IIConnelly Foundation Fellow in Reproductive Biology, Pennsylvania Hospital.

Transcription

1 FERTllJTY AND STERILITY Copyright <> 1984 The American Fertility Society Vol. 42. No.1. July 1984 Prinl d in U.SA. Ultrastructure of ovarian follicles in in vitro perfused rabbit ovaries: response to human chorionic gonadotropin and comparison with in vivo observations* Edward E. Wallach, M.D.t:J: Yuji Okuda, M.D. Hideharu Kanzaki, M.D. Yoshimune Kobayashi, M.D. II Hitoshi Okamura, M.D. Rosemary Santulli, B.A.* Karen H. Wright, M.S.* Department of Obstetrics and Gynecology, Pennsylvania Hospital and University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, and Kyoto University School of Medicine, Kyoto, Japan Ovulation may be achieved and studied in an isolated perfused rabbit ovary upon inclusion of human chorionic gonadotropin (heg) in the perfusion fluid. The ultrastructural features of the rabbit ovarian follicle prior to ovulation in vitro were compared with those in vivo. The perifollicular vasculature was also examined in in vitro perfused rabbit ovaries during the preovulatory interval. Granulosa cells of the preovulatory follicle share many ultrastructural features in vivo and in vitro; however, only small amounts of smooth endoplasmic reticulum (ser) were observed in granulosa cells in vitro after heg. Ovulation after heg in the in vitro preparation tends to occur earlier (6 hours) than in vivo (12 hours). Thus, there may be insufficient time and/or gonadotropin exposure to permit full functional development of granulosa cells, as reflected by reduced amounts of ser. Degradation of collagen fibrils was less prominent in the theca externa and tunica albuginea in vitro than in in vivo. Perifollicular capillaries became dilated after heg, but interendothelial gaps were not observed. Disappearance of surface epithelium in the apex of follicles was similar in vitro and in vivo. Fertil Steril42:127, 1984 Received October 3, 1983; revised and accepted March 12, *Supported by NIH grant HD-05948, the Connelly Foundation. and the Mitchell and Lillian Duberstein Foundation. treprint requests: Edward E. Wallach, M.D., Department of Gynecology and Obstetrics, The Johns Hopkins Medical Institutions, 600 North Wolfe Street, Baltimore, Maryland :j:department of Obstetrics anp Gynecology. Pennsylvania Hospital and University of Pen~sylvania School of Medicine. Department of Obstetrics and. Gynecology, Kyoto University. IIConnelly Foundation Fellow in Reproductive Biology, Pennsylvania Hospital. Local ovarian alterations required for follicle disruption and ovulation have been studied extensively. However, the precise factors responsible for ovulation are incompletely understood. Several humoral factors have been postulated as intermediaries between the preovulatory gonadotropin surge and ovulation itself, including locally released histamine,1-3 prostaglandins,4-7 and proteolytic enzymes. 8 9 Collagenolytic enzymes are thought to be involved in the disrupt jon of the apical region of the follicle wall. lo 11 Ovarian smooth muscle activity has also been implicated Vol. 42, No.1, July 1984 Wallach et al. Ultrastructure of ovarian follicles 127

2 in the process of ovulation.12 Changes in perifollicular vasculature have been observed, including thrombosis within capillaries ofthe periapical region of the preovulatory follicle. 13 The apical region itself displays evidence of ischemia and stasis of blood. Perifollicular capillaries in other regions demonstrate evidence of increased permeability, largely as a result of fenestrations and intercellular gaps; however, no increase has been noted in pinocytotic vesicles. 