- Email: [email protected]
Evidence for a neutrophil–interleukin-8 system in human folliculogenesis
Evidence for a neutrophil–interleukin-8 system in human folliculogenesis R. Jeffrey Chang, MD, Alain Gougeon, PhD, and Gregory F. Erickson, PhD San Diego, California, and Clamart, France OBJECTIVE: This study was conducted to determine whether polymorphonuclear leukocytes (neutrophils) and the potent chemoattractant interleukin-8 are associated with follicle development in the normal human ovary. STUDY DESIGN: We performed a morphometric analysis of neutrophils in 268 human ovarian follicles, of which 199 were preantral and 69 were antral. In each antral follicle the numbers of mitotic, apoptotic, and total granulosa cells were counted to determine healthy and atretic follicles. Interleukin-8 protein and messenger ribonucleic acid were detected by immunohistochemistry and in situ hybridization, respectively. RESULTS: Antral follicles contained relatively large numbers of neutrophils within the theca vasculature. The density of neutrophil was twofold greater (p < 0.05) in atretic versus healthy follicles. The neutrophil index (neutrophils/granulosa cells × 1000) was inversely correlated to the number of granulosa cells per follicle. Immunoreactive interleukin-8 was detected in the theca and granulosa cells of most all antral follicles examined. Interleukin-8 messenger ribonucleic acid was demonstrated in theca and granulosa cells of some but not all follicles examined. CONCLUSIONS: Neutrophils are present in the theca of developing antral follicles, increase in number during atresia, and are associated with expression of interleukin-8 in the follicle wall. (Am J Obstet Gynecol 1998;178:650-7.)
Key words: Folliculogenesis, neutrophil, interleukin-8
There is a growing body of evidence indicating that leukocytes and cytokines may be physiologically important in the mechanisms of ovulation and luteogenesis in female mammals, including the human.1-3 It is well known that large numbers of neutrophils are attracted to the theca compartment of human preovulatory follicles immediately before ovulation.4 During ovulation it also has been documented that follicular fluid concentrations of the cytokine interleukin-8 (IL-8), a potent chemoattractant and activator of neutrophils, are markedly increased.5, 6 In these studies IL-8 messenger ribonucleic acid (mRNA) was detected in granulosa cells of ovulatory follicles, which suggested that the follicular fluid IL-8 was generated, at least in part, by granulosa cells. Thus it has been proposed that parasecretion of IL-8 by granulosa cells may lead to the recruitment or activation of neu-
From the Department of Reproductive Medicine, University of California, San Diego, School of Medicine, and the Institut National de la Santé et de la Recherche Médicale. Supported in part in National Institutes of Health grant No. 1 RO1 HD32158-01. Presented at the Sixteenth Annual Meeting of The American Gynecological and Obstetrical Society, Victoria, British Columbia, Canada, September 4-6, 1997. Reprint requests: R. Jeffrey Chang, MD, Department of Reproductive Medicine, 0674, School of Medicine, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0674. Copyright © 1998 by Mosby, Inc. 0002-9378/98 $5.00 + 0 6/6/89006
650
trophils in the theca compartment of the dominant preovulatory follicle, where they play a role in mediating luteinizing hormone (LH)–human chorionic gonadotropin (hCG)–induced ovulation. Recently we made preliminary observations that neutrophils are present in the theca layers of developing human antral follicles. As yet, the presence and potential involvement of neutrophils and IL-8 in the process of follicle development and atresia has not been studied. The purpose of this research is to examine this issue. Material and methods Tissues. Two hundred sixty-eight follicles were analyzed morphometrically in 18 normal ovaries obtained from reproductive-aged women (26 to 47 years old, median age 44 years) during the follicular phase of the menstrual cycle. All were in good health with a history of regular ovulation and underwent total abdominal hysterectomy and bilateral salpingo-oophorectomy for benign gynecologic disorders. None of the women had recently used hormones or medications known to influence reproductive function. In all subjects there was no evidence of infection or inflammation. In each patient the complete blood cell count and differential were within normal limits. After surgical removal, the ovaries were fixed in 10% buffered formalin and embedded in paraffin for routine histologic studies. The research protocol stipulated the retrieval of archival material for ex-
Volume 178, Number 4 Am J Obstet Gynecol
amination and had been approved by the Committee on Investigation Involving Human Subjects at the University of California, San Diego. Follicle assessment. Whole ovary sections (10 µm) were prepared from each ovary and stained with hematoxylin and eosin. One section from each ovary underwent histologic examination with use of a Zeiss microscope. In each ovarian section quantification of cell numbers was achieved by direct visualization in every follicle. Actual cell counts were performed at a magnification of ×40 by at least two individuals on separate occasions. To determine within-observer variability, cell counts (granulosa cells and neutrophils) in all follicles contained within 12 ovaries were performed twice by each observer. To determine between-observer variability, cell counts in all follicles were compared among each of two to four observers assessing the same follicle. These measurements were separated in time by 4 days and each observer was blinded as to follicle status (healthy vs atretic). The repeated-measures correlation was 0.99 (p < 0.0001) and the interrater reliability was 0.99 (p < 0.0001). In each follicle the numbers of healthy, mitotic, and apoptotic granulosa cells were counted. Apoptotic cells were identified on the basis of classic histologic features.7 In addition to the striking characteristic feature of apoptotic bodies, associated cytoplasmic blebbing, nuclear condensation, and nuclear fragmentation were also noted. These observations were confirmed by identification of fragmented deoxyribonucleic acid (DNA) 3´ ends on random ovarian sections with use of the ApopTag kit (Oncor, Gaithersburg, Md.). Mitotic and apoptotic index values were calculated as previously reported.8, 9 Follicles with an apoptotic index of <2 apoptotic cells per 1000 granulosa cells (<0.2%) were scored as healthy and those with an apoptotic index >0.2% were scored as atretic. Comparative follicle health and atresia was also determined by the number of granulosa cells per follicle normalized to the follicle diameter. Neutrophils were identified histologically by their characteristic polymorphonuclear appearance. In each follicle the number of neutrophils in the perifollicular compartment to a depth of 0.5 mm from the basement membrane was counted. Calculation of the perifollicular area was achieved with the assistance of computerized image analysis. The outline of the basement membrane was traced to determine the follicle perimeter and the perifollicular perimeter. The difference between the two perimeters was used to determine the perifollicular area. In addition, neutrophil counts were determined in at least four areas of ovarian stroma that were comparable and separate from those areas containing antral follicles and served as background counts. Neutrophil accumulation was expressed as either neutrophil density (neutrophils per square millimeter) or neutrophil index (neutrophils per granulosa cell × 1000).
Chang, Gougeon, and Erickson 651
Fig. 1. Comparative measures of follicle viability. A, Average number of granulosa cells in healthy (n = 20) and atretic (n = 49) follicles. B, Apoptotic index values (AI) in granulosa cells of healthy and atretic follicles. Data represent mean ± SEM.
