|Home | About | Journals | Submit | Contact Us | Français|
In patients with poorly controlled type 2 diabetes mellitus (T2DM), hepatic insulin resistance and increased gluconeogenesis contributes to fasting and postprandial hyperglycemia. Since CREB is a key regulator of gluconeogenic gene expression, we hypothesized that decreasing hepatic CREB expression would reduce fasting hyperglycemia in rodent models of T2DM. In order to test this hypothesis, we used a CREB-specific antisense oligonucleotide (ASO) to knock down CREB expression in liver. CREB ASO treatment dramatically reduced fasting plasma glucose concentrations in ZDF rats, ob/ob mice and a STZ-treated high-fat fed rat model of T2DM. Surprisingly, CREB ASO treatment also decreased plasma cholesterol and triglyceride concentrations, as well as hepatic triglyceride content due to decreases in hepatic lipogenesis. These results suggest that CREB is an attractive therapeutic target for correcting both hepatic insulin resistance and dyslipidemia associated with NAFLD and T2DM by down regulation of both lipogenic and gluconeogenic gene expression.
Nonalcoholic fatty liver disease (NAFLD) is strongly associated with hepatic insulin resistance in patients with poorly-controlled type 2 diabetes mellitus (T2DM) (Yki-Jarvinen, 2005). In contrast to our understanding of the mechanism of fat-induced hepatic insulin resistance, the pathogenesis of increased hepatic gluconeogenesis in T2DM is less well-understood. Progressive declines in insulin secretion as well as inappropriate increases glucagon secretion have both been viewed as critical factors responsible for increased rates of hepatic gluconeogenesis (Cherrington et al., 1987; Reaven et al., 1987; Unger and Orci, 1977). Glucagon signals through a canonical G-protein coupled receptor leading to the activation of c-AMP dependent protein kinase A (PKA) (Jelinek et al., 1993). Consequently PKA phosphorylates Ser133 on cAMP response element-binding protein (CREB) and causes subsequent translocation to the nucleus (Gonzalez and Montminy, 1989). CREB is a leucine BH/zipper transcription factor that promotes gene transcription by binding to conserved sequences known as a cAMP responsive element (CRE) (Mayr and Montminy, 2001). CREB is a well known activator of gluconeogenic gene transcription through CRE elements located on key G6Pase, FBPase, and PEPCK. Insulin antagonizes the induction of gluconeogenic enzymes by phosphorylating CREB-binding protein (CBP) (He et al., 2009; Zhou et al., 2004) and transducer of regulated CREB activity 2 (TORC2) (Koo et al., 2005). Thus, CREB is a key nexus where insulin and glucagon signaling converge on the regulation of hepatic gluconeogenesis.
In the present study, we examined the hypothesis that knockdown of CREB would blunt the transcriptional actions of glucagon signaling and improve hepatic glucose metabolism in diabetic rodent models associated with NAFLD and hepatic insulin resistance. In order to assess this hypothesis, we knocked down the expression of CREB in liver and adipose tissue of four different insulin resistant rodent models using an antisense oligonucleotide (ASO) (Samuel et al., 2009). The major advantage of this approach is that it is possible to decrease specific gene expression in a normal adult animal and avoid the confounding compensatory developmental affects often associated with gene knockout mouse models.
To achieve a model of mild diabetes we used an initial streptozotocin injection co-administered with a low dose of nicotinamide. The rats were fed a high-fat diet to induce insulin resistance. This rat model mimics the pathophysiology of hyperglycemia seen in patients with T2DM (Samuel et al., 2009). Fasting (T2DM: 129±4 mg/dl vs. normal: 111±3 mg/dl, P<0.05) and peak intravenous glucose tolerance test (T2DM: 547±24 mg/dl vs. normal: 321±75 mg/dl, P<0.01) plasma glucose concentrations were increased in the T2DM rat model compared to normal rats fed a standard rodent chow. Despite their marked hyperglycemia following the intravenous glucose tolerance test, there was no compensatory hyperinsulinemia as compared to the normal rats (T2DM: 2626±833 µU-min/mL vs. normal: 2280±372 µU-min/mL, P=NS), which is similar to what is observed in patients with T2DM (Samuel et al., 2009).