14, 15 The relative contributions of these various humoral, enzymatic, structural, and vascular factors to ovulation itself have not been established. An in vitro model has been employed to study ovulation.16, 17 Using an isolated perfused rabbit ovary, ovulation can be achieved in response to the addition of human chorionic gonadotropin (hcg) or other stimulatory agents to the chemically defined perfusion medium. Ovulation characteristically occurs in vitro at - 6 hours after exposure to hcg. The ova released are mature and capable offertilizationp-19 Changes responsible for the release of a mature ovum can be examined serially, whereas components of the perfusion medium can be altered. Samples of the perfusate can be collected for biochemical determinations, and ova can be harvested for evaluation of maturity and fertilizability. The present study was undertaken for two purposes: (1) to compare the ultrastructural features of the rabbit ovarian follicle prior to ovulation in vitro with those features observed in vivo, and (2) to examine the perifollicular vasculature during the preovulatory interval in in vitro perfused rabbit ovaries. The latter observation is of particular significance because the perfusion system employed lacks plasma or cellular blood components, either or both of which have been thought to participate in the process of ovulation. MATERIALS AND METHODS New Zealand White female rabbits were caged individually under controlled temperature and light, fed Purina Rabbit Chow ad libitum, and isolated for a minimum of 3 weeks prior to use. Each rabbit underwent laparotomy with sodium pentobarbital anesthesia. Both ovarian arteries were cannulated in situ. The: ovary and its vascular pedicle were removed and placed in a perfusion chamber. The perfusion system used consists of an oxygenator, a pulsatile pump, and an ovarian perfusion chamber. Each ovary was perfused 128 Wallach et al. Ultrostructure of ovarian follicles individually with tissue culture medium 199 (Microbiological Associates, Bethesda, MQ) supplemented with heparin, insulin, streptomycin, and penicillin. The perfusion was carried out in a room with a constant temperature of 37 C. The cannulation and perfusion procedures have been described in detail previously.16,17 With this technique, ovulation can be induced consistently by the direct addition of hcg (100 IU) to the perfusate at the onset of perfusion. Ovulation usually occurs at - 6 hours after this dose of hcg.17 Ovaries are observed continually for follicle growth and evidence of ovulation. Follicle rupture can be detected when an extruded cumulus oophorus is observed on the ovarian surface. Both ovaries of three rabbits were perfused with medium 199 supplemented with 100 IU ofhcg. Both ovaries of a fourth rabbit were perfused with medium 199 without hcg. Perfusion was terminated at 5% hours in order to obtain preovulatory follicles for morphologic study. Using 100 IU hcg, ovulation occurs as early as 2 hours following the onset of perfusion, with a mean time of 5.90 ± 0.12 hours after hcgp No ovulations had yet occurred in the hcg-treated ovaries at 5% hours, although prominent follicles presumed to be preovulatory were noted on the surface. Neither of the unstimulated ovaries contained any follicles considered preovulatory. Each ovary was immediately removed from the perfusion chamber, and follicles 1 to 1.5 mm in diameter were excised using a razor blade and prepared for electron-microscopic examination. The most prominent clear follicles (800 to 900 j.lm in diameter) observed on the surface of the unstimulated ovaries were similarly excised and prepared for study. En bloc fixation was carried out for 30 minutes in chilled glutaraldehyde (6.25%) in 0.1 M cacodylate buffer (ph 7.4) containing 8% sucrose. Follicles were then divided into apical and basal parts, and fixation was continued for an additional 2 hours. Postfixation was carried out in 1% osmium tetroxide solution. Tissue blocks were dehydrated in a series of graded alcohols followed by propylene oxide and embedded in Epon 812 (Ernest F. Fullam, Inc., Schenectady, NY). The blocks were trimmed and thin sections were stained with toluidine blue and examined by light microscopy for orientation to identify the apical region. Using thin sections, granulosa and theca cell layers and capillaries could be localized prior to preparation of ultrathin sections. Ultrathin sections were prepared using Fertility and Sterility r I!

