ROLE OF HEPATOCYTE NUCLEAR FACTOR 4α IN CONTROLLING COPPER-RESPONSIVE TRANSCRIPTION

Abstract

Previous global transcriptome and interactome analyses of copper-treated HepG2 cells identified hepatocyte nuclear factor 4α (HNF4α) as a potential master regulator of copper-responsive transcription. Copper exposure caused a decrease in the expression of HNF4α at both mRNA and protein levels, which was accompanied by a decrease in the level of HNF4α binding to its consensus DNA binding sequence. qRT-PCR and RNAi studies demonstrated that changes in HNF4α expression ultimately affected the expressions of its down-stream target genes. Analysis of up-stream regulators of HNF4α expression, including p53 and ATF3, showed that copper caused an increase in the steady-state level of these proteins. These results support a model for copper-responsive transcription in which the metal affects ATF3 expression, and stabilizes p53 resulting in the down-regulation of HNF4α expression. Additionally, copper may directly affect p53 protein levels. The suppression of HNF4α activity may contribute to the molecular mechanism underlying the physiological and toxicological consequences of copper toxicity in hepatic-derived cells.

INTRODUCTION

Copper is an essential transition metal that has important physiological roles [1, 2]. At supra-physiological concentrations, however, it can damage intracellular components through the generation of reactive oxygen species, which react with lipids, proteins, and DNA. Copper also directly binds to protein sulfhydryl and amino groups to produce structural and functional abnormalities [3–7]. To defend against copper toxicity, cells activate the transcription of a variety of genes whose products detoxify the metal and its reactive by-products, and repair intracellular damage [6, 8, 9]. Transcriptome analysis of HepG2 cells exposed to various concentrations of copper for 4, 8, 12, and 24 h reveals that at low levels of exposure, copper modulates the expression of genes associated with physiological/adaptive responses. Similarly, high levels of exposure affect genes involved in the response to stress [10].

During the analysis of the HepG2 copper transcriptome, using Ingenuity Pathways Analysis and Cytoscape, hepatocyte nuclear factor 4α (HNF4α) modules were identified as significant sub-networks (Fig. 1). This suggests that HNF4α may be an important regulator affecting the transcription of copper-responsive genes, and transmitting copper-induced stress signals to downstream target genes.

Figure 1. HNF4α networks.

(A) Ingenuity Pathway Analysis network of differentially expressed genes from HepG2 cells exposed to 600 µM copper for 24 h. Red indicates up-regulated and green indicates down-regulated genes. Cytoscape HNF4α modules of genes with >1.5-fold increase (B) or >2-fold decrease (C) in expression following an 8 h exposure to 400 µM copper. Data used in these analyses are from Song, et. al [10].

Hepatocyte nuclear factors (HNFs) are a family of transcription factors that regulate liver-specific gene expression and hepatocyte differentiation [11–14]. Some HNFs (HNF1α, HNF4α, HNF6) form core transcriptional regulatory circuitry with other transcription factors including CREB1, USF1, and FOXA2 [15]. HNF4α expression is essential for early embryonic development, hepatocyte differentiation, and liver-specific gene expression [13, 14, 16]. Chromatin immunoprecipitation and promoter microarray analysis indicate that HNF4α controls the expression of 1,829 genes in human hepatocytes [15]. In addition, in the human Biomolecular Interaction Network Database, the HNF4α sub-network consists of 2,313 genes and 2,606 interactions [17]. Potential HNF4α target genes have been identified and include those encoding apolipoproteins, blood coagulation factors, and enzymes involved in lipid, amino acid and glucose metabolism [18–20]. Many of these target genes were also identified in the copper transcriptome [10]. Protein binding microarrays identified more than 1,400 new HNF4α binding sequences and in combination with Support Vector Machine models, >240 new HNF4α human target genes.

