|Home | About | Journals | Submit | Contact Us | Français|
The liver contributes to glucose homeostasis by promoting either storage or production of glucose depending on the physiological state. The cAMP response element binding protein (CREB) is a principal regulator of genes involved in coordinating the hepatic response to fasting, but its mechanism of gene activation remains controversial. We derived CRTC2-(CREB-regulated transcription coactivator 2; previously TORC2) deficient mice to assess the contribution of this cofactor to hepatic glucose metabolism in vivo. CRTC2 mutant hepatocytes showed reduced glucose production in response to glucagon, which correlated with decreased CREB binding to several gluconeogenic genes. However, despite attenuated expression of CREB target genes including PEPCK, G6Pase, and PGC1α, no hypoglycemia was observed in mutant mice. Collectively, these results provide genetic evidence supporting a role for CRTC2 in the transcriptional response to fasting, but indicate only a limited contribution of this cofactor to the maintenance of glucose homeostasis.
Maintenance of blood glucose levels within a relatively tight range is of critical importance, as evidenced by the complications arising from diabetes or hyperinsulinism. The liver, a key target of both insulin and glucagon signaling, contributes to the control of glucose metabolism by facilitating glucose uptake and output depending on the physiologic condition (Pilkis and Granner, 1992). Insulin is released from pancreatic β-cells when nutrient availability is high and instructs peripheral tissues, such as liver and muscle, to remove glucose from the bloodstream and store it for later use, effectively preventing prolonged hyperglycemia. In times of food deprivation, however, the liver is central in the metabolic response to glucagon, epinephrine, and glucocorticoids, which act to stimulate hepatic glucose production via glycogenolysis and gluconeogenesis in order to prevent extended periods of hypoglycemia.
The gluconeogenic program is largely regulated at the level of transcription, a process thought to be coordinated by the cAMP response element binding protein (CREB) and its ability to activate PPARγ coactivator-1α (PGC1α) (Herzig et al., 2001). Two mechanisms have been proposed to explain how CREB responds to glucagon and/or epinephrine stimulation in order to elicit the appropriate changes in hepatic gene expression. Originally, phosphorylation of CREB at serine 133 by the cAMP-responsive protein kinase A (PKA), and the subsequent recruitment of CREB binding protein (CBP) or its paralog p300, was suggested to be the pivotal regulatory event (Chrivia et al., 1993; Gonzalez and Montminy, 1989; Kwok et al., 1994; Mayr and Montminy, 2001). CBP and p300 are cofactors that promote gene transcription through their intrinsic histone deacetylase activity and through direct interactions with the basal transcription machinery (Bannister and Kouzarides, 1996; Goodman and Smolik, 2000; Shiama, 1997). More recently, the simplicity of this model has been challenged with the discovery of the CREB-regulated transcription coactivator (CRTC) family of cofactors (Conkright et al., 2003; Iourgenko et al., 2003). CRTC2, also called transducer of regulated CREB activity 2 (TORC2), is the most abundant CRTC protein in the liver, and has been implicated as the foremost mediator driving CREB target gene expression (Conkright et al., 2003; Koo et al., 2005). Upon glucagon stimulation, cytoplasmic CRTC2 is dephosphorylated and translocates to the nucleus where it engages CREB in a phosphorylation-independent manner to direct the gluconeogenic program (Bittinger et al., 2004; Screaton et al., 2004). In attempts to unify the two models, it has been suggested that CRTC2 is required in addition to CREB phosphorylation for the proper recruitment of CBP and p300 to regulatory complexes at CREB target sequences (Ravnskjaer et al., 2007; Xu et al., 2007). However, it has also been proposed that the transcriptional response to fasting is only transiently dependent on CRTC2 with the demonstration that it is deacetylated and degraded after long-term food deprivation, at which point FOXO1 takes over as the predominant transcriptional regulator (Liu et al., 2008).