3 Figure 1 (A), Electronmicrograph of granulosa cells from follicles of ovaries perfused in vitro without heg showing free ribosomes, rer, and mitochondria with lamellar cristae in cytoplasm and annular nexus (arrow and inset) in cell membrane (x 3200; inset x 10,000). (B), Granulosa cells from control follicles in vivo demonstrate free ribosomes, rer, and mitochondria (x 4000). (C), Granulosa cells after heg treatment in vitro contain ribosomes, rer, mitochondria with few tubular cristae and few ser (x 7500). (D), Granulosa cells after heg treatment in vivo show abundant ser, large mitochondria with tubular cristae, Golgi apparatus (G), and lysosomelike granules (Ly) (x 4200). a glass knife on a Porter-Blum Ultramicrotome (Ivan Sorvall, Inc., Norwalk, CT). Ultrathin sections were stained with saturated uranyl acetate solution and lead citrate and examined by electron microscopy (Hitachi HU-llD, H 600, Tokyo, Japan) at 75 kv. Observations of follicles from in vitro perfused ovaries were compared with those from ovaries removed from intact rabbits before and after administration ofhcg(100 IU) by the marginal ear vein. Eight ovaries were removed from four untreated rabbits, and four ovaries were removed from two rabbits just prio'r to anticipated ovulation (10 hours after hcg administration). The preparation of follicular tissue for histologic and ultrastructural examination was identical to that described above. GRANULOSA LAYER RESULTS Granulosa cells from follicles of ovaries perfused in vitro without hcg (Fig. 1A) were round; intercellular projections were present but appeared less frequently than in those follicles from the in vivo ovaries (Fig. 1B). Abundant free ribosomes, rough endoplasmic reticulum (rer), and mitochondria with lamellar cristae were seen in the cytoplasm of granulosa cells from follicles perfused without hcg. Lipid droplets, Golgi apparatus, and lysosome-like granules were occasionally observed. The appearance of granulosa cells after hcg treatment in vitro (Fig. 1C) was compared with Vol. 42, No.1, July 1984 Wallach et ai. Ultrastructure of ovarian follicles 129

4 Figure 2 (A), Theca interna cells and capillaries from in vitro follicles without hcg. The capillary lumen (L) is narrow. Pericapillary spaces are moderately enlarged. Theca cells show the characteristic features of steroid-producing cells (x 8000). (B), Capillaries in the theca interna before hcg treatment in vivo have narrow lumina (L), and an erythrocyte is present in the lumen (x 8000). (C-l), Capillary lumina (L) in the theca interna after hcg treatment in vitro are dilated with thinner endothelial cells (E) (x 7500). (C-2), Theca interna cells after hcg treatment in vitro show increased ser and Golgi apparatus (G) (x 7500). (D), Capillaries close to the granulosa cells (gr) after hcg treatment in vivo are dilated and filled with platelets Cpl) (x 8000). that in vivo (Fig. ld). The cytoplasm, which was less extensive, contained many free ribosomes, rer, and mitochondria with few tubular cristae. Although smooth endoplasmic reticulum (ser) was rarely seen, well-developed Golgi apparatus, lipid droplets, and lysosome-like granules appeared increased in number after heg (Fig. le). Observations from'in vivo preparations after heg included decreased rer, abundant ser, large mitochondria with tubular cristae, Golgi apparatus, and lysosome-like granules (Fig. ld). Structurally, granulosa cells in vitro appear to be well-maintained. However, differentiation of these granulosa cells into lutein cells seems deficient. 130 Wallach et al. Ultrastructure of ovarian follicles THECA INTERNA LAYER In the in vitro preparation with heg, perifollicular capillaries in the theca intema layer had narrow lumina and were surrounded by basal lamina (Fig. 2A). In the endothelium, free ribosomes, mitochondria with lamellar cristae, and small amounts of rer were observed. Pericapillary spaces were moderately enlarged. There was no remarkable difference in the appearance of capillaries between in vitro and in vivo preparations (Fig. 2B). After the addition of heg to the perfusate, capillaries in the theca intema were dilated, whereas endothelial cells were thinner and blood corpus- Fertility and Sterility

