Localization of leukocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle*

Transcription

1 FERTILITY AND STERILITY Copyright 1994 The American Fertility Society Vol. 61, No, 3, March 1994 Printed on acid-free paper in U. S. A. Localization of leukocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle* Mats Brannstrom, M,D" Ph,D,t Vivian Pascoe, RSc.t Robert J. Norman, M.D.t:J: Neil McClure, M.D. Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, Woodville, South Australia, and Monash University, Clayton, Victoria, Australia Objective: To determine the presence and localization ofleukocyte subsets in the human follicular wall and in the corpus luteum (CL) throughout the menstrual cycle. Design: Tissue was obtained by biopsy at laparoscopy. Frozen sections were labeled with a panel of monoclonal antibodies against leukocyte subtype markers and detected by immunohistochemical methods. Setting: Infertility clinic in large teaching hospital. Patients: Eighteen women who underwent diagnostic laparoscopy. Women with abnormal endocrinology, endometriosis, or ovarian disease were excluded. Results: Macrophages and neutrophilic granulocytes were present in high numbers in the collagen-rich tissues, including the thecal layer, of the follicle wall. At ovulation, there was a marked increase in the density of these cells in the follicle wall, especially in the thecal layer. In the CL, there was a higher density of macrophages (2- to 3-fold) and neutrophils (3- to 5-fold) in the theca lutein area compared with the granulosa lutein area. There were no significant changes in the density of these cell types when early and late luteal phase were compared, but the density of macrophages were markedly elevated in regressing CLs obtained during the subsequent follicular phase. T lymphocytes were present in low numbers in the follicle wall and in the CL, and no variation in the density was detected throughout the menstrual cycle. Conclusion: Macrophages and neutrophilic granulocytes are present in high numbers in the follicle wall and are further increased in the theca at ovulation, suggesting an active role for these leukocyte subtypes in tissue remodeling during the ovulatory process. Macrophages and neutrophils are also abundant in the CL, and an increase in the macrophage density at CL regression could imply a role for these cells in luteolysis. Fertil Steril 1994;61: Key Words: Corpus luteum, follicle, human leukocyte, lymphocyte, macrophage, neutrophilic granulocyte, ovary Received April 23, 1993; revised and accepted October 28, * Supported by a Queen Elizabeth Hospital Fellowships, Woodville, South Australia, and by grant 4982 from the Swedish Medical Research Council, Assar Gabrie Research Foundation, Tore Nilssons Foundation, and Goteborg Medical Society, Goteborg, Sweden. t The University of Adelaide, The Queen Elizabeth Hospital. :j: Reprint requests: Robert J. Norman, M.D., Reproductive Several studies in animal species suggest that tissue-bound leukocytes might be important in the cyclic events in the ovary because of their secretion of a number of inflammatory and immunomodulating substances, including potent connective tissue enzymes and cytokines. There are cyclic variations in the density and regional distribution of leukocytes Medicine Unit, Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, Woodville, South Australia, 5011, Australia. Monash University, 488 Brannstrom et a1. Leukocytes in the human follicle wall and CL Fertility and Sterility

2 in the ovary of several experimental animals (1-3), and a role for these leukocytes has been proposed in the tissue remodeling and regulation of steroidogenesis at ovulation (4), luteinization (5), and luteo lysis (3). Evidence for the participation of leukocytes in ovulation is shown by the increased ovulation rate induced by leukocytes added to the in vitro perfused rat ovary (4) and a marked increase of certain leukocyte subpopulations after the preovulatory LH surge in the rat (2), sheep (1), and other species. In the human, it has been shown that follicular fluid (FF) contains chemotactic activity toward neutrophilic granulocytes (6) and that macrophages (7) and T lymphocytes (8) are present in FF at oocyte pick-up for IVF-ET. Leukocytes, and in particular