Green PCR Master Mix kit (Applied Biosystems, Warrington, UK) and the StepOneTM Real-time PCR system (Applied Biosystems) for qPCR analysis. We utilized gene-specific primers for measuring the Crbn, Btg2, Crebh, Pck1, G6pc, Sesn2, and Shp genes by qPCR15,19,27,28. Gene expression was normalized to ribosomal L32 expression.

Immunoblotting

Proteins were collected from each sample and analyzed according to the methods described previously19,28. BTG2, CREBH, SHP, pAKT, total AKT, β-actin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), CRBN, pAMPK, AMPK, pS6K, S6K (Cell Signaling Technology, Danvers, MA, USA), and SESN2 (Protein Tech, Chicago, IL, USA) antibodies were used as described previously9,14,19,27,28,29,30.

Culture of primary hepatocytes and glucose output assay

Primary hepatocytes were isolated from the livers of male 8-week-old mice using a portal vein collagenase (Sigma‒Aldrich) perfusion method as described previously19. Primary mouse hepatocytes were maintained in M199 medium (Cellgro, Herdon, VA, USA). The cells were utilized for adenoviral infection and FSK (10 μM) or melatonin (500 μM) treatment. Glucose production was performed using a glucose oxidase assay kit (Sigma‒Aldrich) as mentioned previously19.

Plasmids, cell culture, and transient transfections

The reporter plasmid containing the Pck1 and G6pc promoters was previously described19. The Crebh, Btg2, and Shp expression vectors were previously described19,27,28. AML-12 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F-12 medium (Gibco-BRL, Grand Island, NY, USA) with 10% fetal bovine serum (FBS, HyClone, Logan, UT, USA), 1% insulin-transferrin-selenium mixture (ITS, Gibco-BRL), dexamethasone (40 ng/ml, Sigma‒Aldrich), and 1% antibiotics (Gibco-BRL). The cells were maintained in a humidified atmosphere containing 5% CO2 at 37 °C. Transient transfection assays were conducted using AML-12 cells, as previously described19. Briefly, transient transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. Cytomegalovirus-β-galactosidase plasmids were used as an internal standard to adjust for transient transfection efficiency.

Recombinant adenoviruses

Adenoviruses harboring Crbn, Btg2, Crebh, Shp, Ad-siCrebh, Ad-siShp, and green fluorescent protein (GFP) were generated as previously described19,27,28,29. Ad-Sesn2 (Ad-Se2) and Ad-shSesn2 (Ad-shSe2) were purchased from Vector Biolabs (Malvern, PA, USA). A recombinant lentiviral delivery system (Dharmacon, Lafayette, CO, USA) of Btg2-targeted short hairpin (sh)RNA (shBtg2) was prepared as previously described19.

Coimmunoprecipitation (IP) assay

Total liver protein extracts were subjected to immunoprecipitation with antibodies (SHP and CRBH; Santa Cruz Biotechnology). The immunoprecipitated proteins were immunoblotted with specific antibodies as previously described27,29. Membranes were developed using an Amersham Detection kit (GE Healthcare, Piscataway, NJ, USA).

Chromatin immunoprecipitation (ChIP)

The ChIP assay was carried out with corresponding antibodies according to the manufacturer’s protocol (Upstate Biotechnology, Upstate, NY, USA), as previously described19,29. Briefly, crosslinked lysates were prepared with the corresponding beads and salmon sperm DNA. The purified DNA extractions were quantified by qPCR using the proximal region of the Pck1 promoter (F: 5′-CAGACTTTGTCTAGAAGTTT-3′, R: 5′-TCTTGCCTT-AATTGTCAGGT-3′) and nonspecific distal region of the Pck1 promoter (F: 5′-TGCCATGG-CTCACAGTGCCT-3′, R: 5′-GTTACGAAATGACCTGGAGG-3′).

Statistical analysis

Statistical calculations were conducted using GraphPad Prism software (GraphPad, USA). Statistical significance was determined using two-tailed Student’s t test or one-way analysis of variance. All data are presented as the mean±SEM. A p < 0.05 was considered statistically significant.

