Polarized Ovaries of the Long-tongued Bat Glossophaga soricinaA Novel Model for Studying Ovarian Development Folliculogenesis and Ovulation.код для вставкиСкачать
THE ANATOMICAL RECORD 290:1439–1448 (2007) Polarized Ovaries of the Long-Tongued Bat, Glossophaga soricina: A Novel Model for Studying Ovarian Development, Folliculogenesis, and Ovulation CAROLYN M. KOMAR,1* FRANCES ZACHARACHIS-JUTZ,1 CHRIS J. CRETEKOS,2 RICHARD R. BEHRINGER,2 AND JOHN J. RASWEILER IV3 1 Department of Animal Science, Iowa State University, Ames, Iowa 2 Department of Molecular Genetics, University of Texas, M.D. Anderson Cancer Center, Houston, Texas 3 Department of Obstetrics and Gynecology, State University of New York Downstate Medical Center, Brooklyn, New York ABSTRACT Glossophaga soricina is a spontaneously ovulating, monovular, polyestrous bat with a simplex uterus, exhibiting true menstruation. Studies conducted on reproductively active, captive-maintained animals established that G. soricina also has polarized ovaries, with the ovarian surface epithelium (OSE) restricted to the medial side of the ovary, and primordial follicles limited to an immediately adjacent zone. Follicles selected for further development are recruited from the medullary side of this zone, and ovulation is restricted to the portion of the ovary covered by the OSE. To further develop G. soricina as a model for studying ovarian development and physiology, ovaries were collected from fetal, neonatal, and adult females and processed for morphological and immunohistochemical analyses. The latter included staining for factor VIII-related antigen (von Willebrand factor) to assess regional differences in ovarian vascularity. The ovarian structure in fetal and neonatal animals was very similar to that in other species that do not have polarized ovaries at comparable stages of development. This indicated that polarization of the ovary does not occur during fetal development, but rather sometime between the neonatal period and adulthood. Vascular elements were abundant in areas of the ovary surrounding early growing follicles, but sparse in the zone of the ovary containing primordial follicles. The polarized nature of the ovaries in G. soricina suggests that this species might be used as a model to investigate the formation, long-term maintenance, and activation of primordial follicles, and the role of the OSE in ovulation and folliculogenesis. Anat Rec, 290:1439–1448, 2007. Ó 2007 Wiley-Liss, Inc. Key words: ovarian development; ovary; ovarian surface epithelium; ovulation; folliculogenesis Grant sponsor: NSF foundation; Grant number: IBN 0220458; Grant number: GB-6435; Grant sponsor: NIH; Grant number: HD00171; Grant number: TW00143; Grant sponsor: the Population Council. *Correspondence to: Carolyn M. Komar, Division of Functional Biology, West Virginia School of Osteopathic Medicine, Ó 2007 WILEY-LISS, INC. 400 N. Lee Street, Lewisburg, WV 24901. Fax: 304-645-4859. E-mail: email@example.com Received 17 April 2007; Accepted 22 August 2007 DOI 10.1002/ar.20602 Published online in Wiley InterScience (www.interscience. wiley.com). 1440 KOMAR ET AL. Glossophaga soricina is a microchiropteran bat and a member of the family Phyllostomidae (American leafnosed bats). This bat is a common inhabitant of the lowland tropics of the New World from southern Sonora in western Mexico, south to northeastern Argentina and southeastern Brazil. They eat nectar, fruit, ﬂower parts, pollen, and insects (Rasweiler, 1975, 1977; Alveraz et al., 1991). G. soricina is important ecologically; several night-blooming plants rely on them for pollination (Ramirez et al., 1984; Silva and Peracchi, 1999; Zortea, 2003). The physiology of reproduction in G. soricina is similar in many respects to that of anthropoid primates (Hamlett, 1934, 1935; Rasweiler, 1974, 1979a; Rasweiler and Badwaik, 2000). G. soricina is a polyestrous, monovular species with a simplex uterus. Ovulation occurs spontaneously and can take place from either ovary. The ovarian cycle is 22–26 days with a functional luteal phase (Rasweiler, 1972). Nonpregnant cycles are terminated by true menstruation, characterized by extensive breakdown and shedding of the superﬁcial lamina functionalis of the endometrium with associated bleeding. Menstruation is correlated with, and presumably caused by, luteal regression. Implantation of the embryo is interstitial, and ultimately a discoidal, hemochorial placenta is formed. These animals can be easily maintained in captivity, and captive females cycle throughout the year (Rasweiler and de Bonilla, 1972; Rasweiler, 1973). Observations presented herein indicate that the ovaries of G. soricina are also of interest because, in adult animals, they exhibit an unusual