IL-8 immunohistochemistry. A highly specific polyclonal rabbit antibody kindly provided by Dr. Robert Hoch, University of California, San Diego, was used for immunohistochemical localization of IL-8.10 Paraffin sections (6 µm) were mounted on poly-L-lysine–coated slides and deparaffinized in xylene. After hydration in descending concentrations of ethanol, the sections were immersed for 30 minutes with 3% hydrogen peroxide in methanol, washed two times for 3 minutes in Trisbuffered saline solution, and digested for 10 minutes in Tris-buffered saline solution containing 0.2% Triton X100. Subsequently, the slides were blocked with 5% normal goat serum for 30 minutes. After the medium was removed, the sections were incubated overnight at 4° with primary antiserum diluted 1:4000 in Tris-buffered saline solution containing 1% normal goat serum and then washed two times for 3 minutes in Tris-buffered saline solution. Biotinylated goat antirabbit secondary antibody was applied to the sections for 1 hour. After washing in Tris-buffered saline solution, IL-8 antibody complexes were demonstrated with use of an avidin-biotin immunocomplex method (Vectastain ABC Kit, Vector Laboratories, Burlingame, Calif.). Reaction products were developed during a 5- to 10-minute incubation with use of Sigma Fast diaminobenzidine tablets dissolved in water. The color substrate was added to each section for approximately 5 to 10 minutes, washed in Tris-buffered saline solution, and counterstained with hematoxylin. Control sections, composed of acute salpingitic and oophoritic tissues, were incubated without primary antibody and were uniformly negative. IL-8 in situ hybridization. IL-8 mRNA was localized by in situ hybridization as previously described.11 Paraffin sections (10 µm) were cut on a rotary microtome and mounted on poly-L-lysine–coated glass slides, digested with proteinase K (30 minutes at 37° C), acetylated, and dehydrated. The antisense and sense strand RNA probes were prepared with use of the plasmid pHIL8BLSK(+)
652 Chang, Gougeon, and Erickson
April 1998 Am J Obstet Gynecol
Fig. 2. Polymorphonuclear leukocytes (neutrophils) in theca vasculature of atretic follicle. (Original magnification ×20.) GC, Granulosa cells; TC, theca cells. Note that neutrophils (arrows) are present in lumen or appear adherent to endothelial cells.
constructed as follows. The plasmid pHIL8GEX2T in vector pGEX-2T (Pharmacia Biotech, Piscataway, N.J.), which contains the human IL-8 coding region, was digested with BamHI and EcoRI (Boehringer-Mannheim, Indianapolis, Ind.) in the multicloning sites of the vector. It was gel-purified and ligated to BamHI/EcoRI-digested pBluescriptSK(+) (Stratagene, San Diego, Calif.), which contains T3 and T7 promoters, to give pHIL8BISK(+). The antisense strand RNA probe was prepared with use of T7-RNA polymerase and sulfur 35–labeled uridine triphosphate after digestion of pHIL8BISK(+) with BamHI. The sense strand was prepared with use of T3RNA polymerase after digestion with EcoRI. The specific activity of these probes was 2 × 109 counts/min/µg, respectively. Hybridization with the 35S-labeled RNA probe (107 counts/min/ml) was performed at 60° C overnight in a solution containing 50% formamide, 0.3 mol/L sodium chloride, 10 mmol/L Tris (pH 8.2), 1 mmol/L ethylenediaminetetraacetic acid (EDTA), 0.05% yeast transfer RNA, 10 mmol/L dithiothreitol, 1× Denhardt’s solution, and 10% dextran sulfate. After hybridization, sections were treated with ribonuclease A (20 µg/ml, 37° C, 30 minutes) and washed in 15 mmol/L sodium chloride and 1.5 mmol/L sodium citrate at 70° to 75° C for 30 minutes. Dehydrated slides were exposed to x-ray film for several days. After adequate x-ray film images had been obtained, the ovarian sections were defatted in xylene, rinsed in absolute ethanol, air dried, and coated with Kodak NTB-2 liquid autoradiograph emulsion (Eastman Kodak, Rochester, N.Y.). Slides were exposed for 3.5 weeks at 4° C in a desiccated dark box. After exposure the slides were developed (Kodak D19, 3.5 minutes, 14° C), rinsed briefly in distilled water, and fixed (14° C in film-strength Kodak rapid fixer for 6 minutes). After
washing in distilled water for 1 hour, slides were counterstained with hematoxylin. Eosin counterstain was applied to the antisense strand to enhance any expression of IL-8. Statistical method. Data were analyzed with use of the STATVIEW program (Berkeley, Calif.). Nongaussian distributed data were log transformed before analysis. Subsequent normality of distribution was determined by the Kolmogorov-Smirnov goodness of fit. Single comparisons between healthy and atretic antral follicles were analyzed by the unpaired Student t test. Relationships between variables were determined by linear regression analysis with Pearson product-moment correlation or Spearman rank correlation. Data are expressed as mean ± SE. Results Antral follicles. A total of 268 follicles were identified in the 18 ovaries examined. Of these follicles, 69 were at the antral or graafian stage and 199 were preantral. According to the criteria set forth by Gougeon,8, 9 a total of 20 (29%) antral follicles were healthy (apoptotic index <0.2%) and 49 (71%) were atretic (apoptotic index >0.2%). Follicle histologic features. The healthy follicles had multiple intact layers of granulosa cells, which lined the entire circumference of the basal lamina. The granulosa in healthy follicles was usually 4 to 8 cell layers thick. The number of granulosa cells in these follicles was 1947 ± 181. Mitotic figures were seen in the granulosa cells of all healthy follicles and the average mitotic index was calculated to be 8.58 ± 0.99. Correspondingly, apoptotic granulosa cells were very rare (mean apoptotic index 0.60 ± 0.16).
Volume 178, Number 4 Am J Obstet Gynecol
Fig. 3. Neutrophil accumulation in perifollicular environment. A, Neutrophil density (ND, neutrophils per square millimeter) in healthy (n = 20) and atretic (n = 49) follicles. B, Neutrophil index values (NI, neutrophils per 1000 granulosa cells) in healthy and atretic follicles. Data represent mean ± SEM.
The atretic follicles displayed a variety of degenerative changes, particularly in the granulosa. Individual cells were condensed and stained darkly for hematoxylin. As such, the granulosa appeared as compact layers of cells, which were usually one to three layers thick. In some atretic follicles sheets of granulosa cells had dislodged and were floating free in the antral cavity. Not infrequently, this left portions of the inner surface of the basal lamina exposed with no attached granulosa cells. Loss of the entire granulosa cell layer indicated terminal atresia. The number of granulosa cells per atretic follicle was 581 ± 83. Compared with healthy follicles (Fig. 1), the number of granulosa cells per atretic follicle was markedly decreased (p < 0.0001). A striking characteristic of the granulosa within atretic follicles was the presence of apoptotic bodies, which demonstrated that cell death or apoptosis was occurring. In addition, these granulosa cells also exhibited other markers of apoptosis, including cytoplasmic blebbing, nuclear condensation, and nuclear fragmentation. The apoptotic index for atretic follicles was calculated to be 2.57 ± 0.67; this was 42-fold greater (p < 0.01) than that of healthy follicles (Fig. 1). In a few granulosa cells mitotic figures were noted, but the mitotic index (3.09 ± 0.63) of the atretic group was only 36% of that for healthy follicles (p < 0.05). The group of preantral follicles consisted of 80 primordial, 66 primary, and 53 secondary follicles. Evidence of apoptosis was not observed in any preantral follicles. Neutrophils. Surprisingly large numbers of polymorphonuclear leukocytes (neutrophils) were consistently present in all antral follicles, healthy and atretic (Fig. 2). The neutrophils were localized predominantly within the theca vasculature, appearing either free in the lumen or adherent to the endothelial cells of small veins or postcapillary venules. Of interest was the conspicuous lack of neutrophils in the extravascular spaces surrounding these blood vessels. In some highly atretic follicles, a few
Chang, Gougeon, and Erickson 653
Fig. 4. Correlation between neutrophil index (NI) and number of granulosa cells in each follicle (n = 69).