In order to begin to address this complex issue using genetic means, we assessed the hepatic response to fasting in mice harboring a null mutation in the CRTC2 gene. Surprisingly, CRTC2 null mice did not display fasting hypoglycemia despite impairments in the transcription of several metabolic genes. These transcriptional defects were only observed following prolonged fasting indicating that CRTC2 continues to regulate gene expression beyond the short term. We also show that CREB association with target sequences, although unchanged during the initial stages of fasting, is reduced in the livers of 24-hour-fasted CRTC2 null mice, possibly accounting for the observed transcriptional deficiencies. Collectively, our results indicate a supporting role for CRTC2 in the regulation of CREB target genes in response to fasting, but challenge its role as the key mediator of the gluconeogenic program.
To assess CRTC2 function in vivo, we derived a CRTC2 null allele via homologous recombination in mouse embryonic stem cells. The targeting strategy used to construct this mutant allele is shown in Figure 1A. loxP sites were positioned upstream of exon 3 and downstream of exon 11 to force deletion of the sequence spanning exon 4 to exon 11 of the CRTC2 gene upon exposure to Cre recombinase. Crosses to the ubiquitous EIIa-Cre line resulted in recombination between loxP sites in the germline, yielding a global null allele. Western blotting confirmed the effective deletion of the full-length CRTC2 gene product (Figure 1B). CRTC2-/- animals were born at the expected Mendelian frequency, gained weight appropriately with age, and were generally indistinguishable from littermate controls (data not shown). The livers of adult CRTC2 null mice also appeared histologically normal (data not shown). In addition, there were no appreciable differences in most plasma metabolic parameters of fasted CRTC2 mutant mice, compared to controls, except for a small but significant decrease in corticosterone levels (Figure 2E).
CRTC2 has been described as a critical component of the early transcriptional fasting response in the liver (Koo et al., 2005; Liu et al., 2008). When circulating glucose and insulin levels are low, glucagon and/or epinephrine signaling in hepatocytes results in the dephosphorylation of CRTC2 and its translocation to the nucleus where it drives the transcription of several genes important in gluconeogenesis. Following long-term stimulation, however, it has been proposed that CRTC2 is deacetylated and targeted for degradation in the proteosome (Liu et al., 2008). CRTC2 activity is further controlled by insulin signaling which results in CRTC2 phosphorylation and exclusion from the nucleus where it can no longer impact gene expression (Dentin et al., 2007). Thus, in the absence of CRTC2, hypoglycemia due to impaired hepatic glucose production would occur during the initial stages of fasting. To test this hypothesis, blood glucose measurements were taken from CRTC2-/- mice along with littermate controls at regular intervals during a 24-hour-long fasting time course. As expected, no difference in blood glucose levels were observed in mice fed ad libitum (Figure 2A and 2B). Strikingly, we also found no differences between groups at any time point measured during the 24-hour fast (Figure 2A and 2B). Furthermore, glucose production following injection of a pyruvate bolus was not affected in mutant mice (Figure 2C). Finally, hepatic glucose production as determined by radioisotopic tracer techniques before or during hyperinsulinemic-hypoglycemic clamps was not statistically different between CRTC2 mutants and littermate controls (Figure 2D). Collectively, these results indicate that CRTC2 is not required to maintain the appropriate physiological response to fasting.
Next, we examined the hepatic mRNA expression levels of G6Pase, PEPCK, and PGC1α, three cAMP-inducible genes involved in the response to fasting (Yoon et al., 2001). As expected, gene expression levels were indistinguishable in animals that were re-fed following an overnight fast or allowed to feed ad libitum (Figure 3A). Despite the apparently normal glucose homeostasis of CRTC2 null animals, the expression of G6Pase, PEPCK, and PGC1α was moderately but significantly reduced in CRTC2 mutants, but only after a prolonged, 24-hour fast (Figure 3A). In light of the recent report implicating a role for CRTC2 primarily in the early phases of fasting (Liu et al., 2008), it is worth noting that the timing of these transcriptional deficiencies suggest the continued contribution of CRTC2 to the control of gene expression in the long term. In accordance with this finding, total CRTC2 protein remained abundant in the liver even following the 24-hour fast when we expected it to be absent due to degradation (Figure 1B). Protein levels of FOXO1, CREB, and serine-133-phosphorylated CREB, STAT3, HNF4α, C/EBPα, and C/EBPβ were similar in the livers of control and mutant mice indicating that these factors are not likely to compensate for the lack of CRTC2 in the short term (Figure 1B). Furthermore, CRTC1 message levels were unaltered by the loss of CRTC2 activity (Figure 1C). These experiments demonstrate that despite its involvement in the regulation of G6Pase, PEPCK, and PGC1α, CRTC2 is not required for an adequate metabolic response to hypoglycemia.