polarity of structure and function not seen in the more commonly studied mammals. The ovarian surface epithelium (OSE) covers only one part of the ovary, and the primordial follicles are restricted to a zone underneath this epithelial layer. As the follicles grow and mature, they occupy more of the ovarian medulla. Interestingly, ovulation always occurs in the area where the OSE covers the ovary. The polarity of the ovaries in G. soricina suggests that this animal may be a revealing model to study local factors supporting maintenance of primordial follicles in a quiescent state, selection of follicles for further growth and maturation, and development of the eventual site of follicular rupture. The following study was initiated to better deﬁne when the ovaries polarize and to further develop G. soricina as a model for studies of ovarian biology. MATERIALS AND METHODS Source and Maintenance of Captive Bats In part to study ovulation and luteal development, tissues were collected from captive-maintained colonies of G. soricina. Bats worked with in a colony established at Cornell University (Ithaca, New York) had been captured on the West Indian island of Trinidad. Bats worked with in a colony established at the Universidad del Valle (Cali, Colombia) had been captured in rural areas of the Departamento del Valle, Colombia. The animals were housed in bipartite cages of modest dimensions (with enclosed roosting and open feeding areas) and fed fortiﬁed diets using canned peach nectar (Cornell University) or canned guava pulp (Universidad del Valle) as palatable bases (Rasweiler and de Bonilla, 1972: Rasweiler, 1973, 1975). All of the work with captive-maintained animals was conducted before any national laws in the United States or Colombia governing the care and use of laboratory animals. The work was conducted humanely, however, using methods that would now easily meet legislation of this type. Tissue Collection Reproductive tracts were collected from 36 sexually segregated adult females killed at various times during their 22–26 day cycle and 98 mated females killed between days 1 and 32 post coitum (p.c.; Rasweiler, 1972, 1974). Animals were killed for these studies with diethyl ether. Nonpregnant cycles had been timed by using changes in vaginal cytology to identify the approximate time of ovulation. The distribution of cycling specimens examined was as follows, with day 1 being the ﬁrst day on which a newly ovulated oocyte was present in one of the oviducts: day 1 (n 5 2 bats), 2 (1), 3 (1), 4 (1), 5 (1), 6 (1), 7 (2), 11 (2), 12 (2), 13 (3), 14 (2), 16 (1), 17 (2), 18 (1), 20 (3), 21 (2), 22 (2), 23 (4), and 24 (3). Mating was determined by the daily examination of vaginal aspirates for spermatozoa (Rasweiler, 1972). The ﬁrst day of a sperm-positive aspirate (day 1 p.c.) correlated with the presence of a newly fertilized oocyte in one of the oviducts. The distribution of pregnant specimens examined was as follows: day 1 p.c. (n 5 3 bats), 2 (3), 3 (2), 4 (3), 5 (2), 6 (2), 7 (2), 8 (2), 9 (2), 10 (2), 11 (2), 12 (3), 13 (3), 14 (8), 15 (7), 16 (5), 17 (2), 18 (5), 19 (2), 20 (5), 21 (2), 22 (3), 23 (2), 24 (6), 25 (2), 26 (3), 28 (5), 30 (4), and 32 (5). Some mated specimens were found to be nonpregnant upon examination. These were obtained on the following days: day 11 p.c. (n 5 1), 14 (1), 16 (1), 19 (1), 20 (1) 21 (1), 22 (2), and 24 (1). For histological examination, each reproductive tract was removed as a unit, with or without the vagina attached. The tracts were manipulated during dissection and processing primarily by means of the urinary bladder and/ or the cervix and vagina, although a slit was also made in the uterine isthmus to facilitate penetration of the ﬁxative. There was no handling of the ovaries or oviducts. Tissues were ﬁxed in Helly’s, Carnoy’s, or Zenker’s ﬂuid, or acetic acid–mercuric chloride–formalin, and subsequently embedded in parafﬁn. This work was undertaken primarily to conﬁrm reports, based upon ﬁeld-collected specimens, that G. soricina menstruated and exhibited interstitial implantation of a single blastocyst in a simplex uterus (Hamlett, 1934, 1935). This necessitated the development of methods for introducing G. soricina into captivity, maintaining them in a reproductively active state, and timing nonpregnant reproductive cycles and pregnancies. Histological studies were then pursued to (1) characterize major ovarian, oviductal, and endometrial changes during the cycle and early pregnancy; (2) determine the distribution of Graaﬁan follicles and corpora lutea between the two ovaries (because bats frequently exhibit functional asymmetries of the ovaries; Rasweiler and Badwaik, 2000); and (3) characterize embryonic development from fertilization