Mitogen-activated Protein Kinase Regulates Transcription of the ApoCIII Gene

The transcriptional regulation of the apoCIII gene by hormonal and metabolic signals plays a significant role in determining plasma triglyceride levels. In the current work we demonstrate that the apoCIII gene is regulated by the mitogen-activated protein (MAP) kinase signaling pathway. In HepG2 cells, repression of MAP kinase activity by treatment with the mitogen-activated protein kinase/extracellular signal-regulated kinase kinase inhibitor PD98059 caused a 5–8-fold increase in apoCIII transcriptional activity. Activation of MAP kinase by phorbol ester treatment caused a 3–5-fold reduction in apoCIII transcription. The region of the apoCIII promoter responsible for this regulation was mapped in transiently transfected HepG2 cells to a 6-base pair element located at −740. The major protein binding to this site was identified as the nuclear hormone receptor HNF4. An increase in HNF4 mRNA and protein levels was observed in HepG2 cells after treatment with PD98059, indicating that the MAP kinase pathway regulates the expression of the HNF4 gene. These findings demonstrate that the apoCIII gene can be regulated by signals acting through the MAP kinase pathway and that this regulation is mediated, at least in part, by changes in the amount of HNF4.

ApoCIII is a component of very low density lipoprotein and functions as a key regulator of serum triglyceride levels (1). In transgenic animals, overexpression of the apoCIII gene caused hypertriglyceridemia (2), with as little as 30 -40% excess apo-CIII causing a 2-fold increase in triglyceride levels (3). Likewise, mice that did not express apoCIII because of a gene knockout had abnormally low circulating triglyceride levels (4). Genetic studies have demonstrated a key role for apoCIII in determining plasma triglyceride levels in humans. A sequence polymorphism in the 3Ј-untranslated region of the apoCIII gene has been associated with elevated triglyceride levels in several populations (5)(6)(7)(8)(9). In addition, clinical studies have reported that some hypertriglyceridemic patients have elevated apoCIII levels and increased apoCIII production rates (10 -12). ApoCIII modulates serum triglyceride metabolism by reducing both lipolysis and uptake of triglyceride-rich lipoproteins (3,(13)(14)(15)(16).
The apoCIII gene is transcriptionally regulated by a variety of metabolic and hormonal signals. In a previous study, we demonstrated that in a hypoinsulinemic animal model of diabetes, hepatic apoCIII transcriptional activity is regulated by insulin and that these changes correlated with changes in plasma triglyceride levels (17). Further evidence that apoCIII transcriptional activity plays a significant role in determining plasma triglyceride levels comes from the analysis of a genetic polymorphism in the human apoCIII promoter region. Two major sequence variants of the apoCIII promoter can be found in the human population. The most common (wild-type) allele differs from the less common (variant) allele by five single base pair DNA sequence differences. A haplotype of the apoCIII locus containing the variant promoter was associated with hypertriglyceridemia in a genetic study (18) and was defective in its ability to be regulated by insulin in transfected HepG2 cells (19). These results confirm that the apoCIII gene is transcriptionally regulated by insulin and suggest that this regulation plays an important role in determining plasma triglyceride levels. Taken together, these findings support the hypothesis that the rate of apoCIII gene expression is an important determinant of plasma triglyceride levels and suggest that modulation of apoCIII transcription by metabolic and/or hormonal signals is likely to have a direct effect on plasma triglyceride metabolism. Previous work has demonstrated that the transcriptional activity of the apoCIII gene in the liver is dependent on the nuclear hormone receptor HNF4, which interacts with at least two sites in the apoCIII promoter (20 -22). HNF4 is abundant in liver, intestine, and kidney and regulates many genes involved in lipid and glucose metabolism (23,24). The active form of HNF4 is a homodimer, and it does not appear to heterodimerize with other members of the nuclear receptor family (25). Although HNF4 is usually classified as an orphan receptor, recent work has suggested that coenzyme A derivatives of some fatty acids activate the receptor and have been proposed to be endogenous ligands for HNF4 (26). A crucial role for HNF4 in the regulation of metabolism was demonstrated by the recent finding that an inherited form of diabetes (MODY, maturity onset diabetes of the young) is caused by a mutation in the HNF4 gene (27).