Results

CRBN, CREBH, and BTG2 gene expression in the liver specimens of diabetic mice and human patients with DM

We investigated the expression of characteristic genes during prolonged fasting in mice. Mice with prolonged fasting exhibited enhanced expression of the CRBN, BTG2, and CREBH genes and hepatic gluconeogenic genes (Pck1, G6pc) compared to fed mice (Fig. 1a, b). Similarly, increases in the mRNA and protein levels of Crbn, Btg2, Crebh, and hepatic gluconeogenic genes were induced in db/db mice (Fig. 1c, d). We also examined these genes in other diabetic mouse models, including streptozotocin (STZ)- and high-fat diet (HFD)-induced diabetic mice. These genes were significantly increased in STZ- and HFD-induced diabetic mice (Fig. 1e–h). To further evaluate the biological consequence of these genes in human participants with diabetes mellitus (DM), we performed qPCR and immunoblot analysis using liver specimens from healthy donors (NOR) and DM patients. As shown in Fig. 1i, j, the expression of key metabolic genes was considerably higher in the livers of DM patients than in those of healthy donors. Taken together, these findings suggest that the hepatic CRBN, CREBH, and BTG2 genes are upregulated during fasting and in patients with diabetes.

Fig. 1: Expression of Crbn, Crebh, and Btg2 genes in the liver specimens of diabetic mice and diabetic patients.
figure 1

a, b WT mice were fed ad libitum (Fed) and fasted for 9 hr (Fast). a Gene expression analysis using gene-specific qPCR primers. b Western blot analysis using the indicated antibodies. c, d WT and db/db mice were fasted for 9 hr. c Gene expression analysis. d Western blot analysis (FL, full-length; N, nuclear). e qPCR analysis of total RNA from the livers of control (CON) mice and streptozotocin (STZ, 80 mg/kg)-induced diabetic mice for 14 days. f Various protein levels in liver tissue extracts were analyzed by immunoblotting using specific antibodies. g qPCR analysis of total RNA from the livers of chow diet (CD) and high-fat diet (HFD) mice for 14 weeks. h Tissue extracts were analyzed by immunoblotting with the indicated antibodies. n = 4-5 mice per group. i qPCR analysis was performed using liver biopsies from healthy donors (NOR) and diabetes mellitus (DM) patients using the indicated primers. j Tissue extracts were analyzed by immunoblotting with specific antibodies. n = 5-6 per group. *P < 0.05, **P < 0.01vs. fed mice, WT mice, CON mice, CD mice, or healthy donors.

Full size image

CRBN induces hepatic gluconeogenesis

We sought to identify the critical role of CRBN in the regulation of hepatic gluconeogenesis with Ad-Crbn in mice. Ad-Crbn mice had substantially higher expression of the BTG2 and CREBH genes as well as blood glucose levels than Ad-GFP control mice but not the AMPK activity (Fig. 2a–c). Indeed, Ad-Crbn-infected mice showed significant impairment in glucose tolerance (Fig. 2d). Insulin intolerance was exacerbated by Ad-Crbn, indicating decreased insulin sensitivity by Ad-Crbn (Fig. 2e). Next, we investigated the impact of CRBN on gluconeogenic signals in Ad-shCrbn-infected mice. The expression of the CRBN, BTG2, and CREBH genes was enhanced by fasting, and this phenomenon was disrupted when Crbn was silenced (Fig. 2f, g). As expected, an HFD increased the expression of these genes and blood glucose levels, while these phenomena were markedly attenuated in Crbn KO mice (Fig. 2h–j). Collectively, these observations indicate that CRBN functions as a key regulator of hepatic gluconeogenesis.

Fig. 2: CRBN induces hepatic gluconeogenesis.
figure 2

a–e WT mice were injected with Ad-GFP and Ad-Crbn via the tail vein for 10 days. a Gene expression analysis using various gene-specific primers. b Immunoblotting with specific antibodies. c Blood glucose levels on Day 10 in Ad-Crbn-infected mice. d Glucose tolerance test in the indicated groups at Day 4 after adenoviral infection. e Insulin tolerance test at Day 8 after adenoviral infection. f, g WT mice were injected with Ad-Scramble (Ad-Scram) and Ad-shCrbn via the tail vein for 10 days. After injection, these mice were fed ad libitum (fed) and fasted for 9 h (fast). f Total RNA was utilized for qPCR analysis with gene-specific primers. g Tissue extracts were harvested from the indicated group of mice and immunoblotted with specific antibodies. h WT and Crbn null mice were fed a CD and an HFD for 14 weeks. Gene expression was analyzed using qPCR with gene-specific primers. i Tissue extracts were analyzed by Western blot analysis with the indicated antibodies. j Measurement of blood glucose levels in the indicated mouse groups. n = 4-5 mice per group. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. Ad-GFP, Fed infected Ad-Scram mice, CD-fed WT mice. #P < 0.05 vs. Fast infected Ad-Scram mice, HFD-fed WT mice.