neutrophils were observed in the antral cavity, but this was rare. Neutrophils were not detected in the pool of nongrowing and growing preantral follicles. It is noteworthy that in comparable areas of ovarian stroma separate from those occupied by antral follicles the number of neutrophils was sparse. To further analyze the relationship between follicle development and neutrophil accumulation, we compared the neutrophil density in healthy and atretic follicles. In healthy follicles the neutrophil density was 4.1 ± 0.8 neutrophils/mm2, which was clearly distinct from that of the background stroma, 1.0 ± 0.05 neutrophils/mm2. No differences in neutrophil density were detected among healthy follicles that appeared to be either early or more advanced in development. By comparison, the neutrophil density of atretic follicles was 8.8 ± 1.4 neutrophils/mm2 or approximately twofold higher (p < 0.05) than that of healthy follicles (Fig. 3). The neutrophil data were also analyzed with respect to the number of granulosa cells per follicle (neutrophil index). The neutrophil index of healthy and atretic follicles was 6.3 ± 2.0 and 25.1 ± 6.1, respectively (Fig. 3). This fourfold increase in the neutrophil index in the atretic group represented a statistically significant difference (p < 0.05). Neither the neutrophil density nor the neutrophil index of each follicle was correlated to the respective apoptotic index values. However, linear regression analysis revealed an inverse correlation between the neutrophil/granulosa cell ratio and the number of granulosa cells per follicle (r = 0.626), which was highly significant (p < 0.0001) (Fig. 4). IL-8 immunohistochemistry and in situ hybridization. Immunohistochemistry for IL-8 protein and in situ hybridization for IL-8 mRNA were performed on tissue sections from four representative ovaries. After incubation with IL-8 antisera, immunoreactive cells were observed in the theca interna and some granulosa cells of all these follicles. The intensity of the immunostained cells varied from very strong to relatively weak both between and within a given follicle (Fig. 5). Some neutrophils and en-
654 Chang, Gougeon, and Erickson
A
April 1998 Am J Obstet Gynecol
B
Fig. 5. Localization of IL-8 in an antral follicle within human ovary by immunohistochemistry. A, IL-8 protein is localized to follicle wall of atretic follicle. (Original magnification ×5). B, IL-8 immunoreactive cells in theca and granulosa cell layers of atretic follicle. (Original magnification ×20). GC, Granulosa cells; TC, theca cells. Note accumulation of neutrophils (arrows) within adjacent blood vessel of theca compartment.
dothelial cells in the theca blood vessels contained IL-8 immunoreactivity. In addition, immunoreactive IL-8 was detected in the epithelium and tunica albuginea of the ovarian capsule. No immunoreactivity was seen in preantral follicles or ovarian stromal tissue. Control tissues were comprised of fallopian tube and ovary that had been surgically removed for salpingitis and oophoritis. In the absence of primary antibody, IL-8 protein was not detected. Our in situ hybridization studies revealed that IL-8 mRNA message was present in the theca compartment of both healthy and atretic follicles (Fig. 6). For the most part, the IL-8 hybridization signal was weak. Importantly, however, a strong signal was seen in some endothelial cells and neutrophils of the theca. A weak signal for IL-8 was detected in a few granulosa cells. Significant IL-8 message was found in the ovarian epithelium. Similar to that for immunohistochemistry, acute salpingitic and oophoritic tissues were used as controls and IL-8 message was clearly expressed. Compared with the detection of IL-8 message, the sense strand mRNA failed to express any signal. Comment The results of our studies have demonstrated that normal folliculogenesis in the human ovary is associated with the accumulation of neutrophils within the vascula-
ture of the theca compartment of antral follicles, both healthy and atretic. The infiltration of neutrophils was significantly greater in those follicles undergoing atresia compared with healthy follicles. In contrast, preantral follicles were found to be totally unassociated with the presence of surrounding neutrophils. The appearance of these leukocytes in theca blood vessels coincided with the detection of the potent chemoattractant IL-8 in the follicle wall. These findings suggest that (1) the initial acquisition of neutrophils appears to be related to the graafian follicle, (2) subsequent increased accumulation occurred with follicle atresia, and (3) neutrophil attraction in developing follicles is mediated by IL-8. In the ovary the presence of perifollicular neutrophils generally has been recognized as a characteristic feature of the ovulatory process with particular involvement related to follicle rupture.1-3 We have now extended this concept to the population of developing cohort antral follicles. Our finding of neutrophils surrounding small and presumably early antral follicles and not preantral follicles suggests that neutrophils may be attracted to the growing follicle during antrum formation. The acquisition of neutrophils appears to persist throughout the growth, development, and atresia of graafian follicles. Because neutrophil accumulation was apparent in essentially all antral follicles, we suggest that significant num-
Chang, Gougeon, and Erickson 655
Volume 178, Number 4 Am J Obstet Gynecol
A
B
Fig. 6. Localization of IL-8 within follicle wall of antral follicle in human ovary by in situ hybridization. A, IL-8 mRNA (asterisks) is detected by in situ hybridization in some endothelial cells and neutrophils of theca in healthy follicle. (Original magnification ×20.) Eosin counterstain was applied to enhance expression and localization of antisense strand RNA. GC, Granulosa cells; TC, theca cells. B, Corresponding control section (sense strand).
bers of neutrophils in the theca compartment is a distinctive feature of the human graafian follicle. Of particular note was the significantly greater number of neutrophils surrounding atretic follicles compared with healthy follicles. This observation suggests a potential role for these immune cells in the atretic process. It is widely accepted that the principal mechanism for follicular atresia and granulosa cell death is apoptosis.12, 13 In the classic description of apoptosis, it was emphasized that inflammation and inflammatory cells were not involved in this process.7 In addition, evidence to suggest that immunocompetent cells, in particular neutrophils, are capable of inducing apoptosis is lacking. Our data showing no significant relationship between neutrophil density or neutrophil index to the apoptotic index fits with this concept. Nevertheless, the twofold increase in neutrophil number surrounding atretic follicles together with the significant inverse correlation of neutrophil index to follicle viability are noteworthy in that they could suggest a possible link between the aggregation of neutrophils and the process of follicle atresia. The mechanism of leukocyte recruitment to developing antral follicles is unknown, although it has been pro-
posed that neutrophils may be drawn to the follicle by chemotactic mechanisms induced by proinflammatory cytokines or chemokines, the most powerful of which is IL-8.14, 15 The results of our studies have demonstrated the strong expression of immunoreactive IL-8 protein in both theca and granulosa cells of healthy and atretic antral follicles. In addition, our detection of IL-8 mRNA in some of these cells was consistent with the marked immunoreactivity within the follicle wall. In contrast, evidence for the presence of IL-8 in granulosa cells of preantral follicles was completely lacking. Taken together, our data suggest that both theca and granulosa cells of developing antral follicles and not preantral follicles are capable of IL-8 production. These findings support and enlarge on the results of previous studies in which the well-documented invasion of neutrophils into the perifollicular environment just before ovulation appeared to involve IL-8.5, 6 That IL-8 protein and mRNA were expressed in associated neutrophils and endothelial cells of theca blood vessels suggests that local intravascular autocrine or paracrine mechanisms also may contribute to the acquisition of neutrophils. Previous studies have demonstrated
656 Chang, Gougeon, and Erickson
that these cells, when stimulated, are capable of IL-8 production.15 Moreover, both neutrophils and endothelial cells appear to express IL-8 receptor. Thus neutrophils may be able to perpetuate their own recruitment either directly or indirectly by stimulating endothelial cells. Additionally, IL-8 may mediate neutrophil-endothelial cell interaction by inducing binding of neutrophil β2-integrin molecules to endothelial cell intercellular adhesion molecules, resulting in firm attachment between the two cells.16, 17 This would account, at least in part, for the dramatic appearance of neutrophil-endothelial cell adhesion in the theca vasculature of developing follicles. Although neutrophils are thought to participate in the process of follicle rupture at the time of ovulation, the roles for these immune cells in association with antral follicle development or atresia have yet to be addressed. The apparent confinement of neutrophils to theca blood vessels raises uncertainty regarding the traditional proinflammatory role for these immune cells as implied during follicle rupture.1 Whether cytokine production by neutrophils within theca blood vessels has an impact on theca and granulosa cell proliferation or differentiation in the human follicle remains to be ascertained. In summary, we have provided evidence that neutrophils and endogenous IL-8 are expressed in the theca vasculature during folliculogenesis in normal ovulatory women, being particularly abundant in cohort follicles undergoing atresia. The expression of IL-8 and the influx of neutrophils led us to propose that an intrinsic IL8/neutrophil system could function in the mechanism of follicle atresia. The challenges are to understand the regulation of the expression of this novel cytokine system in the follicle wall and how the actions and interactions of IL-8 and neutrophils are coupled to ovarian physiologic features. We thank Holly Kim, Drs. Dan-Mei Li and Toshihide Kubo, Kimberly Best, Henry Ferreyra, and Nazli Ghafouri for their technical assistance and Dr. Rebbeca Baergen at the University of California, San Diego, and the physician staff at Kaiser Hospital, San Diego, for providing tissue samples. REFERENCES
1. Espey LL. Ovulation as an inflammatory reaction—a hypothesis. Biol Reprod 1980;22:73-106. 2. Adashi EY. The potential relevance of cytokines to ovarian physiology: the emerging role of resident ovarian cells of the white blood cell series. Endocrine Rev 1990;11:454-64. 3. Brannstrom M, Norman RJ. Involvement of leukocytes and cytokines in the ovulatory process and corpus luteum formation. Hum Reprod 1993;8:1762-75. 4. Brannstrom M, Pascoe V, Norman RJ, McClure N. Localization of leukocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle. Fertil Steril 1994;61: 488-95. 5. Runesson E, Bostrum E-K, Janson PO, Brannstrom M. The human preovulatory follicle is a source of the chemotactic cytokine interleukin-8. Mol Hum Reprod 1996;4:245-50.
