HNF4α regulates claudin-7 protein expression during intestinal epithelial differentiation

Abstract

The intestinal epithelium is a dynamic barrier that maintains the distinct environments of intestinal tissue and lumen. Epithelial barrier function is defined principally by tight junctions, which, in turn, depend on the regulated expression of claudin family proteins. Claudins are expressed differentially during intestinal epithelial cell (IEC) differentiation. However, regulatory mechanisms governing claudin expression during epithelial differentiation are incompletely understood. We investigated the molecular mechanisms regulating claudin-7 during IEC differentiation. Claudin-7 expression is increased as epithelial cells differentiate along the intestinal crypt-luminal axis. By using model IECs we observed increased claudin-7 mRNA and nascent heteronuclear RNA levels during differentiation. A screen for potential regulators of the CLDN7 gene during IEC differentiation was performed using a transcription factor/DNA binding array, CLDN7 luciferase reporters, and in silico promoter analysis. We identified hepatocyte nuclear factor 4α as a regulatory factor that bound endogenous CLDN7 promoter in differentiating IECs and stimulated CLDN7 promoter activity. These findings support a role of hepatocyte nuclear factor 4α in controlling claudin-7 expression during IEC differentiation.

Figure 1

Increased claudin-7 expression during intestinal epithelial cell (IEC) differentiation. A: IECs proliferate in the base of colonic crypts and migrate toward the lumen while undergoing differentiation. Protein expression changes of claudin-2, -15, -4, and -7 are shown in mouse colonic mucosal cryosections by immunofluorescence labeling and laser confocal microscopy showing crypt–luminal gradients. B: Immunofluorescence labeling of claudin-7 in differentiating Caco-2 cells shows increased expression over time in culture. Claudin-7 (red) colocalizes with zonula occludens protein 1 (green) at the tight junction in differentiated IECs. Reconstituted z-sections are shown under the individual micrographs with the blue line showing the optical section where the micrographs were taken. C: Transepithelial electric resistance (TER) of differentiating Caco-2 monolayers increases over time in culture. The TER in cells treated with small interfering RNA against claudin-7 (siCLDN7 versus nonsilencing RNA) decreases the resistance significantly. D: Claudin-7 increases in Caco-2 model epithelial cells during differentiation. The expression of differentiation markers p21 and Cdx2 also increases in the same samples. In contrast, levels of the crypt base–specific claudin-2 protein decrease during model IEC differentiation. Representative immunoblots are shown from one of three independent experiments. The bar graph shows densitometry (means ± SEM) of claudin-7 immunoblots normalized to β-actin from four independent experiments. E: Claudin-7 protein levels in HT29B6 cells 2 and 5 days after seeding. The bar graph shows the means ± SEM of four independent claudin-7 immunoblots normalized to calnexin. Means ± SEM of three experiments are shown (C). ^∗P < 0.05. Scale bars: 25 μm (A); 10 μm (B).

Figure 2

CLDN7 RNA synthesis increases during differentiation. A: CLDN7 mRNA and heteronuclear RNA (hnRNA) levels were detected in Caco-2 intestinal epithelial cells (IECs) during differentiation. CLDN7 mRNA and hnRNA increases 10- and 20-fold, respectively, after 3 days in culture compared with day 1. The increase is 20- and 40-fold, respectively, after 6 days. qPCR reactions were performed in triplicate, results are represented as the average of the triplicates. B: Schematic representation of the CLDN7 gene (marked gDNA for genomic DNA), nascent hnRNA, and mRNA structure. Forward (Fw) and reverse (Rv) primer locations for PCR detection are indicated on the diagrams. UTR, untranslated region.

Figure 3

Screening for transcription factor binding (TFB) activity during intestinal epithelial cell (IEC) differentiation. A: Changes in junction protein composition as the cells undergo differentiation. B: Left: TF DNA binding activity was assayed in nuclear lysates from 2- and 12-day-old confluent IEC monolayers using a TF/DNA binding array. The array assaying nuclear lysate from nondifferentiated (day 2) IECs is shown in green and the array with differentiated lysate (day 12) is overlaid in red to show differential binding activity of distinct TFs. Right: Two separate probes containing hepatocyte nuclear factor 4α (HNF-4α) binding motifs show increased TFB in differentiated IECs. C:In silico analysis was used to identify potential TFB sites on the CLDN7 promoter. D: The array and in silico data were cross-referenced to identify potential regulators of CLDN7 during IEC differentiation.