CREB ASO reduced both CREB mRNA and protein levels in liver (Supplemental Figure 1A and 1B), and WAT but not muscle (Supplemental Table 2). In normal rats fed a standard rodent chow, CREB ASO treatment had no effect on fasting plasma glucose concentrations (Figure 1A), but increased whole body insulin sensitivity as reflected by a 65% reduction in fasting plasma insulin concentrations (Figure 1B). In T2DM rats, CREB ASO treatment decreased fasting plasma glucose concentrations compared to the Control ASO (Figure 1A). Consistent with our findings in normal rats, CREB ASO treatment also improved whole body insulin sensitivity in this T2DM rat model as reflected by a 66% reduction in fasting plasma insulin concentrations (Figure 1B). Fasting plasma glucagon concentrations were similar in both ASO-treated groups (Supplemental Table 1). Among the gluconeogenic enzymes and their key co-factors—cytosolic phosphoenolpyruvate carboxykinase (PEPCK), mitochondrial PEPCK, and the transcriptional co-activator peroxisomal proliferator activated receptor gamma coactivator-1α (PGC-1α)—mRNA were decreased by 43%, 55%, 54%, respectively, in the liver of the CREB ASO group (Supplemental Table 2). Interestingly, there was no observed difference in catalytic glucose-6-phosphatase unit (G6Pase) and hepatocyte nuclear factor-4α (HNF-4α) even though the DNA promoter sequence encoding these proteins contains known CRE element boxes (Supplemental Table 2). These data suggest that other transcriptional regulatory factors independently regulate G6Pase expression.
To verify the results found with CREB ASO in the T2DM rat model, we used three other models of diabetes. CREB ASO reduced fasting glucose concentrations by 178 mg/dL in Zucker Diabetic Fatty (ZDF) rats (Figure 1C) with significant reductions of CREB in liver but not WAT (Supplemental Figure 1C and 1D). Plasma insulin values tended to be increased due to pancreatic sparing from lower plasma glucose values (Figure 1D) (Harmon et al., 2001). To complement the studies in our T2DM rat model, we substituted the safflower oil with a lard-based diet that is high in hydrogenated fatty acids. Consistent with the other insulin resistant rodent models, CREB ASO reversed insulin resistance in these lard-fed rats as reflected by decreased fasting plasma glucose (Figure 1E) and insulin concentrations (Figure 1F). In addition to the above mentioned rat models, CREB ASO also dramatically reduced plasma glucose concentrations by 144mg/dL in ob/ob mice (Figure 1G), which was associated with ~60% reductions in G6Pase, FBPase, cPEPCK, and pyruvate carboxylase mRNA expression (Figure 1H).
Intrahepatic long chain CoAs and diacylglycerols were both lowered with CREB ASO treatment (Figure 2A and 2B). CREB ASO reduced liver triglyceride content by 57% in the fasted state. After a meal enriched with fat, liver triglycerides increased by 76% in the control ASO-treated rats compared to the fasted state, but did not change in the CREB ASO group (Figure 2C). To test if in vivo lipogenesis was decreased, we gavaged control and CREB ASO T2DM rats with whipping cream enriched with [1-14C] palmitate to examine the rate of incorporation of labeled fatty acids into triglycerides. There was a 73% reduction in the radioactive counts found in hepatic triglycerides with CREB ASO, reflecting decreased hepatic lipogenesis, specifically resterification (Figure 2D). Both C/EBPα and C/EBPβ have been implicated in the regulation of lipogenesis (Matsusue et al., 2004; Pedersen et al., 2007; Qiao et al., 2006; Rahman et al., 2007; Schroeder-Gloeckler et al., 2007) and have potential CRE binding sites when we analyzed the promoter regions of these genes (Niehof et al., 1997). We examined hepatic expression of both of these proteins and found them to be dramatically reduced with CREB ASO (Figure 2E and 2F). Consistent with these observations, DGAT2 and mtGPAT mRNA expression were also decreased, which could explain the decreased hepatic lipogenesis (Figure 2G).
