|Home | About | Journals | Submit | Contact Us | Français|
Endometriosis affects >10% of women during their reproductive years, many of whom report high rates of spontaneous pregnancy loss (SPL). We examined whether gene polymorphisms in apolipoprotein E (APOE), which is involved in lipoprotein metabolism, are associated with endometriosis and/or endometriosis-associated infertility. We conducted a cross-sectional genetic association study of women surgically confirmed to have endometriosis (n = 345) and no surgical evidence of the disease (n = 266). Genotyping of APOE polymorphism (2, 3, 4) was conducted by polymerase chain reaction–restriction fragment length polymorphism followed by visualization of specific patterns by gel electrophoresis. Statistical significance of differences in genotype and allelic frequencies was assessed using Pearson's χ2 test and Risk analysis. Overall, we found no association between APOE genotype and diagnosis of endometriosis. However, patients with endometriosis who reported at least one SPL were three times more likely to be 2 carriers and 2-fold less likely to be 4 carriers. Compared with 3 carriers, patients with endometriosis who were 2 carriers and had at least one live birth reported four times the rate of SPL, while 4 carriers were <0.4-fold less likely to report an SPL. Our data suggest that there may be an association between APOE allelic frequency and SPL in patients with endometriosis, which appears to be independent of mechanisms associated with infertility, an intriguing observation that deserves further investigation.
Endometriosis is a chronic gynecological condition caused by the growth of endometrial glands and stroma at sites outside the endometrial cavity (Sampson, 1927). This condition is thought to result from the retrograde menstrual flow of endometrial cells through the Fallopian tubes into the peritoneum, causing inflammation, pain during menstruation (dysmenorrhea) and chronic pelvic pain (Bulun, 2009). Evidence indicates that endometriosis affects 10% of premenopausal women, with an estimated 7 million women (in the US alone) currently suffering from the condition (Signorello et al., 1997). In Puerto Rico, the estimated prevalence is 2775 per 100 000 women (Flores et al., 2008). Because of the lack of non-invasive specific markers and of knowledge regarding the pathophysiology of this debilitating condition, endometriosis can be diagnosed only through a laparoscopy and still remains an incurable disease (Jacobson et al., 2002). One of the major consequences of endometriosis is infertility, which affects 4 out of 10 patients with endometriosis and is associated with a substantial reduction in quality of life (Abbey et al., 1992; Ballweg, 2004; Lima et al., 2006; Petrelluzzi et al., 2008). Spontaneous pregnancy loss (SPL), also known as a ‘miscarriage’ or a ‘spontaneous abortion’ (loss of pregnancy in first 20 weeks), has been associated with endometriosis; however, these results were based on uncontrolled and retrospective studies (Vercammen and D'Hooghe, 2000; Yenicesu et al., 2010; Zhang and Wu, 2010a). Therefore, further studies are needed to conclusively determine whether there is an association between pregnancy loss and endometriosis as well as to determine what the potential mechanism(s) underlying of such an association may be.
Immunological, environmental and genetic factors have been implicated in the development of endometriosis. The genetic theory for susceptibility to endometriosis has recently been supported by genome wide association studies (Uno et al., 2010; Painter et al., 2011a, b). However, it is still necessary to examine the role of new potential candidate genes that would explain the etiology of infertility associated with endometriosis and would also offer novel therapeutic targets to ameliorate its symptoms and signs. One potential candidate gene is apolipoprotein E (APOE), which plays an important role in the metabolism and redistribution of cholesterol and lipoproteins (Mahley, 1988). Of the three distinct alleles (2, 3 and 4) found in humans, 4 is associated with an increased risk for Alzheimer's Disease (AD), ischemic heart disease and diabetes (Haan et al., 2003; Messier, 2003; Porrata-Doria et al., 2010), while the 2 allele confers a higher risk for cardiovascular disease and hyperlipoproteinemia (de Villiers et al., 1997; Eto et al., 2002).
A relationship between APOE genotype and fertility has been reported in healthy women from two Ecuadorian populations (Corbo et al., 2004a, b, 2007, 2008). The level of fertility was highest in 4 carriers, average in 3 homozygotes and lowest in 2 carriers. In order to understand why specific genotypes may result in a subfertile phenotype, studies have looked at APOE allele frequency in women with recurrent pregnancy loss (RPL) (defined as two or more SPLs), but these studies have yielded conflicting results. However, all of these studies had a small sample that did not allow ample representation of 2 carriers, and APOE allele frequencies for controls were not within the normal expected range, making it impossible to fully interpret results (Bianca et al., 2010; Yenicesu et al., 2010; Zhang and Wu, 2010a, b).
