Hepatic Activation of the FAM3C-HSF1-CaM Pathway Attenuates Hyperglycemia of Obese Diabetic Mice

In the past decades, type 2 diabetes had become a serious public health issue, affecting more than 300 million people worldwide in 2013 (1). Increased hepatic glucose production as a result of insulin resistance had been accepted as being critical for the development of fasting hyperglycemia and type 2 diabetes (2). Protein kinase B (Akt) is the key node molecule of insulin signaling, and Akt activation phosphorylates and inactivates FOXO1 and glycogen synthase kinase 3 (GSK-3) to suppress gluconeogenesis and enhance glycogen synthesis in the liver (2–4). Akt also affects the activities of SREBP-1 and FOXO1 to modulate lipid metabolism (2,5). Under insulin-resistant status, activating Akt via insulin-independent pathway(s) represents a novel strategy for suppressing hepatic gluconeogenesis and ameliorating hyperglycemia (6,7).

The family with sequence similarity 3 (FAM3) gene family contains four members, designated as FAM3A, FAM3B, FAM3C, and FAM3D (8). FAM3A plays important roles in regulation of hepatic and glucose metabolism via ATP-P2 receptor–mediated activation of Akt pathways independent of insulin (7). FAM3B (PANDER) exerts deleterious effects on pancreatic β-cell function (9,10) and hepatic insulin signaling (11–14). To date, FAM3C, also known as interleukin-like epithelial-mesenchymal transition (EMT) inducer (ILEI), has been reported to be important in embryonic development, EMT, and the progression of some cancers (15–17). In situ hybridization and Northern blot analyses revealed that FAM3C was ubiquitously expressed in all tissues (17). Cleaved FAM3C can be secreted by cells, and the secreted FAM3C protein form is associated with autophagy of tumors. Circulating FAM3C level is a biomarker for autophagy and some cancers (15,16,18). Secreted FAM3C protein is important for maintaining retina size and the function of the photoreceptor layer (19). Transgenic FAM3C overexpression protects mice against Alzheimer disease by destabilizing the penultimate amyloid-β precursor (20). FAM3C also regulates osteogenic differentiation and bone metabolism (21). Whether and how FAM3C plays a role in the regulation of glucose and lipid metabolism remains uncharacterized.

The current study revealed that FAM3C expression is decreased in the livers of obese diabetic mice and that hepatic FAM3C restoration attenuates hyperglycemia, insulin resistance, and fatty liver. Mechanistically, FAM3C upregulates HSF1 to directly induce CALM1 gene transcription, elevating calmodulin (CaM) protein level to activate the phosphatidyl inositol 3-kinase (PI3K)–Akt pathway and repress gluconeogenic gene expression in an insulin- and Ca²⁺-independent manner in hepatocytes.

The study used 8- to 10-week-old male C57BL/6 mice and 8- to 10-week-old male db/db mice on a BKS background. The C57BL/6 mice were fed a 45% high-fat diet (HFD) or normal diet (ND) (Medicience Ltd, Jiangsu, China) for 12 weeks (6,7). All procedures involving experimental animals were approved by the Institutional Animal Care and Use Committee of Peking University Health Science Center, which complies with the Guide for the Care and Use of Laboratory Animals.

Anti-FAM3C antibody was purchased from Abcam (Cambridge, U.K.). Anti-phosphorylated (p)Akt (phosphorylation at Ser473 or Thr308 site) and Akt antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). If not indicated specifically, pAkt represented phosphorylation at Ser473 site. Other antibodies were obtained from Santa Cruz Biotechnology, Inc. (Dallas, TX), Cell Signaling Technology, Inc., or other commercial companies.

Adenovirus (Ad) expressing homo FAM3C was constructed by SinoGenoMax Company (Beijing, China) using the homo FAM3C expression plasmid purchased from OriGene (#RG201036). To overexpress FAM3C in the livers of HFD and db/db mice, 1.0 × 10⁹ plaque-forming units of Ad-FAM3C or Ad-green fluorescent protein (GFP) were injected into mice via tail vein in 100 μL volume (6,7). At the 4th and 7th day after virus injection, oral glucose tolerance tests (OGTTs) were performed. On the 9th day, the mice were sacrificed for experimental analysis. The serum was collected for measuring triglycerides (TG) and cholesterol (CHO) levels. The liver was collected for TG and CHO content measurement and other biochemical analyses.

