Comparative analysis of mouse and human placentae across gestation reveals species-specific regulators of placental development

Abstract

An increasing body of evidence points to significant spatio-temporal differences in early placental development between mouse and human, but a detailed comparison of placentae in these two species is missing. We set out to compare placentae from both species across gestation, with a focus on trophoblast progenitor markers. We found that CDX2 and ELF5, but not EOMES, are expressed in early post-implantation trophoblast subpopulations in both species. Genome-wide expression profiling of mouse and human placentae revealed clusters of genes with distinct co-expression patterns across gestation. Overall, there was a closer fit between patterns observed in the placentae when the inter-species comparison was restricted to human placentae through gestational week 16 (thus, excluding full-term samples), suggesting that the developmental timeline in mouse runs parallel to the first half of human placental development. In addition, we identified VGLL1 as a human-specific marker of proliferative cytotrophoblast, where it is co-expressed with the transcription factor TEAD4. As TEAD4 is involved in trophectoderm specification in the mouse, we posit a regulatory role for VGLL1 in early events during human placental development.

Keywords: Comparative study; Cytotrophoblast; Placenta; Placental development; Placental progenitors; Trophoblast stem cells.

Fig. 1.

CDX2 is highly expressed in a subpopulation of human villous cytotrophoblast near the chorionic plate. Immunohistochemistry for CDX2 in early post-implantation human placenta (week 5). (A-C) Nuclear expression was noted in a subpopulation of the villous CTB, concentrated near the chorionic plate (CP), and became less frequent towards the basal plate (BP). No nuclear staining was noted in the syncytiotrophoblast (STB; Ci,ii) or proximal cell column trophoblasts (pCC; Ciii). Squares in A,B indicate the areas magnified in B,C, respectively. CTB, cytotrophoblast; pCC, proximal cell column trophoblast; STB, syncytiotrophoblast; wk, week.

Fig. 2.

ELF5 is expressed in both villous cytotrophoblast and proximal cell column trophoblast in early human placenta. In situ hybridization was performed using an ELF5-specific probe, showing a wider expression pattern in first trimester human placentae than previously reported. (A) ELF5 expression in first trimester villous samples was highest in the CTB. (B) By the second trimester, ELF5 expression was decreased and became restricted to the CTB. (C) Expression of ELF5 in week 5 extravillous trophoblast located in the proximal cell columns (pCCs). In C, the right image is at twice the magnification of the left image; in A,B, the insets are at twice the magnification of the main image. CTB, cytotrophoblast; pCC, proximal cell column; STB, syncytiotrophoblast; wk, week.

Fig. 3.

Comparison of mouse and human placenta gene expression signatures across gestation. (A) Genome-wide expression profiles of mouse placentae across gestation. Samples are color-coded according to gestational age. Heatmap shows expression profiles of 2947 differentially expressed genes (DEGs, v=0.02, a multi-group analysis q<0.05, FC≥2.0). (B) Genome-wide expression profiles of human placentae across gestation. Samples are color-coded according to data-driven groups used for statistical analysis. Heatmap shows expression profiles of 1195 DEGs (v=0.02, a multi-group analysis q<0.05, FC≥2.0). (C) Venn diagram representing overlap of DEGs between mouse and human datasets. (D) Pie charts showing comparison in expression pattern direction between human and mouse common DEGs (517) (heatmaps in Fig. S3A). Central pie chart shows expression pattern in human placentae. Genes with expression pattern other than up- and downregulation across gestation were combined and labeled as ‘Other’. The side charts show expression pattern comparison between the human subclasses up- and downregulated genes, and their expression patterns in mouse. (E) Biological function gene ontology (GO) terms enriched in commonly up- (red) and down- (green) regulated genes in mouse and human placentae. Font size is proportional to the enrichment −logP values, with bigger font size representing higher enrichment.

Fig. 4.

