Hypoxia and loss of GCM1 expression prevent differentiation and contact inhibition in human trophoblast stem cells

Abstract

During the first stages of embryonic development, the placenta develops under very low oxygen tension (∼1%-2% O₂), so we sought to determine the regulatory role of oxygen in human trophoblast stem cells (hTSCs). We find that low oxygen promotes hTSC self-renewal but inhibits differentiation to syncytiotrophoblast (STB) and extravillous trophoblast (EVT). The transcription factor GCM1 (glial cell missing transcription factor 1) is downregulated in low oxygen, and concordantly, there is substantial reduction of GCM1-regulated genes in hypoxic conditions. Knockout of GCM1 in hTSC likewise impaired EVT and STB formation. Treatment with a phosphatidylinositol 3-kinase (PI3K) inhibitor reported to reduce GCM1 protein levels likewise counteracts spontaneous or directed differentiation. Additionally, chromatin immunoprecipitation of GCM1 showed binding near key genes upregulated upon differentiation including the contact inhibition factor CDKN1C. Loss of GCM1 resulted in downregulation of CDKN1C and corresponding loss of contact inhibition, implicating GCM1 in regulation of this critical process.

Keywords: CDKN1C; GCM1; cell column; cytotrophoblast; differentiation; extravillous trophoblast; hypoxia; placenta; placental villi; syncytiotrophoblast; trophoblast stem cell.

Graphical abstract

Figure 1

Reduced and impaired hTSC differentiation in hypoxic conditions (A) Trophoblast stem cells were cultured for 72 h in varying levels of oxygen (20%, 5%, 2% O₂). Flow cytometry plots indicate levels of hTSC (ITGA6 and EPCAM) and EVT (ITGA1 and HLA-G) markers. Note reduction in ITGA1⁺ population in low O₂. (B) ITGA1⁺ HLA-G⁺ population in O₂ and cell line indicated (4 cell lines, n = 3 replicates for each cell line at 3 different passages). Statistical significance was determined via a two-tailed t test. (C) EVT differentiation in 20% O₂ starting with hTSC in oxygen concentration indicated. Successful differentiation is indicated by the upregulation of surface markers ITGA1 and HLA-G and downregulation of EPCAM and ITGA6. (D) STB differentiation in 20% O₂ starting with hTSC in oxygen concentration indicated. STB formation is indicated by loss of TEAD4 and increase in hCGB staining in a cell. (E) EVT differentiation undertaken at oxygen level indicated. (F) STB differentiation undertaken at oxygen level indicated. (G) Principal component analysis (PCA) showing gene expression from hTSC cultured in varying oxygen concentrations. Ovals encompassing all 2%, 5%, and 20% O₂ samples are drawn manually. (H) Hierarchical gene clustering of RNA-seq samples in (G). Red dotted lines indicate the shift in gene expression from 20% O₂ and 2% O₂ labeled as cluster 1 and cluster 2. (I) Gene set enrichment analysis (GSEA) analysis of cluster 1 and cluster 2. (J) Volcano plot showing gene expression differences between TSCs cultured in 20% O₂ to TSCs cultured in 2% O₂. Dashed lines indicate significance and log2 fold change cutoff. (K) Bar graphs showing FPKM of specific genes of interest (same samples as in G, significance indicated corresponds to p_adj values from DESEQ2 analysis, see Table S1). (L) Violin plot showing expression of genes specific to hTSC, EVT, or STB for hTSCs grown in the indicated oxygen concentration. When comparing log₂ fold change between genes in each set, differences between all sets at all oxygen concentrations are significant (p < 0.001) (For analysis in G–L, 3 cell lines; BT2, CT1, CT3; n = 3 replicates for each line in each condition over 3 passages, except BT2 at 20% O₂n = 2).