April 1998 Am J Obstet Gynecol
6. Arici A, Oral E, Bukulmez O, Buradagunta S, Engin O, Olive D. Interleukin-8 expression and modulation in human preovulatory follicles and ovarian cells. Endocrinology 1996;137:3762-9. 7. Wylie AH. Cell death: the significance of apoptosis. Int Rev Cytol 1980;68:251-306. 8. Gougeon A. Qualitative changes in medium and large antral follicles in the human ovary during the menstrual cycle. Ann Biol Biochem Biophys 1979;19:1461-8. 9. Gougeon A. Regulation of ovarian follicular development in primates: facts and hypotheses. Endocrine Rev 1996;17:121-55. 10. Schraufstatter I, Barritt DS, Zenaida MM, Oades G, Cochrane CG. Multiple sites on IL-8 responsible for binding to α and β IL8 receptors. J Immunol 1993;151:6418-28. 11. Simmons DM, Arriza JL, Swanson LW. A complete protocol for in situ hybridization of messenger RNAs in brain and other tissues with radiolabeled single strand RNA probes. J Histotechnol 1989;12:169-81. 12. Hsueh AJW, Billig H, Tsafriri A. Ovarian follicle atresia: a hormonally controlled apoptotic process. Endocrine Rev 1994;15:707-24. 13. Tilly JL, Kowalski KI, Johnson AL, Hsueh AJW. Involvement of apoptosis in ovarian follicular atresia and postovulatory regression. Endocrinology 1991;129:2799-801. 14. Norman RJ, Brannstrom M. Cytokines in the ovary: pathophysiology and potential for pharmacological intervention. Pharmacol Ther 1996;69:219-36. 15. Baggiolini M, Dewald B, Moser B. Interleukin-8 and related chemotactic cytokines—CXC and CC chemokines. Adv Immunol 1994;55:97-179. 16. Crockett-Torabi E, Fantone JC. The selectins: insights into selectin-induced intracellular signaling in leukocytes. Immunol Res 1995;14:237-51. 17. Rot A. Endothelial cell binding of NAP-1/IL-8: role in neutrophil emigration. Immunol Today 1992;13:291-4.
Editors’ note: This manuscript was revised after these discussions were presented. Discussion DR. LINDA C. GIUDICE, Stanford, California. The role of leukocytes and cytokines in the process of ovulation and corpus luteum formation is well known. The appearance of leukocytes in the theca before ovulation is also well known, as is the fact that a cytokine, IL-8, which is a potent chemoattractant for neutrophils, is markedly elevated in follicular fluid at the time of ovulation. The current study takes this cytokine leukocyte system several steps before the process of ovulation per se and examines follicular development. Dr. Chang and colleagues have examined 268 follicles from 18 normal ovaries obtained from reproductive-aged women 26 to 47 years old during the follicular phase of the cycle. A total of 199 of these follicles were preantral, 69 were antral, with 20 being healthy and 49 atretic. Methods used were morphometric analysis on every follicle by two individuals on separate occasions. Apoptotic cells were identified by histologic features and confirmed by APOPTAG kit. Mitotic and apoptotic index values were calculated, as was neutrophil density (numbers of neutrophils per square millimeter)and neutrophil index (numbers of neutrophils per granulosa). Immunohistochemistry was performed for IL-8 immunoreactive protein and in situ hybridization for IL-8 mRNA expression. The authors found no neutrophils around any of the preantral follicles and more neutrophils in atretic versus
Volume 178, Number 4 Am J Obstet Gynecol
healthy follicles. They also have found variable expression of IL-8 mRNA and protein in the follicle, with more in the theca than in the granulosa. The conclusion of the authors is that during follicular development IL-8 is expressed by the granulosa and more so by the theca that attracts neutrophils to the follicle. What happens thereafter? Why do the neutrophils go to the follicles that are destined for atresia more so than those that are destined for continued development? Or, are they attracted to atretic follicles after selection has occurred? What about the activation of these neutrophils? Is there any evidence that there is activation in atretic versus healthy follicles? What is the evidence that supports neutrophils are involved in atresia during follicular development? What is it that the neutrophil secretes that may contribute to the process of atresia and what are the mechanisms? These are the key issues and the pathway that the authors will likely follow in the near future. A key question is: Are neutrophils and their products activators of processes associated with apoptosis? Are the membrane changes that occur before the more classical DNA fragmentation regulated by cytokines from activated neutrophils? Is there any evidence for endonuclease regulation by cytokines from activated neutrophils? And if these are participants in the process of atresia, then how do the neutrophils know that they are needed for the process of ovulation of a mature follicle? These are all very important questions and the finding of neutrophils in human ovarian follicles in the thecal vasculature is likely an important one. Beyond the conceptual basis of this article are some comments on experimental approach and design. The numbers of follicles examined and the data that are reported for IL-8 mRNA and protein detection are not listed as representative of x number. Exactly how many follicles were looked at for IL-8 mRNA and protein? In addition, controls for specificity of the antibody and control for the RNA probe are important to show. Negative controls are also critical and should also be included in this study. Finally, although the neutrophil density does not appear to be related to the size of the follicle, healthy or otherwise, what about IL-8 expression? In conclusion, this study approaches an area of research that has had relatively little investigation, and that is with regard to cytokines and follicular development. The bulk of the investigations that have occurred to date focus on the role of leukocytes and their products, cytokines, in the process of ovulation, including activation of matrix metalloproteinases, angiogenesis, and steroido-
Chang, Gougeon, and Erickson 657
genesis as well as corpus luteum formation and maintenance. This new and exciting area is one that is certainly worthy of further investigation. DR. DAVID L. KEEFE, Providence, Rhode Island. If ovulation involves a process of inflammation, what keeps the oocyte from getting in the line of fire? Specifically, was there any evidence of neutrophil or IL-8 expression? DR. CHANG. Actually, the process of ovulation has been akin to inflammation. It really is not inflammation in the strict sense of the word because of the process of apoptosis occurring. This is a process that is classically not related to inflammation. Why is the oocyte not affected? It’s unclear. We did not look intensely. In fact, we hardly saw any of the follicles that were in the process of ovulation when this very closely coordinated and rapidly acting sequence of events occurred in the cycle. So our studies really did not involve those particular structures, if that’s the question. DR. KEEFE. Were there any neutrophils or IL-8 because of that morphologic observation? DR. CHANG. We didn’t see any around the oocyte specifically. These were two-dimensional sections, and as a result most of the sections did not contain the oocyte; of those that did, we didn’t see that around that specific area, although there may have been a few in which the more peripheral granulosa cell layers did contain it above the protein and the mRNA. DR. JOHN E. BUSTER, Houston, Texas. I have two questions. First of all, have you looked at this model with respect to polycystic ovarian disease and if you have, can you comment on it? Second, have you looked at IL-8 expression and neutrophil migration in the corpus luteum as a lysis? PRESIDENT SCOTT. With manipulation of the intraovarian cytokine, do you think this lends itself to a new method of contraception eventually? DR. CHANG (Closing). It is possible that there could be developed antagonist of these cytokines, which might have therapeutic implications, such as contraception. Although this area has not been explored, the use of cytokine antagonists have been thought to be useful in the treatment of some inflammatory disorders such as rheumatoid arthritis. It appears to provide relief from the symptoms of pain. I should have known that someone would ask about polycystic ovary, given my interest in that particular area. The neutrophils are increased in the follicles of the polycystic ovary. Why did we do the neutrophil index? It’s misleading. The granulosa cells and atresia go down, and the neutrophils may not go up; but the ratio will.
