Loss of hepatocyte-nuclear-factor-4alpha affects colonic ion transport and causes chronic inflammation resembling inflammatory bowel disease in mice

Abstract

Background: Hnf4alpha, an epithelial specific transcriptional regulator, is decreased in inflammatory bowel disease and protects against chemically-induced colitis in mice. However, the precise role of this factor in maintaining normal inflammatory homeostasis of the intestine remains unclear. The aim of this study was to evaluate the sole role of epithelial Hnf4alpha in the maintenance of gut inflammatory homeostasis in mice.

Methodology/principal findings: We show here that specific epithelial deletion of Hnf4alpha in mice causes spontaneous chronic intestinal inflammation leading to focal areas of crypt dropout, increased cytokines and chemokines secretion, immune cell infiltrates and crypt hyperplasia. A gene profiling analysis in diseased Hnf4alpha null colon confirms profound genetic changes in cell death and proliferative behaviour related to cancer. Among the genes involved in the immune protection through epithelial barrier function, we identify the ion transporter claudin-15 to be down-modulated early in the colon of Hnf4alpha mutants. This coincides with a significant decrease of mucosal ion transport but not of barrier permeability in young animals prior to the manifestation of the disease. We confirm that claudin-15 is a direct Hnf4alpha gene target in the intestinal epithelial context and is down-modulated in mouse experimental colitis and inflammatory bowel disease.

Conclusion: Our results highlight the critical role of Hnf4alpha to maintain intestinal inflammatory homeostasis during mouse adult life and uncover a novel function for Hnf4alpha in the regulation of claudin-15 expression. This establishes Hnf4alpha as a mediator of ion epithelial transport, an important process for the maintenance of gut inflammatory homeostasis.

Figure 1. Epithelial specific loss of Hnf4α leads to various signs of chronic colon inflammation.

(A) qRT-PCR of Hnf4α mRNA at 7 days, 1 month and 12 months in mutants (open bars) relative to controls (black bars). Expression levels are shown as mean values (±SEM) relative to controls for each time point (3–8 mice per group) and were normalized with the Tbp housekeeping gene. **P≤.01 and ***P≤.001. (B) Indirect immunofluorescence of Hnf4α performed on controls and Hnf4α colon null mice. A non-specific fluorescent signal was observed in the stroma of both control and null mice. Scale bars, 50 µm. (C) H&E staining of a healthy mucosa in mutants. (D) Mucosa with signs of crypt distortion and crypt loss. (E) Increased crypt length and cell infiltrates. (F) Polyp formation. These photographs are representative of eight 12 month-old individuals all showing these morphological features. The formation of polyps was only observed in 2 independent Hnf4α colon null mice. All scale bars are 100 µm except for panel F that is 200 µm.

Figure 2. Loss of Hnf4α does not compromise goblet cell maturation in healthy colon mucosa but in long term inflamed colon mucosa.

(A) Alcian blue staining of control mice aged of 12 months. (B) Healthy colon region of 12 month-old mutant mice showing normal size and number of goblet cells (Scale bar, 100 µm). Electron microscopy shows a similar maturation level of goblet cells in control (C) and mutant mice (D) (Scale bar, 2 µm). (E) Small size goblet cells in hyperplasic epithelium of 12 month-old mutant mice. (F) Drastic reduction in goblet cell number in hyperplasic crypts (Scale bar, 100 µm). (G) qRT-PCR of Muc2 mRNA and (H) Muc3 mRNA in Hnf4α mutants (Open bars) and controls (Black bars). Expression levels are shown as mean values (± SEM) relative to controls and were normalized with the Hprt gene. **P≤0.01 and ***P≤0.001.

Figure 3. Increased cytokine and chemokine secretion associated with the inflammatory state of Hnf4α colon null mice.

(A) Summarized results (means ± SEM; at least 3 individuals aged of 12 months) of the semi-quantitative cytokine/chemokine array shown in Table S1. (B) ELISA performed on total protein lysates (means ± SEM; n = 6 per group). (C) Labelling of granulocytic cells by MPO immunohistochemistry was negative in controls as opposed to (D) mutants. Scale bars, 100 µm. (E) Western blot of accumulated phospho-IκBα relative to total IκBα and relative quantification (mean ± SEM), *P≤.05.