Previous studies show that endometriosis patients have elevated serum lipoprotein levels normally associated with an increased risk for cardiovascular disease (Crook et al., 1997). In animal studies, apoE protein levels are elevated in endometriotic lesions (Pelch et al., 2010) and its gene is highly expressed in normal human endometrium during the window of implantation (Kao et al., 2002). Since there is an association between APOE gene alleles (2 and 4) and altered lipid profiles (cholesterol, triglyceride, low-density lipoprotein) (Descamps et al., 2005; Hieronimus et al., 2005; Chuang et al., 2009), and these alterations can be detrimental to the mother and fetus (McGladdery and Frohlich, 2001), it has been suggested that the APOE gene may be a potential factor involved in a woman's fertility potential and/or her ability to have a healthy pregnancy. Moreover, apoE plays a role in the regulation of steroid hormone functions; therefore, it could influence the reproduction potential in humans. Thus, we became interested in determining whether there is a relationship between endometriosis, SPL and APOE polymorphisms, which is the main aim of the present studies.
The samples used in this study were obtained in collaboration with the Endometriosis Research Program (ERP) at PMSHS in Ponce, PR. The ERP patient registry consists of 1060 entries from surgically confirmed patients and controls. DNA was available for 616 subjects (345 endometriosis patients and 266 controls/non-endometriosis). Subjects in this registry are aged 16–50 and come from all regions of the island (31% from the north, 37% the south, 12% the east and 13% the west). The study subjects (patients and controls) were recruited by direct referrals from collaborating obstetricians and gynecologists practicing throughout Puerto Rico. The patients were premenopausal women who had been diagnosed with endometriosis via surgery. Controls were women who underwent laparoscopies or laparotomies for benign unrelated gynecologic conditions (e.g. uterine fibroids, dysfunctional uterine bleeding) and who did not have endometriosis, as confirmed by surgery. A Surgery Report Form was completed to allow for both the standardized collection of data and the documentation of disease severity, according to the revised AFS criteria. All study participants filled out a questionnaire that was administered by a research nurse in order to collect demographic information, the gynecologic/obstetric history and the clinical history of each patient, including the number of pregnancies, live births and miscarriages. For patients who reported having had difficulty achieving pregnancy, an additional questionnaire was applied to collect more information on their fertility status. Additional data collected included gestation and parity, time trying to achieve pregnancy, infertility treatments used and whether a reproductive endocrinologist had been consulted. The cohort of subjects analyzed in this study was not recruited based on any obstetrical parameter, including SPL, but rather on the diagnosis of endometriosis (or exclusion of) by laparoscopy. The exclusion criteria for the subjects in our registry are age (16–50, i.e. premenopausal), and presenting with symptoms, but without surgical diagnosis. For previous molecular analysis we have conducted with these subjects we also excluded subjects based on irregular menstrual cycle and current use of hormonal drugs; however, for the present study, we did not exclude subjects based on these factors since they will not change the genotype. A sample of healthy women aged 18–50 (n = 265) representing the general population from Puerto Rico was available from Dr Summer Acevedo's laboratory and was used for calculations of Hardy–Weinberg equilibrium as described below. The study's protocol, informed consent and questionnaires were approved by the PSMHS IRB committee.
Blood samples were collected by venipuncture by a research nurse using standard aseptic procedures. Lymphocytes were isolated from whole blood by centrifugation at 600 g equivalent for 10 min in Histopaque (Sigma, St. Louis, MO, USA). Total genomic DNA was isolated from lymphocytes using the QIAamp DNA Blood Maxi Kit and following the manufacturer's recommendations (QIAGEN, Valencia, CA, USA).
DNA was amplified using polymerase chain reaction (PCR), which was conducted using PCR Master Mix (Promega), Taq 50 U/ml, 400 µM dNTPs and 3 mM MgCl2. The primers used were (F4) 5′-ACAGAATTCGCCCCGGCCTGGTACAC-3′ (1 mM) and (F6) 5′-TAAGCTTGGCACGGCTGTCCAAGG-3′ (1 mM) (Hixson and Vernier, 1990). The PCR cycle consisted of three steps: denature at 95°C for 5 min, followed by 35 cycles of primer annealing at 60°C for 1 min, extension at 70°C for 2 min and denaturation at 95°C for 1 min with a final extension of 10 min at 70°C. After 3 h of digestion at 37°C with CfoI/HhaI (Promega) enzymes, products were run on 4% agarose gel 140 V for 90 min. Figure 1 indicates the bands used to identify APOE genotypes by size according to specific restriction fragment length polymorphism patterns (Wenham et al., 1991).