For the OGTT, mice were fasted from 8 a.m. to 2 p.m. with free access to drinking water in a cage with fresh bedding. Blood glucose was determined after 0 min, and mice were orally administered glucose at the dose of 3 g/kg body weight. Blood glucose levels at 15, 30, 60, 90, and 120 min were monitored. The pyruvate tolerance test (PTT) and insulin tolerance test (ITT) were performed on the 7th day after an injection of Ad-FAM3C or Ad-GFP, as detailed previously (6,7). OGTT, ITT, and PTT were performed using different sets of mice.

Human HepG2 cells were infected with 25 multiplicity of infection of Ad-GFP or Ad-FAM3C for 24 h. For insulin stimulation, infected cells were serum starved for 12 h and then stimulated with 10 nmol/L or 100 nmol/L insulin (Novo Nordisk) for 5 min. The cells were lysed for protein analysis. Infected cells were treated with 50 μmol/L LY294002, 1 μmol/L wortmannin, 100 μmol/L chlorpromazine (CPZ), and 100 μmol/L W-7 for 1 h before being lysed for pAkt analysis. To knock down CALM1 expression, HepG2 cells were transfected with the mixture of four sets of small interfering (si)RNAs against homo CALM1 mRNA (siRNA sequences are provided in Supplementary Table 1) for 6 h, and then the medium replaced with fresh DMEM/high-glucose medium. The cells were infected with Ad-GFP or Ad-FAM3C for 24 h. For depletion of extracellular Ca²⁺, infected cells were treated with Ca²⁺-free DMEM/high-glucose medium plus 0.5 mmol/L EGTA in the absence or presence of 100 μmol/L CPZ for 2 h before pAkt analysis.

Hepatocytes were isolated from mice by nonrecirculating collagenase perfusion through the portal vein, as previously described (7). The isolated mouse hepatocytes were plated on dishes coated with rat collagen type I and cultured in RPMI 1640 containing 10% FBS at 37°C in a 5% CO₂ atmosphere.

HepG2 cells were plated in 6-well plates and transfected with 5 μg CALM1 or HSF1 plasmid using the ViaFect transfection kit (Vigorous Biotechnology, Cat No T001). The transfected cells were incubated for 6 h, and then the transfection medium was replaced with normal culture medium. The cells were cultured for another 24 h, followed by treatment with CPZ or PI3K inhibitor for 1 h before being lysed for pAkt analysis. Plasmids expressing homo CALM1 and HSF1 were purchased from OriGene (CALM1, Cat No SC115829; HSF1, Cat No RG200314).

To knock down FAM3C or HSF1 expression, HepG2 cells were transfected with 50 nmol/L scrambled siRNA or 50 nmol/L siRNAs mixtures against homo FAM3C mRNA or HSF1 (siRNA sequences are provided in Supplementary Table 1). The expression levels of FAM3C, HSF1, CaM, and pAkt were analyzed 24 h after transfection.

Hydrodynamics-based transfection in animals by tail vein administration of naked plasmid DNA was detailed previously (22–24). Briefly, 8- to 10-week-old male C57BL/6 mice were randomly divided into three groups based on OGTT, and 50 μg endotoxin-free pEGFP-C3, pHSF1, or pCALM1 dissolved in sterile saline (the volume of saline was 10% the body weight) was injected into the tail vein in 7 s at room temperature. OGTTs were performed at 72 h after the plasmid injection. Mice were sacrificed 24 h later.

Total RNA of tissues and cells were extracted using RNApure High-purity Total RNA Rapid Extraction Kit (BioTeke Corporation). The complementary DNA was synthesized using RevertAid First Strand cDNA Synthesis Kit (Fermentas, K1622). Target gene mRNA level was normalized to that of β-actin in the same sample. Each sample was assayed in duplicate or triplicate in each experiment. Melting and amplification curves for each PCR action were also analyzed for ensuring the specificity of the PCR amplification. The primer sequences for real-time PCR assays are listed in Supplementary Table 2.

