General transcription factor TAF4 antagonizes epigenetic silencing by Polycomb to maintain intestine stem cell functions

Abstract

Taf4 (TATA-box binding protein-associated factor 4) is a subunit of the general transcription factor TFIID, a component of the RNA polymerase II pre-initiation complex that interacts with tissue-specific transcription factors to regulate gene expression. Properly regulated gene expression is particularly important in the intestinal epithelium that is constantly renewed from stem cells. Tissue-specific inactivation of Taf4 in murine intestinal epithelium during embryogenesis compromised gut morphogenesis and the emergence of adult-type stem cells. In adults, Taf4 loss impacted the stem cell compartment and associated Paneth cells in the stem cell niche, epithelial turnover and differentiation of mature cells, thus exacerbating the response to inflammatory challenge. Taf4 inactivation ex vivo in enteroids prevented budding formation and maintenance and caused broad chromatin remodeling and a strong reduction in the numbers of stem and progenitor cells with a concomitant increase in an undifferentiated cell population that displayed high activity of the Ezh2 and Suz12 components of Polycomb Repressive Complex 2 (PRC2). Treatment of Taf4-mutant enteroids with a specific Ezh2 inhibitor restored buddings, cell proliferation and the stem/progenitor compartment. Taf4 loss also led to increased PRC2 activity in cells of adult crypts associated with modification of the immune/inflammatory microenvironment that potentiated Apc-driven tumorigenesis. Our results reveal a novel function of Taf4 in antagonizing PRC2-mediated repression of the stem cell gene expression program to assure normal development, homeostasis, and immune-microenvironment of the intestinal epithelium.

Fig. 1. Morphological alteration resulting from Taf4 inactivation in the gut endoderm of E18.5 fetuses.

Morphology (HE) and immunohistochemical detection of the indicated proteins in E18.5 control Taf4^los/lox (Ctrl) and Taf4^IEndoC littermates. Bars are 50 µm except for HE and Muc2 where they represent 500 µm and 100 µm, respectively.

Fig. 2. Homeostasis defects induced by Taf4 inactivation in the adult gut epithelium.

A Histology (HE) and immunodetection of the Taf4 protein in the intestine of adult Taf4^IEC and control Taf4^los/lox (Ctrl) mice 10 days after Tamoxifen injection. Bars are 100 µm. B Overall survival of Taf4^IEC (red boxes) and Taf4^lox/lox (Ctrl; white boxes) mice after Tamoxifen injection (n = 28 in each group; LRK Logrank test). Body weight of the 13 Taf4^IEC mice surviving up to the end of the experiment and of 13 mice of the Ctrl group at day 0 and day 120 after Tamoxifen administration. ns not significant; *p = 0.01; **p < 0.02; ****p < 0.0001. Boxes extend from the 25th–75th percentile and whiskers represent mean to max. C Immunodetection of the indicated proteins in the ileum of Taf4^IEC and control mice 10 days after Tamoxifen injection. Bars are 100 µm for Alpi, Aldob, Muc2 and Chga, and 50 µm for Dclk1, Olfm4, Lyz and Sox9. D Ki67 immunostaining and cell counts in the ileum of Taf4^ΔIEC and control mice. Bars are 100 µm. Boxes extend from the 25th–75th percentile and whiskers represent mean to max. *p < 0.05. E Immunofluorescent staining of phospho-Erk1/2 (red) and β-catenin (green) in Taf4^IEC and control mice. Bars are 100 µm. F BrdU detection in the jejunal mucosa of Taf4^IEC and control mice at the indicated days after the single injection of BrdU. Bars are 200 µm.

Fig. 3. Gene expression changes after Taf4 inactivation.

A Venn diagrams representing the down- and up-regulated genes in E17.5 Taf4^IEndoC fetuses and adult Taf4^IEC mice compared to their respective controls. B KEGG ontology enrichment is shown for the downregulated genes in fetuses and adults, ordered according to the p value. C GSEA analysis of the up-regulated and downregulated genes in fetuses identifying respectively hallmarks for E2F targets and for bile acid metabolism, and IFNα and IFNγ response. D GSEA analysis of the up-regulated and downregulated genes in adults identifying respectively hallmarks for Myc targets and for allograft rejection, and IFNα and IFNγ response. E Stromal cell population evaluation from RNA-seq data using the MCP method, expressed as the proportion of each cell type in Taf4^IEC mice compared to controls. F Venn diagrams of the down-regulated genes (left) and up-regulated genes (right) enriched in common in the intestine of E17.5 Taf4^IEndoC fetuses and adult Taf4^IEC mice, and in the Taf4-null liver of 12 days suckling mice. Middle: KEGG ontology enrichment of the 90 genes down-regulated in common in the fetal and adult intestine and in the liver of Taf4-deficient mice.

Fig. 4. Effect of Taf4 inactivation on enteroid morphogenesis.

A Survival and budding activity after early Taf4 gene inactivation. Taf4^IEC enteroids were plated and treated 2 h later with 4-OHT in EtOH (red line) or EtOH alone (black line) for 3 days. In total, 50–100 3D structures were counted at each time point. B Morphology of the Taf4^IEC enteroids at day 5 of culture, that is 2 days after treatment with EtOH (left) or 4-OHT (right). Boxed regions are enlarged below. Nuclei are stained with Dapi (blue) and the actin network with Phalloidin (red). Bars are 50 µm. C Top 6 enriched KEGG pathways in downregulated genes in 4-OHT treated Taf4^IEC vs. Taf4^lox/lox enteroids. D Top 8 enriched KEGG pathways in upregulated genes in 4-OHT treated Taf4^IEC vs. Taf4^lox/lox enteroids. E Immunofluorescence detection of the indicated proteins in Taf4^lox/lox (Ctrl) and Taf4^IEC enteroids treated with 4-OHT at day 5 of culture (2 days after treatment with 4-OHT). Bars are 50 µm.

