Division of Biochemistry and Molecular Biology, Department of Molecular Cell Biology and Samsung Biomedical Institute, Sungkyunkwan University School of Medicine, Suwon, Korea
Division of Biochemistry and Molecular Biology, Department of Molecular Cell Biology and Samsung Biomedical Institute, Sungkyunkwan University School of Medicine, 300 Chunchun-dong, Jangan-gu, Suwon 440-746, Gyeonggi-do, Korea
Potential conflict of interest: Nothing to report.
Postprandial insulin plays a critical role in suppressing hepatic glucose production to maintain euglycemia in mammals. Insulin-dependent activation of protein kinase B (Akt) regulates this process, in part, by inhibiting FoxO1-dependent hepatic gluconeogenesis by direct phosphorylation and subsequent cytoplasmic exclusion. Previously, it was demonstrated that protein arginine methyltransferase 1 (PRMT1)-dependent arginine modification of FoxO1 interferes with Akt-dependent phosphorylation, both in cancer cells and in the Caenorhabditis elegans model, suggesting that this additional modification of FoxO1 might be critical in its transcriptional activity. In this study, we attempted to directly test the effect of arginine methylation of FoxO1 on hepatic glucose metabolism. The ectopic expression of PRMT1 enhanced messenger RNA levels of FoxO1 target genes in gluconeogenesis, resulting in increased glucose production from primary hepatocytes. Phosphorylation of FoxO1 at serine 253 was reduced with PRMT1 expression, without affecting the serine 473 phosphorylation of Akt. Conversely, knockdown of PRMT1 promoted an inhibition of FoxO1 activity and hepatic gluconeogenesis by enhancing the phosphorylation of FoxO1. In addition, genetic haploinsufficiency of Prmt1 reduced hepatic gluconeogenesis and blood-glucose levels in mouse models, underscoring the importance of this factor in hepatic glucose metabolism in vivo. Finally, we were able to observe an amelioration of the hyperglycemic phenotype of db/db mice with PRMT1 knockdown, showing a potential importance of this protein as a therapeutic target for the treatment of diabetes. Conclusion: Our data strongly suggest that the PRMT1-dependent regulation of FoxO1 is critical in hepatic glucose metabolism in vivo. (HEPATOLOGY 2012)
Pancreatic insulin is a major anabolic hormone to maintain glucose homeostasis. Activation of insulin signaling upon feeding conditions promotes glucose uptake in muscle and fat cells by the protein kinase B (Akt)-dependent translocation of the Glut4 glucose transporter to the plasma membrane.1-3 Furthermore, the Akt pathway is critical in reducing glucose production from the liver not only by the promotion of glycogen synthesis, but also by the inhibition of gluconeogenesis at the transcriptional level.4-7 Among the transcription factors that are critical in regulating gluconeogenic genes (e.g., phosphoenol pyruvate carboxykinase [PEPCK], or glucose-6-phosphatase catalytic subunit [G6Pase]), FoxO1 is most tightly linked with the insulin/Akt-dependent regulation of hepatic glucose metabolism.8, 9
FoxO1 belongs to the FoxO subfamily of evolutionally conserved Forkhead transcription factors.10, 11 As an important transcriptional regulator for energy metabolism in mammals, FoxO1 is shown to regulate this pathway in multiple insulin-sensitive tissues, such as skeletal muscle and adipose tissue.12-20 In the liver, FoxO1 is responsible for the activation of gluconeogenesis by the transcriptional activation of gluconeogenic genes, inhibition of the pyruvate dehydrogenase complex (by activation of pyruvate dehydrogenase kinase [PDK]4 expression), and the promotion of glycerol uptake (by activation of aquaporin 9 expression).9, 13, 21-23
When insulin or growth hormone is limiting, FoxO1 resides in the nucleus to activate the transcription of target genes by binding to its response element (i.e., insulin response element; IRE). The activation of insulin/insulin-like growth hormone receptor-mediated signaling cascades promotes the phosphorylation of FoxO1 at threonine 24, serine 253, and serine 316 in a phosphatidylinositol 3-kinase/Akt-dependent manner, which results in the cytoplasmic translocation and subsequent ubiquitin/proteasome-mediated degradation of this factor.8, 24-26