Cells, livers, and other tissues were lysed in fresh Roth lysis buffer on ice. The lysates were centrifuged at 4°C at 12,000 rpm for 10 min, and the pellets were discarded. The protein concentration in the supernatant was determined using bicinchoninic acid assay. Total proteins (80–120 μg) were separated by 10 or 12% SDS-PAGE and then transferred to the Hybond-C Extra membrane (Amersham Biosciences). The Western blot assays were performed as detailed previously (6,7).

HepG2 cells seeded on coverslips were infected with Ad-GFP or Ad-FAM3C for 24 h, loaded with 1 μmol/L Fura-2–acetoxymethyl for 30 min, and then imaged under an Olympus IX71 fluorescence microscope. For depletion of extracellular calcium experiments, infected cells were treated with calcium-free medium plus 0.5 mmol/L EGTA for 2 h before the Fura-2–acetoxymethyl load. The emission intensities under 340 nm and 380 nm illumination were recorded every 1 s, and the average ratio of the emission densities (F340-to-F380) in 300 s reflects the basal intracellular free calcium level (6,7).

For glucose production assay, cells were infected with Ad-GFP or Ad-FAM3C for 24 h and then washed three times by PBS buffer solution. Cells were cultured in glucose and phenol red-free DMEM medium, supplemented with 20 mmol/L sodium lactate and 2 mmol/L sodium pyruvate for 13 h. Insulin (100 nmol/L) was added to the cells for an additional 3 h. Glucose content in the medium was determined using the Glucose Assay Kit (Sigma-Aldrich, GAGO-20) and normalized to total cellular protein content.

The protocol for chromatin immunoprecipitation (ChIP) experiments was detailed previously (25). HepG2 cells were washed twice with PBS and cross-linked with 1% formaldehyde at room temperature for 10 min. The cells were sonicated and centrifuged for 10 min at 4°C. The supernatants were immunocleared using 2 μg sheared salmon sperm DNA, 20 μL preimmune serum, and protein A-Sepharose for 2 h at 4°C or just diluted in dilution buffer as input control. The sheared chromatin was immunoprecipitated with anti-HSF1 antibodies or IgG overnight at 4°C. The eluted immunoprecipitate and the input control were digested with proteinase K. DNA was extracted for PCR amplification with the specific primers spanning the putative HSF1 binding site in homo CALM1 promoter (sense, 5′-acaagcaagcaaagcccttt-3′; antisense, 5′-agccagtaatgtgaacccca-3′).

Cells were seeded on coverslips and infected with Ad-GFP or Ad-FAM3C as previously described (7). The cells were rinsed with PBS and then permeabilized with 0.2% Triton X-100/0.5% PBS for 10 min, followed by washing with PBS. The coverslips were blocked in 1% BSA for 30 min at 37°C and then incubated overnight with anti-FOXO1 antibodies at 4°C. The coverslips were washed with PBS, followed by detecting with goat anti-rabbit Alexa Fluor 594. After nuclear staining with DAPI, coverslips were mounted on glass slides using 70% glycerol in PBS. Mounted coverslips were imaged, and cells were visualized by fluorescence microscopy using a confocal laser scanning microscope.

Data are presented as mean ± SEM. Statistical significance of differences between groups was analyzed by unpaired Student t test or by one-way ANOVA when more than two groups were compared.

A previous study reported that FAM3C protein can be secreted after cleavage of the signal peptide (19). Two FAM3C protein isoforms exist in cells; one is called the full-length form and the other is called the secreted form without the signal peptide (20). Generally, two FAM3C protein isoforms (∼26 and 22 kD) exist in most mouse tissues (data not shown). To determine whether FAM3C is involved in dysregulated glucose and lipid metabolism, its expression in main metabolic tissues of obese diabetic mice was analyzed. The FAM3C mRNA level was decreased in livers but remained unchanged in hypothalamus, pancreas, and epididymal adipose tissue of db/db and HFD-fed diabetic mice (Fig. 1A and C). The FAM3C mRNA level remained unchanged in skeletal muscle of db/db mice but was slightly increased in that of HFD mice (Fig. 1A and C). Both isoforms of FAM3C protein were reduced in the livers of db/db and HFD mice (Fig. 1B and D).