We assessed the effects of decreased liver and adipose CREB expression on hepatic and peripheral insulin sensitivity using hyperinsulinemic-euglycemic clamps (Samuel et al., 2007). Basal hepatic glucose production (Supplemental Figure 1E) and insulin-stimulated peripheral glucose uptake rates (Supplemental Figure 1F) were similar in both rat models and ASO groups. In contrast, hepatic insulin responsiveness was markedly increased in the CREB ASO treated T2DM rat model as reflected by a 45% increase in suppression of hepatic glucose production during the clamp compared to the control ASO treated rats (Figure 3A). Additionally, CREB ASO increased hepatic responsiveness in the lard-fed rats (Supplemental Figure 1G). We have previously shown that NAFLD is associated with accumulation of intracellular diacylglcerol (DAG) content, leading to increased activation of PKCε and subsequent reduction in insulin signaling at the level of IRS-2 tyrosine phosphorylation (Samuel et al., 2004; Samuel et al., 2007). Consistent with these studies, we found that CREB ASO treatment reduced hepatic DAG content, which was associated with lower PKCε membrane translocation in liver (Figure 3B). These changes were associated with improved insulin signaling as reflected by increased insulin stimulated IRS-2 tyrosine phosphorylation (Figure 3C) and Akt2 activity (Figure 3D). Our results are in contrast to previous studies that used a dominant negative polypeptide A-CREB or Ad-CREB RNAi to knock down the expression of CREB in primary rat hepatocytes (Canettieri et al., 2005). While these studies found that it resulted in reduced expression of insulin receptor substrate 2 (IRS-2) (Canettieri et al., 2005), we found no differences in IRS-2 expression (Supplemental Table 2) in the CREB ASO-treated rats compared to the control ASO-treated rats. Regulation of IRS-2 transcription is complicated and dependent on the fed and fasted state, and multiple transcription factors.
CREB ASO treatment decreased plasma triglycerides by 24% in the T2DM rats (Figure 4A). Total plasma cholesterol and HDL cholesterol were also decreased in rats treated with CREB ASO (Supplemental Table 1). FPLC analysis revealed that the decreases in total plasma cholesterol could be attributed to significant decreases in VLDL, LDL, and HDL cholesterol (Figure 4B). Hepatic free cholesterol was unaltered (Figure 4C), but CREB ASO decreased hepatic cholesterol esters as assessed by NMR (Figure 2D). Genes relating to cholesterol synthesis, sterol regulatory element binding protein 2 (SREBP-2), and hydroxy-methyl-glutaryl co-enzyme A (HMG-CoA) were both increased by 50–60% with CREB ASO treatment (Supplemental Table 2). We measured the rate of sterol synthesis in primary rat hepatocytes and found no effect of CREB ASO treatment on the incorporation rate of 14C acetic acid into cholesterol (Figure 4E). Additionally, expression of the low density lipoprotein receptor (LDLR) and scavenger receptor class B type 1 (SR-B1) were unaltered with CREB ASO treatment, suggesting that alterations in cholesterol uptake were not the cause of reduced plasma cholesterol concentrations (Supplemental Table 2). Lastly, we looked at the rate of bile acid synthesis by quantifying the expression of cholesterol 7 alpha-hydroxylase (Cyp7A1). CREB ASO treatment resulted in an approximately two fold increase in Cyp7A1 expression (Supplemental Table 2). Consistent with the increased expression of Cyp7A1, fecal bile content was approximately three fold higher compared to Control ASO treated rats (Figure 4F). This suggests that increased cholesterol export into fecal bile is the most likely explanation for the reduced plasma cholesterol concentrations in the CREB ASO treated rats.
These studies found that decreasing hepatic expression of CREB dramatically improved fasting plasma glucose and insulin concentrations in three different insulin resistant rodent models. In addition, CREB inhibition improved hepatic insulin sensitivity and reduced plasma glucose concentrations in a T2DM rat model associated with fatty liver and hepatic insulin resistance. We found that decreased CREB expression improves hepatic insulin responsiveness by two mechanisms. First, decreased expression of CREB led to reduced intrahepatic DAG content and decreased PKCε activation (Samuel et al., 2007). Previous studies have demonstrated that PKCε binds to the insulin receptor leading to reduced insulin receptor-β kinase activity. Moreover, decreasing expression of PKCε, using a similar ASO approach, protects rats from fat-induced hepatic insulin resistance (Samuel et al., 2007). In addition, knockdown of CREB resulted in reduction of PEPCK and transcriptional gene cofactors involved in the regulation of hepatic gluconeogenesis. It is likely that a combination of these two mechanisms contributed to the decreased glucose production observed in the CREB ASO-treated rats during the hyperinsulinemic-euglycemic clamp.
