Most cases of unexplained recurrent spontaneous abortion (URSA) occur early in pregnancy during the implantation of the human blastocyst (Wang et al., 2010). During implantation and decidualization, the embryo development is coupled with a synchronous increase in endometrial angiogenesis mediated by prostaglandins (PGs, Chakraborty et al., 1996).
The initial step that catalyzes the synthesis of prostaglandins from arachidonic acid is mediated by the enzyme cyclooxygenase (COX) (Matsuzaki et al., 2004). COX is encoded for by two separate genes, COX-1 and COX-2 (Matsuzaki et al., 2004). COX-1 is constitutively expressed in many tissues and has crucial ‘house-keeping functions’ (Wang et al., 2010). COX-2 is an inducible enzyme that catalyzes the conversion of arachidonic acid into PGH2 (Wang et al., 2010). PGE synthase converts PGH2 to PGE2, which regulates vascular physiological processes (St-Louis et al., 2010) and induces the expression of vascular endothelial growth factor (Wang et al., 2010). Furthermore, PGE2, a luteoprotective mediator, opposes the action of PGF2α to maintain the function of the corpus luteum (Ford and Christenson, 1991). According to the aforementioned findings it is speculated that the COX-2-signaling pathway is crucial for blastocyst implantation (Machado et al., 2010) and maintenance of early pregnancy (Ford and Christenson, 1991).
Several studies back up the assumption that COX-2 maneuvers the processes of ovulation, fertilization, implantation, and decidualization. Chakraborty et al. (1996) ascertain the unique expression of COX-2 genes in the peri-implantation tissue in the mouth of the uterus. Several other studies in animal models illustrated that the absence of PG synthesis is coupled with implantation and decidualization defects (Lim et al., 1997; Song et al., 2002; St-Louis et al., 2010; Numao et al., 2011). However, the impact of COX-2 on human implantation and its exact role in URSA is still not as clearly understood. To elucidate a new potential mechanism for URSA, this study investigated the pathophysiological significance of COX-2 protein expression in patients with URSA.
Material and methods
A case–control study was conducted during the period between 1 June 2009 and 30 May 2010. The patients included had a history of recurrent miscarriage (>3) at a young age, normal parent karyotypes, normal anticardiolipin and antiphospholipid antibodies, normal endocrine test results (for thyroid, prolactin, polycystic ovary syndrome, and diabetes mellitus), a normal uterine cavity (on hysterosalpingogram or hysteroscopy), and normal endometrial development (on midcycle transvaginal ultrasonography or midcycle hysteroscopy).
An office endometrial biopsy using a pipelle was performed on days 6–8 after ovulation during a spontaneous menstrual cycle, wherein the patient was protected by barrier contraception and was supported by human chorionic gonadotropin 5000 IU injections every other day throughout the luteal phase. Ovulation was detected on the basis of serial ultrasound scans for folliculometry during the late proliferative phase. All patients were instructed to avoid using medicines that may affect PG synthesis (antiprostaglandins used as analgesics or antipyretics, or PGE1 tablets used for gastritis or peptic ulceration). Patients signed consent forms explaining the nature of the tests and medicines administered and their possible risks.
Histological examination and immunohistochemical staining using commercially available COX-2 antibody were performed. The use of barrier contraception and administration of human chorionic gonadotropin were discontinued and the patients were followed up. Those who conceived during the following 6 months were included in the study.
According to the COX-2 immunohistochemical score, the patients were divided into two groups:
- Group 1: subgroup with low levels of COX-2.
- Group 2: subgroup with high levels of COX-2.
The patients were followed up during pregnancy up to 13 weeks and were all treated with 20 mg/day of clexane and 1×3/day of duphaston starting within the first week after their missed period. The pregnancy was followed up by weekly transvaginal ultrasound scans from the sixth week of pregnancy until the end of the first trimester.
Dilation and curettage biopsy samples were fixed in formalin and embedded in paraffin. Standard 4-μm-thick hematoxylin and eosin-stained microscopic sections were examined.
