Combinatorial Smad2/3 Activities Downstream of Nodal Signaling Maintain Embryonic/Extra-Embryonic Cell Identities during Lineage Priming

Abstract

Epiblast cells in the early post-implantation stage mammalian embryo undergo a transition described as lineage priming before cell fate allocation, but signaling pathways acting upstream remain ill defined. Genetic studies demonstrate that Smad2/3 double-mutant mouse embryos die shortly after implantation. To learn more about the molecular disturbances underlying this abrupt failure, here we characterized Smad2/3-deficient embryonic stem cells (ESCs). We found that Smad2/3 double-knockout ESCs induced to form epiblast-like cells (EpiLCs) display changes in naive and primed pluripotency marker gene expression, associated with the disruption of Oct4-bound distal regulatory elements. In the absence of Smad2/3, we observed enhanced Bmp target gene expression and de-repression of extra-embryonic gene expression. Cell fate allocation into all three embryonic germ layers is disrupted. Collectively, these experiments demonstrate that combinatorial Smad2/3 functional activities are required to maintain distinct embryonic and/or extra-embryonic cell identity during lineage priming in the epiblast before gastrulation.

Keywords: Bmp signaling; Nodal signaling; Smad2; Smad3; TGF-β signaling; cell fate allocation; embryonic stem cells; epiblast-like cells; extra-embryonic; lineage priming.

Graphical abstract

Figure 1

Smad2/3 Repress Expression of Extra-Embryonic and Naive Pluripotency Genes during Lineage Priming (A) WT, Smad2 KO, Smad3 KO, or Smad2/3 DKO ESCs (2iL) were stained for Oct4 and Nanog and counterstained with DAPI. (B) Venn diagrams showing significant changes in gene expression shared by Smad2 KO, Smad3 KO, and Smad2/3 DKO ESCs, relative to WT ESCs, as determined by microarray profiling (n = 3 or 4). Genes uniquely differentially expressed by Smad2 KO or Smad3 KO ESCs were excluded from this analysis. A summary of deregulated genes is presented in Table S1. (C) Pie charts of alkaline phosphatase (AP)-stained WT, Smad2 KO, Smad3 KO, or Smad2/3 DKO ESCs cultured for 5 days in the presence or absence of LIF (n = 3), corresponding to pluripotent, differentiated, or mixed colonies. See also Figure S1F. (D) Scatterplot showing significantly (p < 0.05, Benjamini-Hochberg adjusted) differentially expressed genes in Smad2/3 DKO EpiLCs compared to WT by RNA-seq (n = 3). The cutoff was set to >1.5-fold change. Differentially expressed genes near Smad2/3 ChIP-seq peaks in day 3 EBs (≤50 kb of its transcriptional start site [TSS]) are indicated in yellow. The pie chart indicates the proportion of differentially expressed genes also showing Smad2/3 ChIP-seq peaks (≤50 kb of TSS). (E) Heatmap showing relative expression levels of pluripotency marker genes in Smad2/3 DKO ESCs, EpiLCs, and day 3 EBs compared to WT controls (n = 3 or 4). Profiles of Smad2 KO and Smad3 KO day 3 EBs are shown on the right.

Figure 2

Smad2/3 Influences the Activity of Oct4-Occupied Distal Regulatory Enhancers during Priming (A) Heatmap of regulatory elements with differential chromatin accessibility in Smad2/3 DKO EpiLCs compared to WT EpiLCs, as measured by ATAC-seq (false discovery rate [FDR] < 0.05, fold change > 1.5). Sites with decreased chromatin accessibility (top) and sites with increased chromatin accessibility (bottom) are ranked by decreasing and increasing ATAC signal change. See also Table S3. (B and C) Pie charts indicating (B) distributions of differential accessible sites in Smad2/3 DKO EpiLCs compared to WT overlapping with regulatory elements gained or lost during the ESC-to-EpiLC transition or (C) the distance to known TSS as defined by Genomic Regions Enrichment of Annotations Tool (GREAT). (D) Heatmap read density plots of p300, H3K27ac, and H3K4me1 ChIP-seq signal at regulatory elements with differential accessibility in Smad2/3 DKO EpiLCs (ranked as in A). (E) Heatmap depicting the log2 fold change (log2FC) in gene expression in Smad2/3 DKO EpiLCs relative to WT EpiLCs as determined by RNA-seq. Genes nearest regulatory elements with differential accessibility in Smad2/3 DKO EpiLCs are shown. (F) Genome browser snapshots of RNA-seq and ATAC-seq tracks in Smad2/3 DKO and WT EpiLCs at selected genomic loci. ATAC-seq of WT ESCs, ChIP-seq tracks of Smad2/3 occupancy in ESCs and day 3 EBs, and Oct4 and Otx2 occupancy in EpiLCs are also shown. (G) Heatmap read density plots of WT Smad2/3, Smad1, Oct4, and Otx2 ChIP-seq signal in the indicated cell types at regulatory elements with differential accessibility in Smad2/3 DKO EpiLCs (ranked as in A). ESCs were treated with 10 ng/mL BMP4 for Smad1 ChIP-seq. (H) Motif enrichment analysis of regulatory elements with differential chromatin accessibility in Smad2/3 DKO EpiLCs. Motifs for transcription factors associated with primed or naive, extra-embryonic, and neural cell states were significantly enriched.

Figure 3

Smad2/3 Governs Embryonic Cell Fate Specification (A) Heatmap showing the log2 fold change (log2FC) in expression of the top 20 genes upregulated in day 3 WT EBs relative to WT ESCs (left) in comparison with their expression changes in Smad2/3 DKO (right). See also Table S1. (B) Anti-AP2γ and Oct4 immunofluorescence staining of WT and Smad2/3 DKO day 2 and 4 PGCLCs. (C) WT and Smad2/3 DKO NPCs at day 7 stained with anti-Tuj1 and counterstained with DAPI. (D) Bright-field images of control WT or Smad2/3 DKO NPCs grown in the absence or presence of BMP4 (5 ng/mL) at days 3, 5, and 7. (E) WT and Smad2/3 DKO NPCs grown in the absence or presence of BMP4 (5 ng/mL) stained with anti-Sox1 and Ap2γ and counterstained with DAPI at day 5.

Figure 4

Enhanced Bmp Signaling Caused by the Absence of Smad2/3 Disturbs Embryonic Patterning (A) Heatmap showing the log2FC in expression of selected Bmp target genes and those involved in DNA methylation for Smad2/3 DKO ESCs, EpiLCs, and day 3 EBs compared to WT controls (n = 3 or 4). (B) Western blot analysis of WT and Smad2/3 DKO day 3 EBs treated with DMSO or LDN-193189 (250 nM, 24 hr from day 2 to day 3) or untreated. Blots were probed with the indicated antibodies. (C and D) Anti-p-Smad1/5/8 and Oct4 (C) or Eomes and Otx2 (D) immunofluorescence staining of E5.5 WT or Smad2/3 DKO mouse embryos. (E) Heatmap showing the log2FC in expression of the top 20 upregulated genes (p < 0.05, fold change > 1.5) in Smad2/3 DKO day 3 EBs compared to WT day 3 EBs and their expression changes in Smad2 KO and Smad3 KO day 3 EBs (n = 4, averaged). See also Table S1.