Hnf1alpha (MODY3) controls tissue-specific transcriptional programs and exerts opposed effects on cell growth in pancreatic islets and liver

Abstract

Heterozygous HNF1A mutations cause pancreatic-islet beta-cell dysfunction and monogenic diabetes (MODY3). Hnf1alpha is known to regulate numerous hepatic genes, yet knowledge of its function in pancreatic islets is more limited. We now show that Hnf1a deficiency in mice leads to highly tissue-specific changes in the expression of genes involved in key functions of both islets and liver. To gain insights into the mechanisms of tissue-specific Hnf1alpha regulation, we integrated expression studies of Hnf1a-deficient mice with identification of direct Hnf1alpha targets. We demonstrate that Hnf1alpha can bind in a tissue-selective manner to genes that are expressed only in liver or islets. We also show that Hnf1alpha is essential only for the transcription of a minor fraction of its direct-target genes. Even among genes that were expressed in both liver and islets, the subset of targets showing functional dependence on Hnf1alpha was highly tissue specific. This was partly explained by the compensatory occupancy by the paralog Hnf1beta at selected genes in Hnf1a-deficient liver. In keeping with these findings, the biological consequences of Hnf1a deficiency were markedly different in islets and liver. Notably, Hnf1a deficiency led to impaired large-T-antigen-induced growth and oncogenesis in beta cells yet enhanced proliferation in hepatocytes. Collectively, these findings show that Hnf1alpha governs broad, highly tissue-specific genetic programs in pancreatic islets and liver and reveal key consequences of Hnf1a deficiency relevant to the pathophysiology of monogenic diabetes.

FIG. 1.

Hnf1a deficiency causes tissue-specific gene expression changes in pancreatic islets and liver. (A) Pie charts depict the percentages of genes that are downregulated (green) and upregulated (red) in Hnf1a^−/− pancreatic islets and liver at a 5% FDR. Venn diagrams show the overlap of genes downregulated and upregulated in both tissues. (B) Heat map displaying major functional categories that are enriched among downregulated (green arrows) and upregulated (red arrows) genes in Hnf1a^−/− islets and liver. (C) GSEA of genes involved in glycolysis and gluconeogenesis, the TCA cycle, and the electron transport chain across genes ranked according to their differential expression levels in Hnf1a^−/− pancreatic islets (top) and liver (bottom). Vertical lines beneath the graphs depict the rank positions of each gene in the color-coded gene sets. The results show that the three gene sets are downregulated in Hnf1a^−/− islets yet upregulated in Hnf1a^−/− liver. WT, wild type; KO, knockout.

FIG. 2.

Hnf1α-binding analysis in hepatocytes and pancreatic islets. (A) M-A plots of Hnf1 ChIP hybridization intensities in hepatocytes. Dark dots depict Hnf1-bound genes (M > 0.8 and P < 0.001). (B) Validation of 25 Hnf1 targets in hepatocytes and islets by qPCR. Hnf1 targets in hepatocytes were validated using the Hnf1 antibody and an Hnf1α-specific antibody. Consistent with observations that Hnf1β was not detected in adult hepatocytes (29) (Fig. 6), similar results were obtained with the two antibodies. Gene-specific qPCR signals were calculated as percentages of input DNA and expressed as the enrichment values relative to the mean values for Tbp and Actb. Seven genes were tested due to the presence of a proximal conserved HNF1 motif and are marked with a circle. Genes are grouped according to their tissue expression patterns.

FIG. 3.

Conserved HNF1 motifs are selectively enriched in the immediate 5′ flanking regions of genes downregulated in Hnf1a^−/− islets and liver. (A) Frequency of HNF1 motifs in 100-bp windows relative to the TSSs of all RefSeq genes. We scanned all HNF1 motifs in the mouse genome (yellow line) and conserved motifs in mouse and human aligned genomes (blue line). (B) Frequency of conserved HNF1 motifs in 200-bp windows relative to the TSSs of genes that were downregulated and upregulated in Hnf1a^−/− islets (left) and liver (right). (C) Percentages of genes that were downregulated in Hnf1a^−/− islets and liver among genes containing nonconserved and conserved HNF1 motifs within 200 bp upstream of the TSS. The analysis was restricted to genes that are expressed in each tissue (*, P values of <0.001; **, P values of <0.0001 for comparison to genes that lack an HNF1 motif within 200 bp upstream of the TSS). (D) Venn diagram showing the overlap between Hnf1α-bound genes in ChIP-chip (red) and genes with an HNF1 motif (blue) or a conserved HNF1 motif (green) within 1 kb upstream of the TSS. For the right panel, the same analysis was conducted for genes that were downregulated >2-fold in Hnf1a^−/− liver or islets. Both analyses were restricted to expressed genes represented in the promoter arrays (n = 8,803). (E) Frequencies of Hnf1α binding to genes with nonconserved and conserved HNF1 motifs at different distances from the TSS. The dashed line represents the overall percentage for all RefSeq genes expressed in liver (*, P values of <0.05; **, P values of <0.001; ***, P values of <0.0001 for comparison to the overall Hnf1α-binding frequency).