Immunohistochemical staining was performed using the primary monoclonal mouse antibody against the human COX-2 enzyme (clone cx229, 7 ml prediluted; Cell Marque Corporation, Rocklin, California, USA). Four-micrometer sections were cut from formalin-fixed tissue embedded in paraffin blocks and mounted onto polylysine coated slides. Sections were dewaxed and rehydrated in graded alcohol series. After washing in distilled water, the sections were pretreated with EDTA buffer (pH 9) for 12 min by a microwave antigen-retrieval procedure. The peroxidase blocking reagent was used to block endogenous peroxidase activity. The primary antibody was applied. Biotinylated anti-mouse immunoglobulin and streptavidin conjugated to horseradish peroxidase were then added. Chromogene was applied and counterstained with Mayer’s hematoxylin. Fetal membranes were included in each staining run as a positive control (Slater et al., 1999). Negative controls were done by substituting the primary antibody with saline.
Scoring and analysis of immunostaining
Immunostaining results were analyzed without previous knowledge of the clinicopathological data of the patients. To increase the objectivity of the analysis, the immunohistochemical scoring system developed by Soslow et al. (2000) based on the German ImmunoReactive Score and modified by Wasilewicz et al. (2010) was used. Sections were divided into four nonoverlapping fields, analyzed separately, added, and averaged.
The analysis included the percentage and intensity of COX-2-positive cells.
The intensity of staining was assessed on a 0–3-point scale (‘intensity score’: 0=no staining, 1=weak, 2=moderate, 3=strong). The percentage of positive cells was assessed on 0–4-point scale (‘quantity score’: 0=≤1%, 1=1–25%, 2=26–50%, 3=51–75%, 4=76–100%). The average intensity score and the average quantity score were multiplied by each other to obtain the ‘immunohistochemical score system’ (IHS). Patients were divided into two groups: group 1 with low levels of COX-2 expression (IHS=0–4) and group 2 with high levels of COX-2 expression (IHS=5–12).
Evaluation of the topographic distribution of cyclooxygenase-2
The site of expression of COX-2 was evaluated separately in the surface epithelium, the glandular epithelium, and the stroma.
Statistical analysis was carried out using Microsoft Office Excel (USA), 2003. The results were expressed as mean±SD. Pearson’s χ2-test and Fisher’s exact test were used to analyze the association between COX-2 levels and selected clinicopathological parameters. For a group comparison of quantitative data, a preliminary test for the equality of variances indicated that the variances of the two groups were significantly different. Therefore, a two-sample t-test was performed that does not assume equal variances. P-value less than 0.05 was considered significant and that less than 0.001 was considered highly significant. Pearson’s correlation coefficients were used to establish the strength of the relationship between COX-2 score and clinical outcome.
Twenty-six patients were enrolled in this study. The mean age was 28 (±2.5) years. The average number of previous miscarriages was four. The endometrial biopsy of all patients revealed a pattern characterized by the presence of either late secretory or hypersecretory endometrial glands. The stroma showed a predecidual reaction and was infiltrated by lymphocytes.
Cytoplasmic immunoreactivity for COX-2 was detected in 84.6% (22/26) of the cases (Figs 1 and 2). The mean IHS was 4.47±4.02. Group 1 included 12 cases with a mean IHS of 0.65±0.69 and group 2 included 14 cases with a mean IHS of 7.7±2.6.
The topographic distribution of COX-2-positive cells was compared in the two groups. In group 1, immunostaining for COX-2 was mainly localized to the surface epithelium. The glandular epithelium was almost immunonegative for COX-2 (Fig. 1). In group 2, immunostaining for COX-2 was seen in both surface and glandular epithelia (Fig. 2).