macrophages, also have been suggested to be important during luteolysis in the phagocytosis of luteal cells. Recently, neutrophilic granulocytes were shown to exert luteolytic functions when coincubated with luteal cells (9). In the human corpus luteum (CL), we have reported the presence of several leukocyte subclasses, with a dominance of macrophages and T lymphocytes. Cyclic changes were not reported (10). The present study is apparently the first to examine the distribution ofleukocyte subtypes in the human follicular wall and to describe the distribution of leukocytes in CL obtained at specified times of the menstrual cycle. Description of the type and distribution of leukocytes in the human ovary will help to decipher the role of these cells in the physiology of this reproductive organ. MATERIALS AND METHODS Patients and Tissue Material Eighteen patients undergoing laparoscopy for the investigation of infertility, pelvic pain, or for sterilization gave informed consent for intraoperative biopsy of the wall of the largest ovarian follicle or of the CL. The study received ethical approval from the Institutional Review and Ethics Committee, Epworth Hospital, Melbourne, Australia. Patient exclusion criteria included endometriosis, ovarian cysts (either in isolation or of multi cystic or polycystic appearance), ovarian malignancies, a history of ovarian surgery (apart from adhesiolysis), a self-reported menstrual cycle length :s;; 26 or ~30 days, and the use of any hormonal therapy within the last 3 months. Immediately before the induction of anesthesia, a 10-mL blood sample was taken for the measurement of serum E 2, P, FSH, and LH. During the laparoscopic procedure, the largest follicle or the CL was identified. The ovary was stabilized and then, using the Storz endoscopic biopsy forceps (D-7200; Storz, Tuttlingen, Germany), approximatelya 3 X 1.5-mm biopsy of the follicle/cl wall was taken as close to the superficial margin of the follicle or as far from the stigma of ovulation in the CL as was possible. Bleeding from the biopsy site was excluded after irrigation and was usually not a problem. If necessary, hemostasis was achieved with bipolar diathermy applied at low voltage. The biopsies were immediately immersed in Tissue Tek OCT (Miles Inc., Elkhart, IN) and frozen in isopentane in liquid nitrogen. Specimens were kept at -80 C until further use. Based on hormonal data and menstrual history, the specimens of follicular walls were grouped into midfollicular phase (days -8 to -3), late follicular phase (days -2 to -1), and ovulation time (day 0, with a detectable LH surge at specimen collection). The luteal samples were grouped into early luteal (days 1 to 7), late luteal (days 8 to 14), and regressing luteal (luteal tissue obtained after menstruation). Immunohistochemistry Micrometer frozen sections that were 6 1Lm thick were cut and placed on chrome-alum-gelatincoated slides. The sections were air dried for 12 hours followed by microwave radiation for 15 seconds and then fixed in acetone for 5 minutes. Normal horse or goat serum (Vector Laboratory, Burlingame, CA) was applied to the sections for 20 minutes at room temperature, followed by incubation with primary monoclonal antibody (mab) (Table 1), diluted in phosphate-buffered saline (PBS), for 60 minutes. Sections were washed once with PBS and then incubated with biotinylated horse anti-mouse immunoglobulin (Ig)G or goat antimouse IgM (Vector Laboratory) for 30 minutes, followed by addition of avidin-biotin-peroxidase complex (Vector Laboratory) for 60 minutes. Bound antibody was detected by incubation with 0.05% 3,3' diaminobenzidine-tetrahydrochloride (Sigma Chemical Co., St. Louis, MO) in PBS containing 0.03% H for 7 minutes. The sections were counter stained with hematoxylin. A positive reaction was demonstrated by a brown reaction product. Negative controls were incubated similarly but with a mouse mab against Giardia intestinalis at Vol. 61, No.3, March 1994 Brannstrom et al. Leukocytes in the human follicle wall and CL 489