5 cles were lacking (Fig. 2C-l). These findings are contrary to the appearance of capillaries from in vivo preparations after hcg, which are filled with formed elements (Fig. 2D). Between endothelial cells intracellular gaps were frequently observed in vivo but were not apparent in vitro. There appeared to be no remarkable changes in size, number, and distribution of pinocytotic vesicles after hcg treatment in vitro, consistent with findings in vivo. Theca interna cells in in vitro preparations contained large oval nuclei, with many free ribosomes, rer, and mitochondria with lamellar cristae (Fig. 2A). Although lipid droplets and lysosome-like granules were occasionally observed, ser and Golgi apparatus were not as prominent as in the in vivo specimens. After the addition of hcg to the perfusate, the numbers of round mitochondria with lamellar cristae, lysosqmes, and lipid droplets increased. With an increase of ser and Golgi apparatus, cells in vivo characteristically contain large mitochondria with lamello-vesicular and tubular cristae, abundant ser, and well-developed Golgi apparatus. Structurally, the cells appear to be well-maintained in ovaries perfused in this in vitro system. However, the cytoplasmic organelles involved in steroid production appear to be less pronounced than those of theca cells of ovarian follicles from in vivo preparations. THECA EXTERNA AND TUNICA ALBUGINEA In in vitro preparations without hcg, prominent intercellular spaces containing many collagen fibrils were found in the theca externa and tunica albuginea (Fig. 3A). These fibrils were dissociated in comparison with the dense bundles observed in vivo (Fig. 3B). In the in vitro preparations without hcg, fibroblasts contained moderate amounts ofrer, free ribosomes, mitochondria with lamellar cristae, and occasional lysosomelike granules and Golgi apparatus. Multivesicular structures were not observed in vitro without hcg. Following the addition of hcg to the perfusate, dissociation and fragmentation of collagen fibrils was somewhat more noticefible in both the tunica albuginea and the theca externa than it was in follicles perfused without hcg (Fig. 3C). Degradation of collagen fibrils was not pronounced in vitro, as it was in vivo (Fig. 3D). Multivesicular structures within the fibroblasts, often found in vivo, were infrequently observed in vitro. Surface epithelial cells were oval and more flattened in vitro without hcg (Fig. 3A) than those routinely found in vivo. In the cytoplasm, free ribosomes, rer, and mitochondria with lamellar cristae Were observed. Although a Golgi apparatus was occasionally seen, lysosome-like granules were rarely observed. After the addition of hcg to the perfusate, surface epithelial cells were infrequently seen over the follicle apex. One of the rarely observed surface epithelial cells from ovaries perfused in vitro with hcg is illustrated (Fig. 3C). DISCUSSION The follicular wall of the rabbit ovary perfused in vitro with or without hcg was examined via transmission electron microscopy and compared with in vivo follicles at corresponding stages of follicular development. ser plays a significant role in steroidogenesis in steroid-producing cells. Granulosa cells in the in vitro follicle after hcg revealed small amounts. of ser. This finding contrasts with observations of preovulatory follicles from ovaries in vivo, despite the fact that in vitro and in vivo granulosa cells share many other ultrastructural features when not stimulated by hcg. Ovulation consistently occurs earlier after the addition of hcg (6 hours) to the ovary perfused in vitro than it does in vivo after intravenous hcg administration or following the stimulus of coitus (12 hours).17 The reason(s) for the