Full size image

BTG2 increases hepatic gluconeogenesis via the induction of Crebh

Our previous study reported that gluconeogenic stimuli regulate the expression of BTG2 and its relevant target genes involved in hepatic glucose metabolism19. Thus, we hypothesized that BTG2 plays a crucial role in regulating hepatic gluconeogenesis in mice. We tested this hypothesis by infecting mice with Ad-Btg2 and found significantly increased expression of Crebh mRNA and blood glucose levels (Fig. 3a, b). Furthermore, we performed glucose tolerance tests in Ad-Btg2-infected mice and found that glucose tolerance was markedly impaired in Ad-Btg2-infected mice compared to Ad-GFP control mice (Fig. 3c). To examine the role of Btg2 in hepatic gluconeogenesis in diabetic mice, we administered shBtg2 to db/db mice. High basal levels of gluconeogenic genes, blood glucose, and glucose tolerance were reduced by disrupting Btg2 expression in db/db mice (Fig. 3d–f). Likewise, Ad-siCrebh-infected db/db mice showed attenuated gluconeogenic genes, blood glucose, and glucose tolerance (Fig. 3g–i). We sought to investigate the direct effect of Crebh on Btg2-mediated hepatic gluconeogenesis. Overexpression of Btg2 increased the expression of hepatic gluconeogenic enzyme genes and glucose production in primary hepatocytes (Fig. 3j, k), whereas these phenomena were prominently reduced by silencing of Crebh. Overall, these findings suggest that Btg2 increases hepatic gluconeogenesis via the induction of Crebh.

Fig. 3: Btg2 increases hepatic gluconeogenesis via the induction of Crebh.
figure 3

a–c WT mice were intravenously injected with Ad-GFP and Ad-Btg2 for 10 days. a Gene expression analysis using gene-specific qPCR primers. b Blood glucose levels in these mice at Day 10 after adenoviral infection. c Glucose tolerance test in these mice at Day 4 after adenoviral infection. d–f db/db mice were injected with lentivirus-Scram and lentivirus-shBtg2 via the tail vein for 10 days. d Total RNA was utilized for qPCR analysis with gene-specific primers. e Blood glucose levels in these mice at Day 10 after lentiviral infection. f Glucose tolerance test in these mice at Day 4 after lentiviral infection. g–i db/db mice were injected with Ad-Scram and Ad-siCrebh via the tail vein for 10 days. g Total RNA was utilized for qPCR analysis with gene-specific primers. h Blood glucose levels in mice at Day 10 after adenoviral infection. i Glucose tolerance test in these mice at Day 4 after adenoviral infection. n = 4-5 mice per group. j Mouse primary hepatocytes were infected with Ad-GFP, Ad-Scram, Ad-Btg2, and Ad-siCrebh for 36 h and then analyzed using qPCR with various primers. k Glucose output assay under the indicated conditions. Data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 vs. Ad-GFP, Ad-Btg2, or db/db mice.

Full size image

Melatonin ameliorates hepatic gluconeogenesis in diabetic mice

We investigated the pivotal role of melatonin in glucose metabolism. Melatonin administration significantly increased the phosphorylation of AKT and the expression of the SESN2 and SHP genes in the liver and other tested tissues of mice but not in the lung (Fig. 4a, b, Supplementary Fig. 1). Melatonin-challenged WT and db/db mice exhibited diminished gluconeogenic gene expression and blood glucose levels by stimulating Sesn2 and Shp expression compared to individual controls (Fig. 4c, d). Moreover, impaired glucose tolerance and insulin sensitivity in db/db mice were improved by melatonin (Fig. 4e, f). We observed a similar improvement pattern by melatonin in STZ-induced insulin-deficient mouse models. Melatonin administration to STZ-exposed mice resulted in decreased gluconeogenic gene expression and blood glucose levels by increasing Sesn2 and Shp mRNA expression (Fig. 4g, h). Melatonin-exposed STZ mice had significantly better glucose tolerance and insulin sensitivity relative to individual controls (Fig. 4i, j). Next, we examined whether melatonin could inhibit gluconeogenic gene expression in primary hepatocytes. FSK treatment increased hepatic gluconeogenic genes and glucose output but not Sesn2 and Shp expression. Conversely, melatonin repressed FSK-induced gluconeogenic genes and glucose production by upregulating Sesn2 and Shp (Fig. 4k, l). Collectively, these data demonstrated that melatonin could decrease hepatic gluconeogenesis by inducing SESN2 and SHP expression in diabetic mouse models.

