Chronic Hyperglycemia Enhances PEPCK Gene Expression and Hepatocellular Glucose Production Via Elevated Liver Activating Protein/Liver Inhibitory Protein Ratio

Fao cells, derived from the H4IIE hepatoma cell line, possess a complete gluconeogenic enzyme system that allows the cells to survive and grow in low-glucose (5.5 mmol/l) or glucose-free (0.0 mmol/l) medium. Fao cells are usually maintained in a high-glucose medium (25 mmol/l). Before any procedures in the current study, cells were incubated in Dulbecco’s modified Eagle’s medium (DMEM) (Mediatech, Herndon, VA) with 5.5 mmol/l glucose and 10% fetal bovine serum (FBS) for more than 2 months, which allowed the cells to adapt to growth in a medium with “normal” concentrations of glucose. The medium was changed every day to maintain the glucose supply. Before any testing, the medium was replaced with the same medium containing a normal glucose concentration without FBS and was incubated overnight. Glucose treatment protocols were performed using DMEM and various concentrations of glucose, as indicated in results. Mannitol was used to control for any potential osmotic effects.

Glucose production from Fao cells was measured as described by Duong et al. (13) using a colorimetric glucose oxidase assay (Sigma, St. Louis, MO). Briefly, after 24 h of treatment with different glucose concentrations, the cells were washed three times with PBS. Then the cells were incubated for 6 h at 37°C, 5% CO₂, in glucose production buffer (glucose-free DMEM, pH 7.4, containing 20 mmol/l sodium lactate, 1 mmol/l sodium pyruvate, and 15 mmol/l HEPES, without phenol red). The glucose assays were conducted in duplicate and the intra-assay coefficient of variation was <5%.

Fao cell nuclear protein extracts were isolated as described previously (18). Protein concentration was measured using the Bio-Rad protein assay system (Bio-Rad, Hercules, CA). Nuclear protein extract (30 μg) was separated by 10% acrylamide SDS-PAGE gel, and proteins were transferred to a polyvinylidene difluoride membrane (Bio-Rad) with a Mini Trans Blot cell from Bio-Rad. After blocking with 5% fat-free milk, the membrane was blotted with 1:1,000 dilution of rabbit polyclonal antibody raised against the COOH-terminal of human C/EBPβ (Santa Cruz Biotechnology, Santa Cruz, CA), which cross-reacts with both LAP and LIP, anti-C/EBPα (Santa Cruz Biotechnology), or anti-CREB antibody (Cell Signaling Technology, Beverly, MA). After washing, the membrane was blotted with a 1:10,000 dilution of horseradish peroxidase–conjugated anti-rabbit IgG antibody (Bio-Rad). The peroxidase activity was visualized with ECL-Plus from Amersham Pharmacia Biotech. The specific bands were quantitated using a GEL-DOC density scanner and Quantity One software (Bio-Rad).

Total RNA was prepared from plated Fao cells with Trizol reagent following the manufacturer’s protocol (Invitrogen, Carlsbad, PA). RNA was denatured, separated by agarose gel, and transferred to a positively charged nylon membrane (Roche Molecular Biochemicals, Indianapolis, IN). The cDNA fragment of rodent C/EBPβ, PEPCK, and GAPDH was labeled with [α-³²P]dATP using a Prime-It II Random Primer Labeling kit (Stratagene, Cedar Creek, TX). The membrane was blotted with labeled probes, and radioactive signals were quantified and analyzed with PhosphorImager and Imagequant software (Molecular Dynamics, Sunnyvale, CA).