Surprisingly CREB ASO treatment also decreased hepatic triglyceride content in the T2DM rat model, which we found could mostly be ascribed to decreased hepatic lipogenesis. These findings are in contrast to a previous study, which showed that dominant-negative indirectly increased expression of PPARγ and increased hepatic triglyceride content independent of SREBP by decreasing expression of hairy and enhancer of split 1 (HES-1), a transcriptional repressor of peroxisome proliferator-activated receptor gamma (PPARγ) (Herzig et al., 2003). In our model, we observed no increase in the level of PPARγ expression possibly reflecting differences in the models used or the method of knockdown. We also saw reductions in lipid synthesis as assessed by incorporation of [14C] palmitate into hepatic triglycerides independent of differences in ChREBP, or SREBP1c. Here, we observed decreased C/EBPα and C/EBPβ in the fasted and fed states with CREB ASO. C/EBPβ has been thought to reduce lipogenesis through PPARγ2 (Schroeder-Gloeckler et al., 2007); however, in our study we saw only a trend in decreased PPARγ2 expression. Previous reports have indicated that decreases in C/EBPα are also associated with reductions in lipogenesis through potentially direct binding to key lipogenic genes, and through upstream lipogenic transcription factors (Matsusue et al., 2004; Qiao et al., 2006). It is likely that both of these proteins contributed to the decreased lipogenesis observed with CREB ASO. Immunoneutralization of glucagon and the glucagon receptor (GCGR) ASO have both shown to decrease triglycerides in certain rodent models (Sloop et al., 2004; Sorensen et al., 2006). The reductions of lipids in these models could potentially be mediated via reductions in CREB signaling since glucagon is upstream of CREB. A recent study showed that CREB increased expression of adiponectin through ATF3, resulting in decreased hepatic lipids (Qi et al., 2009). However, in our study we found no differences in plasma adiponectin concentrations. Taken together these data suggest that the reduction in hepatic lipogenesis and improvement in hepatic insulin responsiveness that we observed in the CREB ASO treated animals is most likely due to decreased hepatic expression of CREB, and not due to alterations in adiponectin secretion secondary to reductions in CREB expression in adipocytes. This hypothesis is further supported by our studies in the ZDF T2DM rat model where we found no reductions in adipose tissue CREB expression but dramatic reductions in fasting plasma glucose concentrations.
In summary, this study demonstrates that reduction of CREB expression in liver and fat lowers plasma triglyceride and cholesterol concentrations in a T2DM rat model and protects against fat-induced hepatic steatosis and hepatic insulin resistance by reductions in PEPCK expression and hepatic diacylgycerols. Consistent with these results, CREB ASO proved to lower fasting plasma glucose concentrations in four different rodent models of T2DM and hepatic insulin resistance. This paper is in contrast to previous literature arguing that hepatic CREB would be a poor target for pharmacological inhibition (Canettieri et al., 2005; Herzig et al., 2003). Since T2DM is often associated with hypertriglyceridemia, hypercholesterolemia associated with NAFLD and hepatic insulin resistance, CREB is a potentially attractive therapeutic target for T2DM.
All procedures were approved by the Institutional Animal Care and Use Committee of Yale University School of Medicine. 150g Male Sprague-Dawley rats, ZDF rats, or ob/ob mice were received from Charles River Laboratories and given three days to acclimate. For the T2DM rat model, rats were then given a 175 mg/kg dose of nicotinamide by intraperitoneal injection, after 15 minutes rats were then dosed with a 65 mg/kg dose of streptozotocin (Masiello et al., 1998). Rats were given 4 days of a recovery period and randomized for blood glucose values prior to first ASO injection. Rats with overt diabetes (defined by a glucose value greater than 200mg/dL) were excluded from the study. Rats were housed individually on a 12:12-hour light/dark cycle and received food and water ab libitum. Body weight and food consumption were monitored weekly. The food consumed either consisted of regular rodent chow (60% carbohydrate, 10% fat, 30% protein calories), a high fat diet (26% carbohydrate, 59% fat, 15% protein calories), in which the major constitute is safflower oil, a high fat lard diet (24% carbohydrate, 55% fat, 21% protein), in which the major constitute is hydrogenated vegetable shortening, or in the case of the ZDF rats, the Purina Lab Diet 5008.