Fig. 4: Melatonin ameliorates hepatic gluconeogenesis in diabetic mice.
figure 4

a, b Melatonin (20 mg/kg) was administered to WT mice once daily for 14 days. a Total RNA was utilized for qPCR analysis with gene-specific primers. b Protein levels in liver tissue extracts were analyzed by immunoblotting using specific antibodies. c–f WT and db/db mice were administered melatonin by daily oral gavage for 14 days. c Total RNA was utilized for qPCR analysis with gene-specific primers. d Blood glucose levels in the indicated groups on Day 14 in melatonin-fed mice. e Glucose tolerance test in mice at Day 8 after melatonin administration. f Insulin tolerance test in the indicated groups on Day 12 after melatonin treatment. g–j WT and STZ-induced diabetic mice were administered melatonin by daily oral gavage for 14 days. g Total RNA was utilized for qPCR analysis with gene-specific primers. h Blood glucose levels in these mice at Day 14 after melatonin administration. i Glucose tolerance test in these mice at Day 8 after melatonin administration. j Insulin tolerance test in these mice at Day 12 after melatonin treatment. n = 4-5 mice per group. k, l Mouse primary hepatocytes were treated with FSK (10 μM) and melatonin (500 μM) for 12 hr. k Total RNA was analyzed using qPCR with the indicated primers. l Glucose output assay in mouse primary hepatocytes exposed to FSK and melatonin for 12 h. Data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 vs. the untreated control, WT mice, FSK-treated cells. #P < 0.05 vs. db/db mice, STZ mice.

Full size image

Induction of SESN2 by melatonin inhibits hepatic glucose metabolism

To elucidate the potent role of SESN2 as a negative regulator of hepatic gluconeogenesis, we used Ad-Sesn2 (Ad-Se2)-infected mice. Ad-Se2 elevated Shp mRNA expression and reduced mechanistic target of rapamycin complex 1 (mTORC1)-dependent phosphorylation of ribosomal protein S6 kinase 1 (S6K), leading to decreased blood glucose levels compared to those of the Ad-GFP control groups (Fig. 5a, b). We investigated the crucial role of Sesn2 in glucose metabolism in diabetic mice. Ad-Se2-infected WT and db/db mice exhibited decreased gluconeogenic genes and blood glucose levels by inducing Shp expression (Fig. 5c, d). Ad-Se2-infected db/db mice exhibited improved glucose and insulin tolerance relative to Ad-GFP-infected db/db mice (Fig. 5e, f). Melatonin-fed db/db mice showed attenuation of the phosphorylation status of S6K, gluconeogenic genes, and blood glucose levels, whereas these phenomena were reversed when Sesn2 (Se2) was disrupted (Fig. 5g, h). However, melatonin-induced SESN2 and SHP gene expression was negated by silencing Sesn2, resulting in weak glucose tolerance by Sesn2 knockdown in db/db mice (Fig. 5h, i). FSK-induced hepatic gluconeogenic genes and glucose production were repressed by melatonin, and this phenomenon was rescued by silencing Sesn2. Melatonin-induced SESN2 and SHP protein levels were abolished by silencing Sesn2 (Fig. 5j, k). Taken together, these findings suggest that melatonin-induced Sesn2 inhibits hepatic gluconeogenesis in diabetic mice and primary hepatocytes.