FIG. 4.

Only a subset of Hnf1α-bound genes are affected by Hnf1a deficiency in islets and liver. (A, B) Volcano plot relating the change and statistical significance of gene expression in Hnf1a^−/− (KO) versus wild-type (WT) liver (A) and islets (B). All genes are represented with gray dots. Colored dots are direct Hnf1α targets identified by either ChIP-chip (red) or a proximal conserved HNF1 motif (green). (C) Volcano plot displaying differential H3K4me2 levels in gene promoters in Hnf1a^−/− and wild-type hepatocytes. Red dots depict direct Hnf1α targets as in panel A. (D) Percentage of direct Hnf1α targets as a function of the magnitude of gene expression changes in Hnf1a^−/− islets. “NC” represents all genes lacking differential expression. Direct Hnf1α target genes were defined by Hnf1α binding in ChIP-chip (red), a conserved HNF1 motif (green), or either feature (blue).

FIG. 5.

Functional Hnf1α target genes are downregulated in islets from Hnf1a-haploinsufficient mice. (A) Scatter plot of gene expression changes in Hnf1a^+/− (log₂ heterozygous mutant [HET]/wild-type [WT] values) and Hnf1a^−/− (log₂ knockout [KO]/WT values) islets. Dark spots represent direct Hnf1α targets (identified by either ChIP-chip or proximal conserved HNF1 motifs) that were downregulated >2-fold in Hnf1a^−/− islets. (B) GSEA of Hnf1α-bound genes that were downregulated >2-fold in Hnf1a^−/− islets across genes ranked according to their differential expression levels in Hnf1a^+/− pancreatic islets. The vertical lines beneath the graphs depict the rank positions of the genes.

FIG. 6.

Hnf1α requirement for expression of its direct targets is tissue specific. (A) Venn diagrams showing all Hnf1α direct targets (black ovals) and the subsets that are downregulated >2-fold in Hnf1a^−/− islets (left circle in each oval) and liver (right circle in each oval). Separate analyses were performed for genes containing conserved HNF1 motifs 200 bp upstream of the TSS and those bound in ChIP-chip. (B) Percentages of direct Hnf1α target genes that were downregulated >2-fold in Hnf1a^−/− islets or liver, broken down according to their tissue expression patterns: IL, expression in both islets and liver; I, expression only in islets; L, expression only in liver. (C) Expression patterns for all direct targets that are downregulated >2-fold in Hnf1a^−/− mouse tissues. (D) Enrichment of Hnf1β and H3K4me2 in HNF1 targets in wild-type (WT) and Hnf1a^−/− (KO) hepatocytes. In Hnf1a^−/− hepatocytes, Hnf1β binds to all tested Hnf1α targets that are not silenced in Hnf1a^−/− liver but not to those that become inactive. Genes are grouped according to their tissue expression patterns.

FIG. 7.

Hnf1a deficiency has opposed effects on β-cell and hepatocyte proliferation. (A, B) Heat maps of expression changes (represented as log₂ n-fold change) in Hnf1a^−/− islets and liver for selected genes involved in cell growth, survival, and tumorigenesis (A) or cell cycle progression (B). (C) Low-magnification view of hematoxylin-and-eosin (H&E) staining of pancreases from Hnf1a^+/+ RipTAg and Hnf1a^−/− RipTAg mice. Dotted lines depict normal islets (i), hyperplasic islets (h), and fully developed insulinoma tumors (t). At 3 months of age, insulinomas were invariably identified in Hnf1a^+/+ RipTAg mice but never in Hnf1a^−/− RipTAg mice. (D) Percentages of Ki67⁺ β cells in Hnf1a^+/+ and Hnf1a^−/− RipTAg mice. wt, wild type; ko, knockout. (E) Changes in β-cell number after 7-day culture of dispersed insulinoma β cells from 3-month-old Hnf1a^+/+ RipTAg mice and from a 9-month-old Hnf1a^−/− RipTAg mouse that developed small insulinomas. β cells from Hnf1a^−/− RipTAg insulinomas failed to grow in culture. (F) Immunostaining of Ki67 (red) and insulin (green) in pancreases from Hnf1a^+/+ RipTAg and Hnf1a^−/− RipTAg mice. (G) Percentages of Ki67⁺ Hnf4α-expressing hepatocytes in 1-month-old control and Hnf1a^−/− liver samples. (H) Immunostaining of Ki67 (red) and Hnf4α (green) in liver samples from control and Hnf1a^−/− mice.