through formation of the early embryonic shield in carefully timed pregnancies. Similar studies had not previously been conducted with any other bat in captivity. This strategy required a large number of G. soricina for several reasons: (1) this species’ nonpregnant cycle proved to be relatively long (22– 26 days); (2) early embryonic development was unusu- 1441 POLARIZED OVARIES OF G. SORICINA ally slow (implantation commencing between days 12 and 14 p.c.), and in some respects unusual, when compared with more commonly studied mammals; and (3) there was some variability between animals in the timing of major reproductive events (Rasweiler, 1972, 1974, 1979a,b, 1993; Badwaik and Rasweiler, 2000). The polarized nature of G. soricina’s ovaries was discovered in the course of this work. This played no role in determining the number of animals used. Female G. soricina were also collected from the ﬁeld on the island of Trinidad using techniques described previously for the closely related Carollia perspicillata (with which G. soricina frequently roosts; Evarts et al., 2004). Brieﬂy, females were collected with hand nets in their diurnal roosts during the months of January and May 2004–2005. All procedures for collecting and handling animals in the ﬁeld were approved by the Iowa State University’s Institutional Animal Care and Use Committee. The relative ages and sexual maturity of the animals were determined by observations of pelage (darker for juveniles than adults), area around nipples (if bare, indicating current or recent lactation), and gross examination of the reproductive tract. Neonatal animals were identiﬁed as those still nursing and unable to ﬂy on their own. Animals with large, antral follicles and/or luteal tissue in the ovaries were considered adults. Females carrying young, or with bare nipples from which milk could readily be expressed, were classiﬁed as lactating. These animals were euthanized by cervical dislocation. Intact reproductive tracts were collected from neonatal (n 5 3) and adult (n 5 12) animals, including 2 pregnant bats and their female fetuses. Tissues were ﬁxed in Bouin’s solution or 4% methanol-free formaldehyde buffered with 0.1 M Dulbecco’s phosphate-buffered saline at 48C overnight and then transferred into 70% ethanol. After ﬁxation, tissues were embedded in parafﬁn for morphological or immunohistological analyses. Morphological Analysis Reproductive tracts from ﬁeld-collected G. soricina and ﬁxed in Bouin’s solution were processed to parafﬁn and sectioned at 5 mm. These sections were stained with Gill’s hematoxylin (Fisher Scientiﬁc), or the periodic acid Schiff procedure and counterstained with fast green. A minimum of 4 tissue sections were examined from fetal, neonatal, and adult ﬁeld-caught animals. Tissue sections chosen for analysis were collected from the central portion of the ovary, which represent more of the organ than peripheral sections. Tracts collected from the captive-maintained animals were ﬁxed as noted above, processed to parafﬁn, and serially sectioned at 5–7 mm. The sections were stained with a variety of routine histological and histochemical techniques (Rasweiler, 1972, 1974). All slides were mounted with Permount. Serial sectioning was necessary to (1) adequately characterize advanced follicular development, ovulation, and fertilization; (2) examine the distribution of normally developing follicles, new and regressing corpora lutea, within and between the two ovaries; (3) locate and describe the single embryo at progressive stages of early pregnancy; (4) determine the fate of the zona pellucida, which is usually shed by the blastocyst and then retained within one of the oviducts; (5) attempt to account for all new eggs ovulated by mated females, as well as the remnants of degenerating, ovulated ova typically retained in the oviducts from prior nonfertile cycles; (6) compare corresponding segments of the two oviducts, because G. soricina is unusual in exhibiting their differential stimulation during the periovulatory period; and (7) look for regional differences in endometrial development, because G. soricina apparently exhibits implantation only in very limited and predictable uterine sites (Rasweiler, 1972, 1974, 1979a,b, 1993). Immunohistological Analysis Immunolocalization of factor VIII-related antigen (von Willebrand factor) was used to enhance the visibility of vascular elements and assess the relative vascularity of various zones within the ovaries of G. soricina (n 5 4 adult animals). Parafﬁn-embedded tissues (5 mm) were deparafﬁnized, rehydrated, and endogenous peroxidase activity was quenched with 1.5% hydrogen peroxide in methanol. Sections were rinsed in phosphate buffered saline (PBS; pH 7.2) and