We report here that the transcriptional activity of the apo-CIII gene is regulated by the MAP 1 kinase signaling pathway. Activation of Erk1/2 signaling caused a reduction in apoCIII gene transcription. This effect was dependent on an HNF4 binding site located at Ϫ740 in the apoCIII promoter. The HNF4-dependent MAP kinase-mediated regulation of the apo-CIII gene was caused, at least in part, by changes in the quantity of HNF4 after treatment with modulators of MAP kinase activity.

MATERIALS AND METHODS
ApoCIII Promoter/Luciferase Reporter Constructs-The basic apo-CIII promoter/luciferase reporter construction (pL854) contains human apoCIII promoter sequences from Ϫ854 to ϩ22 linked to the coding sequence of the firefly luciferase gene in the vector pGL-basic (Promega Inc.). The apoCIII promoter sequences were derived from the apoCIII/ chloramphenicol acetyltransferase expression vector pM854 which contains a point mutation at Ϫ126 which generates a unique NdeI site (28). The progressive deletion mutants were generated by digestion of pL854 with SmaI, which cuts the plasmid upstream of the apoCIII promoter at Ϫ866, and with a second restriction enzyme that cuts within the apo-CIII promoter. After repair with T4 DNA polymerase to generate blunt ends, the plasmid was re-ligated with T4 DNA ligase. The enzymes used to generate the deletion series were ApaI (Ϫ782), StuI (Ϫ690), and AspI (Ϫ169). An additional series of promoter deletions was constructed by removal of specific fragments from pL854. pL169h was constructed by inserting the KpnI/StuI fragment (Ϫ862 to Ϫ694) immediately upstream of CIII sequences in pL169. pL169f was generated by inserting the KpnI/ApaI fragment (Ϫ862 to Ϫ784) immediately upstream of CIII sequences in pL169. pL169i was made by inserting the ApaI/StuI fragment (Ϫ784 to Ϫ694) immediately upstream of CIII sequences in pL169.
Point mutations in the pL854 background (pL854C, pL854D, and pL854E) were generated using the GeneEditor in vitro site-directed mutagenesis kit (Promega Inc.) according to the protocol provided by the manufacturer. The sequences of the oligonucleotides used for generating the mutations were: This protocol replaced six nucleotides in the CIII promoter with a unique NcoI site (underlined). Deletion of sequences between the C and E mutations (pL854C/E) and the D and E mutations (pL854D/E) was generated by digestion of the point mutation plasmids (pL854C, pL854D, and pL854E) with NcoI and a second unique enzyme (EcoRI) and recombining the appropriate fragments in vitro. The additional point mutations pL854L, pL854M, and pL854N were generated from the D/E deletion plasmid (pL854D/E). The missing NcoI fragment was replaced with a 53-base pair synthetic double-stranded oligonucleotide containing a unique NheI site at the L, M, and N regions (for exact sequences, see Fig. 5). All mutations were verified by DNA sequencing.
RNA Isolation and Northern Analysis-Total RNA was extracted using the Ultraspec RNA Isolation System (Biotecx, Inc.) from confluent HepG2 cells grown in 100-mm dishes. Total RNA was subjected to gel electrophoresis on a 1.2% formaldehyde gel and transferred onto a nitrocellulose membrane (Bio-Rad). Human apoCIII and rat actin 32 Plabeled cDNA probes were prepared using the Rediprime II labeling system (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Northern analysis hybridization was carried out at 42°C for 16 h with 1 ϫ 10 6 cpm/ml labeled probe in 5 ϫ SSPE, 5 ϫ Denhardt's, 50% formamide, 0.1% SDS, and 100 g/ml salmon sperm DNA. Blots were washed twice at room temperature for 20 min with 1 ϫ SSC and 0.1% SDS and for 20 min at 55°C with 0.2 ϫ SSC and 0.1% SDS and then subjected to autoradiography and PhosphorImage analysis.