Relationship between immunohistochemical score for cyclooxygenase-2 expression and clinical outcome
As shown in Table 1, the mean age in group 1 (IHS≤4) was 28±2.5 years, whereas that in group 2 (IHS≥5) was 28±2.7 years (P>0.05; nonsignificant). No correlation was detected between the patient’s age and COX-2 score (r<−0.3). The average number of previous miscarriages in group 1 was five, whereas that in group 2 was three. The total number of previous miscarriages was highly significantly greater in group 1 compared with group 2 (Student’s t-test, P=0.0002). A strong inverse correlation was found between the COX score and the number of previous miscarriages (r=−0.66).
In the subsequent pregnancy, group 1 had a first trimester miscarriage rate (66.6%) that was statistically significantly higher (P=0.04) than that in group 2 (28.5%). There was no statistically significant difference in the mean gestational age at miscarriage between groups 1 and 2 (7.4±1.2 vs. 9.6±2 weeks, P=0.06). However, the rate of pregnancies reaching the stage of showing a positive fetal pulsation on ultrasound scans was statistically significantly higher in group 2 than in group 1 (100 vs. 66.6%, P=0.01).
Uterine contractions were a common complaint in 58% of the patients in group 1 vs. 14% of the patients in group 2 (P=0.02). The correlation between COX-2 and uterine contractions was r equal to −0.4; that is, low levels of COX-2 protein expression were associated with a higher incidence of uterine contractions. Patients in group 2 were more likely to miscarry by a silent/missed miscarriage rather than by bleeding as in group 1. However, the difference did not reach statistical significance (silent miscarriage rate in group 1 was 8.3 vs. 28.5% in group 2, P=0.1). Patients in group 1 were more likely to develop active bleeding and subchorionic hematoma compared with those in group 2 (58.3 and 25% in group 1 vs. 14.2 and 7.1% in group 2). The correlation was highly significant only in the active bleeding group (P=0.0007). The diagnosis of a blighted ovum tended to be made more frequently in group 1 than in group 2 (25% in group 1 vs. 7.1% in group 2, P=0.1). The rate of developing bleeding that was severe enough to warrant stoppage of clexane was statistically significantly higher in group 1 than in group 2 (25% in group 1 vs. 0% in group 2, P=0.04).
During pregnancy, hormones and other biologically active mediators induce morphological changes in the endometrium (Blitek et al., 2006). These changes are coupled with variations in the local synthesis of PGE2 (Ford and Christenson, 1991). The increase in PGE2 is modulated by the expression of COX-2. Several studies published so far have provided ample evidence for the role of COX-2 in ovulation, implantation, and decidualization in animal models. However, there is a paucity of data on its role in humans. This study has, for the first time, investigated COX-2 expression levels in a human situation designed to mimic the period of early pregnancy, aiming to elucidate the relationship between the expression of this enzyme and URSM.
COX-2 is expressed in the uterus during the peri-implantation phase. This is confirmed in the present study. Expression of COX-2 was detected in 85% of the endometrial biopsies in the pseudopregnant state. This is in agreement with the reports in the literature. The presence of COX-2 has been demonstrated in vitro in the endometrium and choriodecidual tissue during the peri-implantation period (Charpigny et al., 1997; Marions and Gemzell Danielsson, 1999).
The localization of COX-2 primarily to the glandular epithelium supports the in-vitro evidence that glands are the major site of PG synthesis in the endometrium (Lumsden et al., 1984; Smith and Kelly, 1988). The different COX-2 expression levels in luminal and glandular epithelia can possibly be explained by different local regulation of the enzyme (Marions and Gemzell Danielsson, 1999). The presence of neutrophilic infiltrate in biopsies having positive COX-2 immunoreactivity highlighted the previous finding of Colditz (1990). The localization of COX-2 in the endometrium synergizes the action of PGE on interleukin-8, reinforcing the recruitment of neutrophils.
The two groups studied were more or less identical, with no statistically significant differences in the age (mean age, 28±2.5 years). On average, the women in this study reported four previous pregnancy losses. The results constituted evidence that the history of spontaneous miscarriage was strongly inversely related to the endometrial level of COX-2 (r=−0.66). The number of previous miscarriages was significantly higher in group 1 than in group 2 (P=0.0002). In search of an explanation to this finding, Wang et al. (2004) reported that the reproductive abnormalities in COX-2-deficient mice depended upon the genetic background. The absence of a correlation between the age of the study group and the COX-2 score strongly supports an underlying genetic mechanism (r<−0.3).