Many mechanisms, in multiple tissues, are in place to help respond to metabolic challenges. During a fast, blood glucose levels are maintained through the coordinated efforts of glycogenolysis and gluconeogensis, and ultimately the shift from glucose utilization to fatty acids and ketones as substrates for energy production (Wang et al., 2006). Despite having compromised expression of several genes involved in the hepatic response to fasting, no overt glycemic phenotype was observed in CRTC2-/- mice. Indeed, similar outcomes have been observed upon deletion of other metabolic genes, including PEPCK and PGC1α, which also fail to elicit dramatic metabolic phenotypes (Leone et al., 2005; She et al., 2000). Based on our expression analysis indicating that several key genes involved in gluconeogenesis are impacted by CRTC2 deletion, we suspected that hepatocytes in mutant animals might have functional impairments that are masked by compensation in other tissues when we analyzed glucose homeostasis in vivo. To address this point, hepatocytes from control and mutant mice were isolated and assayed in culture. As expected, the induction of G6Pase, PEPCK, and PGC1α mRNA levels in response to glucagon stimulation, as a surrogate for fasting, was blunted in CRTC2 mutant hepatocytes (Figure 3B). Notably, the induced expression of these genes was impaired to a greater extent in culture than we had observed in the liver (Figure 2B), again implicating additional mechanisms, such as activation by glucocorticoids, that may be at play in the whole animal to maintain sufficient hepatic expression of these relevant transcripts. Furthermore, CRTC2 null hepatocytes failed to secrete as much glucose as controls following glucagon stimulation (Figure 3C), indicating a functional impairment that is presumably overcome by additional processes unaffected by CRTC2 deletion in vivo. Indeed, we observed that liver and muscle glycogen levels were significantly reduced in fed CRTC2 mutants (Figure 3D and 3E). This observation could be indicative of an unrecognized role for CRTC2 in regulating glycogen metabolism, but may also represent a mechanism by which hepatocytes and other tissues are able to offset any insufficiencies provoked by the absence of CRTC2.
It has been suggested that CRTC2 imparts regulation on target genes by specifically interacting with CREB and facilitating recruitment of the co-activators CBP and p300 to CREB target sequences (Liu et al., 2008; Ravnskjaer et al., 2007; Xu et al., 2007). We therefore expected that the association of CBP or p300 with PEPCK, PGC1α, and G6Pase control regions would be attenuated in the livers of CRTC2 null mice, thus resulting in their decreased expression. To test this hypothesis, we performed chromatin immunoprecipitation (ChIP) assays on chromatin isolated from the livers of fasted CRTC2-/- and control mice. Strikingly, we found no significant differences in the ability of CBP or p300 to interact with regulatory complexes at any of the promoters tested (Figure 4A and 4B). Furthermore, despite having previously being described as direct targets for CREB activation (Herzig et al., 2001; Schmoll et al., 1999; Waltner-Law et al., 2003; Zhang et al., 2005), CREB binding to these promoter regions was relatively weak, and in the cases of the PGC1α and G6Pase promoters, nearly undetectable (Figure 4C).
