Assistant Professor Oregon State University Corvallis, Oregon, United States
Introduction:: Endometriosis affects approximately 10-15% of menstruating people and is characterized by the growth of ectopic endometrium. This can cause women to suffer from chronic, sometimes debilitating, pain, infertility, and other dysfunction of reproductive organs. While the underlying cause of endometriosis remains unknown, tissue remodeling is critical to the pathogenesis and progression of this disease. Tissue remodeling is a complex and dynamic process, involving both extracellular matrix (ECM) deposition as well as ECM degradation. While individual components of the ECM and ECM-affiliated cytokines have been investigated, the ECM and ECM affiliated proteins form a complex interconnected network of over 1,000 genes known as the matrisome. Thus, in this study we performed a holistic evaluation of the endometriosis matrisome to elucidate specific cues involved in the underlying pathogenesis and perpetuation of endometriosis.
Materials and Methods:: We unified three Gene Expression Omnibus (GEO) DNA microarray datasets containing both endometriosis and healthy samples of eutopic endometrium. We employed a variety of statistical and machine learning methods to explore dysregulation of genes in endometriosis and identify the matrisome genes, gene networks, gene ontology terms, and pathways which have significance for the onset and progression of endometriosis (Figure 1A).
Results, Conclusions, and Discussions:: Results: We found that matrisome gene expression alone correctly identified endometriosis from healthy endometrial tissue with 100% accuracy. Additionally, we found that while menstrual cycle phase accounted for over a third of the matrisome gene expression variance, when we separated the data by menstrual cycle phase before analysis, over 90% of the differentially expressed genes in the early and mid-secretory phases were subsets of the proliferative phase (Figure 1B). From these approaches, we identified 259 matrisome genes that were differentially expressed and predictive of endometriosis stage and classified them by matrisome category (Figure 1C). Among the stage significant DEMGs were 60 secreted factors including CCL3, CCL5, CCL14, CCL21, CX3CL1, CXCL14, IL13, IL15, IL17C, NGF, PDGFA, TGFB1, TNF, and VEGFB, 65 ECM regulators including ADAM metallopeptidase, matrix metallopeptidase, cathepsin, and lysyl oxidase families, 56 glycoproteins including agrin, elastin, fibrillin, laminin, and matrillin families, 41 ECM affiliated proteins including lectin and mucin families, 10 genes related to collagen including the COL4A and COL5A families, and 8 proteoglycans including decorin, podocin, and versican.
Conclusion: This work is one of the most comprehensive omics analyses of endometriosis data currently available, and the only such study which focuses on exploring matrisome dysregulation of endometriosis. Our results highlight the need to explore the matrisome of endometriosis as it relates to the onset and progression of endometriosis and identify matrisome signatures as a potential diagnostic tool.
