With the emergence of new high-throughput technologies such as RNA sequencing, we have recently witnessed a remarkable increase in our knowledge of mammalian transcriptome content and diversity. There has been a particular surge in our understanding of the transcriptome diversity between different tissues and cell types. For example, Wang et al. performed an RNA-Seq analysis of 15 human tissues and cell lines and identified over 22,000 tissue-specific AS events . Other studies have established the association between tissue-specific expression of SFs and genome-wide changes in tissue-specific splicing patterns [42, 45], which underscores a critical role of AS regulation in tissue differentiation and specialization.
The majority of previous gene expression studies of human placental tissue have only provided gene-level insights [6–10], driving the need for higher-resolution analysis to enable a better understanding of the complexity of the placental transcriptome at the level of exon splicing. AS, which has a well-established role in cell differentiation [57, 58], may be critical for the proper functioning of the placenta, an organ composed of a variety of differentiated cell types, each with its own specific functions during pregnancy. Thus, uncovering the complexity of AS in the placental transcriptome will provide a valuable basis for understanding genes with functional and clinical relevance in placental biology and pathophysiology.
In the present study, we used RNA-Seq to characterize the transcriptome of selected compartments of the human placenta from normal term pregnancies. RNA-Seq allows an unbiased and sensitive interrogation of the full repertoire of placental mRNA transcripts. We took a two-step approach to analyze the RNA-Seq data at both the gene-level and the exon-level. First, we investigated differential gene expression between the placental and other human tissues to identify genes that are specifically or abundantly expressed in the placenta. Second, we carried out exon profiling as well as SF expression profiling to find AS events and their potential regulators that are differentially present in the placental versus non-placental tissues.
We have compared placenta-enriched genes to genes with putative functional significance in the placenta using the mouse phenotype data and human PTB association study data. We observed that genes implicated in placental abnormalities and PTB are enriched among the genes with placenta-enriched expression profiles. We note that the mouse phenotype data from MGI were generated independent of any previously known gene expression pattern in the placenta. Among such genes are PRLR and F2R, genes encoding receptors for prolactin and thrombin, respectively, whose levels are precisely regulated during pregnancy [59, 60]. The enrichment of IL1-related genes was also noted, suggesting the importance of IL1 signaling in normal placental function and pregnancy. IGF2, one of the genes associated with abnormal placental phenotypes in mice, is known for its active role in placental and fetal growth [61, 62]. Together, these provide a link between highly expressed placenta-enriched genes and their functional importance in the placenta. Similarly, our work provides evidence suggesting the importance of genes uniquely expressed in the placenta in diverse pregnancy-related processes, with examples including CSH1 in the regulation of fetal growth , CGB in the maintenance of early pregnancy [64, 65], and human leukocyte antigen-G (HLA-G) in feto-maternal immune tolerance [66, 67]. In addition, we observed a significant enrichment of differentially spliced genes in the placenta among genes with placental phenotypes in the mouse, suggesting the importance of tissue-specific AS in placental development and function.
Because the HBM2.0 data all came from adult tissues, it is possible that some placenta-enriched genes identified in our study reflect age-specific expression signatures. Because of the unavailability of RNA-Seq data from other fetal tissues, we assessed this possibility using the GeneAtlas array data . There were 4 fetal tissues (brain, liver, lung, and thyroid) included in the GeneAtlas data. Of the 297 genes with at least 4-fold enrichment in the placenta over adult tissues in both the GeneAtlas array data and our RNA-Seq data, the vast majority (281 genes) were more than 4-fold enriched in the placenta compared with the 4 fetal tissues according to the GeneAtlas array data. This suggests that the placenta-enriched genes identified in our study reflect genuine placenta-associated gene expression signatures. In addition, the strong association of placental expression enrichment with placental disease-related gene sets further supports that most of the placenta-enriched genes found here reflect tissue effect rather than age effect.