Recently, protein arginine methyltransferases (PRMTs) have been emerging as critical components for various cellular processes by modifying arginine residues on both histones and nonhistone substrates. In mammals, a total of nine PRMTs are responsible for the S-adenosyl-L-methionine (SAM)-dependent methylation of arginine residue by adding either one or two methyl groups. Based on their differences in generating dimethylated arginine residues, PRMTs can be classified as either the type 1 subfamily (generating asymmetric dimethyl-arginine; PRMT1, PRMT2, PRMT3, PRMT4/coactivator-associated arginine methyltransferase 1, PRMT6, and PRMT8) or the type 2 subfamily (generating symmetric dimethyl-arginine; PRMT5, PRMT7, and PRMT9/FBXO11), as reviewed elsewhere.27 Notably, as a predominant mammalian member of type 1 PRMTs, PRMT1 was shown to modify arginine 248 and arginine 250 of FoxO1, both in mammalian cell cultures and in nematodes, which promotes increased nuclear retention of this factor by blocking the insulin/Akt-mediated phosphorylation of adjacent serine 253.28, 29
In this study, we further wanted to elaborate the functional role of the PRMT1-mediated arginine methylation of FoxO1 in hepatic gluconeogenesis. We detected a strong interaction between PRMT1 and FoxO1 in primary hepatocytes. Overexpression of PRMT1 promoted an increase in glucose production in primary hepatocytes. Insulin-dependent phosphorylation of FoxO1 was blunted without directly affecting insulin signaling in this setting. On the other hand, short hairpin RNA (shRNA)-mediated depletion of hepatic PRMT1 promoted the inhibition of FoxO1-dependent gluconeogenesis by enhancing the phosphorylation of FoxO1 at serine 253, both in primary hepatocytes and in mice. Furthermore, we observed a similar glycemic phenotype in mouse models of genetic PRMT1 haploinsufficiency or transient depletion of hepatic PRMT1 in diabetic db/db mice, underscoring the critical role of the PRMT1-mediated modification of FoxO1 in the control of hepatic glucose metabolism in vivo. Taken together, our data provide, for the first time, direct evidence for the importance of the PRMT1-mediated modification of FoxO1 on hepatic glucose metabolism in mammals.
Ad, adenovirus; Akt, protein kinase B; bp, base pairs; β-Gal, β-galactosidase; ChIP, chromatin immunoprecipitation; DAF, diaminofluorescein; DEX, dexamethasone; FBS, fetal bovine serum; G6Pase, glucose-6-phosphatase catalytic subunit; GFP, green fluorescent protein; GTT, glucose tolerance test; HA, hemagglutinin; HSP90, heat shock protein 90; IGFBP1, insulin-like growth factor binding protein 1; IP, intraperitoneally; IRE, insulin response element; KO, knockout; NEFA, nonesterified fatty acid; mRNA, messenger RNA; NR, nuclear receptor; PCR, polymerase chain reaction; PDK, pyruvate dehydrogenase kinase; PEPCK, phosphoenol pyruvate carboxykinase; PRMT1, protein arginine methyltransferase 1; qPCR, quantitative polymerase chain reaction; RNAi, RNA interference;; SAM, S-adenosyl-L-methionine; SD, standard deviation; SEM, standard error of the mean; shRNA, short hairpin RNA; TGs, triglycerides; TAG, triacylglycerol; WT, wild type.
Materials and Methods
Culture of Primary Hepatocytes.
Primary hepatocytes were prepared from 8- to 10-week-old C57BL/6 mice by the collagenase perfusion method, as described previously.30
Transient Transfection Assays.
HEK293T cells were maintained with Dulbecco's modified Eagle's medium (HyClone, Logan, UT), supplemented with 10% fetal bovine serum (FBS), 10 units/mL of penicillin, and 10 μg/mL of streptomycin. Transfection was performed with 200 ng of luciferase construct, 50 ng of β-galactosidase (β-Gal) plasmid, and 10-50 ng of expression vector for PRMT1 or FoxO1 using TransIT-LT1 reagent (Mirus, Madison, WI). Cells were treated with or without insulin (10 nM) for 6 hours. Luciferase activity was normalized to β-Gal activity.