Western Blot Analysis-Cells were washed twice with ice-cold PBS and incubated for 10 min on ice in MAP kinase lysis buffer (50 mM glycerol phosphate, 10 mM HEPES pH 7.4, 1% Triton X-100, 70 mM NaCl, 1 mM NaVO 4 , 1 M aprotinin, 1 M leupeptin, 1 M phenylmethylsulfonyl fluoride). The lysates were clarified by centrifugation for 10 min at 10,000 ϫ g at 4°C. Equal amounts of lysate were run on 10% polyacrylamide Tris-glycine gels (Novex Inc.) and transferred onto nitrocellulose membranes by electrophoresis. Membranes were preincubated in blocking buffer, 5% nonfat dry milk in PBST (1 ϫ PBS, 0.1% Tween 20), overnight at 4°C. The blots were incubated for 1 h at room temperature with a 1:1,000 dilution of anti-p44/42 Erk or anti-phospho-p44/42 Erk antibodies (New England Biolabs Inc.) in blocking solution. After washing three times in PBST, blots were incubated with a 1:5,000 dilution of horseradish peroxidase-conjugated anti-rabbit IgG antibody (Amersham Life Sciences) for 1 h at room temperature. The blots were washed three times in PBST and developed using the SuperSignal West Pico kit (Pierce Chemical Co.).
Cell Culture and Transfection-HepG2 cells were maintained in minimum Eagle's medium supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. ApoCIII/luciferase constructions were transfected into HepG2 cells essentially according to the manufacturer's protocol (Life Technologies, Inc.). Cells were transfected at 50 -60% confluence with 2.0 g of DNA/well into 12-well plates using Lipofectin. The internal reference plasmid pCMV␤gal was co-transfected in all experiments. Cells were transfected for 4 h in serum-free minimum Eagle's medium and allowed to recover in minimum Eagle's medium ϩ 10% fetal bovine serum for 18 h. The recovery period was followed by treatment with Me 2 SO, PD98059, or PMA as indicated, in minimum Eagle's medium ϩ 10% fetal bovine serum. Cells were harvested, and luciferase activity was measured using a commercial Dual-Light TM assay system (Tropix Inc.). Luciferase values were normalized by ␤-galactosidase internal reference plasmid.
The HNF4␣ promoter was amplified by polymerase chain reaction from a human liver genomic DNA library (CLONTECH Inc.) using primer pairs derived from the published sequence (27) (GenBank Accession numbers U72959 and U72960). The amplified product was cloned into the pGL3-basic luciferase reporter vector (Promega) and verified by sequence analysis.
Nuclear Extracts and DNA Binding Assays-Nuclear extracts were prepared from HepG2 cells treated with Me 2 SO or PD98059 essentially as described (29). Recombinant human HNF4 was produced in vitro using the Promega TNT Quick Coupled transcription/translation system. DNA binding reactions (20 l, final volume) were carried out in 20 mM HEPES pH 7.9, 60 mM KCl, 3% Ficoll, 0.5 mM MgCl 2 , 0.06% Nonidet P-40, 1 mM dithiothreitol, 1 g of double-stranded poly(dI-dC), with 50,000 cpm of labeled probe (approximately 0.25 ng) and either 3-5 g of nuclear extract or 1-2 l of TNT reaction. Reactions were incubated for 20 min at room temperature and then analyzed on 6.0% DNA retardation gels (Novex Inc.) in 0.5 ϫ TBE at 125 V for 1 h at room temperature. Supershift reactions were carried out by including antibodies in the binding reactions and extending the incubation time to 45 min. After drying, gels were analyzed by autoradiography and Phos-phorImage analysis. The sequences of the probes and competitors used for the gel shift experiments are shown below. W: GCCAGGGATGTTATCAGTGGGTCCAGAGGGCAAAATAG. L: GCCAGGGATGGCTAGCGTGGGTCCAGAGGGCAAAATAG. M: GCCAGGGATGTTATCAGCTAGCCCAGAGGGCAAAATAG.

