Distribution of cyclooxygenase-2 in eutopic and ectopic endometrium in endometriosis and adenomyosis

Transcription

1 Human Reproduction Vol.16, No.3 pp , 2001 Distribution of cyclooxygenase-2 in eutopic and ectopic endometrium in endometriosis and adenomyosis Hirotaka Ota 1,3, Shinichi Igarashi 2, Masato Sasaki 1 and Toshinobu Tanaka 1 1 Department of Obstetrics and Gynecology, Akita University School of Medicine, Akita-city, Akita-ken , and 2 Koto General Hospital, Hachirogata-town, Akita-ken , Japan 3 To whom correspondence should be addressed. otah@obgyn.med.akita-u.ac.jp The objective of this study was to determine the distribution of cyclooxygenase-2 (COX-2) in eutopic and ectopic endometria in endometriosis and adenomyosis. The subjects were 35 patients with endometriosis diagnosed by laparoscopy, 33 patients with histologically confirmed adenomyosis and 50 female controls with normal fecundity. Expression of COX-2 was immunohistochemically investigated in tissues from eutopic endometrium and myometrium and ectopic endometrium of the wall of ovarian chocolate cysts using polyclonal antibody. Surface epithelial cells, endometrial glandular epithelial cells or stromal cells were assessed. Cells were semi-quantitatively assessed on a scale of 1 to 5 using a nomogram created from positive cell count and the degree of staining. COX-2 expression in surface and glandular epithelia of the control group varied markedly during the menstrual cycle. It was lowest in the early proliferative phase and gradually increased thereafter. It remained high throughout the secretory phase. However, in patients with endometriosis, expression of COX-2 in glandular epithelium was higher than that in the control group, though it varied throughout the menstrual cycle. On the other hand, there was no variation in expression of COX-2 in the adenomyosis group during the menstrual cycle, and it was lower than that in the endometriosis group in all phases. Pronounced COX-2 expression was observed in glandular cells from ectopic endometrial tissue of ovarian chocolate cyst walls in all cases regardless of the menstrual phase. In summary, increased COX-2 expression in eutopic and ectopic endometria was believed to be strongly correlated with pathological abnormalities in these disorders. Key words: adenomyosis/cyclooxygenase/endometriosis/endometrium/prostaglandin Introduction intimately involved in reproductive functions. It has been Cyclooxygenase-2 (COX-2) has prostaglandin (PG) hydroperluteum was missing and the uterus was small (Dinchuk 2 from PGG 2. COX-2 proven that COX-2 knock-out mice were infertile, the corpus oxidase activity to synthesize PGH exists in two isoforms, COX-1 and COX-2. The human et al., 1995). COX-1 gene is localized on chromosome 9 and the COX-2 It is known that human endometrium produces PGE 2 and PGF gene is localized on chromosome 1 (Hla and Neilson, 1992; 2α (Smith and Kelly, 1988). It has been reported (Jones et al., 1997) that COX-2 is distributed in endometrial Appleby et al., 1994). They are governed by completely glandular epithelia and vascular endothelia and that it varies different genes, but their protein structures are very similar. during the menstrual cycle. Other studies have also found COX-1 is constitutively present in almost all tissues. It acts that COX-2 is distributed in the human placenta (Wetzka to maintain cell homeostasis (Zweifel et al., 1995). COX-2 et al., 1997) and decidual tissue (Ishihara et al., 1995). A is almost never expressed under normal conditions. It is significant amount of prostaglandins are produced from the induced during inflammation and cell proliferation and endometrium and endometriotic tissues (Lumsden et al., differentiation. It is expressed in macrophages, fibroblasts, 1984). It is interesting to note that low concentrations of vascular endothelial cells, neurons and chondrocytes as a arachidonic acid are only metabolized by COX-2 (Morita result of interleukin-1 (IL-1) (O Banion et al., 1992), human et al., 1995). Accordingly, COX-2 expression may be chorionic gonadotrophin (Sirois and Richards, 1992) or abnormal in endometriosis, but there appear to be no reports serum (Xie et al., 1991) stimulation. COX-2 is not only of investigations of COX-2 kinetics in this disorder. induced by the site of inflammation, but also produced in Therefore, in this study, we investigated the expression of colon cancer tissues (Sheng et al., 1998). COX-2 is also COX-2 in eutopic and ectopic endometria in endometriosis. European Society of Human Reproduction and Embryology 561