Adenoviruses expressing green fluorescent protein (GFP) only and nonspecific RNA interference (RNAi) control (US) were described previously.30 Adenoviruses expressing rat PRMT1 and PRMT1 RNAi (GCAACTCCATGTTTCACAATC, targeting murine PRMT1) were generated by homologous recombination between adenovirus backbone vector pAD-Easy and linearized transfer vector pAD-Track, as described previously.31 For animal experiments, viruses were purified on a CsCl gradient, dialyzed against phosphate-buffered saline buffer containing 10% glycerol, and stored at −80°C.
Western Blotting Analysis.
Western blotting analyses of whole-cell extracts were performed as previously described.32 PRMT1 antibody was purchased from Millipore (Billerica, MA). Antisera against AKT, phospho-Ser473 AKT, FOXO1, and phospho-Ser256 FOXO1 were purchased from Cell Signaling Technologies (Danvers, MA). Antisera against dimethylarginine, asymmetric (ASYM24), was purchased from Millipore. Antibodies against heat shock protein 90 (HSP90; Santa Cruz Biotechnology, Santa Cruz, CA), alpha-tubulin (Sigma-Aldrich, St. Louis, MO), and β-actin (Sigma-Aldrich) were used to assess equal loading.
Quantitative Polymerase Chain Reaction.
Total RNA from either primary hepatocytes or mouse liver was extracted using the RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany). Complementary DNAs generated by Superscript II enzyme (Gibco/Invitrogen, Grand Island, NY) were analyzed by quantitative polymerase chain reaction (qPCR) using a SYBR Green PCR Kit and TP800 Thermal Cycler Dice Real Time System (Takara Bio Inc., Otsu, Japan). All data were normalized to ribosomal L32 expression.
Nuclear isolation, cross-linking, and chromatin immunoprecipitation (ChIP) assays from mouse primary hepatocytes were performed as described previously.33 Mouse primary hepatocytes were infected with adenoviruses for hemagglutinin (HA)-FoxO1 and/or Flag-PRMT1 for 48 hours. Proteins from cell extracts were immunoprecipitated with anti-HA agarose (Sigma-Aldrich). Precipitated DNA fragments were analyzed by PCR using primer sets that encompassed the proximal (−484 to −12 base pairs; bp) region of the mouse PEPCK promoter and (−231 to +57 bp) the region of the mouse G6Pase promoter.
Glucose Output Assay.
Mouse primary hepatocytes were cultured in Medium 199 supplemented by 10% FBS, 10 units/mL of penicillin, 10 μg/mL of streptomycin, and 10 nM of dexamethasone (DEX). Cells were then infected with adenoviruses expressing either GFP or PRMT1 for 40 hours, and media were replaced with Krebs' ringer buffer, supplemented with 20 mM of sodium lactate and 2 mM of sodium pyruvate in the absence or in the presence of 100 nM of insulin, 100 nM of DEX, or 10 μM of forskolin for 8 hours. Glucose concentration was measured by using a QuantiChrom Glucose assay kit (Bioassay Systems, Hayward, CA), according to the manufacturer's instructions.
Male 8-week-old C57BL/6 mice or db/db mice were purchased from Charles River Laboratories International, Inc. (Wilmington, MA). Prmt1+/− mice were received from the EUCOMM consortium and were back-crossed with C57BL/6 mice five times before being used for the experiment. For animal experiments involving adenoviruses, mice were tail-vein–injected with recombinant adenovirus (0.5 × 109 plaque-forming units/mice). Fasting blood-glucose levels were measured from mice that were fasted for 16 or 6 hours with free access to water. For the glucose tolerance test (GTT), mice were fasted for 16 hours and injected intraperitoneally (IP) with glucose (2 g/kg body weight for C57BL/6 mice). For pyruvate challenge, mice were fasted for 16 hours and then injected IP with pyruvate (2 g/kg body weight for C57BL/6 mice and 1 g/kg body weight for db/db mice). Blood-glucose levels were measured from tail-vein blood using an automatic glucose monitor (OneTouch; LifeScan, Inc., Milpitas, CA). All procedures were performed in a specific pathogen-free facility at the Sungkyunkwan University School of Medicine (Suwon, Korea), based on the protocols that were approved by the Sungkyunkwan University School of Medicine Institutional Animal Care and Use Committee.
Measurement of Metabolites.