acceleration of ovulation in response to hcg in vitro is not yet known. Granulosa cells in vitro thus may not have sufficient time and/or gonadotropin exposure to express full functional development prior to ovulation. Similarly, theca interna cells also playa prominent role in steroid production. These cells in the in vitro follicles contained small amounts of ser and lacked mitochondria with tubular cristae. The differences observed between the in vitro and in vivo preparations, as with granulosa cells, may be in part time-related. It is also possible that the theca and granulosa cells in our system do not receive sufficient substrate to sustain long-term steroid production. Progesterone and estradiol both increase in the perfusion effluent to levels similar to those of ovarian vein blood. 20, 21 These Vol. 42, No.1, July 1984 Wallach et al. Ultrastructure of ovarian follicles 131

6 Figure 3 (A), The theca externa and tunica albuginea of a follicle perfused in vitro are illustrated. Many collagen fibers (C), which have been dissociated in comparison with those in vivo, can be seen. Flat surface epithelium (SE) can also be observed (x 6000). (B), Theca externa and tunica albuginea of an in vivo follicle. Collagen fibers (C) are prominent and intact (x 4800). (C), Collagen fibers (C) are dissociated and fragmented after the addition of hcg to the perfusate. Surface epithelium (SE) can be observed (x 5000). (D), Degradation of collagen fibers is pronounced in the follicle after treatment with hcg in vivo. Surface epithelium (SE) can be observed (x 10,000). values may represent acute release of previously synthesized steroids. The relative paucity of organelles associated with steroid-producing cells in vitro may fail to reflect the elaboration of steroids and substrate availability during the relatively short perfusion time. Without hcg, increased intercellular space and dissociation of collagen fibrils were observed in the theca and tunica albuginea of in vitro perfused follicles in contrast to those in vivo. In the presence ofhcg, degradation of collagen fibrils in the theca and tunica albuginea was not as prominent in the in vitro folliclesl as in those in vivo. Although enzymatic degradation of the collagen fibrils and the presence of multivesicular structures are considered essential for ovulation in vivo, the relative absence of multi vesicular struc- 132 Wallach et al. Ultrastructure of ovarian follicles tures in vitro did not preclude the occurrence of ovulation. 22 Although perifollicular capillaries of the perfused follicles in vivo were obviously dilated after the addition of hcg to the perfusate, an increase of interendothelial gaps could not be observed in this study. The disappearance of surface epithelium in the apex of follicles from ovaries perfused in vitro is similar to observations from in vivo studies of the rabbit preovulatory follicle. Bjersing and Cajande~3 demonstrated gradual collapse and disappearance of surface epithelial cells over the apex just prior to ovulation using scanning electron microscopy. In conclusion, many of the ultrastructural features found in preovulatory rabbit follicles in vivo Fertility and Sterility

7 were also observed in vitro. Minor differences, such as reduced ser in granulosa and theca cells, may be related to duration of gonadotropin exposure, time of ovulation, or limited substrate for steroidogenesis. Although degradation of collagen was not as prominent in vitro as in vivo, ovulation nonetheless occurs regularly in the in vitro model. This observation may raise questions as to the role of collagenolysis in the process of ovulation at least in vitro. Vascular changes in vitro, despite the absence of blood components in the perfusate, seem to parallel those in vivo. These latter observations are supported by scanning electron microscopic studies demonstrating increased permeability of the perifollicular microvasculature just prior to ovulation both in vivo and in vitro.24 