Based on copper transcriptome analyses, the molecular mechanism of copper-induced down-regulation in HNF4α expression in HepG2 cells was investigated. A detailed examination of the effects of copper on HNF4α mRNA and protein levels revealed that copper caused decreases in the steady-state levels of HNF4α expression and activity. It has been reported that p53 down-regulates HNF4α expression [21]. The mechanism by which copper-induced stress and p53 and ATF3 activity affects HNF4α expression was also explored. The results support a model in which HNF4α is a master regulator of copper-responsive transcription. In addition, HNF4α down-regulation by copper leads to modulations in expression of its target genes. The present study is the first to demonstrate that the HNF4α is associated with the molecular mechanism of copper toxicity in human hepatocytes.

MATERIALS AND METHODS

Cell culture and RNA interference

HepG2 cells (Human hepatoma cell line, ATCC# HB-8065) were grown in Minimal Essential Medium, supplemented with 10 % heat-inactivated fetal bovine serum, 100 µM nonessential amino acids, 1 mM sodium pyruvate, 100 units/ml penicillin, and 100 µg/ml streptomycin (Invitrogen/Life Technologies, Carlsbad, CA). Cells were maintained in a humidified incubator at 37°C under 5 % CO₂.

For RNA interference (RNAi), HepG2 cells were reverse transfected with single duplex siRNA of ATF3, HNF4α (Dharmacon, Lafayette, CO), p53 (Qiagen, Inc., Valencia, CA), or non-homologous siRNA at final concentrations of 50 nM with Lipofectamine 2000. Briefly, transfection complexes of siRNA and Lipofectamine 2000 were prepared following manufacturer’s instructions (Invitrogen) and dispensed into 6-well plates. HepG2 cells were then added to wells that contained transfection complexes. Following 44 h incubation, cells were treated with 400 µM copper, as copper sulfate (Sigma-Aldrich, St Louis, MO) for 4 h.

RNA isolation and quantitative real-time PCR

For quantitative real-time PCR (qRT-PCR), total RNA from three or more independent transfections was isolated from untreated and treated cells using RNeasy mini kits following manufacturer’s instructions (Qiagen). cDNAs were then generated for two-step qRT-PCR using the SuperScript® First-Strand Synthesis System for RT-PCR according to manufacturer’s instructions (Invitrogen). cDNAs were subsequently used in qRT-PCR using Power SYBR Green RT-PCR kits according to manufacturer’s instructions (Applied Biosystems, Foster City, CA). Quantitative PCR was performed using an ABI 7900 HT Fast Real-Time System (Applied Biosystems). The fold changes in mRNA levels were calculated using the ΔΔCt method with β-actin as reference mRNA [22, 23]. PCR primers for HNF4α, TXNRD1 (thioredoxin reductase 1), G6PC (Glucose-6-phosphatase, catalytic subunit), ATF3 (activating transcription factor 3), LIPA (lipase A), APOC1 (apolipoprotein C1), APOM (apolipoprotein M) and p53 were purchased as QuantiTect® Primer Assays (Qiagen). Primers for UCHL1 (ubiquitin carboxyl-terminal esterase L1) and β-actin (Table 1) were designed using Primer3 [24] and purchased from Integrated DNA Technologies (Coralville, IA).

TABLE 1.

Sequences of primers used for qRT-PCR

Gene	Sequence
UCHL1	Forward	5'-ACCGAGCGTGAGCAAGGAGAAGTC-3'
	Reverse	5'-GAAGGGAAGAGGGGAAATCAGC-3'
β-actin	Forward	5'-GATATCGCTGCGCTGGTCGTC-3'
	Reverse	5'-ACGCAGCTCATTGTAGAAGGTGTGG-3'