The present study clearly supports a role for COX-2 in recurrent miscarriage. Group 1 (COX-2≤4) had a first trimester miscarriage rate that was statistically significantly higher than that in group 2 (COX-2≥5; P=0.04). Moreover, the rate of pregnancies reaching the stage of showing a positive fetal pulsation on ultrasound scans was statistically significantly higher in group 2 than in group 1 (P=0.01). This is confirmatory evidence reinforcing previously demonstrated gene-knockout studies; COX-2-deficient female mice have abnormalities in implantation (Dinchuk et al., 1995; Lim et al., 1997).
COX-2 encodes for PGE2 synthesis, which opposes the function of PGF2α during early pregnancy. PGF2α plays a crucial role in the myometrium during parturition by increasing oxytocin-induced contractions (Fuchs, 1987). In the present study, uterine contractions were a common complaint in patients with low expression of COX-2 protein in endometrial biopsy (P=0.02). Moreover, an inverse correlation between COX-2 levels and uterine contractions was found (r=−0.4). This finding is backed up by the results of St-Louis et al., 2010). They suggest that the augmentation of COX-2 is mainly responsible for the augmentation of PGE2 and downregulation of PGF2α during this period. This mechanism may help prevent early pregnancy loss till implantation is fully established.
Patients with low COX-2 levels (IHS 0–4) were more likely to miscarry and more likely to present at miscarriage with symptoms suggestive of poor implantation – that is bleeding and/or uterine contractions (Table 1). However, those with a high level of COX-2 who miscarried were more likely to have silent miscarriages. The same was also suggested by the tendency toward earlier gestational age at miscarriage (7.4±1.2 weeks in group 1 vs. 9.6±2 weeks in group 2, P=0.06). Yet, this was significantly demonstrated in previous animal models. It was previously shown that PG synthesis inhibitors only affect the production of uterine-derived PGs. They do not influence ovarian or embryonic PGs. This prevents or delays implantation (Smith, 1991; Weitlauf, 1994).
The current study characterized the expression of COX-2 in a human situation designed to mimic the period of early pregnancy. The results provided ample evidence to support the role of COX-2 in the regulation of implantation and the maintenance of early pregnancy. Because this is the first time such a correlation has been investigated, this study can be considered as a pilot study warranting more studies on a larger scale.
Conflicts of interest
There are no conflicts of interest.
Blitek A, Waclawik A, Kaczmarek MM, Stadejek T, Pejsak Z, Ziecik AJ. Expression of cyclooxygenase-1 and -2 in the porcine endometrium during the oestrous cycle and early pregnancy. Reprod Domest Anim. 2006;41:251–257
Chakraborty I, Das SK, Wang J, Dey SK. Developmental expression of the cyclo-oxygenase-1 and cyclo-oxygenase-2 genes in the peri-implantation mouse uterus and their differential regulation by the blastocyst and ovarian steroids. J Mol Endocrinol. 1996;16:107–122
Charpigny G, Reinaud P, Tamby JP, Créminon C, Guillomot M. Cyclooxygenase-2 unlike cyclooxygenase-1 is highly expressed in ovine embryos during the implantation period. Biol Reprod. 1997;57:1032–1040
Colditz IG. Effect of exogenous prostaglandin E2 and actinomycin D on plasma leakage induced by neutrophil-activating peptide-1/interleukin-8. Immunol Cell Biol. 1990;68:397–403