Figure 2

Impaired differentiation upon genetic or chemical reduction in GCM1 level (A) Strategies for mutation of GCM1 using a two-sgRNA CRISPR approach. Lines were generated by deletion of the exon2/intron2 boundary, and by ablation of exon 3, either of which should disrupt the DNA-binding domain of GCM1. (B) Immunofluorescence staining of GCM1 and TEAD4 in control (non-target, NT sgRNA) and GCM1^−/− hTSC. Sporadic GCM1⁺ TEAD4^lo hTSCs are present only in NT control hTSCs. (C) Flow cytometric analysis from EVT differentiation of GCM1 KO1 and NT control TSC. NT hTSC differentiation produced ITGA1^hi/HLA-G^hi cells whereas GCM1^−/− TSC did not. (D) Bar graphs showing formation of ITGA1^hi/HLA-G^hi population from control and GCM1^−/− TSC (2 cell lines, CT1 n = 2 clonal lines, CT3 n = 3 clonal lines for both NT and KO). (E) 3D STB formation of NT and GCM1^−/− hTSC. Control hTSCs form a fluid-filled syncytium while GCM1^−/− hTSCs form a cluster of cells. (F) hCGB ELISA was performed using supernatant from GCM1^−/− and control hTSC (2 cell lines, CT1 n = 2 clonal lines, CT3 n = 3 clonal lines for both NT and KO). Statistical significance was determined via a two-tailed t test. (G) PCA comparing NT and GCM1^−/− hTSC, EVT, and STB3D. Note that GCM1^−/− cells regardless of differentiation state cluster closer to the hTSC population, and similarity of GCM1^−/− lines 1 and 2 (TSC: n = 8 NT, n = 7 KO; EVT: n = 9 NT, n = 8 KO; STB3D n = 7 NT, n = 6 KO clonal replicates). Ovals encompassing WT TSC, STB, and EVT, as well as GCM1 KO STB and EVT, are drawn manually. (H) PCA of control (NT) and GCM1^−/− hTSCs, compared with WT hTSCs grown at different O₂ concentrations. Note that GCM1^−/− hTSCs cluster on principal component axis 1 with WT hTSCs grown at 2% O₂ (n = 8 WT 20% O₂, n = 9 2% O₂, n = 8 NT 20% O₂, n = 7 KO 20% O₂). (I) Scatterplot of genes differentially regulated in hypoxia (same set as Figure 1J) showing their relative expression in 2% and 20% O₂ and their relative expression in GCM1^−/− hTSC and control cells. Examples of placental differentiation genes are shown in blue, while genes involved in glycolysis are shown in red. (J) Bright-field images of GCM1^−/− TB-ORG cultured in mTOM media. (K) Left: Bright-field images of GCM1^−/− TB-ORG culture in mTOM media-CHIR99021. Right: Immunofluorescent staining for trophoblast markers in GCM1^−/− TB-ORG (representative of n = 5 images for NT and KO). (L) NT and GCM1^−/− TB-ORG differentiated to EVT. (M) Flow cytometry of NT and GCM1^−/− TB-ORG differentiated to EVT. GCM1^−/− hTSCs fail to upregulate the EVT marker HLA-G but do upregulate the cell column marker, ITGB6 (representative image, n = 5 for NT and KO). (N) Expression of genes associated with differentiation (CGB, ENDOU) or stemness (TP63), normalized to the housekeeping gene TBP, in steady-state TB-ORG conditions with 5 μM LY294002 or vehicle control (n = 3 cell line replicates). Statistical significance was calculated using a one-tailed t test. (O) Expression of EVT genes upon differentiation to EVT with 5 μM LY294002 or vehicle control (n = 3 cell lines replicates). Statistical significance was calculated using a one-tailed t test.

Figure 3

GCM1 positively regulates differentiation-associated genes (A) Motif analysis of GCM1-binding sites shows very strong enrichment for GCM motif, indicating successful and specific ChIP. (B) GCM1 enrichment over PGF (left) and LMO2 (right). (C) ATAC-seq enrichment in TSC, EVT, and STB over GCM1-binding sites. (D) Plot showing the percentage of genes whose promoters are within a given distance of a GCM1-binding site that show upregulation or downregulation in GCM1^−/− cells. (E) Plot showing the percentage of genes whose promoters are within a given distance of an EVT-specific ATAC-seq site that show upregulation or downregulation in GCM1^−/− cells (left), motif analysis for EVT-specific peaks (right). (F) Plot showing the percentage of genes whose promoters are within a given distance of an STB-specific ATAC-seq site that show upregulation or downregulation in GCM1^−/− cells (left), motif analysis for STB-specific peaks (right).

Figure 4

GCM1 positively regulates CDKN1C and contact inhibition (A) GCM1 enrichment over imprinted KCNQ locus. (B) Hi-C interaction data over the GCM1 locus. Note physical association between GCM1-binding site and CDKN1C promoter. In (A) and (B), the highest GCM1 peak is indicated with a magenta arrow. (C) Expression of genes indicated in plating conditions (cell number and growth time) indicated. Note that plating at higher densities leads to higher expression of GCM1 and CDKN1C (CT1 hTSCs, n = 4 replicates). Statistical significance was determined via a two-tailed t test. (D) GCM1 protein levels increase with higher confluence. (E) Expression of CDKN1C in oxygen concentration indicated. (F) Expression of CDKN1C in control and GCM1^−/− KO1 and KO2 hTSC and differentiated cells (significance marked by p_adj. value from DESEQ2 analysis, see Table S3). (G) Cell number after plating 50k cells and allowing cells to grow for indicated number of days. Note a leveling off in non-targeting cells as cell lines reach confluence, but continued growth in GCM1^−/− hTSCs (CT3 hTSCs, n = 4 replicates, including n = 2 GCM1KO1 and n = 2 GCM1KO2). Statistical significance was determined via a two-tailed t-test.