The FAM3C gene was overexpressed in the livers of obese diabetic mice via tail vein injection of FAM3C-expressing Ads to evaluate its effect on glucose and lipid metabolism. On the 4th and 7th day after viral injection, Ad-FAM3C–transduced HFD mice exhibited significantly improved glucose intolerance and fasting hyperglycemia compared with Ad-GFP–treated control HFD mice (Fig. 2A and C). The global insulin sensitivity was increased at the 7th day after Ad-FAM3C injection as evaluated by the ITT (Fig. 2D) and hyperinsulinemic-euglycemic clamp test (Supplementary Fig. 1A–C). Hepatic glucose production was also significantly suppressed by FAM3C overexpression at the 7th day as evaluated by the PTT in HFD mice (Fig. 2E). Morphological and Oil Red O staining assays indicated that FAM3C overexpression ameliorated the fatty liver phenotype of HFD mice (Fig. 3A). Quantitative assays indicated that FAM3C overexpression reduced TG content but had little effect on CHO content in the livers (Fig. 3B). In contrast, the Ad-FAM3C injection had no significant effect on serum TG and CHO levels in HFD mice (Fig. 3C). In the livers of HFD mice, treatment of Ad-FAM3C reduced the mRNA levels of gluconeogenic and lipogenic genes, including PEPCK, glucose-6-phosphatase (G6Pase), and fatty acid synthase (FAS), but failed to affect that of peroxisome proliferator–activated receptor-γ (PPAR-γ), ATP synthase-β subunit (ATPSβ), and FAM3A (Fig. 3D). Injection of Ad-FAM3C increased both isoforms of FAM3C protein by two- to threefold in HFD mouse livers (Fig. 3E). In contrast, injection of Ad-FAM3C had little effect on the mRNA and protein levels of FAM3C in the pancreas, epididymal white adipose tissue, and skeletal muscle of HFD mice (Supplementary Fig. 2A–G), indicating the specific overexpression of FAM3C in mouse livers via tail vein injection of Ads. FAM3C overexpression elevated pAkt with decreased protein levels of PEPCK, G6Pase, and FAS in HFD mouse livers (Fig. 3E). Similarly, hepatic overexpression of FAM3C attenuated glucose intolerance and fasting hyperglycemia in db/db mice on the 4th and 7th day (Supplementary Fig. 3A–C). FAM3C overexpression ameliorated fatty liver with little effect on serum TG and CHO levels in db/db mice (Supplementary Fig. 3D–F). FAM3C overexpression also increased the pAkt level and reduced the mRNA and protein levels of PEPCK, G6Pase, and FAS in db/db mouse livers (Supplementary Fig. 4A–C).