A 562-bp fragment of the 5′-flanking region of the mouse PEPCK gene, extending from −490 to +73 bp, was generated by PCR using mouse genomic DNA as a template. The PCR product was ligated into the pA3Luc vector at the KpnI and HindIII sites (pA3-PEPCK-Luc). The inserted DNA fragment was verified by restriction analysis and sequencing. The full-length mouse C/EBPβ cDNA was cloned and inserted into the pcDNA3.1 vector, which expresses both LAP and LIP. The LAP expression plasmid (pcDNA-LAP) was generated by mutation of the LIP translation start codon using a QuickChange Site–directed Mutagenesis kit (Stratagene, La Jolla, CA). The sequences of mutagenic primers were as follows: 5′-CGACGCGCCCGCgATcGCGGCCGGTTTCC-3′ and 5′- GGAAACCGGCCGCgATcGCGGGCGCGTCG-3′. The LIP expression plasmid was created by inserting the full-length mouse LIP cDNA into pcDNA3.1 vector at BamHI and EcoRI sites. The PEPCK promoter-luciferase gene reporter construct and pcDNA-LAP or pcDNA-LIP plasmids were cotransfected into Fao cells using the Lipofectamine 2000 reagent (Invitrogen) following the manufacturer’s protocol. A pCMV–β-galactosidase plasmid was cotransfected for correction of transfection efficiency (Clontech Laboratories, Palo Alto, CA). The luciferase activity measurement has been described previously (19).

The association of C/EBPβ protein with the endogenous PEPCK promoter was measured by ChIP assay as described previously (13,20). Fao cells were cross-linked with 1% formaldehyde (Sigma) in serum-free DMEM at RT for 5–10 min. Glycine was added at a final concentration of 125 mmol/l to stop the cross-link reaction. Cells were rinsed with ice-cold PBS. The pelleted cells were resuspended in SDS lysis buffer (1% SDS, 10 mmol/l EDTA, 50 mmol/l Tris-HCl, pH 8.0) and incubated on ice for 10 min. The chromatin was sheared by sonicating on ice for 10-s pulses at a setting of 6 (output of 4–5 watts), yielding chromatin fragments of 100–600 bp in size. Samples were centrifuged at 14,000 rpm for 10 min at 4°C to remove detritus, and the supernatant was divided into aliquots for subsequent 10-fold dilution in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol/l EDTA, pH 8, 16.7 mmol/l Tris-HCl, pH 8.0). Chromatin complexes were immunoprecipitated for 12–18 h at 4°C while rotating with amounts (5 μg) of primary anti-C/EBPβ, Pol II, CBP, acetylated histone (AcH)3, AcH4, histone deacetylase 1, or LAP antibody, or without antibody to provide negative controls. LAP-specific antibody was generated and provided by Dr. Linda Sealy (Vanderbilt University Medical Center), against the COOH-terminus of rat C/EBPβ (13). Immune complexes were collected with 40 μl protein A agarose, while rotating for 3 h at 4°C, followed by centrifugation at 1,000g for 1 min at 4°C. The pellets were washed with the following buffers in succession: low salt, high salt, LiCl wash buffer, and twice with TE (1 mmol/l EDTA, 10 mmol/l Tris-Cl, pH 7.5) buffer. Chromatin complexes were eluted from the beads in three consecutive 30-min rotating incubations with 200 μl elution buffer (1% SDS, 0.1 mol/l NaHCO₃) at room temperature. To reverse cross-linking and digest RNA present in the samples, NaCl (200 mmol/l final concentration) and RNase mixture (Ambion, Austin, TX) was added, and the samples were incubated at 65°C for >6 h. The proteins were digested by incubating at 45°C for 90 min after the addition of the following at their final concentrations: 10 mmol/l EDTA, 40 mmol/l Tris-HCl, pH 6.5, and 50 μg/ml proteinase K. DNA was isolated and purified with the Quick PCR Purification kit (Qiagen, Valencia, CA).

The PEPCK promoter-specific PCR primers were designed with Jellyfish software. These included primers specific for the PEPCK promoter (5′-GAGGCCTCCCAACATTCAT-3′, 5′- CTCAGAGCGTCTCGCCGGA-3′) across the cAMP response element. Two percent of the immunoprecipitate (DNA) was used as template for the PCR, with 1 nmol/l of each primer, 200 μmol/l dNTP mix, PCR buffer, and 2.5 units TaqDNA polymerase (total 50 μl). After 36 cycles of amplification in a thermocycler, PCR products were run on a 2% agarose gel, visualized by ethidium bromide staining, and analyzed using Quantity One software.