Fig. 5: Induction of SESN2 by melatonin inhibits hepatic gluconeogenesis.
figure 5

a, b WT mice were injected with Ad-GFP and Ad-Sesn2 (Ad-Se2) via the tail vein for 10 days. a Tissue extracts were harvested from the indicated group of mice and immunoblotted for the indicated proteins (left panel). Total RNA was utilized for qPCR analysis with gene-specific primers (right panel). b Blood glucose levels on Day 10 in Ad-Sesn2-infeced mice. c–f WT and db/db mice were injected with Ad-GFP and Ad-Se2 for 10 days. c Total RNA was analyzed by qPCR with gene-specific primers. d Blood glucose levels in these mice at Day 10 after adenoviral infection. e Glucose tolerance test in mice at Day 4 after adenoviral infection. f Insulin tolerance test in mice at Day 8 after adenoviral transduction. g–i db/db mice were administered melatonin by daily oral gavage for 10 days and then intravenously infected with Ad-Scram and Ad-shSesn2 (shSe2) for 10 days. g Blood glucose level in the indicated groups on Day 10 after melatonin and adenoviral injection. h Various protein levels in liver tissue extracts were analyzed by immunoblotting using specific antibodies (left panel). Total RNA was utilized for qPCR analysis with gene-specific primers (right panel). i Glucose tolerance test in the indicated groups on Day 4 after melatonin and adenoviral injection. n = 4-5 mice per group. j Mouse primary hepatocytes were transduced with Ad-Scram and Ad-shSe2 for 36 h and then treated with FSK and melatonin for 12 h. Whole-cell extracts were isolated and analyzed using Western blot analysis with the indicated antibodies. k Glucose output assay in the indicated conditions of mouse primary hepatocytes. Data are representative of at least three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 vs. the untreated control, Ad-GFP, Ad-Scram, and FSK-treated cells. #P < 0.05 vs. db/db-infected Ad-GFP mice and db/db- and Mela-treated Ad-Scram mice.

Full size image

Melatonin represses hepatic glucose metabolism by elevating SHP expression

To test the potential role of SHP in melatonin-mediated repression of gluconeogenesis, we examined the effect of SHP on the transcription of hepatic gluconeogenic genes and glucose output with Ad-Shp. FSK-induced increases in Crebh and gluconeogenic gene expression were attenuated by Ad-Shp (Fig. 6a). FSK-induced glucose production was markedly reduced by Ad-Shp (Fig. 6b). The db/db mice were utilized to evaluate the functional contribution of SHP to hepatic glucose metabolism in response to diabetic conditions. Ad-Shp decreased the elevated blood glucose levels and the expression of these genes in db/db mice compared to Ad-GFP-infected db/db mice (Fig. 6c, d). Interestingly, Ad-Shp improved glucose and insulin sensitivity by attenuating db/db-induced glucose and insulin intolerance (Fig. 6e, f). Melatonin reduced blood glucose levels, but this effect was reversed by Shp knockdown in db/db mice (Fig. 6g). Melatonin markedly decreased Crebh, Pck1, and G6pc mRNA expression by upregulating Shp expression, whereas this inhibitory effect of melatonin was recovered by silencing Shp (Fig. 6h), indicating exacerbated glucose intolerance by Shp knockdown (Fig. 6i). The increased expression of these genes and glucose production by FSK were diminished by melatonin, and this phenomenon was reversed by Shp knockdown (Fig. 6j, k). Overall, our findings demonstrated that SHP mediates the deteriorative effect of melatonin on hepatic gluconeogenesis under diabetic conditions.

Fig. 6: Melatonin represses hepatic gluconeogenesis by elevating SHP expression.
figure 6

a, b Mouse primary hepatocytes were infected with Ad-Shp for 36h and then treated with FSK for 12h. a Total RNA was analyzed using qPCR with the indicated primers. b Glucose output assay under the indicated conditions. c–f WT and db/db mice were infected with Ad-GFP and Ad-Shp via the tail vein for 10 days. c Blood glucose levels on Day 10 in Ad-Shp-transduced mice. d Gene expression analysis using gene-specific qPCR primers. e Glucose tolerance test in mice at Day 4 after adenoviral infection. f Insulin tolerance test in mice at Day 8 after adenoviral transduction. g–i db/db mice were administered melatonin by daily oral gavage for 10 days and then intravenously infected with Ad-Scram and Ad-siShp for 10 days. g Blood glucose levels in the indicated groups on Day 10 after melatonin and adenoviral injection. h Total RNA was utilized for qPCR analysis with gene-specific primers. i Glucose tolerance test in the indicated groups on Day 4 after melatonin and adenoviral injection. n = 4-5 mice per group. j, k Mouse primary hepatocytes were infected with Ad-Scram and Ad-siShp for 36 hr and then treated with FSK and melatonin for 12 h. j Gene expression was analyzed using qPCR with various primers. k Glucose output assay performed using samples from the indicated conditions. Data are representative of at least three independent experiments. *P < 0.05, **P < 0.01,***P < 0.001 vs. untreated control, Ad-GFP, Ad-Scram, and FSK-treated cells. #P < 0.05 vs. db/db-infected Ad-GFP mice and db/db- and Mela-treated Ad-Scram mice.

