A transcriptomic comparison of in vitro models of the human placenta

Abstract

Introduction: Selecting an in vitro culture model of the human placenta is challenging due to representation of different trophoblast cell types with distinct biological roles and limited comparative studies that define key characteristics of these models. The aim of this research was to compare the transcriptomes of common in vitro models of the human placenta compared to bulk human placental tissue.

Methods: We performed differential gene expression analysis on publicly available transcriptomic data from 7 in vitro models of the human placenta (HTR-8/SVneo, BeWo, JEG-3, JAR, Primary Trophoblasts, Villous Explants, and Trophoblast Stem Cells) and compared to bulk placental tissue from 2 cohort studies (CANDLE and GAPPS) or individual trophoblast cell types derived from bulk placental tissue.

Results: All in vitro placental models had a substantial number of differentially expressed genes (DEGs, FDR<0.01) compared to the CANDLE and GAPPS placentas (Average DEGs = 10,624), and the individual trophoblast cell types (Average DEGs = 5413), indicating that there are vast differences in gene expression. Hierarchical clustering identified 54 gene clusters with distinct expression profiles across placental models, with 23 clusters enriched for specific KEGG pathways. Placental cell lines were classified by fetal sex based on expression of Y-chromosome genes that identified HTR-8/SVneo cells as female origin, while JEG-3, JAR, and BeWo cells are of male origin.

Discussion: None of the models were a close approximation of the human bulk placental transcriptome, highlighting the challenges with model selection. To enable appropriate model selection, we adapted our data into a web application: "Comparative Transcriptomic Placental Model Atlas (CTPMA)".

Keywords: In vitro; Placenta; RNA sequencing; Transcriptome; Trophoblast.

Fig. 1.

Principal Components Analysis (PCA) for the in vitro placental models compared to A) CANDLE and GAPPS placental tissue, B) FACs Cytotrophoblasts (CTBs), C) FACS Extravillous Trophoblasts (EVTs), and D) Villous tissue digest Syncytiotrophoblasts (STBs). Variance explained by each of the first three PCs is indicated in the axis labels.

Fig. 2.

The number of differentially expressed genes (FDR<0.01) for each placental model after comparison to A) CANDLE and GAPPS placental tissue, B) FACs Cytotrophoblasts (CTBs), C) FACS Extravillous Trophoblasts (EVTs), and D) Villous tissue digest Syncytiotrophoblasts (STBs).

Fig. 3.

A) Hierarchical clustering of expression data from the CANDLE and GAPPS placental tissue model (N = 12,790 genes) identified 54 clusters of genes. B) Average Expression of all genes within each cluster across each placental model. C) Significant KEGG pathways (FDR<0.05) enriched for genes that are members of individual clusters.

Fig. 4.

A) Average expression of genes of interest for placental biology including placenta specific, placenta enriched, and placenta enhanced genes defined by the Human Protein Atlas. 218 genes of 234 placental HPA genes were expressed at sufficient levels (average logcpm>0) in our placental models. B) Average expression of Absorption, Distribution, Metabolism, and Excretion (ADME) genes from PharmaADME. 158 genes of 298 ADME genes were expressed at sufficient levels (average logcpm>0) in our placental models to be included.

Fig. 5.

Fetal sex of placental models as determined by quantification of Y-chromosome gene expression. Placental models with mixed fetal sex (Villous Explants and Primary Trophoblasts) are displayed separately by sex.