Fig. 5. Effect of Taf4 inactivation on enteroid homeostasis.

A Survival and budding activity after late Taf4 gene inactivation. Taf4^IEC enteroids were plated and treated with 4-OHT (red line) or EtOH (black line) at days 5 to 7 of culture. In total, 50–100 3D structures were counted at each time point. B Immunodetection of proliferating cells (Ki67) and apoptotic bodies (activated Caspase-3) at day 8 of culture in 4-OHT-treated Taf4^IEC vs. Taf4^lox/lox (Ctrl) enteroids. White arrowheads show costaining. Bars are 50 µm. C Uniform manifold approximation and projection (UMAP) clustering from scRNA-seq analyses in 4-OHT-treated Taf4^IEC and Taf4^lox/lox enteroids at day 8 of culture. Cell clusters identified in Taf4^lox/lox and Taf4^IEC enteroids were labeled WT0 to WT11 and KO0 to K010, respectively. D Heatmap of the top markers identified by scRNA-seq in the WT0 to WT11 (green) and KO0 to K010 (red) cell clusters. Absorptive: enterocytes; P + G: Paneth and goblet cells, Endocrine: enteroendocrine cells.

Fig. 6. Regulons changes associated with Taf4 loss.

A SCENIC-UMAP of the cell clusters identified by scRNA-seq in 4-OHT-treated Taf4^lox/lox (WT) and Taf4^IEC (KO) enteroids. B Uncertainty-Aware Face Clustering (AUC) of representative highly active regulons in the indicated cell clusters identified by SCENIC in Taf4^lox/lox enteroids.

Fig. 7. Rescue of Taf4 inactivated enteroids by Polycomb complex inhibition.

A Survival and budding activity in Taf4^IEC enteroids treated with EtOH + DMSO (black line), with 4-OHT (2–72 h after plating), or 4-OHT (2–72 h after plating) + EPZ6438 (2 h to 15 days after plating) (red line). In total, 50-100 3D structures were counted at each time point. B Morphology at day 15 of culture of Taf4^lox/lox (ctrl) and Taf4^IEC enteroids treated with EtOH + DMSO, or 4-OHT (2–72 h after plating), or 4-OHT (2–72 h after plating) + EPZ6438 (2 h to 15 days after plating). Bars are 200 µm. C Survival and budding activity in Taf4^IEC enteroids treated with EtOH + DMSO (black line), with 4-OHT (days 5-8 after plating) (red line), or with 4-OHT (days 5-8 after plating) + EPZ6438 (days 5-15 after plating) (blue line). In total, 50–100 3D structures were counted at each time point. D Immunostaining of the indicated proteins in Taf4^IEC enteroids treated 4-OHT or with 4-OHT + EPZ6438 at day 15 of culture. Bars are 50 µm. E Heatmap of the 182 genes of the stem cell signature that are downregulated by Taf4 loss in 4-OHT treated Taf4^IEC enteroids compared to treated Taf4^lox/lox control enteroids, and restored by treatment with the Ezh2 inhibitor EPZ6438.

Fig. 8. Impact of Taf4 inactivation on intestinal tumor development.

A Overall survival of Apc^+/Δ14::Taf4^IEC, Apc^+/Δ14, Taf4^IEC and Taf4^lox/lox (Ctrl) males after Tamoxifen injection; n > 10 for each genotype. LRT AT vs. A: Logrank test between Apc^+/Δ14::Taf4^IEC and Apc^+/Δ14 mice; LRT AT vs. T: Logrank test between Apc^+/Δ14::Taf4^IEC and Taf4^IEC mice. B Tumor number in the small intestine of Apc^+/Δ14::Taf4^IEC and Apc^+/Δ14 mice. ***p < 0.0001. C Histology (HE) and immunodetection of Taf4 protein in ileal tumors of Apc^+/Δ14::Taf4^IEC and Apc^+/Δ14 mice. Bars are 400 µm for HE and 200 µm for Taf4. D Immunohistochemical detection of Ki67, Cdx2 and Hnf4α in Apc^+/Δ14::Taf4^IEC and Apc^+/Δ14 tumors. Bars are 200 µm. E KEGG ontology enrichment is shown for the 699 downregulated genes in Apc^+/Δ14::Taf4^IEC vs. Apc^+/Δ14 mice and ordered according to the p value. F GSEA analysis of the up-regulated and downregulated genes in Apc^+/Δ14::Taf4^IEC vs. Apc^+/Δ14 mice identifying respectively hallmarks for Wnt/β-catenin signaling and for allograft rejection and IFNα, IFNγ, IL2/STAT5 and IL6/STAT3 signaling. G Stromal cell population evaluation from RNA-seq data using the MCP method, expressed as the proportion of each cell type in Apc^+/Δ14::Taf4^IEC compared to Apc^+/Δ14 mice. H KEGG ontology enrichment of the 247 downregulated genes in common between Apc^+/Δ14::Taf4^IEC vs. Apc^+/Δ14 mice and Taf4^IEC vs. wild type mice.