Cell-type Phylogenetics and the Origin of Endometrial Stromal Cells

Abstract

A challenge of genome annotation is the identification of genes performing specific biological functions. Here, we propose a phylogenetic approach that utilizes RNA-seq data to infer the historical relationships among cell types and to trace the pattern of gene-expression changes on the tree. The hypothesis is that gene-expression changes coincidental with the origin of a cell type will be important for the function of the derived cell type. We apply this approach to the endometrial stromal cells (ESCs), which are critical for the initiation and maintenance of pregnancy. Our approach identified well-known regulators of ESCs, PGR and FOXO1, as well as genes not yet implicated in female fertility, including GATA2 and TFAP2C. Knockdown analysis confirmed that they are essential for ESC differentiation. We conclude that phylogenetic analysis of cell transcriptomes is a powerful tool for discovery of genes performing cell-type-specific functions.

The Origin of Animal Multicellularity and Cell Differentiation

“…This endeavor should benefit from ongoing efforts to obtain unbiased transcriptomes and proteomes for animal cell types, notably by single-cell approaches (Achim et al 2015; Wagner et al 2016; Villani et al 2017; Regev et al 2017). Analyzed with refined methods for phylogenetic reconstruction (Liang et al 2015; Kin et al 2015; Musser and Wagner 2015), such datasets should help illuminate our understanding of the origin and evolution of animal cell differentiation.…”

Embryo implantation evolved from an ancestral inflammatory attachment reaction

“…However, there is a significant 40% increase in the level of CD14 during pregnancy; CD14 is primarily produced by macrophages, so this increase suggests there could be some additional macrophage activity. CD14 and its interaction partner lipopolysaccharide-binding protein (LBP) have also been detected in opossum endometrial stromal cells (31). Further, there is a qualitative increase in neutrophil-specific elastase, ELANE, with few mapped transcripts in nonpregnant and midgestation samples but expression of this gene in all late-gestation samples, with a mean level of 1.1 transcripts per million (TPM), suggestive of neutrophil infiltration of the uterus during the attachment reaction.…”

Homology in the Age of Developmental Genomics

“…Fortunately, for mammalian cells the biochemical and the statistical methods are highly congruent and suggest a robust threshold that seems to be independent of cell type (Hebenstreit et al 2011 ;Kin et al 2015 ). It is 1 RPKM or 2-3 tpm for pure cell samples, which corresponds to about one transcript per cell (Hebenstreit et al 2011 ).…”

“…Once we have a tree that is well supported and biologically meaningful, we can use the character tree to identify genes that likely have been recruited at the time of origin of certain character identities. That means that the phylogenetic analysis of character or cell type gene expression data can directly lead to hypotheses about the genetic basis for the origin of the novel character (Kin et al 2015 ). These candidate genes can then be tested for their importance for the development of the derived character identity by using any of the standard knockout or knockdown experimental techniques like morpholinos ® or RNAi and iCrisper.…”

“…Analyzing discretized expression data leads to easily interpreted ancestral state reconstructions. These ancestral state reconstructions lead to hypotheses about the gene expression profi le of the ancestral cell type and can also be used to infer candidate gene recruitment events associated with the origin of novel cell identities (Kin et al 2015 ). Among the candidate gene reconstructions, it is useful to eliminate genes that have expression levels in the gray zone around the threshold of 2 tpm.…”