Acknowledgments. We wish to thank Thomas Henry, Maria Calzonetti, and Ellen Taggart for their technical assistance, and Drs. Santo Nicosia and Hirokatsu Kitai for critical review of the manuscript. REFERENCES 1. Szego CM, Gitin ES: Ovarian histamine depletion during acute hyperaemic response to luteinizing hormone. Nature 201:682, Morikawa H, Okamura H, Takenaka A, Morimoto K, Nishimura T: Histamine concentration and its effect on ovarian contractility in humans. Int J Fertil 26:283, Kobayashi Y, Wright KH, Santulli R, Kitai H, Wallach EE: Effect of histamine and histamine blockers on the ovulatory process in the in vitro perfused rabbit ovary. BioI Reprod 28:385, Armstrong DF, Grinwich DL: Blockade of spontaneous and LH-induced ovulation in fd.ts by indomethacin, an inhibitor of prostaglandin biosynthesis. Prostaglandins 1:21, Tsafriri A, Lindner HR, Zor U, Lamprecht SA: Physiological role of prostaglandins in the induction of ovulation. Prostaglandins 2:1, O'Grady JP, Caldwell BV, Auletta FJ, Speroff L: The effects of an inhibitor of prostaglandin synthesis (indomethacin) on ovulation, pregnancy, and pseudopregnancy in the rabbit. Prostaglandins 1:97, Grinwich DL, Kennedy TG, Armstrong DF: Dissociation of ovulatory and steroidogenic actions of luteinizing hormones in rabbits. with indomethacin, an inhibitor of prostaglandin biosynthesis. Prostaglandins 1:89, Espey LL, Coons PJ: Factors which influence ovulatory degradation of rabbit ovarian follicles. BioI Reprod 14: 233, Cajander S, Bjersing L: Fine structural demonstration of acid phosphatase in rabbit germinal epithelium prior to induced ovulation. Cell Tissue Res 164: Espey LL: Decomposition of connective ti~ue in rabbit ovarian follicles by multi vesicular structures of thecal fibroblasts. Endocrinology 88:437, Fukumoto M, Yajima Y, Okamura H, Midorikawa 0: Collagenolytic enzyme activity in human ovary: an ovulatory enzyme system. Fertil Steril 36:746, Owman CH, Sjoberg N-O, Wallach EE, Walles B, Wright KH: Neuromuscular mechanisms of ovulation. In Human Ovulation, Edited by ESE Hafez. Amsterdam, Elsevier North Holland Biomedical Press, 1979, p Kanzaki H, Okamura Y, Okuda Y, Takenaka A, Morimoto K, Nishimura T: Scanning electron microscopic study of rabbit ovarian follicle microvasculature using resin injection-corrosion casts. J Anat 134:697, Okuda Y, Okamura H, Kanzaki H, Fujii S, Takenaka A, Wallach EE: An ultrastructural study of ovarian perifollicular capillaries in the indomethacin-treated rabbit. Fertil Steril 39:85, Okuda Y, Okamura H, Kanzaki H, Takenaka A: Capillary permeability of rabbit ovarian follicles prior to ovulation. J Anat. In press 16. Lambertsen CJ Jr, Greenbaum DF, Wright KH, Wallach EE: In vitro studies of ovulation in the perfused rabbit ovary. Fertil Steril 27:178, Kobayashi Y, Wright KH, Santulli R, Wallach EE: Ovulation and ovum maturation in the rabbit ovary perfused in vitro. BioI Reprod 24:483, Kobayashi Y, Santulli R, Wright KH, Wallach EE: Fertilizability of ova ovulated and recovered from rabbit ovaries perfused in vitro. Science 213:1127, Kobayashi Y, Santulli R, Wright KH, Wallach EE: In vitro fertilization of rabbit ova ovulated in vitro during ovarian perfusion. J Reprod Fertil 68:41, Wu C, Blasco L, Flickinger GL, Mikhail G: Ovarian function in the preovulatory rabbit. BioI Reprod 17:304, Kitai H, Kobayashi Y, Santulli R, Wright KH, Wallach EE: Unpublished data 22. Espey LL, Coons PJ, Marsh M, LeMaire WJ: Effect of indomethacin on preovulatory changes in the ultrastructure of rabbit graafian follicles. Endocrinology 108:1040, Bjersing L, Cajander S: Ovulation and the mechanism of follicle rupture. Cell Tissue Res 149:301, Kitai H, Yoshimura Y, Wright KH, Santulli R, Wallach EE: Microvasculature of preovulatory follicles. Presented at the Thirty-First Annual Meeting of the Society for Gynecologic Investigation, March 21 to 24, 1984, San Francisco, California Vol. 42, No.1, July 1984 Wallach et ai. Ultrastructure of ovarian follicles 133