Tcf7l2/Tcf4 Transcriptional Repressor Function Requires HDAC Activity in the Developing Vertebrate CNS

Abstract

The generation of functionally distinct neuronal subtypes within the vertebrate central nervous system (CNS) requires the precise regulation of progenitor gene expression in specific neuronal territories during early embryogenesis. Accumulating evidence has implicated histone deacetylase (HDAC) proteins in cell specification, proliferation, and differentiation in diverse embryonic and adult tissues. However, although HDAC proteins have shown to be expressed in the developing vertebrate neural tube, their specific role in CNS neural progenitor fate specification remains unclear. Prior work from our lab showed that the Tcf7l2/Tcf4 transcription factor plays a key role in ventral progenitor lineage segregation by differential repression of two key specification factors, Nkx2.2 and Olig2. In this study, we found that administration of HDAC inhibitors (Valproic Acid (VPA), Trichostatin-A (TSA), or sodium butyrate) in chick embryos in ovo disrupted normal progenitor gene segregation in the developing neural tube, indicating that HDAC activity is required for this process. Further, using functional and pharmacological approaches in vivo, we found that HDAC activity is required for the differential repression of Nkx2.2 and Olig2 by Tcf7l2/Tcf4. Finally, using dominant-negative functional assays, we provide evidence that Tcf7l2/Tcf4 repression also requires Gro/TLE/Grg co-repressor factors. Together, our data support a model where the transcriptional repressor activity of Tcf7l2/Tcf4 involves functional interactions with both HDAC and Gro/TLE/Grg co-factors at specific target gene regulatory elements in the developing neural tube, and that this activity is required for the proper segregation of the Nkx2.2 (p3) and Olig2 (pMN) expressing cells from a common progenitor pool.

Fig 1. The histone deacetylase (HDAC) inhibitor VPA disrupts ventral progenitor boundaries.

(A-F) Sections taken from E3 embryos treated with the indicated dilutions of VPA at E2 and stained for markers (Olig2, Nkx2.2, and FoxA2) that identify the three normally distinct ventral progenitor domains in the spinal cord. Inset schematic at bottom indicates the three progenitor domain boundaries analyzed and the corresponding markers used. pMN = motoneurons progenitor domain, p3 = V3 interneuron progenitor domain, FP = floor plate domain. Note that the number of double-labeled cells in both boundary regions increases significantly with increasing VPA concentrations (arrowheads) (A, D) controls; (B, E) VPA at 1mM; (C, F) VPA at 100mM). (G) Quantification of the results in A-F. *p<0.001.

Fig 2. HDAC activity is required for Tcf repression of Gli2A-induced Nkx2.2 expression in vivo.

(A-E) Sections through E3 chick embryos electroporated with Gli2A (at 2.0 μg/μl) and Gli2A+Tcf4R (at 2.0 μg/μl) in the presence of increasing concentrations of VPA. Right side was transfected in all cases. Note that the number of induced Nkx2.2+ cells increases with increasing VPA, indicating that Tcf4R antagonism becomes less effective at higher concentrations. (A’-E’) Olig2 expression in embryos transfected with Gli2A and Gli2A+TcfR in the presence of increasing concentrations of VPA. No effect is seen on Olig2 expression. (F) Quantification of data in A-E’. Inset at bottom right indicates corresponding bar chart shading. *p<0.001.

Fig 3. Tcf and HDAC1 proteins occupy Nkx2.2 but not Olig2 regulatory sequences.

ChIP-reChIP experiments showing interaction of Tcf4 and HDAC1 with Nkx2.2 but not Olig2 regulatory regions. (A) Left column: Primary IP assays with HDAC1, IgG (negative control), or AcHistone3 (AcHis3) (positive control) antibodies from chromatin prepared from E10.5 mouse embryos. Right column: re-ChIP assays with Tcf3/4, IgG (negative control), or AcHistone3 (positive control) antibodies from primary IP elution. (B) Sequential chromatin immunoprecipitation assay schema. (C) Densitometry analysis of primary IP (left), and re-IP (right) results. The data are expressed as mean± SEM from three independent experiments.

Fig 4. Functional interactions between Grg4 and HDAC/Tcf4 are required to repress Gli2A induction of Nkx2.2 in vivo.

(A) Co-transfection of Tcf4R with Gli2A (both at (at 2.0 μg/μl) suppresses induction of Nkx2.2. (B) Predicted repressor complex for Tcf4R-mediated repression. (C, D) Co-transfection of a full-length Grg4 construct with Gli2A+TcfR also suppresses Nkx2.2 induction. (E-H) Co-transfection of Grg4 deletion constructs with Gli2A+TcfR. Both Grg4-ΔQ and Grg4-Q domain proteins prevent Tcf4R from blocking Gli2A-mediated induction of Nkx2.2 (seen in A). Antibody staining for the Myc-epitope tag was used to detect Grg4 constructs in all figures; GFP expression marks cells transfected with Tcf4R, while Gli activity was monitored by assaying Nkx2.2 expression. (I) Quantification of induced Nkx2.2 cells in each experiment, (J) Model for the regulation of Nkx2.2 by Tcf repressors involving HDAC activity and chromatin remodeling. *p<0.001, **p>0.05.