2 H.Ota et al. Materials and methods Table I. Evaluation of nomogram scores of cyclooxygenase-2 in the surface Patients epithelium during the phases of the menstrual cycle in endometriosis and The subjects consisted of 118 women who were treated at the adenomyosis. Values are presented as mean SEM. Values in parentheses indicate the number of patients examined for the antigen Department of Obstetrics and Gynecology at Akita University Hospital. They were divided into three groups: (i) fertile controls; (ii) the Phase of Study group endometriosis group (n 35) which included women diagnosed by menstrual cycle laparoscopy; and (iii) the adenomyosis group (n 33), who had Control a Endometriosis b Adenomyosis undergone hysterectomy. Endometriosis patients associated with adenomyosis were excluded from the present study and vice versa. Proliferative phase The controls consisted of 50 fertile women with regular and biphasic Early (3) (3) (3) menstrual cycles, 20 in the proliferative phase and 30 in the secretory Middle (8) (8) (6) phase. All of the controls were parous women with clear male factor Late c (9) (4) (8) infertility (mild oligozoospermia or azoospermia). None of controls Secretory phase had identifiable endometriosis confirmed by laparoscopy and adeno- Early (12) (9) (6) Middle (10) (6) (5) myosis by serum CA-125, ultrasonography and/or magnetic reson- Late (8) (5) (5) ance imaging. These women conceived after artificial insemination using hus- a P between the six phases in the control group by the Kruskal band s or donor s semen within three treatment cycles and delivered Wallis test. full-term babies. The mean ages in the fertile control group, the b P 0.05 between the six phases in the endometriosis group by the endometriosis group, and the adenomyosis group were 28.9 years Kruskal Wallis test. P 0.05 among the three groups by the Kruskal Wallis test. (range years), 32.1 years (range years), and 43.1 years (range years) respectively. Before starting any medication in the control group, or just after tetrahydrochloride containing 0.005% hydrogen peroxide in PBS. laparoscopy in the patients with endometriosis or after hysterectomy Finally, the sections were counterstained with Carazzi s haematoxylin. in the patients with adenomyosis, endometrial tissue was obtained Negative controls for immunostaining were prepared by substituting when it became available during any phase of the menstrual cycle. the first antibody with non-immune rabbit serum IgG. In each run, a Ectopic endometrial tissues in ovarian chocolate cysts in endometriosis section of placenta with strong COX-2 staining was routinely included (n 10) or in the myometrium in adenomyosis (n 9) were studied as a positive control. in the identical patients at the same time. In the control and Evaluation of staining the endometriosis group, endometrial specimens were obtained by curettage. The tissues were fixed in neutral-buffered 10% formalin Ten non-overlapping fields of view were examined per biopsy in a solution. The menstrual cycle of the patients was estimated by systematic random sampling pattern (magnification 400). Surface endometrial dating according to previously described criteria (Noyes and glandular epithelia and stromal cells in eutopic and ectopic et al., 1950). Informed consent was obtained in each case, and endometria were evaluated for COX-2 staining. Staining was evaluated approval for the study was granted by the Institutional Review Board. using an evaluation nomogram as previously reported (Ota and Igarashi, 1993). Briefly, each