subsequently blocked with 10% normal goat serum (Zymed Laboratory), followed by overnight incubation at 48C with antiserum directed against factor VIII-related antigen (1:250; rabbit antihuman factor VIII-related antigen; Sigma). The primary antiserum was omitted as a control. After rinsing with PBS, sections were incubated with a biotinylated secondary antibody (1:100; goat anti-rabbit), and subsequently incubated with a peroxidase/streptavidin complex (Vector). To visualize the immunocomplex, sections were treated with 3,3-diaminobenzidine tetrahydrochloride (Zymed Laboratory). Sections were counterstained with Gill’s hematoxylin, dehydrated, cleared, and mounted with Permount. Female reproductive tissues from fetal (n 5 1) and adult (n 5 4) G. soricina were also processed for the immunodetection of cytokeratins. Parafﬁn-embedded tissues were processed as described above with the following modiﬁcations. The primary antiserum used was a mouse monoclonal antibody directed against Type II cytokeratins 1, 5, 6, and 8 from human epidermis (Sigma catalog no. C1801). This antiserum has been shown to cross-react with cytokeratins in tissues from a variety of vertebrates (mammals, chicken, lizard, snake, and carp). Tissues were incubated with the primary antiserum diluted at 1:400 with diluent supplied in the Mouse on Mouse Immunodetection Kit (Vector). For negative controls, the primary antiserum was omitted and tissues were incubated in diluent alone. RESULTS The subject of this study, G. soricina, is a small microchiropteran bat (Fig. 1) that weighs approximately 9–12 grams when adult females are nonpregnant or in early pregnancy (i.e., up to day 32 p.c., the latest stage of pregnancy examined in captive-maintained specimens). G. soricina’s small size is signiﬁcant because it facilitates very thorough morphological studies of the reproductive tract and early embryos in situ that would be impractical with many larger species. These bats are specialized, in part, for hovering over night-blooming ﬂowers and taking nectar with an elongated tongue. 1442 KOMAR ET AL. Fig. 1. An adult, female Glossophaga soricina carrying a neonate, which appears underneath and is gray in color. Figures 2–5 illustrate some major histological features of G. soricina’s ovaries at fetal, neonatal, and adult stages of development. The fetal ovary shown was obtained from a specimen corresponding to stage 22 of embryonic development for the closely-related bat, Carollia perspicillata (approximately 80 days p.c. with a gestation length of 113–120 days; Cretekos et al., 2005). A comparable, carefully-timed staging sequence has not yet been generated for G. soricina beyond the early embryonic disc stage (Rasweiler, 1974). At this stage of development, the OSE (Fig. 2a), luminal epithelia of the oviducts and uterus, and ordinary mesothelia of ovaries, oviducts, and uterus were reactive with the pan-cytokeratin antiserum. The ovary contained an abundance of primordial follicles, located principally in the cortex, and a substantial epithelial component located in the medulla (Fig. 2a,b). The latter included anastomotic elements of the rete ovarii, as well as possibly some medullary tubules and early developing (primary) follicles, containing oocytes. These epithelial cells generally exhibited a high nuclear to cytoplasmic ratio and did not contain much cytoplasm. In some cases, epithelial cells surrounding oocytes had become cuboidal, but the oocytes did not appear to have enlarged or exhibit other signs of activation. The mesenchymal component of the fetal ovary at this stage was relatively inconspicuous. In neonatal females, primordial follicles still predominated in the cortex, and the ovaries were not yet polar- ized (Fig. 3a). In contrast to the fetal ovary, the medulla contained several activated oocytes (denoted by their larger size; Fig. 3a,b), frequently clustered together (Fig. 4). Most of these appeared to be within the rete ovarii, as well as possibly some medullary tubules or primary follicles. Activation of oocytes appeared to occur preferentially in or near elements of the rete ovarii. Particularly in sections stained with the periodic acid–Schiff procedure, zonae pellucidae could also be seen around many of these activated oocytes. During neonatal life, many of the epithelial cells in the rete ovarii and adjacent medullary structures appeared more active than in the fetal ovary (cf., Figs. 2b, 3b). These cells appeared larger, with more cytoplasm and bigger, more euchromatic nuclei than in fetal ovaries. Also, the nuclear to cytoplasmic ratio of these cells generally appeared to be lower than in fetal life. By the time females had reached adulthood, a striking regionalization of the ovaries was apparent (Fig. 