Based on these results, we hypothesized that previously unappreciated control sequences may exist that confer CREB responsiveness to these metabolic genes. To localize these novel elements, we performed ChIP-Seq experiments on chromatin prepared from the livers of fasted mice as an unbiased approach toward the genome-wide identification of CREB target sites. With a focus on the PEPCK, PGC1α, and G6Pase loci, we were able to identify cAMP response element (CRE)-containing sequences strongly bound by CREB in vivo approximately 5.8kb upstream of the PEPCK transcription start site (Figure 4D), and in an intronic region approximately 8kb downstream of the PGC1α promoter (Figure 4E). No new regions bound by CREB were discovered upon examination of the G6Pase locus. Robust CREB occupancy at these novel target sites in the PEPCK and PGC1α genes was confirmed by qPCR in chromatin isolated from livers of fasted control mice (Figure 4C). While CBP and p300 binding at these sites was similarly unaffected by the absence of CRTC2 (Figure 4B), we found that CREB occupancy at the novel PEPCK and PGC1α target sequences was reduced in hepatocytes of CRTC2-/- mice fasted for 24 hours (Figure 4C). In accordance with the gene expression pattern, CREB binding was not affected in the livers of mice fasted for 6 hours, thus revealing a correlation between CREB binding and the transcriptional activity of these genes. These data indicate a direct role for CRTC2 in facilitating CREB binding to select target elements, but also raise questions with regard to the nature of CRTC2-mediated G6Pase regulation. Neither CBP nor p300 recruitment to the G6Pase promoter region was impacted by the CRTC2 deletion, while CREB binding to its presumptive CRE was tenuous, leaving open the possibility of indirect or CREB-independent mechanisms of control. Overall, these results support a role for CRTC2 in regulating the expression of gluconeogenic genes in the liver, but also indicate that the overall contribution of CRTC2 to glucose homeostasis in vivo is limited.
An 18.7kb DNA fragment containing all CRTC2 coding sequences was retrieved from C57BL/6J mouse BAC clone RP23-237D16 via bacterial recombination (Copeland et al., 2001) into plasmid PL253 that contains an Mc1-driven Thymidine Kinase (TK) cassette for negative selection. One loxP site was inserted upstream of exon 3 while a pair of loxP sites flanking a neomycin selection cassette were placed downstream of exon 11, also by bacterial recombination. The linearized targeting vector was electroporated into B6 ES cells (Chemicon) and clones that survived G418 and gancyclovir selection were screened for homologous recombination by Southern blot analysis. Targeted clones were injected into C57BL/6J-derived blastocysts that were then transferred to pseudopregnant females. Male offspring were mated to C57BL/6J females and ES cell-derived offspring were identified by PCR-based genotyping. Mice harboring the targeted insertion of the three-loxP sites in the CRTC2 gene locus were then crossed to the EIIa-Cre line to achieve germline deletion of loxP-flanked sequences (Holzenberger et al., 2000). Mice heterozygous for the CRTC2 deletion (CRTC2+/-) were then mated to C57BL/6J to segregate away the EIIa-Cre transgene. CRTC2+/- mice were then inter-crossed to obtain CRTC2-/- animals. Genotyping of all offspring was performed by PCR using 5′ and 3′ primers for CRTC2 (5′-GCCTGATTTTTCCCCCTATT-3′ and 5′-CATCTCAATGAAGGCCACCT-3′), which yield a 431bp product from the wild-type allele and a 608bp product from the CRTC2 null allele. All PCR was performed for 35 cycles under the following conditions (95°C, 30sec; 60°C, 30sec; 72°C, 60sec).
Whole venous blood collected from the inferior vena cava of CRTC2-/- and control mice was applied to plasma separator tubes with lithium heparin anticoagulant (BD Biosciences), centrifuged for 2 min at 14,000 × g, and the plasma was collected. Metabolic parameters were determined by Ani Lytics, Inc. (Gaithersburg, Md.).
Liver fragments from CRTC2-/- and control mice were homogenized by hand in 200μL of lysis buffer containing 50mM Tris (pH8.0), 150mM NaCl, 1mM EDTA, 1% Triton-X100, 0.5% deoxycholic acid, and 1% SDS, supplemented with protease inhibitor cocktail (Roche) and phosphatase inhibitor cocktails 1 and 2 (Sigma), then sonicated using a Bioruptor (Diagenode). Whole-cell lysates were then centrifuged at maximum speed for 15 minutes to pellet cellular debris, and the supernatant was collected. Protein concentrations were determined by Bradford assay using Protein Assay Reagent (Bio-Rad). Approximately 50μg of whole-cell lysates were separated by SDS-PAGE and transferred to Immobilon-P membranes (Millipore). Membranes were blocked in buffer consisting of 5% milk and 0.1% Tween-20 in 1X PBS (PBST-milk) for 4 hrs, then incubated overnight at 4°C with anti-CRTC2 (Calbiochem ST1099) diluted 1:5000, anti-FOXO1 (Cell Signaling C29H4) diluted 1:1000, anti-CREB (Cell Signaling 48H2) diluted 1:1000, anti-phoshpo-CREB (Cell Signaling 87G3) diluted 1:1000, anti-HNF4α (R&D Systems H6939) diluted 1:1000, anti-STAT3 (Cell Signaling 9132) diluted 1:1000, anti-C/EBPα (Santa Cruz sc-61) diluted 1:500, anti-C/EBPβ (Santa Cruz sc-150) diluted 1:500, or anti-βActin (Cell Signaling 13E5) diluted 1:2000 in PBST-milk. After washing, membranes were incubated with HRP-conjugated goat-anti-rabbit IgG (Amersham) diluted 1:12,500 in PBST-milk for 45 minutes at room temperature. After washing, proteins were visualized using the ECL-plus Western Blotting Detection System (Amersham).
