A systematic review of transcriptomic studies of the human endometrium reveals inconsistently reported differentially expressed genes

Abstract

Genome-wide analysis of gene expression has been widely applied to study the endometrium, although to our knowledge no systematic reviews have been performed. Here, we identified 74 studies that described transcriptomes from whole (unprocessed) endometrium samples and found that these fitted into three broad investigative categories; endometrium across the menstrual cycle, endometrium in pathology, and endometrium during hormone treatment. Notably, key participant information such as menstrual cycle length and body mass index was often not reported. Fertility status was frequently not defined and fertility-related pathologies, such as recurrent implantation failure (RIF) and recurrent pregnancy loss, were variably defined, while hormone treatments differed between almost every study. A range of 1307-3637 reported differentially expressed genes (DEG) were compared in 4-7 studies in five sub-categories; (i) secretory vs proliferative stage endometrium, (ii) mid-secretory vs early secretory stage endometrium, (iii) mid-secretory endometrium from ovarian stimulation-treated participants vs controls, (iv) mid-secretory endometrium from RIF patients vs controls, and (v) mid-secretory eutopic endometrium from endometriosis patients vs controls. Only the first two sub-categories yielded consistently reported DEG between ≥3 studies, albeit in small numbers (<40), and these were enriched in developmental process and immune response annotations. This systematic review, though not PROSPERO registered, reveals that limited demographic detail, variable fertility definitions and differing hormone treatments in endometrial transcriptomic studies hinders their comparison, and that the large majority of reported DEG do not advance the identification of underlying biological mechanisms. Future studies should apply network biology approaches and experimental validation to establish causal gene expression signatures.

Figure 1

Systematic review process and the characteristics of studies meeting the inclusion criteria. (A) Flow diagram of the systematic review process for this study, culminating in 74 studies that met the inclusion criteria. (B) Plot of the number of included studies published in each year of the review period. (C) Plot of the number of participants in each of the included studies. The median number of participants is indicated by a black line. (D) Venn diagram indicating how the included studies fit within three investigative categories. (E) Plot of the frequency of menstrual cycle stages sampled in 72/74 included studies (cycle stage not reported in two studies).

Figure 2

Definitions and characteristics of investigative categories. (A) Pie chart indicating the fertility definitions reported in non-pathologically included studies. (B) Pie chart showing the pathologies in the included studies. (C) Pie chart representing the hormone treatments used in the included studies.

Figure 3

Patterns of differentially expressed gene reporting in common comparisons among studies. (A) The proportion of commonly reported differentially expressed genes (cDEG) across ≥2 studies for each comparison that show consistent (blue) and inconsistent (red) directionality of change in expression. Number of cDEG is indicated in each bar. (B) The number of consistently changed cDEG for each comparison plotted against number of studies in which they are commonly reported. (C–G) Dot plots of consistently changed cDEG fold-change averaged between studies for each comparison. The most commonly reported cDEG gene names are indicated on the plots.

Figure 4

Gene ontology annotations of commonly reported differentially expressed genes. (A) All significantly enriched biological processes within the 39 commonly reported differentially expressed genes (cDEG) from ≥3 studies of secretory vs proliferative endometrium (FDR <0.05). Biological processes are split into broad categories: development process and signalling and metabolic processes. cDEG gene names enriched for these processes are indicated in the bars. (B) All significantly enriched biological processes within the 28 cDEG from ≥4 studies of mid-secretory vs early secretory endometrium (FDR <0.05). Biological processes are split into broad categories: immune response and development process. cDEG gene names enriched for these processes are indicated in the bars.