Key words: Folliculogenesis, neutrophil, interleukin-8
There is a growing body of evidence indicating that leukocytes and cytokines may be physiologically important in the mechanisms of ovulation and luteogenesis in female mammals, including the human.1-3 It is well known that large numbers of neutrophils are attracted to the theca compartment of human preovulatory follicles immediately before ovulation.4 During ovulation it also has been documented that follicular fluid concentrations of the cytokine interleukin-8 (IL-8), a potent chemoattractant and activator of neutrophils, are markedly increased.5, 6 In these studies IL-8 messenger ribonucleic acid (mRNA) was detected in granulosa cells of ovulatory follicles, which suggested that the follicular fluid IL-8 was generated, at least in part, by granulosa cells. Thus it has been proposed that parasecretion of IL-8 by granulosa cells may lead to the recruitment or activation of neu-
From the Department of Reproductive Medicine, University of California, San Diego, School of Medicine, and the Institut National de la Santé et de la Recherche Médicale. Supported in part in National Institutes of Health grant No. 1 RO1 HD32158-01. Presented at the Sixteenth Annual Meeting of The American Gynecological and Obstetrical Society, Victoria, British Columbia, Canada, September 4-6, 1997. Reprint requests: R. Jeffrey Chang, MD, Department of Reproductive Medicine, 0674, School of Medicine, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0674. Copyright © 1998 by Mosby, Inc. 0002-9378/98 $5.00 + 0 6/6/89006
650
trophils in the theca compartment of the dominant preovulatory follicle, where they play a role in mediating luteinizing hormone (LH)–human chorionic gonadotropin (hCG)–induced ovulation. Recently we made preliminary observations that neutrophils are present in the theca layers of developing human antral follicles. As yet, the presence and potential involvement of neutrophils and IL-8 in the process of follicle development and atresia has not been studied. The purpose of this research is to examine this issue. Material and methods Tissues. Two hundred sixty-eight follicles were analyzed morphometrically in 18 normal ovaries obtained from reproductive-aged women (26 to 47 years old, median age 44 years) during the follicular phase of the menstrual cycle. All were in good health with a history of regular ovulation and underwent total abdominal hysterectomy and bilateral salpingo-oophorectomy for benign gynecologic disorders. None of the women had recently used hormones or medications known to influence reproductive function. In all subjects there was no evidence of infection or inflammation. In each patient the complete blood cell count and differential were within normal limits. After surgical removal, the ovaries were fixed in 10% buffered formalin and embedded in paraffin for routine histologic studies. The research protocol stipulated the retrieval of archival material for ex-
Volume 178, Number 4 Am J Obstet Gynecol
amination and had been approved by the Committee on Investigation Involving Human Subjects at the University of California, San Diego. Follicle assessment. Whole ovary sections (10 µm) were prepared from each ovary and stained with hematoxylin and eosin. One section from each ovary underwent histologic examination with use of a Zeiss microscope. In each ovarian section quantification of cell numbers was achieved by direct visualization in every follicle. Actual cell counts were performed at a magnification of ×40 by at least two individuals on separate occasions. To determine within-observer variability, cell counts (granulosa cells and neutrophils) in all follicles contained within 12 ovaries were performed twice by each observer. To determine between-observer variability, cell counts in all follicles were compared among each of two to four observers assessing the same follicle. These measurements were separated in time by 4 days and each observer was blinded as to follicle status (healthy vs atretic). The repeated-measures correlation was 0.99 (p < 0.0001) and the interrater reliability was 0.99 (p < 0.0001). In each follicle the numbers of healthy, mitotic, and apoptotic granulosa cells were counted. Apoptotic cells were identified on the basis of classic histologic features.7 In addition to the striking characteristic feature of apoptotic bodies, associated cytoplasmic blebbing, nuclear condensation, and nuclear fragmentation were also noted. These observations were confirmed by identification of fragmented deoxyribonucleic acid (DNA) 3´ ends on random ovarian sections with use of the ApopTag kit (Oncor, Gaithersburg, Md.). Mitotic and apoptotic index values were calculated as previously reported.8, 9 Follicles with an apoptotic index of <2 apoptotic cells per 1000 granulosa cells (<0.2%) were scored as healthy and those with an apoptotic index >0.2% were scored as atretic. Comparative follicle health and atresia was also determined by the number of granulosa cells per follicle normalized to the follicle diameter. Neutrophils were identified histologically by their characteristic polymorphonuclear appearance. In each follicle the number of neutrophils in the perifollicular compartment to a depth of 0.5 mm from the basement membrane was counted. Calculation of the perifollicular area was achieved with the assistance of computerized image analysis. The outline of the basement membrane was traced to determine the follicle perimeter and the perifollicular perimeter. The difference between the two perimeters was used to determine the perifollicular area. In addition, neutrophil counts were determined in at least four areas of ovarian stroma that were comparable and separate from those areas containing antral follicles and served as background counts. Neutrophil accumulation was expressed as either neutrophil density (neutrophils per square millimeter) or neutrophil index (neutrophils per granulosa cell × 1000).
Chang, Gougeon, and Erickson 651
Fig. 1. Comparative measures of follicle viability. A, Average number of granulosa cells in healthy (n = 20) and atretic (n = 49) follicles. B, Apoptotic index values (AI) in granulosa cells of healthy and atretic follicles. Data represent mean ± SEM.
IL-8 immunohistochemistry. A highly specific polyclonal rabbit antibody kindly provided by Dr. Robert Hoch, University of California, San Diego, was used for immunohistochemical localization of IL-8.10 Paraffin sections (6 µm) were mounted on poly-L-lysine–coated slides and deparaffinized in xylene. After hydration in descending concentrations of ethanol, the sections were immersed for 30 minutes with 3% hydrogen peroxide in methanol, washed two times for 3 minutes in Trisbuffered saline solution, and digested for 10 minutes in Tris-buffered saline solution containing 0.2% Triton X100. Subsequently, the slides were blocked with 5% normal goat serum for 30 minutes. After the medium was removed, the sections were incubated overnight at 4° with primary antiserum diluted 1:4000 in Tris-buffered saline solution containing 1% normal goat serum and then washed two times for 3 minutes in Tris-buffered saline solution. Biotinylated goat antirabbit secondary antibody was applied to the sections for 1 hour. After washing in Tris-buffered saline solution, IL-8 antibody complexes were demonstrated with use of an avidin-biotin immunocomplex method (Vectastain ABC Kit, Vector Laboratories, Burlingame, Calif.). Reaction products were developed during a 5- to 10-minute incubation with use of Sigma Fast diaminobenzidine tablets dissolved in water. The color substrate was added to each section for approximately 5 to 10 minutes, washed in Tris-buffered saline solution, and counterstained with hematoxylin. Control sections, composed of acute salpingitic and oophoritic tissues, were incubated without primary antibody and were uniformly negative. IL-8 in situ hybridization. IL-8 mRNA was localized by in situ hybridization as previously described.11 Paraffin sections (10 µm) were cut on a rotary microtome and mounted on poly-L-lysine–coated glass slides, digested with proteinase K (30 minutes at 37° C), acetylated, and dehydrated. The antisense and sense strand RNA probes were prepared with use of the plasmid pHIL8BLSK(+)
652 Chang, Gougeon, and Erickson
April 1998 Am J Obstet Gynecol
Fig. 2. Polymorphonuclear leukocytes (neutrophils) in theca vasculature of atretic follicle. (Original magnification ×20.) GC, Granulosa cells; TC, theca cells. Note that neutrophils (arrows) are present in lumen or appear adherent to endothelial cells.