Data are expressed as means ± SE. Statistical analysis was performed using the Student’s t test or ANOVA analyses and was followed by a contrast test with Tukey or Dunnett error protection. Differences were considered significant at P < 0.05.

Glucose production from gluconeogenic precursors was assayed in Fao cells and in primary hepatocytes by measuring the amount of glucose released into the medium, which initially contained gluconeogenic precursors but no glucose. In nonstimulated cells (no cAMP), glucose production was not significantly different in cells exposed to either low (5.5 mmol/l) or high (20 mmol/l) glucose for 24 h (Fig. 1A and B). In contrast, cAMP-stimulated glucose production was significantly increased by 45% in cells exposed to high glucose for 24 h. Thus, these data demonstrate that prolonged high-glucose treatment enhances cAMP-stimulated gluconeogenesis and glucose output. Glycogenolysis could contribute to HGP, but it is minor and unlikely to account for the increased glucose production because glycogen levels in these cells are not different and indeed are very low when exposed to 5 or 20 mmol/l glucose.

In a previous study, short-term exposure to high glucose suppressed PEPCK expression independent of insulin (3,5). Therefore, we investigated the time course of glucose regulation of PEPCK gene expression to determine whether this might account for our results with respect to glucose release. Consistent with a previous study (5), high glucose reduced PEPCK mRNA rapidly reaching a nadir after 2 h (Fig. 2). However, PEPCK mRNA levels rebounded and returned to initial levels after ∼20 h (Fig. 2). The differences in PEPCK mRNA over time were not due to differences in the rate of glucose utilization, as this remained linear over the 24-h period (data not shown). Therefore, these data indicate that sustained high-glucose treatment diminishes the inhibitory effect of glucose on PEPCK gene expression in hepatocytes. In Fig. 3, Fao cells were treated with 5.5 or 20 mmol/l glucose for 24 h, the test media was changed to DMEM with 5.5 mmol/l glucose for both groups, and cells were incubated in the presence or absence of 100 μmol/l dibutyryl-cAMP for 2 h. The cAMP-induced PEPCK mRNA levels in 20 mmol/l glucose–treated cells were significantly higher (by 50%) versus 5.5 mmol/l glucose–treated cells (P < 0.05). These data suggest that prolonged high-glucose treatment not only diminishes the suppressive effect of glucose on PEPCK gene expression but also enhances cAMP-induced PEPCK gene expression.

C/EBPβ is an important accessory transcription factor in mediating hormonal regulation of PEPCK gene transcription (4) and has recently been shown to be rapidly induced by glucose deprivation in human HEPG2 hepatoma cells (12). Therefore, we investigated whether the fluctuation of glucose alters C/EBPβ expression in Fao cells. Cells were maintained in DMEM with 5.5 mmol/l glucose. After overnight serum-free incubation, the media were changed to the same media with 5.5 or 20 mmol/l glucose. After 24 h, nuclear protein was extracted and C/EBPβ was measured by Western blot using an antibody that cross-reacts with both LAP and LIP isoforms of C/EBPβ. Total C/EBPβ protein content was reduced in high-glucose–treated Fao cells (Fig. 4A). However, C/EBPα and CREB protein levels were not significantly different between treatments (Fig. 4A). C/EBPβ is a single gene, but several protein isoforms are produced by translation at different translation initiation sites. In Fao cells, two main isoforms can be found: LAP (43 kDa) and LIP (20 kDa). LIP contains only COOH-terminal dimerization and DNA binding domains but lacks a transactivation domain; thus, it has DNA binding activity but not transactivation potential. Therefore, LIP is considered to be a natural dominant-negative isoform of C/EBPβ (15). The LAP/LIP ratio has been reported to be more important than the total C/EBPβ protein level in controlling PEPCK gene transcription in response to insulin (13). To examine the effect of high-glucose treatment on the endogenous LAP/LIP ratio, the LAP and LIP protein levels were measured by Western blot and the LAP/LIP ratio was calculated. After 24 h of high glucose, the LAP/LIP ratio was significantly increased by nearly twofold, despite the reduction of total C/EBPβ protein level (Fig. 4B, P < 0.001).

