Hepatocyte nuclear factor 4 provokes expression of epithelial marker genes, acting as a morphogen in dedifferentiated hepatoma cells

Abstract

We have recently shown that stable expression of an epitope-tagged cDNA of the hepatocyte- enriched transcription factor, hepatocyte nuclear factor (HNF)4, in dedifferentiated rat hepatoma H5 cells is sufficient to provoke reexpression of a set of hepatocyte marker genes. Here, we demonstrate that the effects of HNF4 expression extend to the reestablishment of differentiated epithelial cell morphology and simple epithelial polarity. The acquisition of epithelial morphology occurs in two steps. First, expression of HNF4 results in reexpression of cytokeratin proteins and partial reestablishment of E-cadherin production. Only the transfectants are competent to respond to the synthetic glucocorticoid dexamethasone, which induces the second step of morphogenesis, including formation of the junctional complex and expression of a polarized cell phenotype. Cell fusion experiments revealed that the transfectant cells, which show only partial restoration of E-cadherin expression, produce an extinguisher that is capable of acting in trans to downregulate the E-cadherin gene of well-differentiated hepatoma cells. Bypass of this repression by stable expression of E-cadherin in H5 cells is sufficient to establish some epithelial cell characteristics, implying that the morphogenic potential of HNF4 in hepatic cells acts via activation of the E-cadherin gene. Thus, HNF4 seems to integrate the genetic programs of liver-specific gene expression and epithelial morphogenesis.

Figure 1

HNF4tag expression alters morphology of H5 cells. Phase-contrast micrographs of H5 cells and the HNF4tag-expressing clone HT4-8, cultivated on gelatin-coated plastic dishes either without (top row, co), or with 1 μM Dex for 4 d (bottom row, dex). Only HT4-8 cells express epithelial cell characteristics, which are enhanced by the Dex treatment. Bar, 40 μm.

Figure 2

Dex treatment leads to epithelial morphogenesis only in the HNF4tag-expressing cells. Cellular morphology, expression of tagged proteins, and cytokeratin organization of two independent transfectants of HNF4tag (HT4-2 and HT4-8) were analyzed either in the presence or absence of 1 μM Dex for 4 d. Double- immunofluorescence analysis revealed that only the HNF4tag-expressing cells reexpress cytokeratin proteins and only in these cells does Dex treatment result in epithelial morphogenesis. Bar, 40 μm.

Figure 3

Establishment of epithelial (A and B) and hepatic (C) polarity is restricted to HNF4tag-expressing cells upon Dex treatment. (A) Immunofluorescence analysis of expression and localization of the tight junction–associated protein ZO-1 in untransfected control cells H5 and HNF4tag-expressing clone HT4-8, either untreated (no dex) or treated with 1 μM Dex. (B) Confocal microscopy confirms the absence of ZO-1 protein at the basolateral membrane (basal). The protein is confined to the apical surface as demonstrated by the slice corresponding to the top of the cell layer (apical) and the three dimensional reconstitution (3D). (C) HT4-8 cells occasionally express canalicular-like spaces; the image shown is of cells treated with the demethylating agent azacytidine (Späth and Weiss, 1997). The canalicular-like spaces (arrowhead in phase-contrast micrograph) are monitored by the circular organization of the ZO-1 protein, marking the “kissing zone” of the apical poles of two juxtaposed cells (see immunofluorescence micrograph and the three-dimensional reconstitution by confocal microscopy). Bars, 20 μm.

Figure 4

Dex treatment causes the inhibition of cell growth in cells expressing functional HNF4tag. HNF4del-expressing control cells (HT4del-4) and HNF4tag-expressing cells (HT4-8) were cultivated without Dex (♦) and in the presence of 1 μM Dex (•). Cell growth was analyzed by seeding 10⁵ cells in 10-cm dishes, in the presence or absence of the drug. Cells were counted at the days indicated and the growth curve was obtained by plotting the log of cell number against the days of growth. The medium was changed daily and the counts were performed on triplicate dishes.