constructed as follows. The plasmid pHIL8GEX2T in vector pGEX-2T (Pharmacia Biotech, Piscataway, N.J.), which contains the human IL-8 coding region, was digested with BamHI and EcoRI (Boehringer-Mannheim, Indianapolis, Ind.) in the multicloning sites of the vector. It was gel-purified and ligated to BamHI/EcoRI-digested pBluescriptSK(+) (Stratagene, San Diego, Calif.), which contains T3 and T7 promoters, to give pHIL8BISK(+). The antisense strand RNA probe was prepared with use of T7-RNA polymerase and sulfur 35–labeled uridine triphosphate after digestion of pHIL8BISK(+) with BamHI. The sense strand was prepared with use of T3RNA polymerase after digestion with EcoRI. The specific activity of these probes was 2 × 109 counts/min/µg, respectively. Hybridization with the 35S-labeled RNA probe (107 counts/min/ml) was performed at 60° C overnight in a solution containing 50% formamide, 0.3 mol/L sodium chloride, 10 mmol/L Tris (pH 8.2), 1 mmol/L ethylenediaminetetraacetic acid (EDTA), 0.05% yeast transfer RNA, 10 mmol/L dithiothreitol, 1× Denhardt’s solution, and 10% dextran sulfate. After hybridization, sections were treated with ribonuclease A (20 µg/ml, 37° C, 30 minutes) and washed in 15 mmol/L sodium chloride and 1.5 mmol/L sodium citrate at 70° to 75° C for 30 minutes. Dehydrated slides were exposed to x-ray film for several days. After adequate x-ray film images had been obtained, the ovarian sections were defatted in xylene, rinsed in absolute ethanol, air dried, and coated with Kodak NTB-2 liquid autoradiograph emulsion (Eastman Kodak, Rochester, N.Y.). Slides were exposed for 3.5 weeks at 4° C in a desiccated dark box. After exposure the slides were developed (Kodak D19, 3.5 minutes, 14° C), rinsed briefly in distilled water, and fixed (14° C in film-strength Kodak rapid fixer for 6 minutes). After
washing in distilled water for 1 hour, slides were counterstained with hematoxylin. Eosin counterstain was applied to the antisense strand to enhance any expression of IL-8. Statistical method. Data were analyzed with use of the STATVIEW program (Berkeley, Calif.). Nongaussian distributed data were log transformed before analysis. Subsequent normality of distribution was determined by the Kolmogorov-Smirnov goodness of fit. Single comparisons between healthy and atretic antral follicles were analyzed by the unpaired Student t test. Relationships between variables were determined by linear regression analysis with Pearson product-moment correlation or Spearman rank correlation. Data are expressed as mean ± SE. Results Antral follicles. A total of 268 follicles were identified in the 18 ovaries examined. Of these follicles, 69 were at the antral or graafian stage and 199 were preantral. According to the criteria set forth by Gougeon,8, 9 a total of 20 (29%) antral follicles were healthy (apoptotic index <0.2%) and 49 (71%) were atretic (apoptotic index >0.2%). Follicle histologic features. The healthy follicles had multiple intact layers of granulosa cells, which lined the entire circumference of the basal lamina. The granulosa in healthy follicles was usually 4 to 8 cell layers thick. The number of granulosa cells in these follicles was 1947 ± 181. Mitotic figures were seen in the granulosa cells of all healthy follicles and the average mitotic index was calculated to be 8.58 ± 0.99. Correspondingly, apoptotic granulosa cells were very rare (mean apoptotic index 0.60 ± 0.16).
Volume 178, Number 4 Am J Obstet Gynecol
Fig. 3. Neutrophil accumulation in perifollicular environment. A, Neutrophil density (ND, neutrophils per square millimeter) in healthy (n = 20) and atretic (n = 49) follicles. B, Neutrophil index values (NI, neutrophils per 1000 granulosa cells) in healthy and atretic follicles. Data represent mean ± SEM.
The atretic follicles displayed a variety of degenerative changes, particularly in the granulosa. Individual cells were condensed and stained darkly for hematoxylin. As such, the granulosa appeared as compact layers of cells, which were usually one to three layers thick. In some atretic follicles sheets of granulosa cells had dislodged and were floating free in the antral cavity. Not infrequently, this left portions of the inner surface of the basal lamina exposed with no attached granulosa cells. Loss of the entire granulosa cell layer indicated terminal atresia. The number of granulosa cells per atretic follicle was 581 ± 83. Compared with healthy follicles (Fig. 1), the number of granulosa cells per atretic follicle was markedly decreased (p < 0.0001). A striking characteristic of the granulosa within atretic follicles was the presence of apoptotic bodies, which demonstrated that cell death or apoptosis was occurring. In addition, these granulosa cells also exhibited other markers of apoptosis, including cytoplasmic blebbing, nuclear condensation, and nuclear fragmentation. The apoptotic index for atretic follicles was calculated to be 2.57 ± 0.67; this was 42-fold greater (p < 0.01) than that of healthy follicles (Fig. 1). In a few granulosa cells mitotic figures were noted, but the mitotic index (3.09 ± 0.63) of the atretic group was only 36% of that for healthy follicles (p < 0.05). The group of preantral follicles consisted of 80 primordial, 66 primary, and 53 secondary follicles. Evidence of apoptosis was not observed in any preantral follicles. Neutrophils. Surprisingly large numbers of polymorphonuclear leukocytes (neutrophils) were consistently present in all antral follicles, healthy and atretic (Fig. 2). The neutrophils were localized predominantly within the theca vasculature, appearing either free in the lumen or adherent to the endothelial cells of small veins or postcapillary venules. Of interest was the conspicuous lack of neutrophils in the extravascular spaces surrounding these blood vessels. In some highly atretic follicles, a few
Chang, Gougeon, and Erickson 653
Fig. 4. Correlation between neutrophil index (NI) and number of granulosa cells in each follicle (n = 69).
neutrophils were observed in the antral cavity, but this was rare. Neutrophils were not detected in the pool of nongrowing and growing preantral follicles. It is noteworthy that in comparable areas of ovarian stroma separate from those occupied by antral follicles the number of neutrophils was sparse. To further analyze the relationship between follicle development and neutrophil accumulation, we compared the neutrophil density in healthy and atretic follicles. In healthy follicles the neutrophil density was 4.1 ± 0.8 neutrophils/mm2, which was clearly distinct from that of the background stroma, 1.0 ± 0.05 neutrophils/mm2. No differences in neutrophil density were detected among healthy follicles that appeared to be either early or more advanced in development. By comparison, the neutrophil density of atretic follicles was 8.8 ± 1.4 neutrophils/mm2 or approximately twofold higher (p < 0.05) than that of healthy follicles (Fig. 3). The neutrophil data were also analyzed with respect to the number of granulosa cells per follicle (neutrophil index). The neutrophil index of healthy and atretic follicles was 6.3 ± 2.0 and 25.1 ± 6.1, respectively (Fig. 3). This fourfold increase in the neutrophil index in the atretic group represented a statistically significant difference (p < 0.05). Neither the neutrophil density nor the neutrophil index of each follicle was correlated to the respective apoptotic index values. However, linear regression analysis revealed an inverse correlation between the neutrophil/granulosa cell ratio and the number of granulosa cells per follicle (r = 0.626), which was highly significant (p < 0.0001) (Fig. 4). IL-8 immunohistochemistry and in situ hybridization. Immunohistochemistry for IL-8 protein and in situ hybridization for IL-8 mRNA were performed on tissue sections from four representative ovaries. After incubation with IL-8 antisera, immunoreactive cells were observed in the theca interna and some granulosa cells of all these follicles. The intensity of the immunostained cells varied from very strong to relatively weak both between and within a given follicle (Fig. 5). Some neutrophils and en-
654 Chang, Gougeon, and Erickson
A
April 1998 Am J Obstet Gynecol
B
Fig. 5. Localization of IL-8 in an antral follicle within human ovary by immunohistochemistry. A, IL-8 protein is localized to follicle wall of atretic follicle. (Original magnification ×5). B, IL-8 immunoreactive cells in theca and granulosa cell layers of atretic follicle. (Original magnification ×20). GC, Granulosa cells; TC, theca cells. Note accumulation of neutrophils (arrows) within adjacent blood vessel of theca compartment.