Figure 4. The apoptotic pathway is overly active and the proliferative dynamic is altered in regenerative epithelium.

(A) Increased abundance of cleaved and active forms of caspase-3 in 12 month-old mutants (Open bar) relative to controls (Black bar). (B) Quantification of band intensities of cleaved and active caspase-3 combined (n = 4) normalized to β-actin (Black bars, controls; open bars, mutants). Represented as means ± SEM, * P≤.05. (C) E2f2 and AurkA modulations seen in microarray data were verified at various time points in mutants (Open bars) relative to controls (Black bars). Expression levels are shown as mean values (± SEM) relative to controls for each time point (3–14 mice per group) and were normalized with the Tbp housekeeping gene. (D) Measurements (n = 4 per group) of crypt length. (E) Ki67 immunochemistry of 12 month-old control mice compared to a hyperplasic area seen in mutants of same age. * P≤.05, ** P≤.01 and *** P≤.001.

Figure 5. Claudin-15 and ion transport are modulated in Hnf4α colon null mutants.

(A) qRT-PCR of claudin-4, -8 and -15 at various time points in mutants (Open bars) relative to controls (Black bars). Expression levels are shown as mean values (± SEM) relative to controls at each time point (3–14 mice per group) and were normalized with the Tbp housekeeping gene. (B) Immunoblotting confirmed modulated protein levels of claudin-4, -8 and -15 at 12 months in mutants (n = 3, open bars) and relative controls (n = 3, black bars) normalized to stable expression of E-Cadherin (Cdh1). (C) Quantification of band intensities of each claudin (means ± SEM; n = 3 or 4 for each group) relative to Cdh1. (D) Indirect immunofluorescence (FITC) of claudin-4, -8 and -15 in 3 month-old animals in proximal colon (representative of 4 pairs). Nuclei were counterstained with DAPI (blue) and merged with FITC (green). Scale bars, 50 µm. (E) Claudin-15 expression is lost in the upper half of the colon epithelium in mutants (Open bar) relative to controls (Black bar). (F) Colonic permeability to ⁵¹Cr-EDTA measured in colon tissues from 2 month-old mice (5 mice per group). (G) Baseline short circuit current (Isc) measured in colon tissues from 2 month-old mice (5 mice per group). Represented as mean values ± SEM, * P≤.05, ** P≤.01, *** P≤.001.

Figure 6. The claudin-15 gene is a direct target of Hnf4α.

(A) EMSA of 4 Hnf4α binding sites found in the 1 kb mouse claudin-15 promoter performed with 293T cells transfected with pBabe Hnf4α1 vector and 293T nuclear extracts as negative controls. Shifts (white arrow head) are compared to previously published APOC3 and mutated APOC3 (APOC3m) for control binding site. NS is non-specific binding (grey arrow head). Supershifts (black arrow head) were done with Hnf4α antibody (sc-6556). (B) ChIP assays of the 4 putative sites performed on epithelial cells from adult mouse colon. (C) qRT-PCR of claudin-15 on IEC6 cells infected with pBabe Hnf4α as compared to control vector pBabe (n = 4). (D) Western blot of 3 stable T84 cell populations expressing mouse claudin-15. (E) Conductance measured for three consecutive days after the reach of confluence on three populations of T84 cells expressing mouse claudin-15 or control plenti eGFP vector. Represented as mean values±SEM, *** P≤.001.

Figure 7. HNF4A expression is reduced in IBD patient samples in association with Claudin-15.

(A) qRT-PCR of Hnf4α and claudin-15 expression following 10 days of 2.5% DSS mild treatment in the distal colon of CD1 mice relative to water treated mice. HNF4A expression was retained at 22% (B) and claudin-15 was retained at 37% (C) in ulcerative colitis (UC) patient samples (n = 21) as compared to healthy patient samples (CTRL) (n = 6). Represented as mean values ± SEM, * P≤.05, ** P≤.01, *** P≤.001.