1.8 mL of heavy whipping cream, 0.6 g of sucrose (Sigma Aldrich), 0.3 mL of normal saline, and 30µCi of 1-14C potassium palmitate dissolved in 0.3 mL of EtOH (PerkinElmer) were mixed together. The solution was gavaged into fasted ~380g rats treated with either control ASO or CREB ASO. Plasma blood samples were taken at 0, 60, 180, and 360 minutes for free fatty acid enrichment. Rats were anesthetized with pentobarbital sodium injection (150 mg/kg), and all tissues were extracted and frozen immediately with the use of liquid N2-cooled aluminum tongs. 100 mg of ground liver were used for triglyceride enrichment and extracted using the method from Bligh and Dyer (Bligh and Dyer, 1959). The organic layer was evaporated to dryness and redissolved in 0.5 mL of chloroform. 25 µL of the lipid extract solution was plated on a Silica Gel 60 plate and ran with a 80:20:1 solution of hexane:diethylether:acetic acid. The plate was dried and developed with 0.005% primuline in 80% acetone in water. The triglyceride spots were scraped off the plate and eluted with diethylether in a glass pasteur pipette. The diethylether was evaporated to dryness, and the remaining residue was dissolved in 0.5mL hexanes. 50µL of the hexane mixture was measured using a radioactive scintillation counter to determine total counts.
Hyperinsulinemic-euglycemic clamps were performed as previously described (Erion et al., 2009).
Plasma insulin, glucagon, leptin, and aidiponectin concentrations were determined using RIA kits (Linco). For FPLC analysis, samples from control ASO, and CREB ASO were pooled together and injected on to an Amershan Akta FPLC (Amershan Pharmacia Biotech) and eluted at a constant flow rate of 0.5 mL/min FPLC buffer (0.15M NaCl, 0.01M Na2HPO4, 0.1mM EDTA, pH 7.5).
A separate group of rats that were treated exactly the same as the previous were used to the affect of CREB treatment on insulin signaling. These rats underwent a 20 minute hyperinsulnemic-euglycemic clamp, and immediately following the 20 minute period the liver was harvested and subsequently frozen with the use of N2-cooled aluminum tongs. The activity of AKT was assessed by measuring the incorporation of 32P onto a synthetic AKT substrate as previously described (Choi et al., 2007). PKCε and IRS-2 blots were performed as previously described (Choi et al., 2007). Blots were quantified as amount of PKCε relative to Na-K ATPase located in the membrane fraction (arbitrary units).
PCR was performed as previously described (Erion et al., 2009). For western blot analysis 20mg of powdered tissue in 200µL of homogenization buffer LDS Laemmli sample buffer. After centrifugation for 15 min at 10,000g, 40µg of crude protein was then separated on a 4–12% gradient polyacrylamide gel in MOPS buffer system (Invitrogen). Subsequently, membranes were transferred to nitrocellulose membranes, and membranes were incubated in blocking buffer (5% milk) for 1 hr and immunoblotted with anti-CREB antibody (Cell Signaling) overnight.
All values are expressed as the average S.E.M. The significance between the mean values for each study was evaluated by two-tailed unpaired Student’s t test or one-way ANOVA followed by post hoc analysis using the Tukey’s Honestly Significant Differences (HSD) test.
We thank Jianying Dong, Yanna Kosover, and Toddy May for expert technical assistance and advice with the studies. We also thank Aida Groszmann for performing the hormone assays. This work was supported by grants from the US Public Health Service (R01 DK-40936 and P30 DK-45735 to G.I. Shulman) and a Distinguished Clinical Scientist Award from the American Diabetes Association. G.I. Shulman is an investigator of the Howard Hughes Medical Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest: Xing Xian Yu, Susan F. Murray, Sanjay Bhanot, and Brett P. Monia own stock and/or hold stock options in Isis Pharmaceuticals.