Western immunoblot analysis

Whole cell lysates were prepared for Western immunoblot analysis by adding SDS sample buffer (62.5 mM Tris–HCl (pH 6.8), 2 % (w/v) SDS, 10 % (v/v) glycerol and 50 mM DTT) to PBS-washed cells and sonicating the mixture for 10–15 sec. The lysate was then centrifuged for 3 min at 14,000 × g at 4°C, and the supernatant was collected. Nuclear extracts were prepared using Active Motif (Carlsbad, CA) nuclear extract kits following manufacturer’s instructions. Protein concentrations were determined using the Bradford dye-binding assay (Bio-Rad Laboratories, Hercules, CA). Proteins were resolved on 4–12 % bis-Tris gels (NuPAGE Novex, Invitrogen) and then transferred to PVDF membranes. Immunoblotting was performed using polyclonal antibodies against HNF4α, p53, β-actin, tubulin (Cell Signaling Technology, Danvers, MA), ATF3 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), and histone 3 (Millipore, Billerica, MA). HPR-conjugated secondary antibodies were obtained from Cell Signaling Technology. Antigen-antibody complexes were visualized using enhanced chemiluminescence (GE Healthcare, Piscataway, NJ) following exposure to X-ray film. Target protein levels were measured by densitometry and quantified using Image J software [25]. Data were expressed as fold-change of treated samples compared to untreated cells, normalized to the level of β-actin, tubulin, or histone 3. Each experiment was repeated at least three times.

HNF4α binding assay

HepG2 cells were grown to 70 – 80 % confluence and then treated with 100, 200, 400, or 600 µM copper for 0.5, 1, 2, 4, or 8 h and nuclear extracts were prepared. Protein binding to the HNF4α consensus binding motif (5′-TGGACTTAG-3′) was measured using TransAM HNF family transcription factor assay kit according to the manufacturer’s instructions (Active Motif). Briefly, 8 µg of nuclear protein was added to wells containing a mixture of immobilized oligonucleotides that contained HNF-1, HNF-3, and HNF-4 consensus binding sequences and incubated at room temperature. Wells were washed and then anti-HNF4α antibody was added. Following a 1 h incubation and washing, HRP-conjugated secondary antibody was added and allowed to incubate for 1 h at room temperature. Wells were washed, and the developing solution was added. Reactions were terminated and then the absorbance at 450 nm measured using a FLUOstar OPTIMA plate reader (BMG Labtech, Durham, NC). To confirm the specificity of HNF4α binding, competition studies were performed using oligonucleotides with wild-type (Active Motif) or mutant (Santa Cruz Biotechnologies, Inc) HNF4α consensus sequences, according to manufacturer’s instructions.

Statistical Analysis

All statistical analyses were performed using StatView software (SAS Institute Inc., Cary, NC). The results are presented as the mean ± standard error. The significance of mean differences was detected by analysis of variance (ANOVA) followed by Fisher's Protected Least Squares Differences post hoc test for individual comparisons.

RESULTS

Effect of HNF4α and copper on HNF4α-regulated gene expression

The effect of decreased HNF4α expression on steady-state mRNA levels of several cognate target genes was determined using qRT-PCR. Specific HNF4α- regulated genes were selected based on the previous analysis of the HepG2 copper transcriptome [10]. Target genes based on pathway and network analysis included UCHL1, G6PC and TXNRD1; and pathway mapping and literature mining included LIPA, APOC1, and APOM [18].

Transcriptome analysis showed that LIPA, APOC1, and APOM were significantly down-regulated following copper treatment [10]. Consistent with previous microarray data, mRNA levels of LIPA, APOC1 and APOM were 66 % (p = 0.0048), 67 % (p = 0.0093) and 76 % (p = 0.0085) of that observed in non-treated cells, respectively, following a 4 h treatment with 400 µM copper (Table 2). In addition, knocking down expression of HNF4α using RNAi resulted in significant decreases in steady-state mRNA levels of LIPA, APOC1, and APOM; 28 % (p < 0.0001), 60 % (p = 0.0018), and 52 % (p < 0.0001) relative to cells transfected with a non-homologous control siRNA, respectively (Table 2). A combination of HNF4α siRNA and copper further down-regulated LIPA, APOC1, and APOM mRNA levels to 12 % (p = 0.0048), 43 % (p = 0.0093), and 36 % (p = 0.0085) of that observed in control cells, respectively. These levels were significantly lower than those observed in cells exposed to copper alone (Table 2).