Blood was sampled from the tail vein of mice allowed to feed ad libitum (fed) or fasted for 6, 9, 12, or 24 hours. Glucose levels were measured with a OneTouch Ultra Glucometer (Lifescan). For gene expression analyses of re-fed animals, mice were fasted overnight and then allowed to feed for two hours before sacrifice. For pyruvate challenge experiments, animals were fasted overnight, and a blood glucose measurement was taken and considered time zero. Mice were then injected intraperitoneally with 2g of pyruvate (Sigma-Aldrich) per kg of body weight. Glucose levels were then measured at 15, 30, 60, 90, and 120 min postinjection.
Glucose homeostasis was measured in 12-15 week old male control and CRTC2-/- mice using insulin clamp and radioisotopic tracer techniques (Alenghat et al., 2008; Varela et al., 2008). An indwelling catheter was inserted in the right internal jugular vein under sodium pentobarbital anesthesia and extended to the right atrium. Four days after recovery, the mice were fasted overnight for 16 hours. They were placed in restrainers and administered a bolus injection of 5 μCi of [3-3H] glucose, followed by continuous intravenous infusion at 0.05 μCi/min. Baseline glucose kinetics were measured for 60 min, followed by hyperinsulinemic-hypoglycemic clamp for 120 min. A priming dose of regular insulin (16 mU/kg, Humulin; Eli Lilly, Indianapolis, IN) was given intravenously, followed by continuous infusion at 2.5 mU/kg-1/min-1. Blood glucose was maintained at 60-70 mg/dL via a variable infusion rate of 20% glucose. The mice were euthanized and liver samples were excised, frozen immediately in liquid nitrogen and stored at −80°C. The rates of basal glucose turnover and whole body glucose uptake are measured as the ratio of the [3H] glucose infusion rate (dpm) to the specific activity of plasma glucose. Hepatic glucose production (HGP) during clamp was measured by subtracting the glucose infusion rate (GIR) from the whole body glucose uptake (Rd).
Tissue samples from mice fed ad libitum were homogenized in 6% PCA (500μL/100μg tissue). Precipitates were pelleted and the supernatant was collected. One volume of H2O was added and the solution was adjusted to pH 6.5 with 10N KOH. A fraction of each sample was then incubated with five volumes of amyloglucosidase (1mg/mL in 0.2M, pH 4.8 acetate buffer) at 40°C for 2 hours. Glucose concentrations were then determined using the Amplex Red glucose/glucose oxidase assay (Invitrogen). Samples incubated in the absence of amyloglucosidase were used as baseline controls.
Total cellular RNA was extracted from liver fragments or cultured primary hepatocytes using RNeasy Kit (Qiagen), then assayed for quantity and quality with the Agilent 2100 Bioanalyzer (Agilent Technologies). Approximately 1μg of total RNA was reverse transcribed using oligo (dT) and Superscript II reverse transcriptase (Invitrogen). The resulting cDNA samples were the template for qPCR experiments performed with SYBR GreenER qPCR Supermix (Invitrogen) and the SYBR Green (with dissociation curve) program on the Mx3000 Multiplex Quantitative PCR System (Stratagene). Reactions were performed in triplicate and normalized relative to the ROX reference dye. Median cycle threshold values were determined and used for analyses. Expression levels were normalized to those of hypoxanthine-guanine phosphoribosyltransferase (HPRT) as the internal control. Primer information is available at http://www.med.upenn.edu/kaestnerlab/index.shtml.