3 Table 1 Reactivity, Dilutions, and Source of MAbs Against Human Leukocyte Determinants MAb Dilution Predominant reactivity Source Anti-Leu M1 1:50 Monocyte (CD15) in tissue; granulocyte cells in blood Anti-Leu M5 1:40 Macrophage (CDllc) only * Anti-Leu 4 1:50 T lymphocyte (CD3) only * Anti-Leu 2a 1:40 T-cytotoxic/suppressor lymphocyte (CD8) and some NK lymphocytes * Anti-Leu 14 1:5 B lymphocyte (CD22) only * Anti-Leu 7 1:20 NK cell and a subset T cells * MCA 149 1:800 Neutrophilic granulocyte t Anti-Giardia intestinalis 1:4 Giardia intestinalis (negative control) :j: * Becton Dickinson and Co., Mountain View, California (detailed cross-reacting antigens shown in catalog). t Serotec Ltd., London, United Kingdom. :j: Gift from Graham Mayrhofer, Ph.D., Department of Immunology and Microbiology, University of Adelaide, Australia. 1:4 dilution. Sections of tonsil and granuloma tissue were used as positive controls. Quantification of Positively Stained Cells The number of positive cells in tunica albuginea, stroma, theca, theca-lutein, and granulosa lutein were counted within parts of a lined grid area covering an area of the section of mm 2 The granulosa cell layer in follicles was only present in a few samples. In the germinal epithelium, the number of positive cells over a length of 1.5 mm was counted. In each area, three to six randomly chosen areas were counted. Counting was performed independently by two observers blinded to the specimen details, and the mean of the two observations was used. The interobserver variation was <10%. Statistics Statistical differences were calculated with the nonparametric Mann -Whitney U-test. A P value of <0.05 was considered significant. RESULTS General Morphology and Tissue Types Six samples were obtained from follicle walls before ovulation. Two of these were taken during the LH surge, just before ovulation (ovulation, day 0). Three samples were from the midfollicular phase (cycle days -8 to -3) and 1 sample from late follicular phase (cycle days -2 to -1) Twelve samples were from luteal tissue, and two of these samples were of structurally regressing CL obtained during the subsequent follicular phase. Five ofthe 12 luteal samples were from early luteal phase (cycle days 1 to 7) and 5 from late luteal phase (cycle days 8 to 14). The germinal epithelium was partly present in all follicular wall samples and was lacking in some CL samples. In these latter samples, the border be- tween the tunica albuginea and germinal epithelium could be clearly identified. The tunica albuginea was defined as the outermost layer of the ovarian cortex, made up of a narrow layer of densely arranged fibroblast-like cells aligned parallel with the germinal epithelium (Fig. la). The stromal layer was recognized by its position between tunica albuginea and the thecal blood vessels and was composed of irregularly arranged fibroblast-like cells. Occasionally, primordial and primary follicles were seen in this layer. The thecal layer was obvious by the presence of thecal blood vessels and the regular organization of the tissue. No distinction could be made between theca interna and theca externa in the frozen sections of the present study. Granulosa cells were only occasionally present in the samples, and the material was not sufficient to assess this layer. In the CL, the granulosa lutein area was easily identified by the cells' larger rounded nuclei and the enlarged cytoplasm. The thecal-lutein area was peripheral to the granulosa lutein, and the cells were more densely arranged (Fig. IB). Distribution of Monocytes and Macrophages There were, in general, very few monocytes present in the tissue sections. In the walls of the follicles during the midfollicular phase, monocytes were scarce and mostly present in the stroma. At ovulation, some monocytes were found in the theca of the preovulatory follicles (2 and 3.3 per grid area) where they were situated around the blood vessels. After ovulation, monocytes could occasionally be observed in the granulosa lutein and theca lutein areas (0 to 3.3 per grid area), and in a majority of the samples the density was greater in the theca lutein area than in the granulosa lutein area. Macrophages were present in much greater density than monocytes. In the follicle walls, 490 Brannstrom et al. Leukocytes in the human follicle wall and CL Fertility and Sterility