TABLE 2.

Effect of copper and decreased HNF4α expression on gene expression

Gene	Expression Level (%)a
	Control siRNA+ 400 µM Copper	HNF4α siRNA	HNF4α siRNA + 400 µM Copper	HNF4α leveld
LIPA	66 ± 14b	28 ± 5c	12 ± 3b,c	22 ± 2
APOM	76 ± 8b	52 ± 8c	36 ± 8b,c
APOC1	67 ± 9b	60 ± 11c	43 ± 10b,c
G6PC	23 ± 3b	33 ± 8c	7 ± 2b,c	18 ± 3
UCHL1	324 ± 51b	155 ± 57	580 ± 135b	33 ± 8
TXNRD1	296 ± 30b	54 ± 6c	171 ± 23 b,c	34 ± 7

^aValues are expressed as percent ± SE relative to mRNA levels measured in cells transfected with non-homologous control siRNA

^bSignificantly different from sample without copper (p < 0.05)

^cSignificantly different from non-homologous control (p < 0.05)

^dThese values present the mean percent of HNF4α mRNA remaining following RNAi treatment for the indicated target gene(s).

The promoter of G6PC, a key enzyme in glucose homeostasis, contains an HNF4α consensus binding sequence. It has been suggested that polyunsaturated fatty acyl CoA suppresses G6PC transcription by modulating HNF4α DNA binding activity [26]. Decreased HNF4α expression caused a significant reduction in steady-state G6PC mRNA levels, 33 % relative to cells transfected with non-homologous control siRNA (p < 0.0001). Exposure to copper decreased G6PC expression to 23 % relative to control (p < 0.0001) and the combination of HNF4α siRNA and copper significantly reduced G6PC mRNA levels to 7 % of that observed in control cells (p = 0.0003) (Table 2).

HNF4α was reported to be a negative regulator of UCHL1 transcription in human hepatoma HuH7 cells [27]. Thus, decreased expression of HNF4α in HepG2 cells should produce an increase in the steady-state mRNA level of UCHL1. UCHL1 expression showed a non-significant (p = 0.0586) 155 % increase, relative to control cells when HNF4α expression was reduced (Table 2). However, HepG2 copper transcriptome analysis showed that copper significantly up-regulated UCHL1 expression more than 1.5-fold in 8 out of 16 treatment conditions [10]. Exposure to 400 µM copper for 4 h caused a statistically significant (p = 0.0005) 324 % increase in UCHL1 mRNA levels. A combination of reduced HNF4α expression and copper treatment caused a significant 580 % increase in UCHL1 expression. These results confirmed that UCHL1 is a copper-responsive gene however, the effect of HNF4α on UCHL1 expression was not clear.

In cells exposed to 400 µM copper for 4 h, TXNRD1 was identified as a significant sub-network of the HNF4α module [10] (Fig. 1B). Inhibition of HNF4α expression caused a significant (p = 0.0002) decrease in TXNRD1 expression, 54 % relative to control cells (Table 2). Treatment with copper alone caused a 296 % increase in TXNRD1 expression (p < 0.0001) while both decreased HNF4α expression and copper treatment produced a 171 % increase in TXNRD1 expression (p = 0.0506). These results suggest that copper affects TXNRD1 expression by a mechanism that may be independent of HNF4α.

Effect of copper on HNF4α expression

The previous results indicate that HNF4α contributes to the regulation of selected target genes and that copper also modulates the expression of these genes. To directly assess the effect of copper on HNF4α expression, HepG2 cells were exposed to various concentrations of copper (0, 100, 200, 400 and 600 µM) for 4 h. Exposure to copper concentrations above 200 µM produced significant decreases in HNF4α mRNA levels (Fig. 2A). Exposure to 200, 400, and 600 µM copper reduced HNF4α mRNA levels to 67 %, 46 %, and 33 % of that observed in non-exposed cells, respectively.

