Hepatocyte nuclear factor 4alpha orchestrates expression of cell adhesion proteins during the epithelial transformation of the developing liver

Abstract

Epithelial formation is a central facet of organogenesis that relies on intercellular junction assembly to create functionally distinct apical and basal cell surfaces. How this process is regulated during embryonic development remains obscure. Previous studies using conditional knockout mice have shown that loss of hepatocyte nuclear factor 4alpha (HNF4alpha) blocks the epithelial transformation of the fetal liver, suggesting that HNF4alpha is a central regulator of epithelial morphogenesis. Although HNF4alpha-null hepatocytes do not express E-cadherin (also called CDH1), we show here that E-cadherin is dispensable for liver development, implying that HNF4alpha regulates additional aspects of epithelial formation. Microarray and molecular analyses reveal that HNF4alpha regulates the developmental expression of a myriad of proteins required for cell junction assembly and adhesion. Our findings define a fundamental mechanism through which generation of tissue epithelia during development is coordinated with the onset of organ function.

Fig. 1.

Hepatocyte-specific loss of E-cadherin does not affect the formation of cell junctions in the liver. (A) RT-PCR showed loss of E-cadherin (E-cad) mRNA in livers of Cdh1^loxP/loxP;AlfpCre mice compared with control Cdh1^loxP/+;AlfpCre and WT littermates. Hnf4a levels were unchanged, and hypoxanthine guanine phosphoribosyl transferase 1 (Hprt1) confirmed equal loading. (B) Immunoblot analysis of liver extracts indicated that E-cadherin (E-cad) protein is undetectable in the Cdh1^loxP/loxP;AlfpCre livers compared with controls (Cdh1^loxP/+;AlfpCre and WT). Total protein levels of the tight junction protein OCLN were unchanged, and β-actin (ACTB) demonstrated equal loading. (C) Immunohistochemistry detected E-cadherin between hepatocytes in control livers (Top Left) but not between hepatocytes in Cdh1^loxP/loxP;AlfpCre livers (Top Right) (Inset is higher magnification). Confocal immunofluorescence microscopy was used to detect TJP1 (also known as ZO1) at the apical surface of the hepatocytes in both control (Cdh1^loxP/+;AlfpCre, Middle Left) and Cdh1^loxP/loxP;AlfpCre (Middle Right) livers. Junctional complexes (indicated by brackets) were identified in both control (Cdh1^loxP/+;AlfpCre, Bottom Left) and Cdh1^loxP/loxP;AlfpCre (Bottom Right) livers by transmission electron microscopy. Asterisks indicate bile canaliculi, which confirm that hepatocytes are polarized in the absence of E-cadherin. High-resolution electron microscopy images are provided in Fig. 5, which is published as supporting information on the PNAS web site.

Fig. 2.

HNF4α is required for expression of cell junction and adhesion proteins in fetal mouse liver. (A) Immunoblots revealed a loss of or reduction in expression of proteins required for the formation of adherens junctions [E-cadherin (E-CAD)], tight junctions (CLDN1, F11R, and OCLN), and desmosomes (DSC2) in 18.5-dpc fetal livers lacking HNF4α (Hnf4a^loxP/loxP;AlfpCre) compared with control livers (Hnf4a^loxP/+;AlfpCre). β-Actin (ACTB) was used as a loading control. (B) Confocal microscopy demonstrated that the expression and localization of the tight junction protein CLDN1 and GJB1 (also known as connexin 32) were disrupted in 18.5-dpc fetal livers lacking HNF4α (Hnf4a^loxP/loxP;AlfpCre) compared with control livers (Hnf4a^loxP/+;AlfpCre).

Fig. 3.

Loss of HNF4α in hepatocytes disrupts diverse pathways in the developing liver, including those associated with cell adhesion and junction formation. RT-PCR analysis of control Hnf4a^loxP/+; AlfpCre and mutant Hnf4a^loxP/loxP; AlfpCre 18.5-dpc livers confirmed that the mRNA levels of multiple cell junction and adhesion genes were down-regulated in mutant livers. Genes with decreased expression by RT-PCR of ≥2.5-fold in HNF4α-null livers compared with control livers are shown. Fold changes predicted from the array analysis using dchip software are listed. All genes tested, with the exception of Gjb2 and Cldn2, were predicted by dchip to have fold changes ≥2.0-fold; P ≤ 0.05. The P value of the fold changes predicted for Gjb2 and Cldn2 exceeded 0.05. Hprt1 was used as a standard to normalize loading.

Fig. 4.

HNF4α occupies sites in several cell junction and adhesion genes requiring HNF4α for their expression. ChIP showed that HNF4α occupies sites in 8 of the 18 genes assayed. We performed ChIP using chromatin isolated from independent 18.5-dpc WT mouse brains and livers and antibodies that immunoprecipitate either HNF4α (anti-HNF4α) or a nonrelated protein, PES1 (anti-Pescadillo). Input samples confirmed that equivalent amounts of chromatin were used in each ChIP reaction. ChIP of a known HNF4α binding site from the Apoc3 gene is shown as a positive control, and ChIP of a sequence lacking an HNF4α binding site from the Hprt1 gene is shown as a negative control. Two sites in the Lgals9 gene, H4.94 and H4.41, were treated as a single site because their proximity to each other prevents them from being discerned by ChIP. With the exception of site H4.191 in the Rhpn2 gene, which provides an example of a predicted HNF4α site we scored as negative by ChIP, only sites found to be occupied by HNF4α are shown.