section was graded according to the Reagents frequency of positive cells and intensity of staining in endometrium. The frequency was defined as 1, 2 or 3 when the number of The goat polyclonal antibody for human COX-2 (sc-1745) raised positive cells in the endometrium in each section was 10%, 10 against a peptide mapping at the carboxy terminus of COX-2 of 50% or 50% respectively. Intensity was defined as 3 when staining human origin was obtained from Santa Cruz Biotechnology Inc. of the cells was as strong as that observed in the positive controls, (Santa Cruz, CA, USA). The second antibody [rabbit anti-sheep as 1 when staining was weakly positive but distinct from the immunoglobulins (Ig) H L horseradish peroxidase conjugate; 6155 negative controls, and as 2 when the staining was between 1 and 04] was obtained from Southern Biotechnology Associates Inc. 3. The sections were ranked from 1 to 5 according to the evaluation (Birmingham, AL, USA). nomogram. Sampling and grading of each specimen were done by Staining two different observers blinded as to the specimen source. Sections were assigned a score by a first observer, and confirmed by a The endometrial tissue samples were cut into blocks (~1 cm 3 ). Serial second observer. 3 mm sections of tissue were cut, deparaffinized, and rehydrated through ethanol, as in routine histology. The sections were stained Statistical analyses using the indirect method. Before the staining, a microwave antigen The results are expressed as the mean SEM where applicable. retrieval technique was utilized, whereby the sections were heated in Statistical analysis was performed by the Kruskal Wallis test. P sodium citrate buffer (ph 6.0) for 5 min at high power (500 W), and 0.05 was considered to be statistically significant. this was repeated four times. First, the sections were incubated in 3% hydrogen peroxide in phosphate-buffered saline solution (PBS; 0.1 mol/l) for 15 min to block endogenous peroxidase activity. Nonspecific Results background staining was reduced by treating the sections Variation in COX-2 expression during the menstrual cycle with non-immune 10% swine serum (809 10; Cosin Bio, Sakato, in eutopic endometrium Japan) in PBS. Then, the polyclonal antibody ( 50 dilution) was added and the sections were incubated overnight at 4 C. After washing Expression of COX-2 in the surface epithelium in the control them with PBS, the second antibody ( 400 dilution) was added and group varied markedly during the menstrual cycle (Figure 1, they were incubated for 1 h at 37 C. After washing the sections with Table I). It was lowest in the early proliferative phase and PBS, they were stained with 0.2 mg/ml 3,3 -diaminobenzidine (DAB) gradually increased thereafter. It remained high throughout the 562

3 COX-2 in endometriosis Table II. Evaluation of nomogram scores of cyclooxygenase-2 in the glandular epithelium during the phases of the menstrual cycle in endometriosis and adenomyosis. Values are presented as mean SEM. Values in parentheses indicate the number of patients examined for the antigen Phase of menstrual cycle Study group Control a Endometriosis b Adenomyosis Proliferative phase Early (3) (3) (3) Middle c (8) (8) (6) Late d (9) (4) (8) Secretory phase Early (12) (9) (6) Middle (10) (6) (5) Figure 1. Evaluation nomogram score of cyclooxygenase-2 in the Late (8) (5) (5) surface epithelium during the menstrual cycle in the fertile controls and in the endometriosis group. d control group, s a P between the six phases in the control group by the Kruskal endometriosis group. Wallis test. b P 0.01 between the six phases in the control group by the Kruskal Wallis test. c P 0.05 between the six phases in the endometriosis group by the Kruskal Wallis test. d P 0.01 among the three groups by the Kruskal Wallis test. Table III. Evaluation of nomogram scores of cyclooxygenase-2 in the stromal cells during the phases of the menstrual cycle in endometriosis and adenomyosis. Values are presented as mean SEM. Values in parentheses indicate the number of patients examined for the antigen