Blood triglycerides (TGs), liver TGs, and nonesterified fatty acid (NEFA) were measured by colorimetric assay kits (Wako, Osaka, Japan). Insulin was measured by a mouse insulin enzyme-linked immunosorbent assay kit (ALPCO Diagnostics, Windham, NH).
In Vivo Imaging.
Male 8-week-old mice were infected with adenovirus expressing wild-type (WT) G6Pase (−231/+57) luciferase. Five days postadenoviral injection, 18-hour-fasted mice were injected IP with 100 mg/kg of sterile firefly D-luciferin (Gold Biotechnology, St. Louis, MO). After 10 minutes, mice were anesthetized and imaged by using the IVIS Luminar XR Imaging System (Caliper Life Sciences, Hopkinton, MA).
Results are shown as mean ± standard deviation (SD) or ± standard error of the mean (SEM), as indicated in the figure legends. The comparison of different groups was carried out using the two-tailed unpaired Student t test, and differences at or under P < 0.05 were considered statistically significant and are reported as in the figure legends.
PRMT1 Promotes FoxO1-Dependent Glucose Production in Primary Hepatocytes.
Previously, PRMT1 was shown to enhance the FoxO1-dependent transcription of genes that are critical in relieving oxidative stress.29 PRMT1 catalyzes the asymmetric dimethylation of FoxO1 at arginines 248 and 250, thus interfering with Akt-dependent phosphorylation at adjacent serine 253, both in mammalian transformed cells in culture and in Caenorhabditis elegans.29 Because FoxO1 plays a critical role in the transcriptional regulation of hepatic gluconeogenesis in mammals, we wanted to test a potential role of PRMT1 in FoxO1-mediated glucose metabolism in the liver. Indeed, we were able to confirm the interaction between flag-tagged PRMT1 and HA-tagged FoxO1 in HEK293T cells by coimmunoprecipitation assay (Fig. 1A). Furthermore, we observed a strong association of endogenous PRMT1 and FoxO1 in primary hepatocytes (Fig. 1B), suggesting that PRMT1 may modify FoxO1 in the liver in a manner similar to the one in other cell types or in nematodes. Indeed, PRMT1 overexpression increased the nuclear retention of FoxO1 in hepatocytes (Supporting Fig. 1), and we were able to confirm the PRMT1-dependent modification of FoxO1 by the in vitro methylation assay (Supporting Fig. 2). Coexpression of PRMT1 significantly enhanced FoxO1-dependent IRE promoter activity and partially blocked insulin-mediated repression of the transcriptional activity of FoxO1 (Fig. 1C).
To further test the role of PRMT1-dependent regulation of FoxO1, we tested a recombinant adenovirus expressing rat PRMT1 in primary hepatocytes. Indeed, overexpression of PRMT1 reduced the insulin-mediated repression of FoxO1 target genes, such as G6Pase and insulin-like growth factor binding protein 1 (IGFBP1; Fig. 1D). Occupancy of FoxO1 on the gluconeogenic promoters was also enhanced with the coexpression of PRMT1 in primary hepatocytes (Fig. 1E). Furthermore, PRMT1 overexpression blunted the inhibitory effects of insulin on glucose production from hepatocytes (Fig. 1F). These data strongly suggest that PRMT1-dependent regulation of FoxO1 could enhance a potential for gluconeogenesis in the liver.
Knockdown of PRMT1 Induces an Increase in FoxO1 Phosphorylation and a Reduction in Hepatic Gluconeogenesis.