Figure 4

Dex treatment causes the inhibition of cell growth in cells expressing functional HNF4tag. HNF4del-expressing control cells (HT4del-4) and HNF4tag-expressing cells (HT4-8) were cultivated without Dex (♦) and in the presence of 1 μM Dex (•). Cell growth was analyzed by seeding 10⁵ cells in 10-cm dishes, in the presence or absence of the drug. Cells were counted at the days indicated and the growth curve was obtained by plotting the log of cell number against the days of growth. The medium was changed daily and the counts were performed on triplicate dishes.

Figure 5

Expression of HNF4tag is paralleled by the activation of hepatic and epithelial marker genes. (A) Northern blot, prepared with 15 μg total RNA per lane. Differentiated hepatoma FGC4 RNA provides a positive control. Untransfected H5 cells and transfectant HT4-8 were either untreated (−) or treated with 1 μM Dex for 4 d (dex). The blot was hybridized with probes for HNF4tag, HNF1, and the epithelial marker genes CK18, CK8, and E-cadherin. (B) Quantitative immunoprecipitation monitoring the de novo synthesis of CK8 and CK18 using total protein extracts of in vivo–labeled FGC4 (positive control), H5 (negative control), and HT4-8 cells, either untreated (−) or treated with 1 μM Dex for 4 d (dex). The Cα subunit of cAMP-dependent protein kinase A demonstrates that equal amounts of labeled extract were analyzed. The signals specific for CK8 (55 kD) and CK18 (45 kD) are indicated. (C) Western Blot analysis for E-cadherin expression in control extracts (co; extracts purchased with the antibody), FGC4 cells, H5 cells, and the transfectant HT4-8. The cells were not treated with Dex. The signal specific for E-cadherin (120 kD) is indicated.

Figure 6

Phase-contrast micrographs and immunofluorescence staining of untreated cells of FGC4 (positive control), H5 (negative control), and the HNF4tag transfectant HT4-8 for expression and localization of pan-cadherin (the field corresponds to the respective phase-contrast image), E-cadherin, α-catenin, and β-catenin. Note the nuclear localization of β-catenin in both H5 and HT4-8 cells. Antibodies were purchased and diluted as described in Materials and Methods. Bar, 20 μm.

Figure 7

Somatic hybrids of Fao and HT4-8 are insulated from extinction of the epithelial phenotype only when HNF4tag expression is retained. (A) Northern blot analysis using 15 μg of total RNA of the differentiated parental cell line Fao, dedifferentated H5 cells (negative control), HNF4tag-expressing parental cell line HT4-8, and two (Fao × HT4-8) hybrid clones (HTF). The epithelial expression pattern is restricted to the HTF-2 hybrid clone that maintains the expression of HNF4tag. (B) Immunostaining of differentiated Fao cells and two hybrid clones, representing differentiated (HTF-2) and dedifferentitated (HTF-11) morphology, for the expression and localization of E-cadherin, and β-catenin. Note that abundant nuclear localization of β-catenin is restricted to the dedifferentiated cells. Bar, 20 μm.

Figure 8

Northern blot analysis of 15 μg total RNA of FGC4 cells and of pools and clones of hygromycin-resistant H5 cells stably transfected either with the vector alone (H5-Hyg) or with the vector driving expression of the mouse E-cadherin cDNA uvomorulin (H5uvo). The blot was hybridized with the probes indicated. Note that the endogenous E-cadherin transcripts in FGC4 cells are not visible because of its slower mobility compared to the exogenous E-cadherin/uvomorulin from which the 3′ untranslated region has been deleted.

Figure 9

E-cadherin–expressing H5 cells undergo epithelial morphogenesis. Phase-contrast micrographs of E-cadherin–transfected cells of H5uvo8 and immunostaining for expression and localization of E-cadherin, α-catenin, β-catenin, cytokeratin, and ZO-1. Bar, 20 μm.