Figure 4

Deletion analysis of CLDN7 promoter activity. A: Representation of the pGL4C7 CLDN7/luciferase (luc) reporter. The sequence includes the transcription start site and the 5′ untranslated region (UTR) of CLDN7. B:CLDN7 promoter activity increases during intestinal epithelial cell (IEC) differentiation. C: Right: Truncated CLDN7 promoter luciferase reporter constructs were used to determine the minimal promoter region required for CLDN7 promoter activity during IEC differentiation. Left: Representation of the truncated CLDN7 luciferase reporters. Representative graphs of three independent experiments are shown for B and C.

Figure 5

ChIP analysis of CLDN7 promoter binding. A: Representation of the CLDN7 promoter. Predicted transcription factor binding (TFB) sites are marked with the symbol of the TF above the promoter, their respective distance from the start ATG is indicated below their symbol. PCR products detecting specific TF/DNA interactions in chromatin immunoprecipitation (ChIP) samples are indicated under the promoter and numbered 1 to 3 and C1 to C2. B: ChIP was performed using HNF-4α, PU.1, and Oct2.1 antibodies, and PCR was performed using primers flanking the predicted binding sites of Oct2.1, HNF-4α, and PU.1. A representative gel image of two independent experiments is shown. C: ChIP was performed using an antibody to HNF-4α and immunoprecipitated DNA was analyzed by real-time PCR. Data are expressed relative to DNA isolated from the same amount of input chromatin. Note the enrichment of HNF-4α at primer sets spanning the predicted binding motifs of HNF-4α compared with parallel samples precipitated using a nonspecific IgG. Three independent experiments were performed in triplicate. Data shown are the average ± SD from a representative experiment. D: HNF-4α is enriched at the CLDN7 promoter in differentiated compared with nondifferentiated Caco-2 cells. Data shown are the means ± SD of one experiment assayed in triplicate. E: HNF-4α shows a gradient similar to claudin-7, with the highest expression in differentiated cells in differentiating intestinal epithelial cells (IECs). Western blots of Caco-2 cell lysates are shown. Analogous to model IECs, gradients were observed in the expression of claudin-7 and HNF-4α in cryosections of human colonic mucosa by immunofluorescence labeling and laser confocal microscopy. Claudin-7 and HNF-4α are shown in green in F and H, and G and I, respectively. Notice the higher expression of both proteins on the surface (F and G) compared with the crypt base (H and I). Scale bar = 50 μm (F–I). F-actin was labeled with Alexa phalloidin in the HNF-4α images. IP, immunoprecipitation; UTR, untranslated region.

Figure 6

CLDN7 regulation by HNF-4α, mutation analysis. A: HNF-4α binding motifs at −2524 bp (Mut1) or −861 bp (Mut2) from the start codon were mutated on the CLDN7 luciferase reporter. The mutations were introduced by PCR primers containing mismatches as indicated by bold letters. B: Luciferase assays were performed in Caco-2 model intestinal epithelial cells (IECs) to test the effect of mutated HNF-4α binding sequences on CLDN7 promoter activity (white bars). Mutation of one or the other HNF-4α binding sites does not change CLDN7 promoter activity. Simultaneous mutation of both HNF-4α binding motifs (Mut1+2) diminishes CLDN7 promoter activity. At the same time, HNF-4α is expressed exogenously in Caco-2 model IECs (black bars). Luciferase assay shows up-regulated CLDN7 promoter activity in response to increased HNF-4α expression compared with control vector–treated cells. This increase was reduced in the single mutant construct Mut2, whereas the double mutant abolished the exogenous HNF-4α–induced CLDN7 promoter activity. The data shown are average ± SEM of six independent experiments. C: Small interfering RNA (siRNA) knockdown of HNF-4α in Caco-2 cells results in a 20%, statistically significant, decrease in CLDN7 promoter activity. Data shown are the average ± SEM of three independent experiments. D: Luciferase assays performed in HT29/B6 model IECs yielded results similar to those in Caco-2 cells. Data are presented as the average ± SEM of four experiments for overexpression and three experiments for knockdown of HNF-4α. E: Exogenous HNF-4α increases claudin-7 protein levels by twofold in nondifferentiated Caco-2 IECs compared to empty vector (EV) treated cells. Representative blots and densitometry analysis (means ± SEM) of five blots for overexpression and three blots for knockdown are shown. ^∗P < 0.05, ^∗∗P < 0.01. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; n.s., nonspecific control RNA; UTR, untranslated region.