FAM3C overexpression promoted Akt activation in HepG2 cells without insulin stimulation (Supplementary Fig. 5A–D). To further validate that FAM3C-induced Akt activation is independent of insulin, cells infected with Ad-FAM3C or Ad-GFP were serum starved for 12 h and then stimulated with 10 nmol/L or 100 nmol/L insulin for 5 min before the pAkt level was analyzed. In Ad-GFP–treated HepG2 cells, 100 nmol/L insulin significantly activated Akt (Fig. 4A), whereas 10 nmol/L insulin failed to induce Akt activation (Supplementary Fig. 5E and F). In Ad-FAM3C–treated cells, Akt was significantly activated without insulin stimulation (Fig. 4A). Insulin at 10 nmol/L or 100 nmol/L failed to further augment Akt activation in Ad-FAM3C–treated cells (Fig. 4A and Supplementary Fig. 5E and F). FAM3C clearly activated Akt independent of insulin. FAM3C-induced Akt activation was completely blocked by inhibition of PI3K using LY294002 (Fig. 4B) and CaM using CPZ (Fig. 4C). Similarly, FAM3C-induced Akt activation was also abolished by the different PI3K inhibitor wortmannin and CaM inhibitor W-7 (Supplementary Fig. 6A and B). Overall, FAM3C promoted Akt activation via the CaM-PI3K pathway independent of insulin stimulation. The roles of extracellular calcium influx and internal calcium release in FAM3C-induced activation Akt were further evaluated. Unexpectedly, inhibition of P2 receptors, inositol 1,4,5-trisphosphate receptor (IP3R), L-type calcium channel, and transient receptor potential cation channel 4 (TRPV4) failed to affect FAM3C-induced Akt activation in HepG2 cells (data not shown). In support, depletion of extracellular calcium failed to affect FAM3C-induced Akt activation (Fig. 4D). Moreover, FAM3C-induced Akt activation was dependent on CaM in the presence or absence of extracellular calcium (Fig. 4E). FAM3C overexpression reduced the cellular calcium level, and depletion of extracellular calcium further decreased the cellular calcium level without affecting Akt activation (Fig. 4E and F). Clearly, FAM3C induced Akt activation through CaM independent of calcium. As a positive control, FAM3A overexpression elevated the cellular calcium level (Fig. 4F) (7). Overall, these findings revealed that FAM3C-induced activation of the CaM-PI3K-Akt pathway is independent of calcium and insulin.

To determine the CaM-dependent mechanism of FAM3C-induced Akt activation, whether FAM3C regulated CaM expression was evaluated. There are three different calmodulin genes (CALM1, CALM2, and CALM3) that collectively yield five transcripts, and all of them encode one identical CaM protein (26). FAM3C overexpression increased CALM1, but not CALM2 and CALM3 mRNA levels, and CaM protein in HFD mouse livers (Fig. 5A). The CALM1 mRNA and CaM protein levels were both reduced in the livers of HFD and db/db mice (Supplementary Fig. 7A–D). FAM3C overexpression also upregulated the CALM1 mRNA and CaM protein levels in HepG2 cells (Fig. 5B). CALM1 plasmid transfection increased the cellular CaM protein level and activated Akt without insulin stimulation. Moreover, Akt activation induced by CALM1 overexpression was blocked by inhibitors of PI3K and CaM (Fig. 5C). The effect of CALM1 silencing on FAM3C-induced Akt activation was further determined. siCALM1 treatment significantly reduced the CALM1 mRNA and CaM protein levels in HepG2 cells (Supplementary Fig. 8A and B). CALM1 silencing inhibited FAM3C-induced Akt activation (Fig. 5D). Clearly, FAM3C increased the CaM protein level through enhancement of CALM1 transcription to activate the PI3K-Akt pathway in hepatocytes.

To further study the mechanism of FAM3C-induced enhancement of CALM1 transcription, the potential binding sites of certain key transcriptors in the promoter regions of mouse and human CALM1 genes were analyzed. The results indicated that human and mouse CALM1 promoters both contain putative binding sites for Hairy, v-Myb, HSF1, and Oct-1 (Supplementary Fig. 9). The mRNA level of Hairy was increased, whereas the levels of HSF1 and v-Myb were decreased in HFD mouse livers (Supplementary Fig. 10A). The HSF1 protein level was also decreased in HFD mouse livers (Supplementary Fig. 10B). Similarly, the mRNA and protein levels of HSF1 were decreased in db/db mouse livers (Supplementary Fig. 10C and D). FAM3C overexpression increased the mRNA and protein levels of HSF1 in the livers of HFD mice (Fig. 6A) and db/db mice (data not shown) and in HepG2 cells (Fig. 6B). HSF1 overexpression upregulated the CaM protein level (Fig. 6C) and induced Akt activation via the CaM-PI3K pathway in HepG2 cells (Fig. 6D). Consistently, HSF1 silencing reduced the mRNA level of CALM1 and the protein levels of CaM and pAkt in HepG2 cells (Fig. 6E and F). FAM3C silencing reduced the mRNA levels of HSF1 and CALM1 and the protein levels of HSF1, CaM, and pAkt in HepG2 cells (Fig. 6G and H). In support, FAM3C overexpression upregulated the protein levels of HSF1, CaM, and pAkt in primary cultured mouse hepatocytes without insulin stimulation (Fig. 6I). Moreover, inhibition of HSF1 using KRIBB11 (2,6-pyridinediamine trifluoroacetate salt) blocked FAM3C-induced elevation in CaM and pAkt protein levels in HepG2 cells (Fig. 7A). ChIP assay further revealed that HSF1 was directly bound to the potential site of human CALM1 promoter (Fig. 7B and Supplementary Fig. 9).