4 Figure 1 Photomicrographs of ovarian biopsies. (A), Section of follicle wall from midfollicular phase. The different layers are marked in the photograph. GE, germinal epithelium; TA, tunica albuginea; ST, stroma. (B), Section of CL from late luteal phase. The theca lutein (TL) and granulosa-lutein (GL) areas are shown. Magnification was x60. (C), Macrophage (arrow) in contact with germinal epithelial cells in follicle wall of midfollicular phase. Magnification was X100. (D), Macrophages in a follicle wall at ovulation. Magnification was X60. (E), Macrophages in a CL. Magnification was x60. (F), T lymphocytes (arrows) in follicle wall of midfollicular phase. Magnification was x60. (G), Neutrophilic granulocytes in follicle just after ovulation. Note the high density of neutrophilic granulocytes on both sides of the rupture point. Magnification was x40. (H), Neutrophilic granulocytes in CL. Magnification was x60. macrophages were observed in close proximity to germinal epithelium cells (1 to 3 macrophages per 1.5 mm; Fig. lc). The macrophages were scattered among the fibroblast-like cells in tunica albuginea and stroma, and the density of these cells was simi1ar in all preovulatory samples, with a marked increase in the density of the two samples at ovulation (Fig. 2A). There was also a pronounced increase in macrophage density in the theca immediately before ovulation (Fig. ID). Vol. 61, No.3, March 1994 In the capsule of the CL, macrophages were present in the germinal epithelium and tunica albuginea in numbers between 1 and 6 per 1.5 mm germinal epithelium length and per grid area, respectively. No time-related changes were observed. Stromal tissue was only obtained in four CL samples with a density of macrophages below 1.3 per grid area. The density of macrophages was significantly higher in the theca lutein than in the granulosa lutein area both during early and late Brannstrom et al. Leukocytes in the human follicle wall and CL 491

5 10 no. of positive cells/grid area A no. of positive cells/grid area B 10.=TA.=TA.=ST.=ST , TA 0 0 MF LF 0 MF LF 0 no. of positive cells/grid area no. of positive cells/grid area =TH =TH MF LF 0 MF LF 0 Figure 2 Macrophages and neutrophils in ovarian biopsies. (A), Number of macrophages in follicle wall sections of midfollicular (MF) phase, late follicular (LF) phase, and at ovulation (0). Top:..t., values in tunica albuginea (TA); e, values in stroma (ST). Bottom: Values in theca (TH). Note the difference in scale between top and bottom. (B), Number of neutrophilic granulocytes in follicle wall sections from midfollicular (MF) phase, late follicular (LF) phase, and at ovulation (0). Top:..t., values in tunica albuginea (TA); e, values in stroma (ST). Bottom: Values in theca (TH). Note the difference in scale between top and bottom. luteal phases (Fig. 3). Macrophages were evenly distributed in the granulosa lutein and commonly aggregated around blood vessels in theca lutein area (Fig. IE). In the two samples of morphologically regressing CL, no distinction between granulosa lutein and theca lutein could be made. The overall density of macrophages was similar in early and late phase and was dramatically increased in the two regressing CL (Fig. 3). Distribution of Lymphocytes B lymphocytes were extremely rare in the sections and were occasionally noted only in the luteal tissue (1 to 2 positive cells per section). Natural killer (NK) cells also were found in very low numbers. In follicle walls, they were occasionally identified in the theca and in one sample in the germinal epithelium. In CL tissue, NK cells were found in two of the five tissue sections from the late luteal phase where they were identified solely in the theca lutein tissue. T lymphocytes, identified with Leu 4 antibody against CD3, were present in the tunica albuginea, stroma, and theca tissue of all preovulatory follicular walls (Fig. IF) with numbers below 5 positive cells per grid area. No preovulatory changes in density were observed. There were few T lymphocytes in the granulosa lutein and theca lutein areas, and no obvious increase was seen in regressing CL (Fig. 3). 492 Brannstrom et al. Leukocytes in the human follicle wall and CL Fertility and Sterility