The reduction in C/EBPβ protein level after sustained hyperglycemia might lead to reduced availability of C/EBPβ to bind to DNA. However, the increase in the LAP/LIP ratio suggested there could be a shift in the proportion of LAP relative to LIP. We therefore investigated the effects of glucose on total C/EBPβ binding to the PEPCK promoter region using the ChIP assay. Incubation of Fao cells under either low or high glucose did not affect total C/EBPβ binding to the PEPCK promoter in the absence or presence of cAMP (Fig. 5). However, because the antibody we used recognizes both LAP and LIP, the results do not discriminate between LAP and LIP binding; therefore, a LAP-specific antibody was used to determine LAP-specific binding to the PEPCK promoter under high-glucose conditions. As shown in Fig. 6, high-glucose treatment increased LAP binding to the PEPCK promoter, whereas Pol II and CBP increased slightly, and AcH3 and binding were unchanged. Thus, the data indicate that the total C/EBPβ protein bound to the endogenous PEPCK promoter CRE site is similar; however, high-glucose treatment increased LAP association with the PEPCK promoter, which may be responsible for cAMP-induced higher PEPCK expression in the cells under high-glucose conditions.

The above results indicate that high glucose suppresses total C/EBPβ protein levels but increases the LAP/LIP ratio in Fao cells. To verify the importance of the LAP/LIP ratio in controlling PEPCK gene transcription, we measured PEPCK promoter activity in response to different LAP/LIP ratios, which were achieved by manipulating the amounts of cotransfected LAP or LIP expression plasmids (Fig. 7). Because the expression of LAP and LIP was driven by the cytomegalovirus promoter in the expression plasmids, the expressed LAP or LIP protein level was closely correlated to the amount of transfected plasmid under current assay conditions (data not shown). Therefore, we used the pcDNA-LAP/pcDNA-LIP plasmid DNA ratio to represent the LAP/LIP ratio in transfected cells (Fig. 7), while keeping the total amount of LAP and LIP constant. When only pcDNA-LIP was cotransfected, the luciferase activities were suppressed 75% compared with the cells cotransfected with empty vector pcDNA (P < 0.001, Fig. 7). In contrast, increased LAP expression increased luciferase activity up to 25-fold (P < 0.0001). There was a linear relationship between the pcDNA-LAP/pcDNA-LIP ratio and the reporter luciferase activity. Thus, these data demonstrate that the PEPCK promoter activity is precisely regulated by the LAP/LIP ratio.

Increased LIP protein or a reduction of the LAP/LIP ratio inhibits the formation of transcription factor complexes on the PEPCK promoter region, thus inhibiting hormone-activated PEPCK gene transcription (13). Our above studies demonstrated that sustained high glucose increased the LAP/LIP ratio. To investigate whether an increased LAP/LIP ratio alters hormone regulation of PEPCK gene expression, Fao cells were studied in response to overexpression of LAP. As shown in Fig. 8, there was a slight increase in PEPCK mRNA when LAP was overexpressed in Fao cells in the absence of hormones or cAMP but was not statistically different (P > 0.05). However, in the presence of cAMP, dexamethasone, or both, PEPCK mRNA levels were significantly higher in cells transfected with LAP (i.e., increased LAP/LIP ratio) versus vector alone (Fig. 8, P < 0.05). Interestingly, insulin treatment reduced cAMP-induced PEPCK mRNA levels in both control and pcDNA-LAP–transfected Fao cells, but the PEPCK mRNA level was still significantly higher in pcDNA-LAP–transfected Fao cells (P < 0.05). Together, therefore, these data indicate that increased LAP/LIP ratio enhances both cAMP- and dexamethasone-stimulated PEPCK gene transcription. Furthermore, the increased LAP/LIP ratio also appears to interfere with repression of PEPCK gene expression in response to insulin.