To examine group (non-endometriosis versus endometriosis) differences in demographic information and prevalence of symptoms, we used Student's t-test. Hardy–Weinberg analysis was conducted using Chi χ2 P-values report to determine variation of allelic frequency from the general population. An analysis of variance or with a covariate (ANCOVA) with age as a covariate was conducted for all the symptoms and demographic information to compare genotype differences. The effect of APOE allelic frequency was determined using cross tabulation with Pearson's χ2 test and Risk analysis. Power calculations showed that a sample size of 240 subjects could detect statistically significant differences in allelic frequencies within groups with an alpha of 0.05 and an effect size of 0.2 (http://www.stat.uiowa.edu/~rlenth/Power/), detecting (Cohen, 1988). Analyses were conducted using SPSS statistical software version 17.0; P< 0.05 was considered statistically significant for all analyses.
All women in this cross-sectional study displayed at least one symptom related to a menstrual, uterine or fertility disorder requiring laparoscopic surgery. The women with surgical confirmation of endometriosis were significantly younger (33.0 ± 0.04) than were controls (37.0 ± 0.05; P < 0.001). After adjusting for age, we observed higher rates of dysmenorrhea, incapacitating pain, dyspareunia, perceived infertility, ovarian cysts and family history of endometriosis in cases versus controls (Table I). In the endometriosis group, there were fewer women who had been pregnant and fewer pregnancies per patient (Table I). One of the most common reasons that women seek medical attention and are examined for endometriosis is SPL. Therefore, it is not surprising that 25–27% of the women in both groups reported at least one pregnancy loss, and the number of RPLs was similar in both groups (Table I). There were no differences among groups for age at menarche, prevalence of uterine fibroids or civil status (Table I). There was no difference in genotype or allelic frequency between the endometriosis and non-endometriosis subjects.
We observed no differences in APOE genotype or allelic frequency associated with any of the symptoms or clinical characteristics in patients with endometriosis versus controls (Table II). Hardy–Weinberg χ2 analysis indicated that both the endometriosis patients and controls were within expected values when compared with a general population of healthy women from Puerto Rico (Table II).
However, when we considered only women with at least one pregnancy, 2 allelic frequency was higher in those reporting SPLs—15% versus 5.9% (F = 5.58, P< 0.02)—and the number of RPLs for 2 carriers was also higher (F = 7.12, P< 0.008) than 3 homozygotes. Similarly, in women with at least one live birth, 2 allelic frequency was increased in women who reported RPLs (F = 7.61, P< 0.006) and the number of RPLs was higher (F = 4.90, P< 0.03) compared with 3 homozygotes. Results also showed that women who had been pregnant at least once and who had suffered an SPL were less likely to be 4 carriers (F = 3.43, P= 0.06) than 3 homozygotes. Among women with at least one live birth, 4 allelic frequency was 13% for those reporting at least one SPL, and 20.5% for those who had not reported an SPL, which is similar to the expected population frequency.
When we compared patients with endometriosis with controls, we observed that women in the endometriosis group who had experienced at least one SPL were 3-fold more likely to have an 2 allele (P< 0.02) compared with 3 homozygotes (Table III). The risk of SPL increased to more than 4-fold in 2 carriers (P< 0.004) who had had one successful pregnancy. Women in this group were 2-fold less likely to be 4 carriers than 3 homozygotes (Table III). These data suggest that there is an association between APOE allelic frequency and SPL, particularly in women with endometriosis, via mechanisms that are probably unrelated to the infertility defects seen in this patient population.