6 No. of positive cells/grid area 300;=~~----~--~ ~--~ 250 EL * LL * x x RL x X X 'M T N Figure 3 Number (mean ± SEM) of macrophage~ (M), :r lymphocytes (T), and neutrophilic granulocytes (N) m? gnd area of mm" in granulosa lutein (_) and theca-lutem (II1II) area of early luteal (EL) and late luteal (LL) phase. Obs~rv~tions in the luteal area of regressing (RL) CL area are also mdicated. *Significantly higher than granulosa-lutein area of the same phase. T lymphocytes of the cytotoxic/suppressor subtype were only occasionally present in preovulatory follicular wall tissue. In CL, the number of cytotoxic/suppressor T lymphocytes was between 60% and 80% of the number of total T lymphocytes. No significant changes were found in the theca lutein and granulosa lutein areas when late luteal was compared with early luteal. Distribution of Neutrophilic Granulocytes Neutrophilic granulocytes were present in higher numbers than macrophages in germinal epithelium (6 to 12 per 1.5 mm) and tunica albuginea (Fig. 2B) of follicle walls. The distribution was scattered. Low numbers of neutrophilic granulocytes were observed in the stroma. There was a marked preovulatory increase in the neutrophil density in the theca concomitant with the LH surge (Fig. 2B). The density in the theca was in the same range as the density of macrophages in this area. Just after follicle rupture, a large number of neutrophils could be seen close to the rupture point (Fig. 1G). Neutrophilic granulocytes were also present in high number inside the CL (Fig. 1H) and in the tunica albuginea surrounding the CL. The cells were more numerous in the theca lutein than granulosa lutein, and no differences were observed when early luteal and late luteal were compared (Fig. 3). No obvious increase was detected in regressing CL.. DISCUSSION x During the development and life span ofthe follicle and CL, ovarian signals may attract certain leukocyte subpopulations to perform special functions such as breakdown of extracellular tissue and regulation of function of steroidogenic cells (11). Studies in rodents and domestic species have shown an influx of leukocytes at ovulation and that leukocytes are located in close proximity to luteal cells. The presence of leukocytes also increases the LHinduced ovulation rate in in vitro perfused rat ovaries (4), while macrophages (5), as well as neutrophils (9), alter the P production of luteal cells. The limited number of studies that have examined the presence and distribution of leukocytes in the human ovary have mostly used conventional histologic methods that do not permit classification of leukocytes into subtypes. By contrast, immunohistologic techniques employing mabs to specific antigens allow great accuracy in leukocyte identification and classification. Also, because of the limited availability of tissue, only one study on the human CL (12) employing immunohistochemistry against fibroblasts and macrophages has attempted to quantitate temporal changes in density. In the present study, we have examined the distribution of leukocytes in the follicle wall. We have, for the first time, demonstrated that leukocytes, particularly neutrophilic granulocytes and macrophages, are present in most layers of the follicle wall and that there is an increase in the densities of these cells in the thecal layer immediately after the onset of the LH surge. In the CL, leukocytes are localized in both the granulosa lutein and the theca lutein area with a predominance in the theca lutein area. No changes in the densities of these cells were apparent during the luteal phase, but a distinct increase in macrophage density was seen in two samples from regressing CL obtained after menstruation. The macrophage is the best established ovarian cell of lymphohematopoietic origin and has been localized in the ovary of several species (10, 12, 13). In the rat, it was found that macrophage density increased fivefold in the thecal layer at ovulation (2). Studies in the human have shown the presence of macrophages in the ovarian stroma (13), the CL (10,12), and in FF aspirates oflvf (7). The marked increase in the macrophage population at ovulation, identified in this study, is in accordance with our previous findings in the tat ovary (2). Activated macrophages are a well-established rich source ofthe multifunctional proinflammatory cytokines, tumor necrosis factor-a (TNF-a), and interleukin-1 (IL-1). It has been demonstrated that the levels of messenger ribonucleic acid of IL-1{j Vol. 61, No.3, March 1994 Brannstrom et ai. Leukocytes in the human follicle wall and CL 493