To further test the importance of PRMT1 in hepatic gluconeogenesis, we generated an adenovirus expressing shRNA for PRMT1 and tested the efficacy of this virus in primary hepatocytes. An 80% reduction of PRMT1 expression enhanced the insulin-dependent phosphorylation of FoxO1 at serine 253 and promoted proteasome-dependent degradation without affecting the phosphorylation of Akt, in agreement with previous results in the transformed cell line (Fig. 2A and Supporting Figs. 3 and 4). Similarly, chronic knockdown of PRMT1 reduced the arginine dimethylation of FoxO1 in cultured cells (Supporting Fig. 5, left). The level of lysine acetylation of FoxO1, however, was not affected by PRMT1 knockdown (Supporting Fig. 5, right). Next, we attempted to delineate the role of PRMT1 in vivo by tail-vein injection of either adenovirus (Ad)-PRMT1 shRNA (PRMT1i) or control Ad-shRNA adenovirus (US). Interestingly, knockdown of PRMT1 displayed reduced fasting blood-glucose levels (Fig. 2B), with a slight decrease in insulin levels (Supporting Fig. 6). No significant changes were observed in plasma triacylglycerol (TAG) or NEFA levels between control and PRMT1 knock-down mice (Supporting Fig. 6). Phosphorylation of FoxO1 was increased with PRMT1 knockdown without there being changes in the phosphorylation of Akt (Fig. 2C). Indeed, the insulin tolerance test revealed no significant changes in insulin sensitivity between control and PRMT1 knock-down mice (Supporting Fig. 7). Gluconeogenic gene expression was reduced with a depletion of PRMT1 in the liver, further accentuating the critical role of PRMT1 in regulating FoxO1-dependent gluconeogenesis (Fig. 2D). Improved glucose tolerance was also evident with PRMT1 knockdown (Fig. 2E). These data collectively suggest that PRMT1 is critical in regulating glucose homeostasis by modulating the insulin-dependent regulation of FoxO1 activity in the liver.
Chronic Haploinsufficiency of PRMT1 Lowers Blood-Glucose Levels in C57BL/6 Mice.
Having observed the effect of the acute knockdown of PRMT1 in vivo, we next wanted to assess the effect of chronic depletion of PRMT1 using mouse models. To this end, we obtained knock-out (KO) mice for the PRMT1 gene in the C57BL/6 background from the Sanger Institute (Hinxton, UK). Because PRMT1 homozygous KO mice are embryonically lethal, we utilized heterozygous null mice (Prmt1+/−) for our study. Haploinsufficiency of PRMT1 resulted in lower fasting glucose levels (Fig. 3A). Although not statistically significant, serum insulin levels, as well as plasma TG, liver TG, and NEFA levels, were slightly reduced in Prmt1+/− mice, compared to those in WT mice (Fig. 3B,C). Furthermore, hepatic gluconeogenesis might be impaired with the depletion of PRMT1, as evidenced by reduced glucose production upon injection of pyruvate (Fig. 3D) and reduced hepatic G6Pase promoter activity in the optical in vivo imaging experiment (Fig. 3E and Supporting Fig. 8) in Prmt1+/− mice, compared to WT mice.
Because Prmt1+/− mice might display the glucose phenotype because of the nature of global haploinsufficiency, we attempted to test the direct effect of chronic depletion of PRMT1 in the liver by preparing primary hepatocytes from either WT or Prmt1+/− mice. Indeed, insulin-mediated repression of gluconeogenic genes was more pronounced in Prmt1+/− hepatocytes than in the control (Fig. 4A-D). The phosphorylation level of FoxO1 was higher in Prmt1+/− hepatocytes than in Prmt1+/+ hepatocytes, without changes in the insulin-dependent phosphorylation of Akt (Fig. 4E). In addition, glucose production was significantly impaired in Prmt1+/− hepatocytes (Fig. 4F), suggesting that loss of PRMT1 expression in the liver could be directly responsible for the phenotypes shown in Prmt1+/− mice.
Hepatic Depletion of PRMT1 Ameliorates Hyperglycemic Phenotypes in Diabetic db/db Mice.
Type 2 diabetic individuals suffer hyperglycemia, in part, as a result of uncontrolled glucose production from the liver. To test whether depletion of PRMT1 in the liver may ameliorate hyperglycemia in this setting, we utilized db/db mice, a mouse model of type 2 diabetes. Indeed, knockdown of hepatic PRMT1 resulted in reduced blood-glucose levels, compared to the control (Fig. 5A), without significant changes in body weight or plasma insulin levels (Fig. 5A and Supporting Fig. 9). Depletion of hepatic PRMT1 significantly repressed the gluconeogenic potential of db/db mice, as shown by reduced glucose levels in response to pyruvate challenge (Fig. 5B) or reduction in the expression of gluconeogenic genes (Fig. 5C). Acute knockdown of PRMT1 in db/db mouse liver may partially revert other phenotypes associated with type 2 diabetes, because reduced expression of genes involved in lipogenesis were also observed between hepatic RNAs from control mice and PRMT1 knock-down mice (Supporting Fig. 10). We also observed increased phosphorylation of FoxO1 in livers of Ad PRMT1 shRNA-infected mice, compared to the control, whereas no drastic changes were observed in the phosphorylation levels of Akt between the two groups (Fig. 5D). These data collectively display the importance of PRMT1 in the control of FoxO1-dependent gluconeogenesis, and suggest that reduction in PRMT1 activity in the liver might be beneficial in the control of glycemia in diabetes.
