Changes in eutopic endometrial gene expression during the progression of experimental endometriosis in the baboon, Papio anubis

Abstract

Endometriosis is associated with aberrant gene expression in the eutopic endometrium of women with disease. To determine if the development of endometriotic lesions directly impacts eutopic endometrial gene expression, we sequentially analyzed the eutopic endometrium across the time course of disease progression in a baboon model of induced disease. Endometriosis was induced in baboons (n = 4) by intraperitoneal inoculation of autologous menstrual endometrium. Eutopic endometria were collected during the midsecretory phase (Days 9-11 postovulation) at 1, 3, 6-7, 10-12, and 15-16 mo after disease induction and compared with tissue from disease-free baboons. RNA was hybridized to Human Genome U133 Plus 2.0 Arrays, and data were extracted using Gene-Chip Operating Software. Subsequently, both Gene Set Enrichment Analysis and Ingenuity Pathways Analysis were used to find biological states that have a statistically significant enrichment concomitant with pairwise comparison of human endometriosis arrays. Within 1 mo of induction of the disease, 4331 genes were differentially expressed (P < 0.05). Hierarchical clustering revealed self-segregation into two groups-a) 1, 3, and 10-12 mo and b) 6-7 and 15-16 mo-together with controls. Clustering analysis at each stage of disease validated dysregulation of several signaling pathways, including Nodal-like receptor, EGF, ERK/MAPK, and PI3/AKT. Sequential analysis of the same animals during disease progression demonstrated an early disease insult and a transitory dominance of an estrogenic phenotype; however, as the disease progressed, a progesterone-resistant phenotype became evident. Furthermore, we demonstrate a 38.6% differential gene expression overlap with endometrial samples in the midsecretory phase from women with endometriosis, concomitant with similar dysregulation in human disease candidate genes Fos, Nodal, Suclg2, and Kras, among others. Molecular changes in the eutopic endometrium, associated with endometriosis, are directly impacted by endometriotic lesions, providing strong evidence that it is the disease rather than inherent defective endometrium that results in aberrant gene expression in the eutopic endometrium. Furthermore, this baboon model provides a powerful means whereby the early events associated with the pathology of disease and the resulting infertility may be elucidated.

FIG. 1

Endometriosis was experimentally induced in female Papio anubis (baboons) with documented regular menstrual cycles by intraperitoneal inoculation with menstrual endometrium. Endometrium was harvested on Days 1–2 of menses immediately prior to laparoscopy. At the subsequent mense, the animals underwent a second laparoscopy and endometrial reseeding at the same ectopic site, the uterine fundus, the cul de sac, the ovaries, and the pouch of Douglas. Following each laparoscopy, a laparotomy was performed, and eutopic endometrial tissue was harvested. The progression of disease was monitored in each animal by consecutive laparoscopies at 1 (n = 2), 3 (n = 4), 6 (n = 4), 9–12 (n = 4), and 15–16 (n = 4) mo after inoculation, during the window of uterine receptivity (Days 9–11 postovulation [PO] in the baboon). At 15–16 mo following the second inoculation, the animals were euthanized, and a necropsy was carried out to obtain all of the associated reproductive tissues within the peritoneal cavity.

FIG. 2

A) PCA of the average expression levels at each time point. The first two principal components, the ones that account for the largest variance between average samples, were used to visualize the data. Samples that have similar gene expression profiles cluster close together by disease length. Using all genes on the Human Genome Affymetrix array, samples self-segregated into two major clusters. The y-axis is the second PCA component, wherein the DF controls and the early and transitional stages are separated from the M10–12 and M15–16 of disease. B) Unfiltered hierarchical clustering and heat mapping demonstrated similarities in eutopic endometrial transcriptome between animals at 1, 3, and 10–12 mo of disease. Pairwise comparisons between dysregulated transcripts at 1, 3, and 10–12 mo of disease revealed striking similarities. The second cluster was comprised of samples from animals 6–7 and 15–16 mo following induction of disease, together with those from DF control animals. C) Hierarchical clustering of each animal (sample), including baboons with spontaneous disease.

FIG. 3

The Venn diagram depicts differentially expressed genes through each transitional time point of disease and overlap between these transitions. Genes are differentially expressed with an adjusted P value <0.05. Note that between the month (M)1–(M)3 disease points, no genes were significantly differentially expressed (Table 1).

FIG. 4

After laparoscopic entry, a complete systemic survey of the abdomen and pelvic cavity was performed, and the number of each visible lesion (A) and lesion location (B) was documented. At 15–16 mo, a hysterectomy and necropsy were performed, and all lesions were documented. The total number of lesions at each time point was not significantly different until the 15–16-mo time point (**P < 0.05), which is the time point at which all lesions are visualized and documented in the laboratory following complete removal of all tissues.