Expression profile analysis of the inflammatory response regulated by hepatocyte nuclear factor 4α

Abstract

Background: Hepatocyte nuclear factor 4α (HNF4α), a liver-specific transcription factor, plays a significant role in liver-specific functions. However, its functions are poorly understood in the regulation of the inflammatory response. In order to obtain a genomic view of HNF4α in this context, microarray analysis was used to probe the expression profile of an inflammatory response induced by cytokine stimulation in a model of HNF4α knock-down in HepG2 cells.

Results: The expression of over five thousand genes in HepG2 cells is significantly changed with the dramatic reduction of HNF4α concentration compared to the cells with native levels of HNF4α. Over two thirds (71%) of genes that exhibit differential expression in response to cytokine treatment also reveal differential expression in response to HNF4α knock-down. In addition, we found that a number of HNF4α target genes may be indirectly mediated by an ETS-domain transcription factor ELK1, a nuclear target of mitogen-activated protein kinase (MAPK).

Conclusion: The results indicate that HNF4α has an extensive impact on the regulation of a large number of the liver-specific genes. HNF4α may play a role in regulating the cytokine-induced inflammatory response. This study presents a novel function for HNF4α, acting not only as a global player in many cellular processes, but also as one of the components of inflammatory response in the liver.

Figure 1

Knock-down of HNF4α in HepG2 cells. HepG2 cells were transfected with non-specific shRNA control or HNF4α shRNA plasmid. mRNA and whole cell lysates were prepared for real-time PCR (A) and Western blots (B), respectively. The results shown in (A) represent the relative mRNA expression level normalized to GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of 4 replicates. An HNF4α antibody (sc-6556, Santa Cruz Biotechnology) and β-actin (Sigma) antibody, used as an internal loading control, were utilized for Western blot (B).

Figure 2

The clustering expression profiles of up- and down-regulated genes. A global transcriptional view of HepG2 cells in response to the treatment with cytokines and HNF4α knock-down alone (2 and 3) or in combination (4) is shown. Each group has 4 replicates. Relative expression values are expressed as a color code (bar color chart on the bottom, red up- and blue down-regulation). (A), Category A contains genes that are significantly regulated by HNF4α knock-down, but not by cytokines; (B), Category B contains genes that respond to cytokine treatment, but not to HNF4α knock-down; (C), Category C contains genes whose expression is dependent on both treatments with cytokines and HNF4α knock-down in an independent manner; (D), Category D has genes that exhibit an expression pattern dependent on both treatments in an interactive manner.

Figure 3

Inflammatory response genes are enriched in two clusters within the interactive category (Category D). Expression profiles of up- and down-regulated genes in the different groups treated by cytokines and HNF4α knock-down alone (2 and 3) or in combination (4) are shown. Each group has 4 replicates. Relative expression values are expressed as a color code (bar color chart on the bottom, red up- and blue down-regulation). Inflammatory response genes (listed at the right side of the graph) extracted from GO are highly enriched in two different clusters (A and B). Gene abbreviations: VNN1, vanin 1; SAA1/SAA2, serum amyloid A1/serum amyloid A2; C5; complement component 5; CX3CL1, chemokine (C-X3-C motif) ligand 1; LBP, lipopolysaccharide binding protein; PTX3, pentraxin-related gene, rapidly induced by IL-1 beta; PTAFR, platelet-activating factor receptor; NMI, N-myc (and STAT) interactor; ORM1/ORM2, orosomucoid 1/orosomucoid 2; C4A/C4B, complement component 4A/complement component 4B; CEBPB, CCAAT/enhancer binding protein (C/EBP), beta; SERPINA3, serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 3; CCL20, chemokine (C-C motif) ligand 20; HIF1A, hypoxia-inducible factor 1, alpha subunit (basic helix-loop-helix transcription factor).

Figure 4

Confirmation of microarray results using quantitative real-time PCR. (A), Raw intensity values were measured using microarray. (B), Real-time PCR results for the same genes are expressed as the relative mRNA expression level normalized by GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of 4 replicates. Ctr, Control; Cyto, Cytokines; shRNA, HNF4α shRNA; shRNA+Cyto, HNF4α shRNA plus Cytokines. *p < 0.05 and **p < 0.01 indicate a significant difference compared to control. Gene abbreviations are the same as described in Figure 3 legend.

Figure 5

Gene ontology analysis. Twenty six broad categories of biological processes were analyzed for genes that exhibit HNF4α-dependent expression. Significance levels are plotted as-log (p value). One-tail p-values were calculated using the Fisher Exact test. Threshold (line) denotes the p = 0.002 level, which is the threshold for significance after Bonferroni correction.

Figure 6

Sequence logos of ETS transcription factor binding sites. Sequence logos of consensus DNA binding sites for the three ETS transcription factors enriched in genes regulated by HNF4α. Y axis indicates amount of information at each position in the motif. These logos were generated from information obtained from the JASPAR core database [42]. (A), ELK1 binding motif; (B), ELK4 binding motif; and (C), GABPA binding motif.

Figure 7

The ELK1 expression in HepG2 cells with HNF4α or ELK1 knock-down. HepG2 cells were transfected with non-specific siRNA control (siControl), HNF4α shRNA plasmid (siHNF4) or siELK1, and then treated with or without cytokines. The expression of ELK1 was measured by real-time PCR. The results represent the relative mRNA expression level normalized to GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of three different experiments. *p < 0.05 and **p < 0.01 indicate a significant difference compared to siControls.

Figure 8

A decrease in ELK1 expression leads to elimination of the regulatory effect of HNF4α knock-down on a subset of genes. HepG2 cells were treated with non-specific siRNA control (siControl), HNF4α shRNA plasmid (siHNF4), siELK1 or both siHNF4 and siELK1. The expressions of COL4A1 (collagen, type IV, alpha 1), ZNF175 (zinc finger protein 175), MMP15 (matrix metallopeptidase 15) and SEC24A (SEC24 family, member A) genes were determined by real-time PCR. The abundance of mRNA in the siControl was set at 1. Data are presented as mean ± SD of three different experiments. *p < 0.05 and **p < 0.01 indicate a significant difference compared to siControl. #p < 0.01 indicates a significant difference compared to siHNF4α.

Figure 9

The reduction of ELK1 level decreases the ability of HNF4α to interact with the promoters of a subset of genes. (A), HepG2 cells were transfected with siControl (siCtr) or siELK1. Protein interaction of HNF4α and ELK1 was determined by ChIP assay with either antibody against HNF4α or goat normal IgG (IgG). Chromatin-immunoprecipitated DNA was analyzed by PCR with primers specific for the ELK1 binding sites in the promoter regions of COL4A1, ZNF175, MMP15 and SEC24A genes. The result shown in (A) is a representative experiment, replicated three times with similar results. (B), Histograms show densitometric analyses of relative binding abilities. Values represent mean ± SD of three separate experiments. The relative quantitative analysis was carried out by comparison of siELK1 with siControl, and the siControl was set at 1. *p < 0.05 and **p < 0.01 indicate a significant difference compared to siCtr. IP, immunoprecipitation.