Full size image

SHP inhibits hepatic gluconeogenesis via the competition of BTG2 and the interaction of CREBH

To determine whether the SHP-mediated reduction in hepatic gluconeogenesis is mediated by the BTG2-CREBH axis, we examined the effect of SHP on BTG2- and CREBH-mediated hepatic gluconeogenic gene expression and glucose production in primary hepatocytes. Ad-Btg2-mediated increases in Pck1 and G6pc expression and glucose production were reduced by Ad-Shp (Fig. 7a, b). Increased glucose production and expression of gluconeogenic genes by Ad-Crebh and/or Ad-Btg2 were strikingly attenuated by Ad-Shp (Fig. 7c, d). We confirmed that melatonin regulates the transcriptional activity of gluconeogenic genes. Pck1 and G6pc promoter activities were synergistically increased by Crebh and Btg2 but not by melatonin treatment (Fig. 7e). We also verified the competition between Btg2 and Shp for the transcriptional activity of Crebh on gluconeogenic gene promoters. Notably, Btg2 increased and Shp repressed Crebh-mediated activation of gluconeogenic gene promoter activities. Cotransfection of Shp attenuated the synergistic activation of gluconeogenic promoter activities by Btg2 and Crebh (Fig. 7f). We examined whether SHP modulates CREBH transcriptional activity via physical interaction. Thus, we performed co-IP assays to test the endogenous association between SHP and CREBH in WT and db/db mice. There was a clear increase in the association of SHP and CREBH in vivo in melatonin-exposed WT and db/db mice compared to their respective unexposed controls (Fig. 7g). These findings indicate that the transcriptional regulation of hepatic gluconeogenesis by Btg2 and Shp may be accomplished by competition and direct interaction between these factors for the association with Crebh on gluconeogenic gene promoters.

Fig. 7: SHP inhibits hepatic gluconeogenesis by interfering with the BTG2 and CREBH interaction.
figure 7

a, b Mouse primary hepatocytes were infected with Ad-GFP, Ad-Btg2, and Ad-Shp for 36 h. a Gene expression analysis using gene-specific primers. b Glucose output assay under the indicated conditions. c, d Mouse primary hepatocytes were infected with Ad-GFP, Ad-Crebh, Ad-Btg2, and Ad-Shp for 36 h. c Gene expression was analyzed using qPCR with various primers. d Glucose output assay in the indicated conditions of mouse primary hepatocytes. *P < 0.05, **P < 0.01 vs. Ad-GFP, individual- or coinfected cells. e AML-12 cells were transiently transfected with Crebh, Btg2, and the indicated reporter genes for 36 hr. After transfection, cells were treated with melatonin for 12 h. f AML-12 cells were cotransfected with Crebh, Btg2, Shp, and the indicated reporter genes for 36 h. Luciferase activity was normalized to β-galactosidase activity to correct for transfection efficiency. Data are representative of at least three independent experiments. *P < 0.05 vs. the untreated control, individual transfected or cotransfected cells. g Coimmunoprecipitation (IP) assays from liver extracts showed the functional interaction between CREBH and SHP under the observed conditions. Total protein extracts were utilized for immunoprecipitation and/or immunoblotting with specific antibodies. h ChIP assay for the occupancy of CREBH on the Pck1 gene promoter. db/db mice were exposed to Mela and then infected with Ad-Scram and Ad-siShp under the indicated conditions. Purified DNA samples were subjected to PCR using specific primers binding to the specific proximal (Pro) and nonspecific distal (Dis) regions on the Pck1 gene promoter. Ten percent of the soluble chromatin was used as input. i ChIP assay showing the occupancy of BTG2 and SHP on the Pck1 promoter under the indicated conditions. Soluble chromatin was immunoprecipitated with BTG2 and SHP antibodies. n = 4-5 mice per group. *P < 0.05 vs. untreated control, db/db mice. #P < 0.05 vs. db/db- and Mela-treated Ad-Scram mice.