To further confirm the role of the HSF1-CaM pathway in Akt activation and hepatic glucose and lipid metabolism, HSF1 and CALM1 were overexpressed in HFD mouse livers via tail vein plasmid injection. At 72 h after the HSF1 and CALM1 plasmid injection, glucose intolerance and fasting hyperglycemia were significantly improved compared with mice treated with GFP plasmid (Fig. 7C and D). HSF1 overexpression upregulated the mRNA level of CALM1 and reduced the mRNA levels of PEPCK, G6Pase, and FAS in HFD mouse livers (Fig. 7E). HSF1 overexpression increased the protein level of CaM and pAkt and reduced that of PEPCK, G6Pase, and FAS in HFD mouse livers (Fig. 7F). CALM1 overexpression reduced the mRNA levels of PEPCK, G6Pase, and FAS (Fig. 7E) but increased the pAkt level with a decrease in PEPCK, G6Pase, and FAS protein levels in HFD mouse livers (Fig. 7G). HSF1 and CaM overexpression both reduced TG and CHO content in HFD mouse livers (Supplementary Fig. 11A–C).

FOXO1, a key transcriptor controlling the transcription of gluconeogenic genes PEPCK and G6Pase, is phosphorylated and inactivated by Akt activation. Consistent with Akt activation, FAM3C overexpression promoted nuclear exclusion of FOXO1, which was blocked by inhibiting CaM (Fig. 8A). In support of FOXO1 repression, FAM3C overexpression reduced the protein levels of PEPCK and G6Pase in HepG2 cells (Fig. 8B) and in primary mouse hepatocytes (Fig. 8C). FAM3C overexpression also suppressed glucose production in HepG2 cells (Fig. 8D). Moreover, FAM3C-induced suppression on glucose production was completely reversed by inhibition of HSF1 using KRIBB11 in HepG2 cells (Supplementary Fig. 12). As expected, FAM3C overexpression similarly induced Akt phosphorylation at Thr308 site in HepG2 cells. FAM3C overexpression reduced phosphorylated mechanistic target of rapamycin (mTOR) complex 1 (pmTORC1), SREBP-1c, and FAS protein levels in HepG2 cells (Supplementary Fig. 13A and B). In support, FAM3C overexpression also induced Akt phosphorylation at Thr308 site and reduced pmTORC1, SREBP-1c, and FAS protein levels in HFD mouse livers (Fig. 3E and Supplementary Fig. 13C and D).

The current study revealed that FAM3C plays important roles in regulation of hepatic glucose and lipid metabolism via insulin-independent activation of the Akt pathway. Although FAM3A and FAM3C both activate Akt via CaM-dependent pathways, they display distinct mechanisms. FAM3A-induced Akt activation depends on elevation of the cytosolic calcium level resulting from ATP/P2 receptor–mediated extracellular calcium influx and internal calcium release (7). In contrast, FAM3C overexpression had little effect on extracellular ATP content (Supplementary Fig. 14). FAM3C upregulated HSF1 to increase the CaM level, which activated Akt in an insulin- and calcium-independent manner. Overall, these findings together suggested that activating the insulin-independent CaM-Akt pathway may represent a potential strategy for the treatment of type 2 diabetes and nonalcoholic fatty liver disease with severe insulin resistance.