Hepatocytes were isolated from 3-month-old mice. The vena cava was cannulated and the liver was perfused with 37°C Liver Perfusion Buffer (Gibco) for 2 min, then 50mL of 37°C Liver Digestion Buffer (Gibco) supplemented with 30mg of collagenase D (Roche). The liver was then excised, dispersed in DMEM containing 10% FBS and penicillin/streptomycin, filtered through a 100 micron mesh, pelleted, and subjected to a Percoll gradient (45% Percoll in DMEM) to isolate viable cells. After two washes in DMEM containing 10% FBS and penicillin/streptomycin, cells were plated in 12-well, collagen I-coated plates (BD BioCoat) at a density of 150k cells/well in DMEM containing 10% FBS and penicillin/streptomycin. For gene expression analysis, cells were cultured for 48 hours, then stimulated with 50nM glucagon for 2 hours before harvesting for RNA isolation. For glucose production measurements, cells were cultured for 24 hours, then switched to glucose-free medium containing 20mM sodium lactate and 2mM sodium pyruvate before stimulating for 5 hrs with 50nM glucagon. The culture medium was collected for glucose concentration determination via Amplex Red glucose/glucose oxidase assay (Invitrogen).
Chromatin from control and CRTC2-/- livers was prepared as previously described (Tuteja et al., 2008). Immunoprecipitations were also performed as previously described (Tuteja et al., 2008), except that the herring sperm DNA was excluded from the agarose bead blocking step. Sheared chromatin (10μg) was incubated overnight at 4°C with 2μg of anti-CREB (Santa Cruz sc-186X), anti-CBP (Santa Cruz sc-369X and sc-583X), or anti p300 (Santa Cruz sc-584X and sc-585X), and immune complexes were recovered with protein-G conjugated agarose beads for 45min at 4°C. Enrichment of target DNA fragments was determined via qPCR using SYBR GreenER qPCR Supermix (Invitrogen) and the SYBR Green (with dissociation curve) program on the Mx3000 Multiplex Quantitative PCR System (Stratagene). qPCR experiments were performed in triplicate on both input (sheared genomic DNA) and ChIP sample DNA, and median cycle threshold values were used for analysis. Enrichment of bound regions was calculated using the myelin basic protein (MBP) promoter as a reference for nonspecific binding, and comparing to amplification values obtained from input DNA.
For ChIP-Seq experiments, the remainder of the immunoprecipitated DNA was modified for sequencing following the manufacturer’s protocol (Illumina). Briefly, DNA samples were blunted with a combination of T4 DNA polymerase, Klenow polymerase, and T4 PNK, then a single 3′-end “A” base was added using Klenow exo (3′ to 5′ exo minus). Adapters provided by Illumina were then ligated to the ends of the modified DNA before size selection of ~200bp fragments via PAGE extraction. The isolated DNA was then amplified by 18 cycles of PCR as described. PCR products were column purified with the QIAquick PCR Purification Kit (Qiagen) and assayed for quantity and quality with the Agilent 2100 Bioanalyzer (Agilent Technologies), then diluted to a concentration of 10nM. The library was sequenced on an Illumina 1G Genome Analyzer, at a concentration of 3-4pM. The sequencing output was analyzed using the Genome Analyzer pipeline provided by Illumina. Sequence tags that aligned uniquely to the mouse genome build MM8 with zero, one, or two mismatches, according to the ELAND alignment algorithm, were used for further analysis.
We thank Dr. Joshua Curtin for assistance in performing the primary hepatocyte culture experiments, Dr. Hong Fu for the ES cell work, Beth Helmbrecht and Karrie Brondell for her help in managing the mouse colony, and Dr. Russell Miller for help with the glycogen assays. We also thank Olga Smirnova and Alan Fox for their technical assistance in processing the ChIP-Seq libraries and the University of Pennsylvania Radioimmunoassay and Biomarkers Core for performing the glucagon assays. This work was supported by NIH grant DK049210, the University of Pennsylvania Training Grant T32 DK007314 in Diabetes, Metabolism and Endocrine Diseases, and the American Diabetes Association (Award No: 7-08-MN-28).
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.