Phase of menstrual cycle Study group Control Endometriosis Adenomyosis Figure 2. Evaluation nomogram score of cyclooxygenase-2 in the glandular surface epithelium during the menstrual cycle in the fertile controls and in the endometriosis group. d control group, s endometriosis group. more pronounced than that in the control group. On the other hand, there was no variation in expression of COX-2 in the adenomyosis group during the menstrual cycle, and COX-2 expression was lower than that in the endometriosis group in all phases. Expression of COX-2 in stromal cells in the control group was weaker than that in the surface and glandular epithelia, and there was no variation during the menstrual cycle (Table III). In patients with endometriosis, unlike that in the surface and glandular epithelia, there was no variation during the menstrual cycle, but COX-2 expression tended to be slightly higher than that in the control group. In patients with adeno- myosis, there was no difference from the control group. 563 secretory phase. In patients with endometriosis, however, it varied throughout the menstrual cycle, but the range of the variation was narrower than that in the control group. In the late proliferative phase, expression of COX-2 was significantly more pronounced than that in the control group. On the other hand, there was no variation in expression of COX-2 in the adenomyosis group during the menstrual cycle. COX-2 expression in the glandular epithelium in the control group showed a similar tendency to the shift in expression in the surface epithelium (Figure 2, Table II). In other words, it was lowest in the early proliferative phase and gradually increased thereafter. It peaked at the early secretory phase, and the same amount of expression was maintained up to the late secretory phase. In patients with endometriosis, however, there was a significant variation in COX-2 expression during the menstrual cycle, but, like the surface epithelium, the difference in expression was reduced. In the mid- to late proliferative phases, expression of COX-2 was significantly Proliferative phase Early (3) (3) (3) Middle (8) (8) (6) Late (9) (4) (8) Secretory phase Early a (12) (9) (6) Middle b (10) (6) (5) Late (8) (5) (5) a P 0.05 among the three groups by the Kruskal Wallis test. b P 0.02 among the three groups by the Kruskal Wallis test.

4 H.Ota et al. Figure 3. Cells stained for cyclooxygenase-2 antigen. Eutopic endometrium in (A) the late proliferative phase in the control group (evaluation nomogram score, 2) and (B) in the early secretory phase in the control group (evaluation nomogram score, 3). Eutopic endometrium in (C) the early secretory phase in the endometriosis group (evaluation nomogram score, 4) and (D) surface epithelium in the late secretory phase in the adenomyosis group (evaluation nomogram score, 4). (E) Ectopic endometrial tissue in the late secretory phase in the endometriosis group (evaluation nomogram score, 3) and (F) ectopic endometrial tissue in the late secretory phase in the adenomyosis group (evaluation nomogram score, 3). (A D) 100 magnification; (E and F) 50 magnification. COX-2 expression in ectopic endometrium Expression of COX-2 was observed in ectopic endometrial tissue in all cases. Mean evaluation nomogram levels in glandular epithelium and stromal cells in ovarian chocolate cyst wall were and respectively. They were higher than that in the secretory phase of the control group. On the other hand, evaluation nomogram scores in glandular epithelium and stromal cells in adenomyosis tissue 564 were and respectively, approximately the same as that in the early to mid-proliferative phases in the eutopic endometrium. Histological findings COX-2 formed a granular pattern or texture in the cytoplasm of glandular cells and cells were stained almost diffusely (Figure 3). Pronounced localization was not found in cell