dothelial cells in the theca blood vessels contained IL-8 immunoreactivity. In addition, immunoreactive IL-8 was detected in the epithelium and tunica albuginea of the ovarian capsule. No immunoreactivity was seen in preantral follicles or ovarian stromal tissue. Control tissues were comprised of fallopian tube and ovary that had been surgically removed for salpingitis and oophoritis. In the absence of primary antibody, IL-8 protein was not detected. Our in situ hybridization studies revealed that IL-8 mRNA message was present in the theca compartment of both healthy and atretic follicles (Fig. 6). For the most part, the IL-8 hybridization signal was weak. Importantly, however, a strong signal was seen in some endothelial cells and neutrophils of the theca. A weak signal for IL-8 was detected in a few granulosa cells. Significant IL-8 message was found in the ovarian epithelium. Similar to that for immunohistochemistry, acute salpingitic and oophoritic tissues were used as controls and IL-8 message was clearly expressed. Compared with the detection of IL-8 message, the sense strand mRNA failed to express any signal. Comment The results of our studies have demonstrated that normal folliculogenesis in the human ovary is associated with the accumulation of neutrophils within the vascula-
ture of the theca compartment of antral follicles, both healthy and atretic. The infiltration of neutrophils was significantly greater in those follicles undergoing atresia compared with healthy follicles. In contrast, preantral follicles were found to be totally unassociated with the presence of surrounding neutrophils. The appearance of these leukocytes in theca blood vessels coincided with the detection of the potent chemoattractant IL-8 in the follicle wall. These findings suggest that (1) the initial acquisition of neutrophils appears to be related to the graafian follicle, (2) subsequent increased accumulation occurred with follicle atresia, and (3) neutrophil attraction in developing follicles is mediated by IL-8. In the ovary the presence of perifollicular neutrophils generally has been recognized as a characteristic feature of the ovulatory process with particular involvement related to follicle rupture.1-3 We have now extended this concept to the population of developing cohort antral follicles. Our finding of neutrophils surrounding small and presumably early antral follicles and not preantral follicles suggests that neutrophils may be attracted to the growing follicle during antrum formation. The acquisition of neutrophils appears to persist throughout the growth, development, and atresia of graafian follicles. Because neutrophil accumulation was apparent in essentially all antral follicles, we suggest that significant num-
Chang, Gougeon, and Erickson 655
Volume 178, Number 4 Am J Obstet Gynecol
A
B
Fig. 6. Localization of IL-8 within follicle wall of antral follicle in human ovary by in situ hybridization. A, IL-8 mRNA (asterisks) is detected by in situ hybridization in some endothelial cells and neutrophils of theca in healthy follicle. (Original magnification ×20.) Eosin counterstain was applied to enhance expression and localization of antisense strand RNA. GC, Granulosa cells; TC, theca cells. B, Corresponding control section (sense strand).
bers of neutrophils in the theca compartment is a distinctive feature of the human graafian follicle. Of particular note was the significantly greater number of neutrophils surrounding atretic follicles compared with healthy follicles. This observation suggests a potential role for these immune cells in the atretic process. It is widely accepted that the principal mechanism for follicular atresia and granulosa cell death is apoptosis.12, 13 In the classic description of apoptosis, it was emphasized that inflammation and inflammatory cells were not involved in this process.7 In addition, evidence to suggest that immunocompetent cells, in particular neutrophils, are capable of inducing apoptosis is lacking. Our data showing no significant relationship between neutrophil density or neutrophil index to the apoptotic index fits with this concept. Nevertheless, the twofold increase in neutrophil number surrounding atretic follicles together with the significant inverse correlation of neutrophil index to follicle viability are noteworthy in that they could suggest a possible link between the aggregation of neutrophils and the process of follicle atresia. The mechanism of leukocyte recruitment to developing antral follicles is unknown, although it has been pro-
posed that neutrophils may be drawn to the follicle by chemotactic mechanisms induced by proinflammatory cytokines or chemokines, the most powerful of which is IL-8.14, 15 The results of our studies have demonstrated the strong expression of immunoreactive IL-8 protein in both theca and granulosa cells of healthy and atretic antral follicles. In addition, our detection of IL-8 mRNA in some of these cells was consistent with the marked immunoreactivity within the follicle wall. In contrast, evidence for the presence of IL-8 in granulosa cells of preantral follicles was completely lacking. Taken together, our data suggest that both theca and granulosa cells of developing antral follicles and not preantral follicles are capable of IL-8 production. These findings support and enlarge on the results of previous studies in which the well-documented invasion of neutrophils into the perifollicular environment just before ovulation appeared to involve IL-8.5, 6 That IL-8 protein and mRNA were expressed in associated neutrophils and endothelial cells of theca blood vessels suggests that local intravascular autocrine or paracrine mechanisms also may contribute to the acquisition of neutrophils. Previous studies have demonstrated
656 Chang, Gougeon, and Erickson
that these cells, when stimulated, are capable of IL-8 production.15 Moreover, both neutrophils and endothelial cells appear to express IL-8 receptor. Thus neutrophils may be able to perpetuate their own recruitment either directly or indirectly by stimulating endothelial cells. Additionally, IL-8 may mediate neutrophil-endothelial cell interaction by inducing binding of neutrophil β2-integrin molecules to endothelial cell intercellular adhesion molecules, resulting in firm attachment between the two cells.16, 17 This would account, at least in part, for the dramatic appearance of neutrophil-endothelial cell adhesion in the theca vasculature of developing follicles. Although neutrophils are thought to participate in the process of follicle rupture at the time of ovulation, the roles for these immune cells in association with antral follicle development or atresia have yet to be addressed. The apparent confinement of neutrophils to theca blood vessels raises uncertainty regarding the traditional proinflammatory role for these immune cells as implied during follicle rupture.1 Whether cytokine production by neutrophils within theca blood vessels has an impact on theca and granulosa cell proliferation or differentiation in the human follicle remains to be ascertained. In summary, we have provided evidence that neutrophils and endogenous IL-8 are expressed in the theca vasculature during folliculogenesis in normal ovulatory women, being particularly abundant in cohort follicles undergoing atresia. The expression of IL-8 and the influx of neutrophils led us to propose that an intrinsic IL8/neutrophil system could function in the mechanism of follicle atresia. The challenges are to understand the regulation of the expression of this novel cytokine system in the follicle wall and how the actions and interactions of IL-8 and neutrophils are coupled to ovarian physiologic features. We thank Holly Kim, Drs. Dan-Mei Li and Toshihide Kubo, Kimberly Best, Henry Ferreyra, and Nazli Ghafouri for their technical assistance and Dr. Rebbeca Baergen at the University of California, San Diego, and the physician staff at Kaiser Hospital, San Diego, for providing tissue samples. REFERENCES
1. Espey LL. Ovulation as an inflammatory reaction—a hypothesis. Biol Reprod 1980;22:73-106. 2. Adashi EY. The potential relevance of cytokines to ovarian physiology: the emerging role of resident ovarian cells of the white blood cell series. Endocrine Rev 1990;11:454-64. 3. Brannstrom M, Norman RJ. Involvement of leukocytes and cytokines in the ovulatory process and corpus luteum formation. Hum Reprod 1993;8:1762-75. 4. Brannstrom M, Pascoe V, Norman RJ, McClure N. Localization of leukocyte subsets in the follicle wall and in the corpus luteum throughout the human menstrual cycle. Fertil Steril 1994;61: 488-95. 5. Runesson E, Bostrum E-K, Janson PO, Brannstrom M. The human preovulatory follicle is a source of the chemotactic cytokine interleukin-8. Mol Hum Reprod 1996;4:245-50.