Dinchuk JE, Car BD, Focht RJ, Johnston JJ, Jaffee BD, Covington MB, et al. Renal abnormalities and an altered inflammatory response in mice lacking cyclooxygenase II. Nature. 1995;378:406–409
Ford SP, Christenson LK. Direct effects of oestradiol-17β and prostaglandin E-2 in protecting pig corpora lutea from a luteolytic dose of prostaglandin F-2α. J Reprod Fertil. 1991;93:203–209
Fuchs AR. Prostaglandin F2alpha and oxytocin interactions in ovarian and uterine function. J Steroid Biochem. 1987;27:1073–1080
Lim H, Paria BC, Das SK, Dinchuk JE, Langenbach R, Trzaskos JM, et al. Multiple female reproductive failures in cyclooxygenase 2-deficient mice. Cell. 1997;91:197–208
Lumsden MA, Brown A, Baird DT. Prostaglandin production from homogenates of separated glandular epithelium and stroma from human endometrium. Prostaglandins. 1984;28:485–496
Machado DE, Berardo PT, Landgraf RG, Fernandes PD, Palmero C, Alves LM, et al. A selective cyclooxygenase-2 inhibitor suppresses the growth of endometriosis with an antiangiogenic effect in a rat model. Fertil Steril. 2010;93:2674–2679
Marions L, Gemzell Danielsson K. Expression of cyclo-oxygenase in human endometrium during the implantation period. Mol Hum Reprod. 1999;5:961–965
Matsuzaki S, Canis M, Pouly JL, Wattiez A, Okamura K, Mage G. Cyclooxygenase-2 expression in deep endometriosis and matched eutopic endometrium. Fertil Steril. 2004;82:1309–1315
Numao A, Hosono K, Suzuki T, Hayashi I, Uematsu S, Akira S, et al. The inducible prostaglandin E synthase mPGES-1 regulates growth of endometrial tissues and angiogenesis in a mouse implantation model. Biomed Pharmacother. 2011;65:77–84
Slater D, Dennes W, Sawdy R, Allport V, Bennett P. Expression of cyclo-oxygenase types-1 and -2 in human fetal membranes throughout pregnancy. J Mol Endocrinol. 1999;22:125–130
Smith SK. The role of prostaglandins in implantation. In: Seppälä M, editor. Baillieres Clin Obstet Gynaecol. 1991;5:73–79
Smith SK, Kelly RW. The release of PGF2a and PGE2 from fatty acids separated cells of human endometrium and decidua. Essent Fatty Acids. 1988;33:91–96
Song H, Lim H, Paria BC, Matsumoto H, Swift LL, Morrow J, et al. Cytosolic phospholipase A2α deficiency is crucial for ‘on-time’ embryo implantation that directs subsequent development. Development. 2002;129:2879–2889
Soslow RA, Dannenberg AJ, Rush D, Woerner BM, Nasir Khan K, Masferrer J, et al. COX-2 is expressed in human pulmonary, colonic and mammary tumors. Cancer. 2000;89:2637–2645
St-Louis I, Singh M, Brasseur K, Leblanc V, Parent S, Asselin E. Expression of COX-1 and COX-2 in the endometrium of cyclic, pregnant and in a model of pseudopregnant rats and their regulation by sex steroids. Repr Biol Endocrinol. 2010;8:103
Wang H, Ma WG, Tejada L, Zhang H, Morrow JD, Das SK, et al. Rescue of female infertility from the loss of cyclooxygenase-2 by compensatory up-regulation of cyclooxygenase-1 is a function of genetic makeup. J Biol Chem. 2004;279:10649–10658
Wang Y, Zhao AM, Lin QD. Role of cyclooxygenase-2 signaling pathway dysfunction in unexplained recurrent spontaneous abortion. Chin Med J. 2010;123:1543–1547
Wasilewicz MP, Kołodziej B, Bojułko T, Kaczmarczyk M, Sulzyc-Bielicka V, Bielicki D. Expression of cyclooxygenase-2 in colonic polyps. Pol Arch Med Wewn. 2010;120:313–320
©2012Egyptian Journal of Pathology
Weitlauf HMKnobil E, Neill JD. Biology of implantation. The physiology of reproduction. 19942nd ed. New York Raven Press:391–440