The liver is a critical organ in maintaining blood glucose concentration both via its ability to supply glucose to the circulation via glycogenolysis and gluconeogenesis in the postabsorptive state and remove glucose from the circulation after meal ingestion (1). These processes are coordinately regulated by hormones and nutrients (1,2). However, in both type 1 and type 2 diabetes, the gluconeogenic pathway is inappropriately active and supplies a relatively larger amount of glucose into the circulation, which contributes to both fasting and postprandial hyperglycemia (21–23). Despite a wealth of literature characterizing the molecular regulation of gluconeogenesis by hormones and nutrients, little is known regarding the specific molecular basis for inappropriate gluconeogenesis at the level of the PEPCK promoter. It is well known that sustained hyperglycemia impairs hepatic insulin sensitivity and results in reduced suppression of HGP by insulin (24). However, recent studies indicate that chronic hyperglycemia itself also plays an important role in the inappropriate gluconeogenesis, a phenomenon referred to clinically as “glucose toxicity” (6–8). The present study was undertaken to examine the molecular events that might account for this effect of hyperglycemia.

The regulation of PEPCK by hormones is largely at the transcription level, and most of the hormone response units in the PEPCK promoter region have been well characterized. Several transcription factors appear to be essential factors or important coactivators for regulation of PEPCK transactivation, including the C/EBPs, CREB, peroxisome proliferator–activated receptor-γ coactivator 1α, and forkhead box O1 (4,9,11,25,26). Of these, C/EBPβ is highly expressed in the liver and contributes to the transcriptional regulation by several gluconeogenic hormones and appears to be highly responsive to changes in glucose levels.

There are several regions on the PEPCK promoter that bind C/EBPβ. Of these, the first region containing a CRE, which is ∼60 bp 5′ from the TATA box (14), has been shown to bind members of the leucine zipper family of transcription factors, including C/EBPβ, C/EBPα, and CREB (14). In region 2, there is a P3(I) site, which exclusively binds with C/EBP transcription factors. P3(I) is required for the liver-specific expression of the PEPCK gene (27). Deletion of this site from the PEPCK promoter decreases cAMP-stimulated PEPCK transcription by 60% (27). Region 3 contains the glucocorticoid response unit composed of two glucocorticoid regulatory elements and three glucocorticoid accessory factor binding sites (28,29). It has been demonstrated that C/EBPβ mediates glucocorticoid-, glucagon-, and insulin-regulated PEPCK gene transcription through these units (4,9,13,14,26,29). Generally speaking, in response to hormonal stimulation, there is a change in protein level or location of specific transcription factors or coactivators and a change of binding of the transcription factor to the hormone response unit. For example, upon cAMP stimulation, there is a rapid (<30 min) increase in C/EBPβ protein and increased C/EBPβ binding at sites in the PEPCK promoter region, which increase PEPCK gene transcription. However, there are two main isoforms of C/EBPβ in hepatocytes: LAP and LIP. LIP is a functional dominant-negative isoform of C/EBPβ, which competes with LAP in regulating PEPCK gene expression (15). Thus, the LAP/LIP ratio may play an important role in controlling target gene transcription, independently of total transcription factor protein level. For example, insulin rapidly upregulates C/EBPβ expression; however, this increase in total C/EBPβ protein does not increase PEPCK gene transcription but instead results in suppression of glucocorticoid-mediated transactivation because LIP is preferentially induced compared with LAP (13). In the current study, we have further defined the importance of the LAP/LIP ratio in controlling PEPCK gene transcription using a transient transfection system. Specifically, the data demonstrated a direct relationship between the LAP/LIP ratio and PEPCK promoter activity.