Figure 2. Effect of copper on HNF4α mRNA and protein expression.

(A) Steady-state HNF4α mRNA levels in HepG2 cells exposed to 0, 100, 200, 400 and 600 µM copper for 4 h. *indicates significant difference from control (p<0.005; n = 4). (B) Steady state levels of HNF4α mRNA, measured using qRT-PCR (gray bars; n = 10), and protein levels, determined by Western immunoblotting of whole cell lysates (open bars; n = 3). Data are expressed as mean percent ± standard error (SE) relative to cells not exposed to copper. * indicates significant difference relative to cells transfected with control siRNA (p < 0.0001) # indicates significant difference relative to cells transfected with control siRNA and exposed to copper (p < 0.0001). (C) Representative Western immunoblot of whole cell lysates probed with antibodies to HNF4α and β-actin.

Transfection of HepG2 cells with HNF4α siRNA reduced the steady-state HNF4α mRNA level to 19 % of that observed in cells transfected with non-homologous control siRNA (Fig. 2B). The level of HNF4α mRNA in HNF4α siRNA transfected cells exposed to 400 µM copper for 4 h was only 9 % of that observed in control cells. In these cells, the level of HNF4α mRNA expression was significantly lower than that observed in cells treated with copper or transfected with HNF4α siRNA alone.

Comparable results were obtained when HNF4α protein levels were measured by Western immunoblotting. Exposure of HepG2 cells to 400 µM copper significantly reduced HNF4α protein levels to 40 % of that observed in non-copper treated cells (Fig. 2B and C). Reduction of HNF4α expression using siRNA (14 % of control levels) significantly lowered HNF4α protein levels to 5.9 % of that observed in cells transfected with control siRNA. The addition of 400 µM copper to HNF4α siRNA treated cells resulted in a further significant reduction of HNF4α protein levels to 3.3 % of that observed in control cells (Fig. 2B and C). These results indicate that treatment with copper reduces both HNF4α protein and mRNA levels.

Effect of copper on HNF4α DNA binding activity

DNA binding assays were performed to determine if copper exposure would lead to a decrease in HNF4α binding to its consensus binding motif. HepG2 cells were exposed to copper (0, 200, 400 and 600 µM) for 2 h. This exposure time was chosen based on preliminary binding assay data (Suppl. Fig. 1). The level of HNF4α binding decreased in a dose-dependent manner: HNF4α binding in cells exposed to 400 and 600 µM copper was 53 % and 18 % of that observed in control cells, respectively (Fig. 3). HNF4α binding also decreased in a time-dependent manner at 400 and 600 µM copper (Suppl. Fig. 1). The specificity of HNF4α binding in this assay was confirmed by competition with oligonucleotides that contained wild type and non-functional HNF4α consensus sequences (results not shown).

Figure 3. Effect of copper on HNF4α binding activity.

Nuclear extracts were prepared from HepG2 cells exposed to 0, 200, 400, and 600 µM copper for 2 h. Data are expressed as means ± SE relative to cells not exposed to copper (n = 3) * indicates significantly different from non-treated cells p < 0.001.

Effect of copper on p53-mediated HNF4α expression

The transcription of human HNF4α is reported to be negatively regulated by p53 [21]. To confirm this observation in HepG2 cells the steady-state level of HNF4α mRNA was determined following knock-down of p53 expression using siRNA. An 84 ± 3.5 % reduction in p53 expression resulted in a significant 197 ± 29.5 % increase in the level of HNF4α mRNA. This result confirmed that p53 regulated HNF4α expression in HepG2 cells.