Full size image

Finally, we evaluated the effect of SHP on the recruitment of endogenous CREBH to the Pck1 promoter. CREBH directly binds to the proximal region (Pro) of the Pck1 promoter, and this interaction was negated in melatonin-exposed db/db mice. However, melatonin failed to abolish CREBH occupancy on the Pck1 promoter by silencing Shp (Fig. 7h). To confirm whether BTG2 and SHP affect CREBH occupancy on the Pck1 gene promoter, we performed a ChIP assay in db/db mice. As shown in Fig. 7i, induction of BTG2 in db/db mice and melatonin-induced SHP recruitment to the proximal region of the Pck1 promoter were observed. Interestingly, the BTG2 increase in db/db mice promoted CREBH binding to the proximal region of the Pck1 promoter, and this recruitment was completely ablated in melatonin-exposed mice. Conversely, melatonin-stimulated SHP binds to the BTG2-CREBH complex to inhibit its binding to the proximal region of the Pck1 promoter. However, knockdown of Shp inhibited melatonin-mediated recruitment of CREBH to the proximal region of the Pck1 promoter. Overall, these findings suggest that melatonin-induced SHP inhibits hepatic gluconeogenesis via the competition of BTG2 and the interaction of CREBH.

Discussion

Our goal was to identify the novel molecular mechanism of melatonin in the regulation of hepatic gluconeogenesis. We established that gluconeogenic stimuli enhance hepatic gluconeogenesis mediated by the CRBN-BTG2-CREBH signaling pathway. Melatonin attenuates fasting- and diabetes-induced hepatic gluconeogenesis by stimulating the SESN2-SHP axis. Upregulation of SHP by melatonin repressed hepatic gluconeogenic gene transcription via the competition of BTG2 and the interaction of CREBH. Therefore, the melatonin-SESN2-SHP signaling network ameliorates hepatic glucose homeostasis by decreasing the CRBN-BTG2-CREBH axis in diabetic mice and primary hepatocytes. Based on these findings, we propose that the melatonin-SESN2-SHP signaling pathway may provide a novel molecular mechanism underlying the regulation of hepatic gluconeogenic gene expression and glucose metabolism, including glucose production, glucose excursion, and insulin sensitivity.

Lee et al. demonstrated that Crbn KO mice showed highly activated endogenous AMPK activity in diverse tissues31. HFD-fed Crbn KO mice showed lower expression of lipogenic genes and decreased gluconeogenesis, resulting in improved glucose intolerance and insulin resistance15. This study suggested that Crbn deficiency ameliorates lipogenesis and gluconeogenesis by stimulating AMPK and ACC activity in HFD-fed mice. Our current findings revealed that CRBN expression was increased during fasting, diabetes, and HFD and in DM patients (Fig. 1), and Ad-Crbn impaired glucose clearance and insulin sensitivity when compared to those of the Ad-GFP control groups. However, the disruption of Crbn resulted in decreased fasting-induced BTG2 and CREBH gene expression, and we also observed this phenomenon in Crbn KO mice (Fig. 2). These observations clearly suggested the functional role of CRBN in hepatic gluconeogenesis. Our observation implies that CRBN is highly expressed in diabetic mice and human patients with DM, and CRBN positively regulates hepatic glucose metabolism by stimulating BTG2 and CREBH as well as attenuating AMPK.

Next, we investigated the mechanism by which BTG2 induces CREBH in diabetic mice. Hepatic gluconeogenesis is mediated by transcriptional regulation of the CREB, PCK1, and G6PC genes via increased BTG2 expression, and CREBH has a crucial role in glucose and lipid metabolism in the liver and small intestine18,21,32. These results are consistent with our findings that Ad-Btg2 increased Crebh gene expression and glucose production, resulting in impaired glucose tolerance (Fig. 3a–c), whereas Btg2- and Crebh-silenced db/db mice exhibited decreased gluconeogenic gene expression and reduced glucose production and glucose intolerance (Fig. 3d–i). Collectively, these results suggest that transcription of the Pck1 and G6pc genes is enhanced by Btg2 through the induction of Crebh in diabetic mice, which is a coregulator of BTG2 in the liver during diabetic conditions. Although Pck1 and G6pc gene expression was strongly increased in Ad-Btg2-infected primary hepatocytes and mice, this phenomenon was negated by silencing Crebh. However, the regulatory mechanism of the BTG2-CREBH signaling pathway on hepatic gluconeogenesis has not been fully established. Further studies could identify a detailed underlying molecular mechanism by utilizing the hallmark of genetically engineered animal models, such as tissue-specific Btg2 and Crebh knockout and knock-in mice.