7 increase fourfold to fivefold in the thecal layer of the rat ovary after ovulation induction by hcg (14). Both TNF -a and IL-l{3 have been shown to induce synthesis of the ovulatory mediators prostaglandins and P in incubated rodent preovulatory follicles (15, 16) as well as significantly potentiating the ovulatory effect of LH in the in vitro perfused rat ovary model (17). Macrophages also secrete a number of other factors such as plasminogen activator, collagenase, eicosanoids, and platelet-activating factor (P AF) that are essential in connective tissue breakdown and vascular changes in ovulation (18). Macrophages have been described earlier to be present in high numbers inside the human CL with a predominance in the theca lutein area (10). In the present study, we confirmed these observations. Lei and co-workers (12) found an increase in macrophage density toward later stages of the luteal phase. We could not establish a difference in macrophage density between early and late luteal phase but detected an increase in regressing CL. This tissue had a general histology resembling CL of the late luteal tissue, described earlier. These findings parallel those reported in the rabbit CL at luteal regression (3). An active role for these cells in the structural changes of luteolysis has been suggested, based on histologic observations of macrophage phagocytosis of apoptopic luteal cells in the guinea pig. The presence of high numbers of neutrophilic granulocytes in the human follicle wall and in CL has not been described before. Several animal studies have described a preovulatory influx of neutrophils to the follicle (1, 2), and the presence of neutrophils in the rat CL has recently been demonstrated. In contrast to macrophages, which were increased in all layers of the preovulatory follicle, neutrophils were only increased in the thecal layer, where they were present in a similar density to the macrophages. This increase could be due to the presence of a chemotactic factor for neutrophilic granulocytes. Such a factor has been reported in human FF (6). Neutrophils are also multifunctional cells secreting a number of ovulatory mediators such as eicosanoids, PAF, and collagen-degrading enzymes (19). The types of collagen present in the human follicle wall are collagen types I and III (thecal externa, tunica albuginea) and type IV (basal membranes), and it has been demonstrated that human neutrophils contain and secrete specific collagenases active against these collagen types (20). Neutrophilic granulocytes were observed in fairly large numbers in the CL with a concentration in the theca lutein area. There were no indications of time-dependent changes in neutrophil density. Because neutrophilic granulocytes have a short survival time in the tissue (19), the continuous presence of these cells in the CL tissues suggests an active migration of these cells into CL throughout the life of the CL. T lymphocytes and NK cells were present in considerably lower numbers than macrophages and neutrophils, whereas B lymphocytes were not found in any specimen. The low numbers of T lymphocytes in ovarian tissue is in accord with recent findings in the rat (2) and rabbit (3). However, T lymphocytes are very rich sources of a variety of cytokines such as IL-2 and interferon-,)" which have been shown to affect steroidogenesis of human granulosa-lutein cells (21, 22), and a functional role of T lymphocytes cannot be ruled out based on the lower numbers. In the present study, a major portion of the T lymphocytes in the CL appeared to be of the CD8 cytotoxic/suppressor subtype. This is a similar distribution to that found in FF of women at IVF but different from the CD4:CD8 ratio of 2:1 found in circulating human blood (23). The predominance of suppressor/cytotoxic T lymphocytes in ovarian tissue indicates a selective recruitment of these cells to the ovary. In conclusion, neutrophilic granulocytes and macrophages are present in high numbers in the follicle wall at ovulation and in the CL, suggestive of functional roles in the ovulatory process and in regulation of CL function. Acknowledgments. We