Affinity propagation analysis of mouse and human DEGs. (A) Affinity propagation (AP) analysis of the 2947 DEGs in the mouse dataset distinguished seven clusters of co-regulated genes. Heatmap represents expression values of genes; line graphs (right) represent the average trace for each cluster. (B) Table showing percentage distribution of mouse placental DEGs (2947 genes) across the seven AP clusters (‘Placenta dataset’) and of the DEGs upregulated in mTSC at day 0 (‘Up at d0’, 765 genes) or in day 7-differentiated mTSC (‘Up at d7’, 604 genes). Clusters in which genes are enriched in either mTSC d0 or d7 over the placenta dataset are highlighted in red, whereas those in which there is a relative depletion are highlighted in green. (C) AP analysis of the 1195 DEGs in the human dataset distinguished six clusters of co-regulated genes. Heatmap represents expression values of genes; line graphs (right) represent the average trace for each cluster. (D) Table showing percentage distribution of human placental DEGs (1195 genes) across the six AP clusters (‘Placental dataset’) and of the DEGs upregulated in the early (first trimester) CTB (week 8-10) compared with the mature CTB (week 12-39) (197 genes) (see analysis in Fig. S5A-C). Clusters in which genes are enriched in the early CTB over the placenta dataset are highlighted in red, whereas those where there is a relative depletion are highlighted in green.

Fig. 5.

Comparison of expression profiles of best-correlated mouse and human co-expression clusters. (A) Graphs showing the AP clusters from the HL dataset (which included data from term human placentae) and HS dataset (which did not include data from term human placentae) that showed the highest similarity to each mouse cluster. Data from mouse clusters are shown in red and data from human clusters are shown in blue. The Euclidean distance (E.d.) value for each pair of curves is shown. (B) Gene Ontology enrichment analysis shows overlapping function between the genes included in the cluster pairs with closely fitting curves, Ms_1/HS_2 and Ms_2/HS_5. # indicates the number of genes in common between the mouse and human clusters. Each rectangle is a representative GO term from the enrichment analysis. The representatives are joined into ‘super-groups’ of loosely related terms, visualized with different colors. Size of the rectangles is proportional to the P value. Overlapping functions between mouse and human clusters are marked with matching colors.

Fig. 6.

In situ hybridization for Cdx2 and Tead4 in mouse placentae confirms upregulation with increasing gestational age. (A) In situ hybridization for Tead4. Tead4 expression was observed in the chorion (ch) at E7.5, and in both the labyrinth (lab) and the junctional zone (jz) at E11.5 and E16.5. (B) In situ hybridization for Cdx2. Cdx2 expression was observed in the chorion (ch) and primitive streak (PS) at E7.5, and in the labyrinth (near the chorionic plate, lab) and junctional zone (jz) later in gestation. PAS staining in a consecutive section at E16.5 indicates Cdx2 expression in glycogen cells.

Fig. 7.

VGLL1 and TEAD4 are co-expressed in the human villous cytotrophoblast across gestation. (A) In early post-implantation human placenta, VGLL1 and TEAD4 are co-expressed in the villous cytotrophoblast (CTB). Extravillous trophoblast in the proximal cell column (pCC) expresses VGLL1 but not TEAD4. (B) In second trimester placentae, the villous CTB (now discontinuous) maintains expression of both VGLL1 and TEAD4. (C) At full term, VGLL1 expression is visible both in the persistent villous CTB (term villi) and also in the mature extravillous trophoblast (term basal plate). TEAD4 expression is restricted to the villous CTB. CTB, cytotrophoblast; pCC, proximal cell column; STB, syncytiotrophoblast; wk, week.

Fig. 8.

VGLL1 is expressed in primary and hESC-derived CTB in vitro. (A) Western blot analysis of VGLL1 and β-actin (loading control) in week 6 whole placenta (P) and primary first trimester CTB (#1, week 13; #2, week 11). Expression was analyzed both in freshly isolated CTB (d0) and in differentiated CTB (d4). (B) qRT-PCR analysis of VGLL1, TEAD4, CTB markers (TP63 and CDX2) and a pluripotency marker (OCT4) in undifferentiated hESCs (U) and in hESC-derived CTB-like cells (d4) following step 1 of our two-step BMP4-based trophoblast differentiation protocol (Horii et al., 2016) (n=3). (C) Western blot analysis of VGLL1 and β-actin (loading control) in undifferentiated H9 hESCs (U) and in hESC-derived CTB-like cells (d4). (D) Double immunostaining of undifferentiated H9 hESC (U) and hESC-derived CTB-like cells (d4) with VGLL1 and either OCT4 (top) or pan-trophoblast marker KRT7 (bottom). (E) qRT-PCR analysis of VGLL1, TP63 and OCT4 in hESC clones stably expressing either scrambled or VGLL1-targeting shRNA. n=3, *P<0.05 compared with scramble d4.