Proteomic analysis of endometrium from fertile and infertile patients suggests a role for apolipoprotein A-I in embryo implantation failure and endometriosis

Abstract

Pregnancy is dependent upon the endometrium acquiring a receptive phenotype that facilitates apposition, adhesion and invasion of a developmentally competent embryo. Surface-enhanced laser desorption/ionization time-of-flight mass spectrometry of mid-secretory endometrial biopsies revealed a 28 kDa protein peak that discriminated highly between samples obtained from women with recurrent implantation failure and fertile controls. Subsequent tandem mass spectroscopy unambiguously identified this peak as apolipoprotein A-I (apoA-I), a potent anti-inflammatory molecule. Total endometrial apoA-I levels were, however, comparable between the study and control group. Moreover, endometrial apoA-I mRNA expression was not cycle-dependent although there was partial loss of apoA-I immunoreactivity in luminal and glandular epithelium in mid-secretory compared with proliferative endometrial samples. Because of its putative anti-implantation properties, we examined whether endometrial apoA-I expression is regulated by embryonic signals. Human chorionic gonadotrophin (hCG) strongly inhibited apoA-I expression in differentiating explant cultures but not when established from eutopic endometrium from patients with endometriosis. Pelvic endometriosis was associated with elevated apoA-I mRNA levels, increased secretion by differentiating eutopic endometrial explant cultures and lack of hCG-dependent down-regulation. To corroborate these observations, we examined endometrial apoA-I expression and its regulation by hCG in a non-human primate model of endometriosis. As in humans, hCG strongly inhibited endometrial apoA-I mRNA expression in disease-free baboons, but this response was entirely lost upon induction of pelvic endometriosis. Together, these observations indicate that perturbations in endometrial apoA-I expression, modification or regulation by paracrine embryonic signals play a major role in implantation failure and infertility.

Figure 1

Normalized SELDI-TOF-MS of mid-secretory endometrial biopsies from fertile women (A) and RIF patients (B) reveal a discriminatory 28 063 mass-to-charge (m/z) protein peak. Sample numbers are indicated on the right. The X-axis denotes the m/z values, whereas the Y-axis is a relative intensity scale. The mass spectra were obtained at pH 5.0 on Q10 chips, which capture positively charged proteins due to its negatively charged surface.

Figure 2

Relative intensities of the four discriminatory protein peaks with m/z values of 11 987 (A), 28 063 (B), 15 867 (C) and 16 075 (D). Open circles represent the values of individual samples. Horizontal bars denote the median of normalized intensities.

Figure 3

Endometrial apoA-I expression and plasma levels in fertile and infertile patients during the mid-luteal phase of the cycle. (A) Nineteen of the 25 endometrial samples used for SELDI-TOF-MS were resolved on SDS–PAGE and immunoblotted for apoA-I. Control samples from fertile patients are indicated as C1-10, whereas biopsies from patients with RIF are denoted RIF1-9. β-Actin served as loading control. (B) Circulating apoA-I and HDL levels at the time of biopsy were determined in serum samples of fertile women (controls, n = 15) and patients with RIF (n = 10); P > 0.05.

Figure 4

apoA-I expression and tissue distribution in proliferative and secretory endometrium. (A) RTQ-PCR analysis of apoA-I mRNA expression throughout the cycle. The relative expression level of apoA-I transcripts, normalized to β-actin mRNA, was determined in 33 endometrial samples, spanning the entire menstrual cycle. The results are presented on a logarithmic scale. (B) Western blot analysis of apoA-I expression in proliferative (P1-5) and secretory (S1-6) endometrial samples. β-Actin served as loading control. (C) apoA-I immunolocalization in proliferative and secretory endometrium. apoA-I staining, performed on five proliferative endometrial samples (cycle days 8–12), was equally prominent in surface epithelium, glandular epithelium and the stromal compartment (a, b). In surface epithelium, apoA-I was expressed predominantly at the apical border of the cells (a). Immunostaining decreased markedly in the glandular but not stromal compartment in mid-secretory endometrium (cycle days 19–24; n = 6), whereas apoA-I expression in luminal epithelial cells appeared highly variable (c, d).

Figure 5

Regulation of endometrial apoA-I expression by hCG. (A) Undifferentiated human endometrial stroma cells (ESCs) and cultures decidualized with 8-Br-cAMP and MPA were co-treated with or without hCG for the indicated time-points. Total protein lysates were subjected to western blot analysis for apoA-I. The membranes were then probed for α-tubulin, a loading control. (B) Endometrial explants established from patients with and without endometriosis were cultured for 24 h in the presence of 8-Br-cAMP and MPA with or without hCG. Parallel cultures were harvested for western blot or RTQ-PCR analyses. (C) The supernatant of explant cultures established from eutopic endometrial samples from patients with and without endometriosis were analysed for apoA-I using ELISA. The levels were normalized to the weight of the explants. *P < 0.01. (D) apoA-I mRNA levels determined by RTQ-PCR in snap-frozen, luteal-phase, eutopic, endometrial samples from patients with (n = 5) and without pelvic endometriosis (n = 5); *P < 0.01.

Figure 6

Endometriosis is associated with aberrant endometrial apoA-I expression in human and baboon endometria. (A) apoA-I staining in mid-secretory endometrium (cycle days 8–12) in disease-free controls (a–c; n = 8) and patients with endometriosis (d–f; n = 8). (B) RTQ-PCR analysis of endometrial RNA obtained from control animals (n = 4) and baboons with endometriosis (n = 5) following hCG infusion. Different letters above the error bars indicate that those groups are significantly different from each other at P< 0.05. (C) Sequential serum samples were obtained from baboons (n = 4) at 3 and 12 months following the induction of endometriosis and circulating apoA-I levels compared with those from control disease-free baboons (n = 4) at the same point during the menstrual cycle, i.e. between d 9 and 11 PO. Different letters above the error bars indicate that those groups are significantly different from each other at P< 0.05.