April 1998 Am J Obstet Gynecol
6. Arici A, Oral E, Bukulmez O, Buradagunta S, Engin O, Olive D. Interleukin-8 expression and modulation in human preovulatory follicles and ovarian cells. Endocrinology 1996;137:3762-9. 7. Wylie AH. Cell death: the significance of apoptosis. Int Rev Cytol 1980;68:251-306. 8. Gougeon A. Qualitative changes in medium and large antral follicles in the human ovary during the menstrual cycle. Ann Biol Biochem Biophys 1979;19:1461-8. 9. Gougeon A. Regulation of ovarian follicular development in primates: facts and hypotheses. Endocrine Rev 1996;17:121-55. 10. Schraufstatter I, Barritt DS, Zenaida MM, Oades G, Cochrane CG. Multiple sites on IL-8 responsible for binding to α and β IL8 receptors. J Immunol 1993;151:6418-28. 11. Simmons DM, Arriza JL, Swanson LW. A complete protocol for in situ hybridization of messenger RNAs in brain and other tissues with radiolabeled single strand RNA probes. J Histotechnol 1989;12:169-81. 12. Hsueh AJW, Billig H, Tsafriri A. Ovarian follicle atresia: a hormonally controlled apoptotic process. Endocrine Rev 1994;15:707-24. 13. Tilly JL, Kowalski KI, Johnson AL, Hsueh AJW. Involvement of apoptosis in ovarian follicular atresia and postovulatory regression. Endocrinology 1991;129:2799-801. 14. Norman RJ, Brannstrom M. Cytokines in the ovary: pathophysiology and potential for pharmacological intervention. Pharmacol Ther 1996;69:219-36. 15. Baggiolini M, Dewald B, Moser B. Interleukin-8 and related chemotactic cytokines—CXC and CC chemokines. Adv Immunol 1994;55:97-179. 16. Crockett-Torabi E, Fantone JC. The selectins: insights into selectin-induced intracellular signaling in leukocytes. Immunol Res 1995;14:237-51. 17. Rot A. Endothelial cell binding of NAP-1/IL-8: role in neutrophil emigration. Immunol Today 1992;13:291-4.
Editors’ note: This manuscript was revised after these discussions were presented. Discussion DR. LINDA C. GIUDICE, Stanford, California. The role of leukocytes and cytokines in the process of ovulation and corpus luteum formation is well known. The appearance of leukocytes in the theca before ovulation is also well known, as is the fact that a cytokine, IL-8, which is a potent chemoattractant for neutrophils, is markedly elevated in follicular fluid at the time of ovulation. The current study takes this cytokine leukocyte system several steps before the process of ovulation per se and examines follicular development. Dr. Chang and colleagues have examined 268 follicles from 18 normal ovaries obtained from reproductive-aged women 26 to 47 years old during the follicular phase of the cycle. A total of 199 of these follicles were preantral, 69 were antral, with 20 being healthy and 49 atretic. Methods used were morphometric analysis on every follicle by two individuals on separate occasions. Apoptotic cells were identified by histologic features and confirmed by APOPTAG kit. Mitotic and apoptotic index values were calculated, as was neutrophil density (numbers of neutrophils per square millimeter)and neutrophil index (numbers of neutrophils per granulosa). Immunohistochemistry was performed for IL-8 immunoreactive protein and in situ hybridization for IL-8 mRNA expression. The authors found no neutrophils around any of the preantral follicles and more neutrophils in atretic versus
Volume 178, Number 4 Am J Obstet Gynecol
healthy follicles. They also have found variable expression of IL-8 mRNA and protein in the follicle, with more in the theca than in the granulosa. The conclusion of the authors is that during follicular development IL-8 is expressed by the granulosa and more so by the theca that attracts neutrophils to the follicle. What happens thereafter? Why do the neutrophils go to the follicles that are destined for atresia more so than those that are destined for continued development? Or, are they attracted to atretic follicles after selection has occurred? What about the activation of these neutrophils? Is there any evidence that there is activation in atretic versus healthy follicles? What is the evidence that supports neutrophils are involved in atresia during follicular development? What is it that the neutrophil secretes that may contribute to the process of atresia and what are the mechanisms? These are the key issues and the pathway that the authors will likely follow in the near future. A key question is: Are neutrophils and their products activators of processes associated with apoptosis? Are the membrane changes that occur before the more classical DNA fragmentation regulated by cytokines from activated neutrophils? Is there any evidence for endonuclease regulation by cytokines from activated neutrophils? And if these are participants in the process of atresia, then how do the neutrophils know that they are needed for the process of ovulation of a mature follicle? These are all very important questions and the finding of neutrophils in human ovarian follicles in the thecal vasculature is likely an important one. Beyond the conceptual basis of this article are some comments on experimental approach and design. The numbers of follicles examined and the data that are reported for IL-8 mRNA and protein detection are not listed as representative of x number. Exactly how many follicles were looked at for IL-8 mRNA and protein? In addition, controls for specificity of the antibody and control for the RNA probe are important to show. Negative controls are also critical and should also be included in this study. Finally, although the neutrophil density does not appear to be related to the size of the follicle, healthy or otherwise, what about IL-8 expression? In conclusion, this study approaches an area of research that has had relatively little investigation, and that is with regard to cytokines and follicular development. The bulk of the investigations that have occurred to date focus on the role of leukocytes and their products, cytokines, in the process of ovulation, including activation of matrix metalloproteinases, angiogenesis, and steroido-
Chang, Gougeon, and Erickson 657
genesis as well as corpus luteum formation and maintenance. This new and exciting area is one that is certainly worthy of further investigation. DR. DAVID L. KEEFE, Providence, Rhode Island. If ovulation involves a process of inflammation, what keeps the oocyte from getting in the line of fire? Specifically, was there any evidence of neutrophil or IL-8 expression? DR. CHANG. Actually, the process of ovulation has been akin to inflammation. It really is not inflammation in the strict sense of the word because of the process of apoptosis occurring. This is a process that is classically not related to inflammation. Why is the oocyte not affected? It’s unclear. We did not look intensely. In fact, we hardly saw any of the follicles that were in the process of ovulation when this very closely coordinated and rapidly acting sequence of events occurred in the cycle. So our studies really did not involve those particular structures, if that’s the question. DR. KEEFE. Were there any neutrophils or IL-8 because of that morphologic observation? DR. CHANG. We didn’t see any around the oocyte specifically. These were two-dimensional sections, and as a result most of the sections did not contain the oocyte; of those that did, we didn’t see that around that specific area, although there may have been a few in which the more peripheral granulosa cell layers did contain it above the protein and the mRNA. DR. JOHN E. BUSTER, Houston, Texas. I have two questions. First of all, have you looked at this model with respect to polycystic ovarian disease and if you have, can you comment on it? Second, have you looked at IL-8 expression and neutrophil migration in the corpus luteum as a lysis? PRESIDENT SCOTT. With manipulation of the intraovarian cytokine, do you think this lends itself to a new method of contraception eventually? DR. CHANG (Closing). It is possible that there could be developed antagonist of these cytokines, which might have therapeutic implications, such as contraception. Although this area has not been explored, the use of cytokine antagonists have been thought to be useful in the treatment of some inflammatory disorders such as rheumatoid arthritis. It appears to provide relief from the symptoms of pain. I should have known that someone would ask about polycystic ovary, given my interest in that particular area. The neutrophils are increased in the follicles of the polycystic ovary. Why did we do the neutrophil index? It’s misleading. The granulosa cells and atresia go down, and the neutrophils may not go up; but the ratio will.