In this study, we assessed the possible association of endometriosis and its symptoms or outcomes, including SPL, with APOE genotype. We observed no association between APOE genotype and having a diagnosis of endometriosis. Frequencies of APOE genotype in endometriosis patients were within the reported average in the general population (i.e. 8–10% for 2 and 15–20% for 4; Corbo and Scacchi, 1999). We did observe that women with endometriosis reported higher rates of perceived infertility and a lower number of pregnancies than did women without endometriosis. However, there was no difference in the incidence (~20%) of self-reported SPL among groups. This study shows that, as previously published, SPLs are not necessarily associated with having endometriosis (Vercammen and D'Hooghe, 2000). However, we observed significant differences between women with and without endometriosis and self-reported SPL in regards to APOE genotype. Specifically, the odds of having an SPL in 4 carriers were half the expected number in endometriosis patients, and those with an 2 allele were 3-fold higher than were 3/3 genotype. Moreover, among patients with endometriosis who had at least one live birth, those who were 2 carriers had four times higher rates of SPL, while 4 carriers were <0.4-fold less likely to report an SPL compared with 3 carriers. This suggests that there may be a link between apoE protein levels function and SPL in patients with endometriosis via mechanisms unrelated to the infertility defects seen in this patient population, and that the 2 genotype could be associated with a subfertile phenotype.
Both infertility and SPL are possibly related to the presence of altered cholesterol and lipoprotein levels during pregnancy which result in endothelial and possibly endocrine dysfunction (Crook et al., 1997; Descamps et al., 2005). During pregnancy, the changes in the levels of apoE, which plays a role in lipoprotein and cholesterol metabolism and transport, are consistent with metabolic changes that take place during gestation (Chikosi et al., 2000a; McGladdery and Frohlich, 2001; Descamps et al., 2005). However, distinct defects in lipid profiles have been associated with particular genotypes: 4 is associated with higher cholesterol, higher triglyceride and higher low-density lipoprotein (LDL) cholesterol levels than those associated with 3 homozygotes. In addition, 2 carriers have higher triglyceride and lower LDL cholesterol levels during pregnancy than do non-2 carriers (Chikosi et al., 2000a). Patients with 2 and mutations in a lipoprotein lipase gene are frequently reported to suffer from hypertriglyceridemia, hyperlipoproteinemia or dyslipidemia during pregnancy, any of which can be detrimental to the mother and fetus if not detected (Hieronimus et al., 2005; Chuang et al., 2009). In addition, there is a 100-fold increase in APOE gene expression in the normal human endometrium during the window of implantation (Kao et al., 2002). These results suggest that APOE may be a factor involved in a woman's ability to have a healthy pregnancy. Considering that apoE protein levels are associated with other disorders that can occur during pregnancy and that can lead to fetal mortality (de Villiers et al., 1997; Nagy et al., 1998; Chikosi et al., 2000a, b; Eto et al., 2002; Zetterberg et al., 2002), it is important to examine the potential effects of these levels on SPL. In addition, our data suggest that hyperlipoproteineima could be a potential confounder in studies addressing subfertility in the context of the endometriosis patient, one that should be accounted for in future studies.
The findings of this study need to be validated in other racial/ethnic populations and with women who suffer other types of menstrual disorders in order to more clearly understand the association between APOE polymorphisms and SPL. Additional studies using animal models could also help advance this field. Lower rates of fertility have been observed in ApoE knock-out mice (unpublished, author observation), and endometriosis can be surgically induced in this model. Such studies will provide a great tool for studying the interaction of the mechanism(s) underlying the observed association. In sum, there appears to be a relationship between APOE and SPL that should be investigated further. Knowledge of an individual's APOE genotype and identification of all of the other possible factors involved will aid in the development of successful interventions to enhance fertility potential in patient populations at higher risk for pregnancy loss.
M.S.C. conducted genotyping worked on writing and revising the paper and T.P.-D. conducted genotyping and revised the paper. I.F. worked on study design and data analysis, was the principal investigator for the ERP cohort and revised the manuscript. S.F.A. was the principal investigator for this project, designed the study, conducted and supervised genotyping, conducted data analysis and manuscript preparation and revisions.
We acknowledge the support of the RCMI Molecular Biology Core Lab (G12 RR003050), and the Endometriosis Research Program (NIH-NICHD R01 HD050559; NIH-MBRS S06-GM 08239). Finally, thanks go to Bob Ritchie of the RCMI Publications Office (G12 RR003050) for his editing of the manuscript.
Authors report no conflict of interest.
We would like to thank Sonia Abac, RN, for extracting blood samples; Abigail Ruiz, BS, for processing genetic samples; and Diana M. Morales for technical assistance. In addition, we are grateful to Jessica Fourquet, MPH, who enters and maintains all of the information in the Patient Registry of the ERP.