Given the heterogeneous tissue composition of the placenta, we have characterized the transcriptome profiles of the placenta not only at the whole-organ level, but also at the sub-organ level. It should be noted that the placental samples used in our study (amnion, chorion, and decidua) may not be completely pure, containing minor contamination with other placental components. Nonetheless, our study demonstrated that they are highly enriched for the corresponding tissue types, displaying compartment-specific expression profiles and splicing patterns. The amnion is the innermost layer of the fetal membranes lining the amniotic cavity and is composed of an epithelial cell layer on top of a basement membrane and an avascular matrix [68, 69]. Consistent with these histological properties of the amnion, we have detected enrichment of genes involved in cell/focal adhesion and observed that the epithelial splicing regulator ESRP1 was highly expressed. Our splicing analysis of the amnion using RNA-Seq and RT-PCR revealed 20 and 12 known ESRP1 target exons, respectively, with differential splicing activities in the amnion. It should be noted that ESRP1 is a master cell-type-specific splicing regulator critical for maintaining the epithelial cell identity and has been implicated in a variety of developmental and disease processes . The ESRP1 target exons are strongly enriched in genes involved in the regulation of cell adhesion such as the exon in MINK1, that was found to be differentially spliced in the amnion compared to other human tissues by RNA-Seq and validated by RT-PCR. These data support a role of the ESRP1 splicing regulatory network in the amnion. The chorion, the outer layer of the fetal membranes in contact with the decidua, consists of the reticular layer, the basement membrane, and the trophoblast layer . Similar to the amnion, genes with a role in cell/biological adhesion are also enriched in the chorion, which may be important for the adherence of the trophoblast layer to the decidua . The enrichment of genes involved in vascular-related processes in the chorion may be explained by velamentous vessels traversing the extraplacental membranes or maternal vessels in interdigitating decidua processed along with the chorion. Unlike the two fetal membranes, the decidua is of maternal origin [1, 10]. It is noteworthy that genes related to female pregnancy were significantly enriched in this compartment of the placenta, further supporting the crucial role of this tissue in pregnancy. Of note, we observed significant differential expression of a splicing factor MBNL3 in the decidua. In future studies, it would be useful to examine how MBNL3 globally impacts gene splicing and function in the decidua.
We also examined potential interactions among genes highly expressed and differentially spliced in the placenta compared to other human tissues by constructing FI networks composed of sub-network modules enriched for specific gene categories and functional pathways. Analysis performed separately on each of the three placental tissues revealed enrichment of set of pathways commonly enriched in all three compartments, for example, regulation of cytoplasmic and nuclear SMAD2/3 signaling and TGF-beta receptor signaling. These pathways are known to be involved in a wide range of cellular processes , which reflects the versatile function of the placenta that can be achieved through diverse cellular activities occurring in different parts of the placenta. Among its other main functions, the placenta plays an important role as an immune barrier, protecting the fetus from the mother's immune system . This function is reflected by the enriched expression of transcription factors (TFs) involved in immune regulation such as GATA3 and IRF7 as well as the differential splicing of REL, a member of the Rel/NFKB family and NFATC2, a member of the nuclear factors of activated T cells transcription complex. HIF-1 alpha TF network is another pathway that was enriched in module 0 of all the three FI networks. The placenta, during its development, is exposed to different oxygen environments and tight regulation of oxygen homeostasis is necessary for proper placental development and function, which requires active involvement of the HIF-1 alpha TF network . These findings suggest: (1) the common importance of these pathways in the functioning of the different parts of the placenta examined in the present study; and (2) the importance of the regulation of gene expression and AS as critical mechanisms underlying anatomical, developmental, and functional specialization of the placenta. When the analysis was performed on all of the tissues combined, we observed the overrepresentation of ECM-related gene sets such as integrin signaling pathway, ECM-receptor interaction, focal adhesion, and integrin cell surface interactions. These results provide evidence for the role of ECM in placental development and placental cell proliferation as demonstrated in earlier studies [73, 74].