Previously, PRMT1-dependent modification of FoxO1 of arginines 248 and 250 was shown to be critical in blocking the Akt-dependent phosphorylation of adjacent serine 253.29 In a study performed in HEK293 cells or HeLa cells, Fukamizu et al. provided evidence showing that this novel modification could serve as one of the critical codes to dictate the fate of FoxO1, which enables this factor to escape from the insulin/Akt-mediated phosphorylation/nuclear exclusion and enhances prolonged expression of target genes that induce cell death in response to oxidative stress. As in the case of histone H3 modification, methylation and phosphorylation could serve as mutually exclusive modifications to determine the function of this nonhistone nuclear protein. They further suggested that PRMT1-mediated regulation of FoxO1 could be also critical in other areas of signaling cascades known to be transcriptionally controlled by the FoxO1 and Akt axis.
Recently, this notion was further supported by the same group in nematode C. elegans, where diaminofluorescein (DAF)-2 (an ortholog of Akt)-dependent regulation of DAF-16 (an ortholog of forkhead proteins, including FoxO1) constitutes a critical component for the response of this organism to environmental cues in mediating stress resistance or dictating longevity.28 As in the case of its mammalian counterpart, nematode PRMT1 catalyzes the asymmetric dimethylation of arginine residues of DAF-16 and blocks DAF-2-dependent phosphorylation. The physiological importance of this modification on DAF-2 was verified in genetic experiments showing that loss of the prmt-1 gene reduces the lifespan of this animal, which was restored by the reintroduction of functionally active prmt-1. Not all DAF-16 activity was affected by loss of PRMT1 in nematodes, however, particularly in neurons where PRMT1 is hardly expressed. Thus, other known post-translational modifiers could be sufficient to preserve the intact activity of DAF-16 in that regard.
Because FoxO1 is one of the major regulators for fasting-induced hepatic gluconeogenesis, we wanted to investigate the physiological relevance of PRMT1-dependent regulation of FoxO1 on hepatic glucose metabolism in this study. Indeed, PRMT1 was shown to be critical in blocking the inhibitory effect of insulin on FoxO1-dependent glucose production and gluconeogenic gene expression in the liver. Because depletion of PRMT1 did not promote changes in Akt phosphorylation, PRMT1-dependent modification may not directly affect insulin/Akt-dependent signaling per se as described in cultured cells or nematodes. Furthermore, we were able to show that knockdown of PRMT1 reversed the hyperglycemic phenotype in diabetic db/db mice (Fig. 5). These data suggest that, together with decreased insulin signaling, intact PRMT1-dependent modification of key arginine residues might contribute to the resulting increase in FoxO1-dependent hepatic glucose production in this setting. Interestingly, PRMT1 was also shown to promote arginine methylation of other members of FoxO transcription factors, including FoxO6, an isoform that was recently linked to hepatic glucose metabolism (Supporting Fig. 11).29 Further study is necessary to fully understand the effect of PRMT1-dependent modification associated with insulin resistance.
Classes of drugs, such as biguanides, meglitinides, thiazolidinediones, sulfonylureas, alpha-glucosidase inhibitors, and DPP-4 inhibitors, are currently utilized to maintain constant blood-glucose levels in the management of type 2 diabetes, with a varying degree of efficacies associated with certain side effects. Our data indicate that hepatic PRMT1 serves as a key regulator for FoxO1-dependent hepatic glucose production not only in the physiological state, but also in pathological states, such as insulin resistance or type 2 diabetes. Thus, it is intriguing to speculate whether the reagents that reduce PRMT1 activity might be alternative means to sustain euglycemia during the treatment of diabetes. Further studies will ensure the potential benefits to control hepatic PRMT1 activity in the regulation of blood-glucose levels in type 2 diabetes patients.
The authors thank Sun-Myong Park (Sungkyunkwan University) for technical assistance.