Previous reports have shown that melatonin dysregulates hepatic gluconeogenesis by activating the SIRT1-STAT3 signaling pathway and hypothalamic PI3K-AKT axis in rats9,10. We hypothesized that the increase in SESN2 by melatonin regulates hepatic gluconeogenesis. This hypothesis raised intriguing questions regarding whether SESN2 functionally regulates hepatic gluconeogenic gene expression and glucose production. We found that melatonin challenge protects against hepatic gluconeogenesis by stimulating SESN2 and SHP expression in WT and diabetic mice (Fig. 4). Moreover, Ad-Sesn2 and Ad-Shp reduced gluconeogenic gene expression, blood glucose, glucose excursion, and insulin sensitivity via the induction of SHP, whereas this phenomenon was ablated by Sesn2 and Shp knockdown in diabetic mice (Figs. 5, 6). Therefore, melatonin-stimulated SESN2 may increase SHP gene expression in WT and diabetic mice, resulting in decreased hepatic gluconeogenesis and subsequently improved insulin resistance. Based on these results, we speculate that melatonin alleviates hepatic glucose metabolism via enhancement of the SESN2-SHP signaling pathway in response to diabetic conditions. Future studies may explain whether SESN2 also modulates the role of SHP in regulating hepatic gluconeogenesis in tissue-specific conditional knockout of the SESN2 gene.

Our previous studies have demonstrated that SHP alone or metformin-induced SHP represses gluconeogenic signal- and growth hormone-induced PCK1/G6PC gene expression33,34,35. However, the exact molecular mechanism linking SHP and CREBH in diabetes-mediated hepatic gluconeogenesis remains unknown. Based on these findings, we suggest that the melatonin-SHP signaling pathway can explain the molecular mechanism of diabetes-induced gluconeogenic gene transcription. First, we demonstrated that melatonin and SHP attenuate the induction of gluconeogenic gene promoter activities by CREBH and BTG2 (Fig. 7e, f). Second, melatonin and SHP decreased gluconeogenic gene transcription by physical association of SHP with CREBH (Fig. 7g). Third, diabetes-stimulated CREBH recruitment to the Pck1 gene promoter was diminished in melatonin-fed mice but was significantly restored by silencing Shp (Fig. 7h). Finally, melatonin-induced SHP binds to the BTG2-CREBH complex to repress its recruitment to the Pck1 promoter, whereas this phenomenon was rescued by Shp knockdown (Fig. 7i). Our current findings strongly validated a novel molecular mechanism between the melatonin-SHP signaling pathway and diabetic hepatic gluconeogenesis. However, further detailed studies are required to identify more novel molecular networks linked with melatonin-induced SHP expression and hepatic gluconeogenic gene transcription under diabetic conditions by modulating several unknown events, including chromatin remodeling, protein degradation, and microRNAs.

In conclusion, our current study suggests that gluconeogenic stimuli and diabetes deteriorate hepatic gluconeogenesis through the CRBN-BTG2-CREBH axis. Melatonin ameliorates hepatic glucose metabolism by enhancing the SESN2-SHP signaling pathway, while this protective effect of melatonin was negated by silencing Sesn2 and Shp during fasting and diabetic conditions. We propose that the melatonin-SESN2-SHP axis protects against glucose homeostasis by inhibiting the CRBN-BTG2-CREBH signaling pathway under fasting and diabetic conditions, as described in Fig. 8. Our current findings provide the fundamental framework for a therapeutic strategy for the prevention of metabolic dysfunction in response to melatonin and a potential therapeutic approach for the treatment of metabolic diseases such as diabetes.

Fig. 8: Regulation of hepatic gluconeogenesis by gluconeogenic signals and melatonin.
figure 8

Gluconeogenic signals (forskolin and fasting) and diabetes increase hepatic gluconeogenesis by upregulating the CRBN-BTG2-CREBH axis. Melatonin attenuates hepatic gluconeogenesis by stimulating the SESN2-SHP signaling pathway in diabetic mice and primary hepatocytes.

Full size image