5 COX-2 in endometriosis membranes. Polarity in the basal and apical sides was not Myometrial movement in patients with endometriosis was particularly pronounced. No localization was seen in nuclei. observed using ultrasonography (Leyendecker et al., 1996) Cytoplasm of interstitial cells was slightly positive. and they found that hyperperistalsis and dysperistalsis are frequently observed in patients with endometriosis, and speculated that abnormal peristalsis may inhibit normal sperm Discussion transport in the uterus. Secondly, excessive prostaglandins Expression of COX-2 varied with menstrual cycle. In a may be impairing implantation of fertilized oocytes. The previous report (Jones et al., 1997), the expression of COX-2 implantation rate was said to have decreased significantly in in endometrial glandular epithelium peaked at the menstrual an experimental endometriotic model using rabbits that were phase, bottomed out in the ovulatory phase and gradually subsequently artificially inseminated (Hahn et al., 1986). began to increase again in the secretory phase. These results, COX-2 expression was found in endometriotic tissues of however, differed slightly from theirs. In this study, it was the ovarian chocolate cyst wall in all cases, and expression of low in the early proliferative phase and gradually increased COX-2 in glandular epithelial cells was almost as pronounced thereafter. It peaked at the early secretory phase, and the same as that in the secretory phase in the eutopic endometrium. In amount of expression was maintained into the late secretory fact, there have been many reports that large amounts of PGs phase. Possible reasons for this difference include a variation are produced in ascites and endometrial tissues in endometriosis in controls (women with underlying disease versus healthy (Badawy et al., 1985; De Leon et al., 1986; Moon et al., 1981). women with fecundity), parameters of assessing staining Over-expression of COX-2 in ectopic endometrial tissue, (intensity only versus evaluation nomogram score), difference and prostaglandins that is produced as a result, may be involved in the immunized animal for the first antibody (rabbit versus in the proliferation and differentiation of cells and malignant goat) and staining method. In addition, the volume of prostaglandins transformation. For example, it was reported that when COX- secreted from endometrium differs during the men- 2 was over-expressed in epithelial cells of mouse small strual cycle. The volume of PGE 2 and PGF 2α secreted from intestine, expression of adhesion molecules such as laminin endometrial glandular epithelial cells was investigated in vitro increased, while apoptosis was suppressed (Tsujii and DeBois, (Lumsden et al., 1984). According to their findings, PGE 2 in 1995). It has also been reported that the incidence of colon the proliferative phase was 1.5 times that in the secretory cancer and the number and size of colorectal polyps decreased phase, whereas PGF 2α was 3.4 times higher in the secretory in people who took aspirin for an extended length of time phase. Consequently, increased expression of COX-2 in the (Giovannucci et al., 1995; Marcus, 1995). It is said that secretory phase may contribute mainly to production of PGF 2α. apoptosis is reduced in the eutopic and ectopic endometria in This is the first study to look at the expression of COX-2 endometriosis compared with normal women (Dmowski et al., in eutopic and ectopic endometria in endometriosis. Expression 1998; Imai et al., 2000). Furthermore, apoptosis is further of COX-2 in eutopic endometrium varied significantly during reduced in the ectopic endometrium compared with the eutopic the menstrual cycle. Expression was higher in the secretory endometrium (Gebel et al., 1998). In fact, loss of heterozygosity phase than in the proliferative phase, but it was higher than was frequently observed on chromosomes in endometriotic that in the control throughout the menstrual cycle. Overexpression tissue from ovarian chocolate cysts (Jiang et al., 1996). It is of COX-2 leads to increased production of prosta- concluded that COX-2 expression in endometriosis appears to glandins. The mechanism of over-expression of COX-2 in be intimately