CaM protein level was reduced in obese diabetic mouse livers, and FAM3C overexpression increased CaM protein in mouse livers and cultured hepatocytes. FAM3C overexpression upregulated CALM1 mRNA but not CALM2 and CALM3 mRNAs in HFD mouse livers, suggesting that FAM3C likely increased the CaM protein level by enhancing CALM1 transcription. Sequence analyses revealed that human and mouse CALM1 gene promoters both contain potential binding sites for the transcriptors Oct-1, Hairy, HSF1, and v-Myb. A previous study reported that HSF1 activates the calcium/CaM-dependent protein kinase II (CaMKII) pathway in cardiomyocytes (27). HSF1 also protects mice against HFD-induced obesity, insulin resistance, and fatty liver via induction of PPAR-γ coactivator 1-α (PGC1α) expression in adipose tissues (28). Thus, the role of HSF1 in FAM3C-induced increase in CaM and pAkt protein levels was further determined. In diabetic mouse livers, the mRNA and protein levels of HSF1 were reduced concomitant with a decrease in FAM3C and CaM expression. FAM3C overexpression increased HSF1 and CaM expression in diabetic mouse livers and cultured hepatocytes. FAM3C silencing reduced HSF1, CaM, and pAkt protein levels in cultured hepatocytes. HSF1 directly bound to the putative site in the human CALM1 promoter. Inhibition or overexpression of HSF1 reduced or increased the CaM protein level and Akt activity in hepatocytes. Importantly, inhibition of HSF1 reversed FAM3C-induced suppression on glucose production in liver cells. HSF1 overexpression activated the CaM-Akt pathway to improve hyperglycemia and fatty liver in obese diabetic mice. Overall, FAM3C upregulated HSF1 to enhance CALM1 transcription, elevating the CaM protein level to activate the Akt pathway and suppress gluconeogenesis in hepatocytes (Fig. 8E). These findings also provided a novel mechanism of HSF1 in regulation of hepatic glucose and lipid metabolism by inducing CALM1 expression beyond its regulatory effect on PGC1α expression in adipose tissues (28). FAM3C overexpression repressed FAS expression in the liver and improved global insulin sensitivity in obese diabetic mice, which likely together contributed to the attenuation of fatty liver. The mechanism(s) of FAM3C repression on FAS expression in obese mouse livers and cultured hepatocytes remains unclear at present. Activation of mTOR increases SREBP-1c expression to induce FAS expression in liver cells (29,30). Therefore, it is possible that FAM3C represses FAS expression by inhibiting the mTOR–SREBP-1c signaling pathway. It had been reported that Akt and mTOR can activate each other in various cell types (31,32), so research is still needed to clarify the distinct mechanism of FAM3C on Akt activation and mTOR repression. Another possibility is that FAM3C inhibits FAS expression by repressing FOXO1 activity. Previous studies reported that adenoviral or transgenic overexpression of FOXO1 induced FAS expression and lipid deposition in mouse livers (12,33) beyond its well-known role in controlling gluconeogenic gene expression.

Overall, our findings for the first time revealed an important role of the FAM3C-HSF1-CaM pathway in regulating Akt activity and glucose/lipid metabolism in hepatocytes independent of insulin and calcium. Because FAM3C overexpression upregulated v-Myb mRNA level in HFD mouse livers, the possibility that FAM3C also induced CALM1 transcription through v-Myb cannot be ruled out. So far, the role of v-Myb in hepatic glucose and lipid metabolism remains unknown. Regarding the effect of liver FAM3C overexpression on glucose intolerance, insulin resistance, and hepatic glucose production, it should also be noted that although Ad-FAM3C–treated mice had a lower fasting blood glucose level, they exhibited comparable response to glucose, insulin, and pyruvate during OGTT, ITT, and PTT compared with Ad-GFP–treated mice. The changes of areas under the curve (AUCs) observed in these tests may have been the result of the lower initial blood glucose level in FAM3C-overexpressing mice compared with control mice.

