Cellular junction and mesenchymal factors delineate an endometriosis-specific response of endometrial stromal cells to the mesothelium

Abstract

Endometriosis is a debilitating gynecologic disorder that affects ∼10% of women of reproductive age. Endometriosis is characterized by growth of endometriosis lesions within the abdominal cavity, generally thought to arise from retrograde menstruation of shed endometrial tissue. While the pathophysiology underlying peritoneal endometriosis lesion formation is still unclear, the interaction between invading endometrial tissue and the peritoneal mesothelial lining is an essential step in lesion formation. In this study, we assessed proteomic differences between eutopic endometrial stromal cells (ESCs) from women with and without endometriosis in response to peritoneal mesothelial cell (PMC) exposure, using single-cell cytometry by time-of-flight (CyTOF). Co-cultured primary eutopic ESCs from women with and without endometriosis with an established PMC line were subjected to immunostaining with a panel of Maxpar CyTOF metal-conjugated antibodies (n = 28) targeting cell junction and mesenchymal markers, which are involved in cell-cell adhesions and epithelial-mesenchymal transition. Exposure of the ESCs to PMCs resulted in a drastic shift in cellular expression profiles in ESCs derived from endometriosis, whereas little effect by PMCs was observed in ESCs from non-endometriosis subjects. The transcription factor SNAI1 was consistently repressed by PMC interactions. ESCs from endometriosis patients are unique in that they respond to PMCs by undergoing changes in adhesive properties and mesenchymal characteristics that would facilitate lesion formation.

Keywords: Cellular adhesion; CyTOF; Endometriosis; Mass cytometry; Mesenchymal factors.

Fig. 1.

PMC co-culture results in a shift in ESC expression profiles in endometriosis but not in endometriosis-free subjects. (A) Experimental scheme shows control ESC culture and ESC-PMC co-culture (10:1 ratio). After 24 h incubation, the cells were subjected to CyTOF analysis. Gating on ESCs for CD10⁺KRT8/18⁻ profile (indicated by the outlined area in each plot) was used to eliminate LP9 PMCs from the analysis: Effective exclusion of LP9 PMCs is illustrated in lone LP9 cell culture (bottom left panel), and inclusion of stromal cells is illustrated in lone ESC cultures (bottom middle panel). (B) Dimensionality reduction principal component analysis (PCA) was used to align the three normal/endometriosis ESC sets (167/163; 171/169; and 172/170), in the presence or absence of LP9 PMC co-culture, based on expression profiles of the 28 factors tested relative to the PC1 and PC2 coordinates. A shift in coordinates of distinct ESC subpopulations was observed in the endometriosis ESCs (163, 171, 172) with LP9 PMC co-culture (within dotted circles) relative to solo cultures. This shift was less evident in the normal samples (167, 171, and 172). (C) The divergence of endometriosis ESCs due to LP9 PMC co-culture (yellow) from solo cultures of ESCs [control (red) or endometriosis (green)] or co-cultures with control ESCs (blue)], all of which largely superimpose, is demonstrated in expression profile trajectories from pseudotime analysis, combining all samples.

Fig. 2.

Mesenchymal markers are decreased in the major cell clusters from endometriosis ESCs (subject 163) when exposed to PMCs compared to normal ESCs (subject 167). (A) T-SNE cell clustering, based on the expression of the 28 factors, delineated 22 distinct cell clusters. (B) Cell clustering of differentially expressed mesenchymal transcription regulators SNAI1 and ZEB2, tight junction associated proteins ZO1 and ZO2, gap junction protein Cx43, N-cadherin (CDH2) and caveolin (CAV2) in ESCs is shown as a heatmap. For ease of visualization, the clusters were ordered based on those showing overall lowest average expression (bright blue) on the left and overall highest average expression (bright red) on the right. Grey bars/areas in the heatmaps indicate lack of cells with a particular expression profile in the sample (i.e., no cell cluster). Balloon plots indicate the size of the cell clusters (i.e., % of total cells within a particular cluster).

Fig. 3.

Mesenchymal markers are decreased in the major cell clusters from endometriosis ESCs (subject 169) when exposed to PMCs compared to normal ESCs (subject 171). (A) T-SNE cell clustering, based on the expression of the 28 factors, delineated 32 distinct cell clusters. (B) Cell clustering of differentially expressed mesenchymal and adhesion markers SNAI1, TWIST1, CAV2, TGFBR2 and NOTCH1 in ESCs is shown as a heatmap, and as balloon plots, both plotted as described in Fig. 2.

Fig. 4.

Mesenchymal markers are decreased in the major cell clusters from endometriosis ESCs (subject 170) when exposed to PMCs compared to normal ESCs (subject 172.) (A) T-SNE cell clustering, based on the expression of the 28 factors, delineated 21 distinct cell clusters. (B) Cell clustering of differentially expressed mesenchymal and adhesion markers SNAI1, ZEB2, CAV2, FN1, EGFR and TGFB1 in ESCs is shown as a heatmap, and as balloon plots, both plotted as described in Fig. 2.

Fig. 5.

SNAI1 levels consistently decrease in endometriosis ESCs exposed to LP9 PMC co-culture. (A, B, C, respectively) Violin plots showing expression distribution of factors that were differentially expressed due to LP9 PMC exposure in endometriosis but not endometriosis-free samples in each of the 3 normal/endometriosis sets compared: (A) 167/163; (B) 171/169; and (C) 172/170. (Violin plots for all the 28 factors for these pairings are shown in Supplemental Figs. 2, 3, and 4, respectively. SNAI1 levels were significantly decreased (p < 0.05) in the endometriosis samples exposed to LP9 PMCs in all 3 sets. Significance was determined by Duncan’s multiple range test and statistical differences (P < 0.05) between samples, as indicated by different letters (e. g. a vs. b; b vs. d). Samples with the same letter (e.g., a vs. a) indicate no statistical significance. In addition, ‘a’ indicates the highest mean expression and ‘d’ indicates the lowest mean expression value for each factor (i.e. a>b > c > d).