involved in its pathology. Further investigations eutopic endometrium in endometriosis is unclear. It has been into the involvement of COX-2 are required. indicated that an increased autoimmune response occurs in the endometrium in endometriosis. For instance, T cells and B cells are increased (Witz et al., 1994; Ota et al., 1996a), Acknowledgement macrophage is activated (Haney et al., 1981), and immuno- This manuscript was in part supported by Grants-in-Aid for Scientific globulin and complements are deposited in endometrial glandu- Research from the Ministry of Education, Japan (No ). lar epithelial cells (Weed and Arquembourg, 1980; Ota and Maki, 1990). Furthermore, human leukocyte antigen expression in glandular epithelial cells is increased (Ota and Igarashi, References 1993) and adhesion molecules are abnormally expressed (Les- Appleby, S.B., Ristimaki, A., Neilson, K. and et al. (1994) Structure of the sey et al., 1992; Ota et al., 1996b). In addition, expression of human cyclooxygenase-2 gene. Biochem. J., 302, Badawy, S.Z.A., Marshall, L. and Cuenca, V. (1985) Peritoneal fluid free radical-related enzymes such as nitric oxide synthase (Ota prostaglandins in various stages of the menstrual cycle: role in infertile et al., 1998), superoxide dismutase (Ota et al., 1999a), and patients with endometriosis. Int. J. Fertil., 30, glutathione peroxidase is increased (Ota et al., 1999b; Ota De Leon, F.D., Vijayakumar, R., Brown, M. et al. (1986) Peritoneal fluid et al., 2000). Consequently, a variety of cytokines are secreted volume, estrogen, progesterone, prostaglandin, and epidermal growth factor concentrations in patients with and without endometriosis. Obstet. Gynecol., from these immune cells and macrophages, including IL-1 and 68, IL-2. These cytokines may very well initiate COX-2 expression. Dinchuk, J.E., Car, B.D., Focht, R.J. et al. (1995) Renal abnormalities and an Over-expression of COX-2 and the accompanying increase in altered inflammatory response in mice lacking cyclooxygenase II. Nature, 378, prostaglandin production likely initiate an abnormal state in the Dmowski, W.P., Gebel, H. and Braun, D.P. (1998) Decreased apoptosis and uterus. Endometriosis is often accompanied by dysmenorrhoea. sensitivity to macrophage mediated cytolysis of endometrial cells in Abnormal uterine contraction is also observed in endometriosis. endometriosis. Hum. Reprod. Update, 4,

6 H.Ota et al. Gebel, H.M., Braun, D.P., Tambur, A. et al. (1998) Spontaneous apoptosis of components C3 and C4 in endometriotic tissue or endometrium in women endometrial tissue is impaired in women with endometriosis. Fertil. Steril., with adenomyosis or endometriosis. Med. Sci. Res., 18, , Ota, H., Igarashi, S., Hayakawa, M. et al. (1996a) Effect of danazol on the Giovannucci, E., Egan, K.H., Hunter, D.J. et al. (1995) Aspirin and the risk immunocompetent cells in the eutopic endometrium in patients with of colorectal cancer in women. N. Engl J. Med., 333, endometriosis: a multicenter cooperative study. Fertil. Steril., 65, Hahn, D.W., Carraher, R.P., Foldesy, R.G. et al. (1986) Experimental evidence Ota, H., Igarashi, S. and Tanaka, T. (1996b) Expression of γδt cells and for failure to implant as a mechanism of infertility associated with adhesion molecules in endometriotic tissue in patients with endometriosis endometriosis. Am. J. Obstet. Gynecol., 155, and adenomyosis. Am. J. Reprod. Immunol., 35, Haney, A.F., Muscato, J.J. and Weinberg, J.B. (1981) Peritoneal fluid cell Ota, H., Igarashi, S., Hatazawa, M. et al. (1998) Endothelial nitric oxide populations in infertility patients. Fertil. Steril., 35, synthase in the endometrium during the menstrual cycle in patients with Hla, T. and Neilson, K. (1992) Human cyclooxygenase-2 cdna. Proc. Natl endometriosis and adenomyosis. Fertil. Steril., 69, Acad. Sci. USA, 89, Ota, H., Igarashi, S., Hatazawa, M. et al. (1999a) Immunohistochemical Imai, A., Takagi, A. and Tamaya, T. (2000) Gonadotropin-releasing hormone assessment of superoxide dismutase expression in the endometrium in analog repairs