As reviewed by Berchtold and Villalobo (34), CaM generally is bound and activated by increased cellular free calcium to interact with target proteins and exert different functions. CaM also interacts with target proteins via Ca²⁺-independent mechanisms. CaM was previously reported to interact with insulin receptor substrate (IRS) 1/2, PI3K, and Akt (35–37). The interaction between CaM and IRS proteins in rat soleus muscle is enhanced in insulin-resistant status induced by dexamethasone, and overexpression of CaM in Chinese hamster ovary cells induces insulin resistance (35). Activation of CaMKII induces insulin resistance and Akt inhibition in the obese condition (38); however, CaM also mediates Akt activation in mouse mammary carcinoma cells (37). Silencing of CaMKII inhibits H₂O₂-induced Akt activation in vascular smooth muscle cells (39). Muscle-specific activation of CaMKIV increases global insulin sensitivity in mice (40). FAM3A or ATPSβ activates the CaM-PI3K-Akt pathway to suppress gluconeogenic and lipogenic gene expression in liver cells (6,7). The current study further revealed that FAM3C upregulated HSF1 to induce CALM1 transcription. An elevation in CaM protein level suppressed hepatic glucose production, improved insulin resistance and hyperglycemia, and attenuated fatty liver in type 2 diabetic mice. FAM3C overexpression decreased cellular calcium in HepG2 cells. Depletion of extracellular calcium further reduced cellular calcium but failed to affect FAM3C-induced Akt activation in liver cells. Clearly, FAM3C activates the CaM-Akt pathway independent of calcium. In support, direct CaM overexpression activated Akt to suppress gluconeogenic and lipogenic gene expression in obese diabetic mouse livers. The FAM3C-HSF1-CaM-Akt signaling axis represents a novel model of CaM action independent of calcium. Restoring hepatic FAM3C expression to activate the HSF1-CaM-Akt pathway is beneficial for type 2 diabetes with severe insulin resistance.

Regarding the role of FAM3C in promoting Akt activation, several issues should be noted. Akt also plays important roles in cell proliferation and some cancers (41). The adverse effects of some anticancer drugs inhibiting Akt activity include global insulin resistance and increased hepatic gluconeogenesis (42). Although mesenchymal transdifferentiation induced by transforming growth factor-β (TGF-β) is associated with upregulated Dab2 and FAM3C expression, the corelationship between FAM3C expression and Akt activation remains unrevealed in the previous studies (43,44). FAM3C repression by factors such as fatty acids contributes to dysregulated glucose and lipid metabolism, but long-term FAM3C overactivation by TGF-β may trigger tumorigenesis in various cell types (15,43,44). It had been believed that dysregulated glucose metabolism is involved in various cancers. Gluconeogenesis is reduced in rat liver with Walker-256 tumor (45). Dexamethasone attenuates hepatocellular carcinoma in mice with restored gluconeogenesis in malignant hepatocytes (46). However, upregulation of PEPCK expression also promotes the growth of colon cancer (47). Metformin prevents or reduces the risk of hepatocellular carcinoma in humans and animals with diabetes with the suppression of hepatic gluconeogenesis (48,49). Metformin also increases the pAkt level in diabetic animal livers (50). FAM3C may represent a unique molecule determining the balance between dysregulated gluconeogenesis and tumorigenesis. The upstream signals such as TGF-β may play critical roles in determining whether FAM3C promotes tumorigenesis. Actually, transgenic overexpression of FAM3C reduces brain amyloid-β deposition and ameliorates the memory deficit without triggering tumorigenesis in a mouse Alzheimer disease model (20).

In summary, we presented novel data revealing that the insulin- and calcium-independent FAM3C-HSF1-CaM-Akt pathway plays important roles in regulating glucose and lipid metabolism in the liver (Fig. 8E). Under obese conditions, a decrease in hepatic FAM3C expression will lead to the inhibition of the HSF1-CaM-Akt pathway and contribute to dysregulated glucose and lipid metabolism. Restoring the hepatic FAM3C-HSF1-CaM-Akt pathway is beneficial for correcting dysregulated glucose and lipid metabolism under severe insulin resistance.

Funding. This study was supported by grants from the Ministry of Science and Technology (2016YFC1304800) and the Natural Science Foundation of China (81471035/81670748/91339106/81322011/81422006/81390351).

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

Author Contributions. Z.C. and L.D. wrote the manuscript. Z.C., L.D., W.Y., and J.W. researched data and contributed to the discussion. Z.C., L.D., Y.G., and J.Y. designed the study and revised and edited the manuscript. L.C., Y.C., B.G., and Q.C. contributed to the discussion and reviewed and edited the manuscript. J.Y. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.