RESULTS
To determine if apoCIII gene expression is regulated by the MAP kinase signaling pathway, HepG2 cells were treated with PD98059, an inhibitor of the upstream activator of Erk1/2 (30), and apoCIII mRNA levels were measured. The results (Fig. 1A) demonstrate that apoCIII mRNA levels increased with time after the addition of PD98059, starting at 4 h and rising to a maximum (20-fold above control) 16 h after addition of the inhibitor. Western analysis of phosphorylated Erk1/2 (the active form of the kinase) was carried out to determine the effect of PD98059 on the activity of the MAP kinase pathway in HepG2 cells. The results of this analysis (Fig. 1B) demonstrate that HepG2 cells growing in normal media contained relatively high levels of activated Erk1/2 and that PD98059 treatment rapidly reduced the amount of activated Erk1/2 to undetectable levels. To determine if MAP kinase-dependent changes in apo-CIII mRNA levels are caused by activation of transcription, HepG2 cells were transfected with an apoCIII promoter/reporter construct and treated with PD98059. The results (Fig.  1C) demonstrate that inhibition of Erk1/2 caused an increase in apoCIII transcriptional activity with a time course similar to that seen with the endogenous gene. Taken together, these results indicate that a reduction in MAP kinase signaling causes an increase in apoCIII gene transcription.
Because repression of MAP kinase signaling caused an increase in apoCIII transcription, we tested whether PMA-mediated MAP kinase activation would have the opposite effect on apoCIII expression. As shown in Fig. 2, treatment of HepG2 cells for 24 h with PMA caused a strong activation of Erk1/2 and a 5-fold reduction of apoCIII mRNA levels. This downregulation of apoCIII expression was blocked by PD98059, in-dicating that the effect of PMA on apoCIII expression was mediated by MAP kinase. The reduction of apoCIII expression by PMA was also observed with a transfected apoCIII/luciferase plasmid, indicating that the regulation was at the transcriptional level (Fig. 2C). These results are consistent with the hypothesis that apoCIII transcriptional activity is regulated by the MAP kinase signaling pathway.
To identify the region of the apoCIII promoter which mediates transcriptional regulation by MAP kinase, we tested a series of deletions for their ability to be activated by PD98059 in HepG2 cells. Deletion of sequences between Ϫ782 and Ϫ690 dramatically reduced the effect of PD98059 on apoCIII transcriptional activity (Fig. 3). Sequences between Ϫ854 and Ϫ782 and between Ϫ690 and Ϫ169 did not appear to contribute to  regulation by PD98059. A construction that linked the Ϫ782 to Ϫ690 region to the proximal part of the apoCIII promoter (pL169i, Fig. 3) was also responsive to PD98059 treatment. The residual 2-fold activation seen with constructions missing the Ϫ782 to Ϫ690 sequences appears to be mediated by sequences downstream of Ϫ169. The PD98059-responsive region (Ϫ782 to Ϫ690) was analyzed further by introduction of several point mutations within this region (Ϫ767, Ϫ755, and Ϫ704). Although none of these point mutations affected apoCIII transcription, deletion of sequences between Ϫ755 and Ϫ704 abolished the response to PD98059 treatment (Fig. 4).
To identify the specific sequences required for PD98059 regulation, three mutations were introduced between Ϫ755 and Ϫ704. These mutations, designated L, M, and N, replaced 6 base pairs in the full-length promoter with an NheI site at Ϫ746, Ϫ740, and Ϫ712, respectively. When these constructions were transiently transfected into HepG2 cells, only the M mutation completely abolished the responses to PD98059 and PMA (Fig. 5), indicating that this element is required for regulation by MAP kinase. The L mutation, which is immediately adjacent to the sequences mutated in M, partially reduced the MAP kinase response, whereas the N mutation had essentially no effect. These findings demonstrate that the activation of apoCIII transcription by PD98059 treatment in HepG2 cells requires a regulatory element located at Ϫ740 in the apoCIII promoter. We have designated this response element C3MK. We have also observed C3MK-dependent regulation of apoCIII transcription in transiently transfected Caco2 cells (results not shown). This human intestinal epithelial cell line expresses endogenous apoCIII (31) and can be transfected with the same constructions used in the HepG2 experiments. These findings suggest that the regulation of apoCIII by MAP kinase signaling is not particular to HepG2 cells.