reduced endometrial cell apoptosis in endometriosis in vitro. endometriosis and adenomyosis. Fertil. Steril., 72, Am. J. Obstet. Gynecol., 182, Ota, H., Igarashi, S., Hatazawa, J. et al. (1999b) Endometriosis and free Ishihara, O., Matsuoka, K., Kinoshita, K. et al. (1995) Interleukin-1β- radicals. Gynecol. Obstet. Invest., 48 (Suppl.), stimulated PGE 2 production from early first trimester human decidual cells Ota, H., Igarashi, S., Kato, N. et al. (2000) Aberrant expression of glutathione is inhibited by dexamethasone and progesterone. Prostaglandins, 49, peroxidase in eutopic and ectopic endometrium in endometriosis and Jiang, X., Hitchcock, A., Bryan, E.J. et al. (1996) Microsatellite analysis of adenomyosis. Fertil. Steril., 74, endometriosis reveals loss of heterozygosity at candidate ovarian tumor Sheng, H., Shao, J., Morrow, J.D. et al. (1998) Modulation of apoptosis and suppressor gene loci. Cancer Res., 56, bcl-2 expression by prostaglandin E 2 in human colon cancer cells. Cancer Jones, R.L., Kelly R.W. and Critchley, H.O.D. (1997) Chemokine and Res., 58, cyclooxygenase-2 expression in human endometrium coincides with Sirois, J. and Richards, J.S. (1992) Purification and characterization of a leukocyte accumulation. Hum. Reprod., 12, novel, distinct isoform of prostaglandin endoperoxide synthase induced by Lessey, B.A., Damjanovich, L., Coutifaris, C. et al. (1992) Integrin adhesion human chorionic gonadotropin in granulosa cells of rat preovulatory follicles. molecules in the human endometrium. J. Clin. Invest., 60, J. Biol. Chem., 267, Lumsden, M.A., Brown, A. and Baird, D.T. (1984) Prostaglandin production Smith, S.K. and Kelly, R.W. (1988) The release of PGF 2α and PGE 2 from from homogenates of separated glandular epithelium and stroma from separated cells of human endometrium and decidua. Prostagland. Leuk. human endometrium. Prostaglandins, 23, Ess. Fatty Acids, 33, Leyendecker, G., Kunz, G., Wildt, L. et al. (1996) Uterine hyperperistalsis Tsujii, M. and DuBois, R.N. (1995) Alterations in cellular adhesion and and dysperistalsis as dysfunctions of the mechanism of rapid sperm transport apoptosis in epithelial cells overexpressing prostaglandin endoperoxide in patients with endometriosis and infertility. Hum. Reprod., 7, synthase 2. Cell, 83, Marcus, A.J. (1995) Aspirin as prophylaxis against colorectal cancer. N. Engl. Weed, J.C. and Arquembourg, P.C. (1980) Endometriosis: can it produce an J. Med., 333, autoimmune response resulting in infertility? Clin. Obstet. Gynecol., 23, Moon, Y.S., Leung, P.C.S., Yuen, B.H. et al. (1981) Prostaglandin F in human Wetzka, B., Nüsing, R., Charnock-Jones, D.S. et al. (1997) Cyclooxygenaseendometriotic tissue. Am. J. Obstet. Gynecol., 141, and -2 in human placenta and placental bed after normal and pre-eclamptic Morita, I., Smith, W.L., DeWitt, D.L. et al. (1995) Expression-activity profiles pregnancies. Hum. Reprod., 12, of cells transfected with prostaglandin endoperoxide H synthase measured Witz, C.A., Montoya, I.A., Dey, T.D. et al. (1994) Characterization of by quantitative fluorescence microscopy. Biochemistry, 34, lymphocyte subpopulations and T cell activation in endometriosis. Am. J. Noyes, R.W., Hertig, A.T. and Rock, J. (1950) Dating the endometrial biopsy. Reprod. Immunol., 32, Fertil. Steril., 1, Xie, W., Chipman, J.G., Robertson, D.L. et al. (1991) Expression of a mitogen- O Banion, M.K., Winn, V.D. and Young, D.A. (1992) cdna cloning and responsive gene encoding prostaglandin synthase is regulated by mrna functional activity of a glucocorticoid-regulated inflammatory cyclo- splicing. Proc. Natl Acad. Sci. USA, 88, oxygenase. Proc. Natl Acad. Sci. USA, 89, Zweifel, B.S., Colburn, S.M., Orgberg, R.L. et al. (1995) The role of Ota, H. and Igarashi, S. (1993) HLA-DR expression in endometriotic tissue cyclooxygenase-2 in inflammation. FASEB J., 9, A110. in patients with endometriosis and adenomyosis. Fertil. Steril., 60, Ota, H. and Maki, M. (1990) Content of immunoglobulin G and complement Received on September 4, 2000; accepted on November 17,