In the HepG2 copper transcriptome, metal exposure did not significantly affect p53 mRNA levels at any concentration or exposure time examined [10]. These results are consistent with previous observations that demonstrated p53 activation results from protein stabilization and post-translational modifications rather than changes in gene expression [reviewed in 28]. To investigate the role of p53 in copper-induced down-regulation of HNF4α expression, the effect of copper on p53 protein expression was determined. Whole cell lysates and nuclear extracts were prepared from HepG2 cells treated with copper (0, 100, 200, 400, or 600 µM) for 4 h. The level of p53 increased in a dose-dependent manner in both whole cell lysates and nuclear extracts (Fig. 4A and B, respectively). In nuclear extracts, HNF4α protein decreased 1.28-, 2.27-, and 2.44-fold when exposed to 200, 400, and 600 µM copper, respectively (Fig. 4C; Supplemental Table 1). The levels of p53 increased 1.67-fold relative to controls following exposure to 400 µM copper and 2.56-fold after exposure to 600 µM copper (Fig. 4C; Supplemental Table 1). These results support a model where p53 mediates copper-induced down-regulation of HNF4α.

Figure 4. Effect of copper on HNF4α, ATF3, and p53 – Concentration response.

Representative Western immunoblot of (A) whole cell lysates or (B) nuclear extracts prepared from HepG2 cells exposed to 0, 100, 200, 400, and 600 µM, or 0 and 400 µM copper for 4 h, respectively. (C) Target nuclear protein levels were quantified by densitometry using Image J software, and normalized against those of histone 3. Data are expressed as fold change of treated samples over that of non-treated sample. * indicates significantly different from non-treated sample (p < 0.01).

The steady-state level of p53 is regulated in part by ATF3 [29]. To further define a mechanism by which copper can affect the levels of HNF4α, the effect of copper exposure on ATF3 protein expression was determined. ATF3 protein levels in both whole cell lysates and in nuclear extracts increased in a dose-dependent manner when HepG2 cells were exposed to copper for 4 h (Fig. 4A and B). In nuclear extracts, ATF3 protein levels significantly increased 15.5- and 20.4-fold following exposure to 400 µM and 600 µM copper, respectively (Fig. 4C; Supplemental Table 1). This is in agreement with previous expression profiling results showing that copper significantly up-regulated ATF3 mRNA levels [10].

To further investigate the role of ATF3 in mediating p53 protein activation/stabilization in response to copper, the effect of exposure time on ATF3, p53 and HNF4α protein levels was determined. ATF3 and p53 protein levels remained constant until 4h where significant increases in protein levels were observed. A significant decrease HNF4α protein level was only observed following 8h exposure (Fig. 5). These results support a model in which copper can affect HNF4α-mediated transcription through ATF3 and then p53:

Figure 5 . Effect of copper on HNF4α, ATF3 and p53 – Time response.

(A) Whole cell lysates wee prepared from HepG2 cells exposed to 400 µM copper for 1, 2, 4 and 8 h. ATF3 (■), p53 (◆), HNF4α (▲) protein levels were quantified and normalized against those of tubulin. Data are expressed as fold change of treated sample over that of non-treated sample of each exposure time. * (p < 0.05) and # (p < 0.001) indicate significantly different from 1 h-treated sample. (B) Representative Western immunoblot of whole cell lysates probed with antibodies to ATF3, p53, HNF4α, and tubulin.

Effect of ATF3 on HNF4α and p53

To further explore the role of ATF3 in p53 activity, the effect of ATF3 RNAi on copper-mediated activation of p53 and HNF4α expression was examined. Transfection of HepG2 cells with ATF3 siRNA resulted in a significant (50 ± 6.1 %) decrease in ATF3 and a non-significant (40 %) increase in p53 protein levels relative to controls. Treatment with ATF3 siRNA produced a non-significant 20 % increase in HNF4α protein levels. Similarly, ATF3 siRNA treatment did not significantly affect HNF4α mRNA levels (results not shown).