Over the past few years, several different models of PEPCK gene transcription have been proposed (13,25,30,31). Although none of these models affords the global mechanism defining regulation of PEPCK gene transcription in response to hormone stimulation, C/EBPβ has been suggested to be involved in several sites of action within the promoter affecting hormone sensitivity. Duong et al. (13) recently proposed a new model that characterized the role of the LAP/LIP ratio in mediating glucocorticoid-, insulin-, and cAMP-regulated PEPCK gene transcription. In this model, C/EBPβ constitutively binds to the PEPCK promoter region and functions to recruit activating factors in response to hormonal signals. Whereas basal transcription is maintained by recruitment of the Pol II holoenzyme, several factors are recruited in response to hormone stimulation. In the present study, sustained high-glucose treatment decreased total C/EBPβ protein level in hepatoma cells but increased the LAP/LIP ratio. Although the association of C/EBPβ with the PEPCK promoter was not altered by high-glucose treatment, LAP association appeared to increase after 24-h exposure to high glucose. We tried looking for CREB immunoprecipitation using ChIP, but unfortunately the commercially available CREB antibodies were very poor at immunoprecipitation, and thus we were unable to perform this experiment. We previously showed that diabetes (hyperglycemia) in mice does not change the levels of two other CRE-binding proteins, C/EBPα or CREB, or their binding to the PEPCK promoter using gel-shift assays of mouse liver nuclear extracts and PEPCK CRE (16). In the present report, we also saw no change in the levels of CREB or C/EBPα in Fao cells exposed to chronic high glucose, as shown in Fig. 4. Although we cannot rule out a possible additional effect of CREB on the cAMP response, our current data and previously published results lend credence to the idea that hyperglycemia has specific effects on LAP/LIP levels and functions, at least in part, to increase the PEPCK promoter’s ability to interact with other proteins via cAMP to affect transcription. These data indicate that high glucose alone increases the in vivo occupancy of LAP on the PEPCK promoter but does not increase assembly of Pol II, CBP, or acetylated histone binding to the transcriptional machinery. Rather, the increase in LAP association under hyperglycemic conditions may serve as an important metabolic switch, allowing greater assembly of these factors in response to hormonal stimulation.

The physiological importance of these observations lies in the fact that patients with hyperglycemia not only have hepatic insulin resistance at the level of the insulin receptor but also experience hyperglucagonemia (32,33). Therefore, the present research suggests that an important manifestation of prolonged high glucose in liver may be an increase in the LAP/LIP ratio, which may further potentiate the effect of glucagon at the level of the PEPCK promoter, thereby contributing further to hyperglycemia in a vicious cycle. The present study focuses on mechanisms for chronic hyperglycemia. The molecular mechanism for acute inhibition of PEPCK gene expression by high glucose is unknown. The effect could involve several transcription control regions of the PEPCK promoter, but these were not addressed by the studies.

In summary, the present study identifies a novel molecular mechanism for glucose regulation of the PEPCK gene and glucose production. Our data suggest that chronic hyperglycemia may induce a new steady-state level of the LAP/LIP proteins in addition to a rearrangement of the chromatin environment, or a greater recruitment of activating factors available to the DNA. Sustained hyperglycemia diminishes the ability of glucose to inhibit gluconeogenesis and PEPCK gene expression and augments hormone-induced PEPCK gene expression. The study further shows that although chronic hyperglycemia decreases C/EBPβ expression in liver cells, the increase in the LAP/LIP ratio potentiates hormone-induced PEPCK gene expression. Therefore, the current study, using an in vivo system, has identified an important molecular mechanism that mediates hyperglycemia-induced dysregulation of HGP at the level of the PEPCK promoter.

This work was supported by grants from the American Diabetes Association (1-04-JF-44, to J.S.) and National Institutes of Health (DK-42938, to J.E.F.).

We gratefully acknowledge Paul MacLean for reading the manuscript and for helpful discussion. We thank T. Jackson and A. Bradford for technical assistance with the luciferase assay. We also thank the members of the Friedman laboratory for helpful discussion.