In a previous publication a binding site for the nuclear hormone receptor HNF4 was identified in the region of the apoCIII promoter which contains the C3MK element described above (20,32). To determine if HNF4 interacts with the MAP kinase response element, gel mobility shift experiments were carried out with DNA probes containing wild-type and mutant versions of the C3MK element. Recombinant human HNF4 produced in an in vitro transcription/translation system bound strongly to the wild-type probe and to the probe containing the L mutation (Fig. 6A); however, the M mutation, which abolished MAP kinase-mediated transcriptional regulation, was completely defective for HNF4 binding. When the same probes were used with HepG2 nuclear extracts, a DNA binding protein that co-migrated with recombinant HNF4 and supershifted with anti-HNF4 antibodies also failed to bind to the M mutation (Fig. 6A, right lanes). Competition experiments with excess cold oligonucleotides representing the wild-type, L, and M sequences or a known HNF4-binding element confirmed that this band was HNF4 (Fig. 6B). Comparison of the gel shift pattern on the mutant probe with patterns seen with the competition and supershift experiments indicates that the major C3MKbinding protein in HepG2 nuclear extracts is HNF4. These results demonstrate that HNF4 binds to the C3MK element and present the possibility that it may mediate the effect of changes in MAP kinase activity.
To explore the possibility that HNF4 is modified as a result of MAP kinase activity, we examined the gel shift pattern of HNF4 in nuclear extracts prepared from HepG2 cells treated for various times with PD98059. Although no additional new bands were observed in treated extracts, there was a noticeable increase in the amount of HNF4 binding activity (Fig. 7A). Although it is difficult to quantify bands on mobility shift gels accurately, it appeared that the amount of HNF4 DNA binding activity was increased about 2-fold over control levels. These findings suggest that either the amount of HNF4 was increased after 98059 treatment or that the inhibition of MAP kinase activity caused HNF4 to be modified in such a way that it bound to DNA with a higher affinity.
To determine if treatment of HepG2 cells with PD98059 caused HNF4 levels to change, Western blot analysis with anti-HNF4 antibodies was performed on whole cell lysates prepared from 98059-treated cells. The results presented in Fig. 7B demonstrate that the quantity of HNF4 rises during PD98059 treatment by about 2-fold relative to control values. Increased HNF4 amounts could be caused by reduced rates of protein degradation or increased rates of HNF4 gene expression. Northern analysis of mRNA isolated from HepG2 cells treated with PD98059 demonstrated that HNF4 mRNA levels increased approximately 2-fold (Fig. 7B). The magnitude and time course of this increase were consistent with the change observed in HNF4 protein levels (Fig. 7B, lower panels). Together, these results indicate that treatment of HepG2 cells with PD98059 caused an increase in HNF4 gene transcription leading to increased levels of HNF4 protein. To explore this possibility, the HNF4 gene promoter was isolated and linked to a luciferase reporter gene. When HepG2 cells were transfected with this construction and treated with PD98059, a 2-fold increase in HNF4 promoter activity was observed (Fig. 8). These results confirm that the expression of the HNF4 gene is regulated by the MAP kinase signaling pathway. DISCUSSION An important determinant of the plasma triglyceride level is the quantity of apoCIII in circulation. The amount of apoCIII produced is controlled to a large degree at the level of apoCIII gene transcription. This suggests that the regulation of apoCIII transcription by metabolic and hormonal signals plays a significant role in controlling plasma triglyceride levels. In the current study we have demonstrated that activation of the Erk1/2 MAP kinase pathway negatively regulates apoCIII transcriptional activity. The MAP kinase response element (C3MK) was mapped to a site located at Ϫ740 in the apoCIII promoter, and we identified the main transcription factor binding to this site as the orphan nuclear receptor HNF4.