DISCUSSION

Several studies have examined changes in global gene expression associated with exposure to elevated levels of copper. These studies have lead to a better understanding of the molecular mechanisms underlying the physiological and toxicological processes affected by copper [23, 30, 31]. A common observation from toxicogenomic analyses is the ability of chemicals and toxicants to affect the expression of hundreds to thousands of genes. In the transcriptome of HepG2 cells, copper affected the steady state level of 2,257 mRNAs: 1,088 with increased and 1,169 with decreased levels of expression [10]. The expression of many of these genes could be directly linked to metal exposure (e.g., metallothioneins) or metal-induced stress (heat shock proteins, glutamate cysteine ligase). For a majority of the genes, however, direct links between copper exposure and affected genes were not apparent. This may be due to the ability of metals to affect the activity of central transcription factors, which regulate the expression of hundreds of genes. Bioinformatic analysis of copper-responsive genes identified HNF4α as a potential central transcription factor. The HNF4α module for copper-responsive genes contained 242 up-regulated and 498 down-regulated genes (Fig. 1). This represented ~30 % of the total genes in the HNF4α module [10]. A map of the transcriptional regulatory circuitry of human hepatocytes shows that HNF4α is one of the six master regulators [15].

RNAi and Western immunoblot analyses demonstrated that copper could suppress HNF4α protein expression, which ultimately affected the transcription of several target genes. The ability of copper to suppress HNF4α expression was partially controlled by p53. Over expression of p53 suppresses HNF4α protein levels [21]. In the present study, immunoblot analysis showed that copper exposure caused a significant increase in p53 protein levels. In MCF-7 cells, p53 protein levels also increased following copper exposure [32].

Copper-inducible expression of p53 may be regulated by ATF3. ATF3 is a member of the ATF/CREB family of transcription factors and is a stress-inducible gene [33]. ATF3 positively regulates p53 levels by increasing protein stability [29]. ATF3 transcription is induced by a variety of agents including growth-stimulating factors, cytokines, genotoxic agents, signaling molecules, bacterial products, and viral proteins [reviewed in 34, 35]. Copper significantly increased ATF3 mRNA levels 1.2 to 19.3-fold relative to non-exposed HepG2 cells, which confirmed ATF3 as a copper-responsive gene [10]. Western analysis of copper-treated HepG2 cells showed ATF3 protein levels increased as a function of metal concentration and exposure time. p53 protein levels showed a similar pattern of increase, although ATF3 protein levels appeared to be ~10-fold higher than those of p53 at identical copper concentrations. Time-course analysis showed ATF3 and p53 protein levels remained constant until 4 h after copper exposure; after which levels significantly increased. HNF4α protein levels increased only after ATF and p53 protein accumulated. These results suggest a kinetic relationship between ATF3 and p53 expression, and down-regulation of HNF4α. These observations and previous reports support a model in which copper exposure causes an increase in ATF3 protein levels that results in the stabilization of p53. The elevation in p53 causes a decrease in HNF4α levels, which ultimately affects transcription of its downstream target genes (Fig. 6).

Figure 6. Model describing the mechanism by which copper can affect HNF4α-mediated transcription.

In this model exposure to copper causes an increase in ATF3 activity/levels or induces DNA damage that produces an increase in p53. Elevated p53 down-regulates HNF4α levels which ultimately affects the transcription of its target genes.

Reduction of ATF3 expression did not significantly affect HNF4α and p53 levels (data not shown). This may be due to the low efficiency (<50 %) in the suppression ATF3 expression. An alternative explanation was that copper affected p53 levels via an ATF3-independent mechanism (Fig. 6). Exposure of cells to elevated concentrations of copper leads to oxidative DNA damage that could activate p53 [6, 36, 37]. Alternatively, copper could directly promote post-translational modifications of p53. p53 is stabilized by different stresses through post-translational modifications, including phosphorylation, acetylation, methylation, ubiquitination and sumoylation [reviewed in 38]. Toxic concentrations of copper produce intracellular stress that could induce some of these post-translational modifications. Transcriptome analysis of copper-treated HepG2 cells showed that p300 was significantly up-regulated (1.4 – 1.7 fold) by 400 and 600 µM copper [10]. Considering that p300 can acetylate p53 [38], toxic concentration of copper could induce p53 acetylation at through p300. Copper may also directly act on p53 to affect its activity. In vitro studies, using in vitro translated p53 and cell-free systems, demonstrated that copper could induce conformational changes in p53 and inhibit its DNA-binding activity at physiological concentration (< 30 µM) [39, 40].