thank Mrs. Helen Holmes, Department of Obstetrics and Gynaecology, the University of Adelaide, Woodville, South Australia, for expert secretarial assistance and Rani Barua, M.D., Department of Pathology, The Queen Elizabeth Hospital, Woodville, South Australia, for assistance with histologic analysis of specimens. We acknowledge the gift of mab of Giardia intestinalis from Graham Mayrhofer, Ph.D., University of Adelaide. REFERENCES 1. Cavender JL, Murdoch WJ. Morphological studies of the microcirculatory system of periovulatory ovine follicles. Bioi Reprod 1988;39: Brannstrom M, Mayrhofer G, Robertson SA. Localization of leukocyte subsets in the rat ovary during the periovulatory period. Bioi Reprod 1993;48: Bagavandoss P, Wiggins Re, Kunkel SL, Remick DG, Keyes PL. Tumor necrosis factor production and accumulation of inflammatory cells in the corpus luteum of pseudopregnancy and pregnancy in rabbits. Bioi Reprod 1990;43: Brannstrom et a1. Leukocytes in the human follicle wall and CL Fertility and Sterility

8 4. Hellberg P, Thomsen P, Janson PO, Briinnstriim M. Leukocyte supplementation increases the luteinizing hormoneinduced ovulation rate in the in vitro-perfused. BioI Reprod 1991;44: Halme J, Hammond MG, Syrop CH, Talbert LM. Peritoneal macrophages modulate human granulosa-luteal cell progesterone production. J Clin Endocrinol Metab 1985;66: Herriot DM, Warnes GM, Kerin JF. Pregnancy-related chemotactic activity of human follicular fluid. Fertil Steril 1986;45: Loukides JA, Loy RA, Edwards R, Honig J, Visitin I, Polan ML. Human follicular fluids contain tissue macrophages. J Clin Endocrinol Metab 1990;71: Droesch K, Fulgham DL, Liu H-C, Rosenwaks Z, Alexander NJ. Distribution of T cell subsets in follicular fluid. Fertil Steril 1988;50: Pepperell JR, Wolcott C, Behrman HR. Effects of neutrophils in rat luteal cells. Endocrinology 1992;130: Wang LJ, Pascoe V, Petrucco OM, Norman RJ. Distribution of leukocyte subpopulations in the human corpus luteum. Hum Reprod 1992;7: Robertson SA, Briinnstriim M, Seamark RF. Cytokines in rodent reproduction and the cytokine-endocrine interaction. Curr Opin Immunol 1992;4: Lei ZM, Chegini N, Rao CHV. Quantitative cell composition of human and bovine corpora lutea from various reproductive states. BioI Reprod 1991;44: Katabuchi H, Fukumatsu Y, Okamura H. Immunohistochemical and morphological observations of macrophages in the human ovary. In: Hirshfield AN, editor. Growth factors and the ovary. New York: Plenum Press, 1989: Hurwitz A, Ricciarelli E, Bolero L, Rohan RM, Hernandez ER, Adashi EY. Endocrine- and autocrine-mediated regulation of rat ovarian (theca-interstitial) interleukin-l/'l gene expression: gonadotropin-dependent preovulatory acquisition. Endocrinology 1991;129: Roby KF, Terranova PF. Effects of tumor necrosis factor a in vitro on steroidogenesis of healthy and atretic follicles in the rat: theca as a target. Endocrinology 1990;126: Briinnstriim M, Wang L, Norman RJ. Effects of cytokines on prostaglandin production and steroidogenesis of incubated preovulatory follicles of the rat. BioI Reprod 1993;48: Briinnstriim M, Wang L, Norman RJ. Ovulatory effects of interleukin-l/'l on the perfused rat ovary. Endocrinology 1993;132: Briinnstriim M, Janson PO. The biochemistry of ovulation. In: Hillier SG, editor. Ovarian endocrinology. Oxford: Blackwell Scientific Publications, 1991: Parker CWo Neutrophil mechanisms. Am Rev Respir Dis 1991;143: Tscheshe H, Fedrowitz J, Michaelis J, Macartney HW. Matrix degrading proteinases from human granulocytes: type I, II, III collagenase, gelatinase and type IV collagenase. Folia Histochem Cytobiol 1986;61: Wang L, Robertson S, Seamark RF, Norman RJ. Lymphokines including interleukin-2, alter gonadotropin-stimulated progesterone production and proliferation of human granulosa lutein cells in vitro. J Clin Endocrinol Metab 1991;72: Wang HZ, Lu SH, Han XJ, Zhou W, Sheng WX, Sun ZD, et al. Inhibitory effect of interferon and tumor necrosis factor on human luteal function in vitro. Fertil Steril 1992;58: Hill JA, Barbieri RL, Anderson DJ. Detection of T8 (suppressor/cytotoxic) lymphocytes in human ovarian follicular fluid. Fertil SteriI1987;47: Vol. 61, No.3, March 1994 Brannstrom et al. Leukocytes in the human follicle wall and CL 495