At least part of the explanation for the MAP kinase-mediated change in apoCIII expression appears to be an indirect effect on the expression of the HNF4 gene. Inhibition of the Erk1/2 MAP kinase pathway caused an approximately 2-fold increase in HNF4 mRNA and protein levels. It is difficult to predict how much of the fairly dramatic change in apoCIII transcriptional activity which was observed after PD98059 treatment (for example, see Fig. 2C) is the result of this relatively small increase in HNF4 protein levels. Previously we have shown that cotransfected wild-type HNF4 can increase transcription of an apoCIII/luciferase reporter by about 5-fold (21). These results suggest that changes in HNF4 levels could at least contribute to the effect on apoCIII expression seen after PD98059 treatment. On the other hand, most of the transcriptional effect caused by increased HNF4 levels mapped to a proximal HNF4 site (designated C3P) at Ϫ86 which is distinct from the C3MK. This is in contrast to the results reported in the mapping experiments presented in Figs. 3 and 4, which demonstrate that deletion of the C3MK reduces the PD98059 response from 6 -8-fold to about 2-fold. Taken together, these findings suggest that the two HNF4 sites differ in their responses to MAP kinase inhibition. The distal C3MK site responds to changes in MAP kinase activity, but the proximal C3P site does not. On the other hand, the C3P element seems to be more sensitive to changes in the quantity of HNF4 protein in the cell.
Part of the impact on apoCIII transcription could be a direct result of the MAP kinase effect on the transcriptional activity of HNF4. An obvious possibility is that Erk1/2 phosphorylates HNF4 and reduces its transcriptional activity. Phosphorylation of HNF4 could modify its activity by changing its affinity for transcriptional co-activators or co-repressors. For example, phosphorylation of the AF1 domain of the estrogen receptor by MAP kinase increases its affinity for the co-activator SRC-1 (33). Another possibility is that phosphorylation of HNF4 modifies its interaction with other transcription factors bound to the template. This could provide an explanation for the apparent differential response of the proximal and distal HNF4 elements to MAP kinase signaling. Previous studies have identified several transcription factors that interact with sequences near the C3MK element (20,32). One of these proteins is ATF-2, which has been shown to be phosphorylated by MAP kinase family members (34). Two lines of evidence indicate that ATF-2 is not mediating the effect of MAP kinase on apoCIII transcription. The first is that phosphorylation of ATF-2 by c-Jun NH 2 -terminal kinase (JNK/SAPK) or p38 subgroups of MAP kinase causes an increase in ATF-2-mediated transcriptional activity (34,35) rather than the decrease observed in our studies. Second, we have observed that purified ATF-2 can bind to the nonresponsive M mutation as well as to the wild-type sequence (results not shown). It is, however, possible that HNF4 interacts with ATF-2 in a phosphorylation-specific manner to contribute to the transcriptional effects of MAP kinase.
What is the physiological role of MAP kinase in the regulation of apoCIII transcription? Although apoCIII expression is regulated by a variety of hormonal and metabolic signals, most of them do not signal through MAP kinase. For example, the regulation of apoCIII transcription by insulin is not dependent on the Erk1/2 MAP kinase pathway. 2 On the other hand, we have observed regulation of apoCIII expression in the livers of animals treated with endotoxin. 3 This response is probably mediated by inflammatory cytokines and could potentially be mediated by MAP kinase signals. Another possibility is that this pathway is relevant to apoCIII gene expression during liver development or regeneration. ApoCIII expression is a trait of fully differentiated hepatocytes. It has been reported that MAP kinase levels are low in the fully differentiated liver (36), which would favor apoCIII expression. This might explain why apo-CIII expression is elevated in terminally differentiated hepatocytes.
A single null allele of HNF4 causes an inherited form of type II diabetes (maturity onset diabetes of the young) in the human population (27). This form of diabetes is characterized by an inability of the pancreas to secrete the proper amount of insulin in response to a glucose stimulus. This observation suggests that reduced levels or activity of HNF4 in the pancreas could contribute to the development of a pancreatic dysfunction and diabetes. It is therefore of great interest to know if the regulation of HNF4 activity by MAP kinase which we have observed in liver and intestinal cells also occurs in pancreatic ␤-cells. This issue is currently under investigation. FIG. 8. PD98059 stimulates HNF4␣ promoter activity. HepG2 cells transfected with an HNF4/luciferase reporter construction were treated with 20 M PD98059. HNF4␣ promoter activity was measured at the indicated time points. Luciferase activity was normalized by ␤-galactosidase activity with increases in transcription expressed as -fold change over a parallel Me 2 SO vehicle-treated group. Data points indicate mean Ϯ S.E. for three replicate treatments.