Under normal conditions humans maintain homeostatic levels of copper. Individuals with Wilson Disease, however, accumulate toxic levels of hepatic copper. Wilson disease is caused by mutations in the gene coding the ATPase copper pump, ATP7B, which transports copper out of the cytoplasm [41]. As a result, copper accumulates to toxic levels in hepatocytes leading to steatosis, inflammation, cirrhosis, and ultimately liver failure [reviewed in 42, 43]. Recent molecular mechanistic studies on the pathology of Wilson disease indicate a link between copper overload and a disruption in lipid metabolism [44, 45]. Copper selectively down-regulates genes associated with lipid metabolism and transport in the Wilson disease model Atp7b^−/− mice. Affected genes include low density lipoprotein receptor (LDLR) and HMG-CoA reductase, both of which are regulated by SREBP-2 (sterol regulatory-element binding protein 2). Huster and Lutsenko (2007) suggested that cross-talk between copper and lipid metabolism may involve several mechanisms including modulation of SREBP-2 activity [44]. This suggests that elevated copper may inhibit SREBP-2 activity. SREBP-2 activity is affected by HNF4α. Overexpression of HNF4α causes an increase in SREBP-2 activity and both proteins were found to directly interact with each other [46, 47]. Thus, the ability of copper to negatively affect lipid metabolism may be due to copper-induced inhibition of HNF4α.

Further support of this mechanism for copper-induced disruption of lipid metabolism comes from studies of Hnf4α^−/− mice [18]. These mice present liver steatosis and reduced serum cholesterol levels, which are phenotypes similar to those observed in individuals with Wilson disease and Atp7b^−/− mice [42, 43, 45, 48]. Hnf4α^−/− mice also show reduced expression of genes involved in several pathways of lipid metabolism and transport [18]. qRT-PCR and RNAi results clearly demonstrated that decreased HNF4α expression and copper down-regulated the expressions of LIPA, APOC1 and APOM, which are involved in maintaining lipid homeostasis (Table 2). In addition, transcriptome analysis of HepG2 cells exposed to toxic levels of copper showed down-regulation of gene associated with metabolism and transport of biomolecules including lipids [10]. Taken together, these results indicate that copper down-regulates genes associated with lipid metabolism and transport through the down-regulation of HNF4α. Furthermore, HNF4α may be an important factor linking copper overload to the disruption in lipid metabolism.

Several lines of evidence support the hypothesis that HNF4α is a master regulator of copper-responsive transcription. HNF4α down-regulation by copper leads to altered expression of hundreds of target genes. This in turn may define the molecular mechanism underlying the physiological and toxicological consequences of copper toxicity. In addition, these results suggest that the physiological changes in lipid metabolism associated with copper overload/toxicity may be controlled by HNF4α. These observations provide novel insights into the molecular mechanism of human copper toxicity.

Supplementary Material

01. Supplemental Figure 1. Effect of copper on HNF4α binding activity.

HNF4α binding assays were performed with TransAM HNF family transcription factor assay kit. Nuclear extracts were prepared from HepG2 cells following exposure to copper 0 (◆), 200 (■), 400 (▲) and 600 (●) µM) for 0.5, 1, 2, 4, or 8 h. Data are expressed as means ± SE.

Abbreviations

HNF4α

Hepatocyte Nuclear Factor 4α

ATF3

Activating Transcription Factor 3

IPA

Ingenuity Pathway Analysis

RNAi

RNA interference

siRNA

small interfering RNA

qRT-PCR

quantitative real-time PCR

SREBP-2

sterol regulatory-element binding protein 2