5). Only a cap-shaped region on the medial surface of each ovary was covered by the OSE (Fig. 6). This was equivalent to the mesothelial-derived, surface epithelium as described by Mossman and Duke (1973). The remainder of the ovary was covered by a less specialized, squamous mesothelium (Fig. 6). It was highly unlikely that any of the OSE might have been artifactually removed during dissection and processing, because the ovaries were never mechanically manipulated and were also well-protected within bursal sacs (albeit, each with an opening; Rasweiler, 1972). The ovarian surface epithelial cap was opposite the ovarian hilus. Primordial follicles were limited to a broad zone immediately adjacent to the OSE (Figs. 5, 6). In all histological specimens examined, follicles that had been selected for further growth and development were observed only on the medullary side of the primordial follicle zone. Follicles containing multiple oocytes were also occasionally observed in ovaries from adult G. soricina. Despite that preovulatory follicles had large antra and developed well within the ovaries of G. soricina (Rasweiler, 1972), ovulation ultimately took place only in the region of the ovary covered by the OSE. This ﬁnding was determined by observing follicle rupture points in serial sections of the ovaries from a large series of nonpregnant, cycling, or early pregnant bats (Figs. 5, 7). In females examined on days 1 or 2 of their cycle or pregnancy that had recently ovulated, rupture points (stigmata) could be readily identiﬁed on the ovarian surface. Immediately after ovulation, the still partially swollen cortex around the ruptured follicle was responsible for some attenuation of the overlying surface epithelium (Fig. 7); however, the presence of primordial follicles in these swollen areas, to either side of the stigmata, indicated that this epithelium was most likely OSE (rather than ordinary mesothelium). In sections containing areas where the OSE had healed, the shape and proximity of the corpus luteum to the OSE indicated the probable rupture point (Figs. 5, 7). Often the corpus luteum had a tapering neck, or was funnel-shaped, on this side. In no case was there any suggestion that ovulation had occurred through the extensive portion of the ovaries covered by an ordinary, squamous mesothelium or their broad mesenteric attachments. It is interesting to note that this situation prevailed, even though preovulatory Graaﬁan follicles in G. soricina became very large and 1443 POLARIZED OVARIES OF G. SORICINA Fig. 2. Sections of an ovary from a fetal Glossophaga soricina. a: A fetal ovary and adjacent tissues processed for immunolocalization of cytokeratins 1, 5, 6, and 8, counterstained with hematoxylin. As depicted by the brown reaction product, the pan anti-cytokeratin antiserum labeled the ovarian surface epithelium (OSE) and ordinary mesothelium. Much of the ovarian cortex is occupied by primordial fol- licles. Arrowheads denote epithelial components in the medulla. b: The medulla and adjacent cortex depicting the anastomotic, epithelium-lined channels of the rete ovarii (arrow heads). Although some oocytes (arrows) in this region are surrounded by a cuboidal epithelium, none exhibit clear signs of having been activated. Section was stained with hematoxylin. Scale bars 5 250 mm in a, 50 mm in b. sometimes closely approached this ordinary mesothelium (see Fig. 1 in Rasweiler, 1972). All of the ovaries from the adult, captive-maintained G. soricina contained an abundance of atretic follicles, many of which were large and antral. Three large follicles in relatively early stages of atresia are evident along the bottom of the ovary shown in Figure 7a. Each contained some granulosa cells with pyknotic nuclei, or which were necrotic and had been sloughed off into the antrum. In the case of the middle follicle, the cumulus had also been completely lost from around the oocyte. These are some of the generally accepted signs of atresia in other mammals (Ingram, 1962; Peters and McNatty, 1980). The cumulus was still present in large, normal Graaﬁan follicles approaching ovulation (see Fig. 1 in Rasweiler, 1972). It also deserves to be noted that the ovary shown in Figure 7a had been obtained on day 1 p.c., soon after ovulation. G. soricina normally ovulates only one follicle per cycle. All antral follicles persisting in the ovaries on that day, or during the ﬁrst couple of weeks of the nonpregnant cycle and early pregnancy, were invariably atretic (Rasweiler, unpublished observations). Immunolocalization of factor VIII-related antigen was used to visualize regional differences in the density of vascular elements within ovarian tissue from adult G. soricina. Vascular elements were quite abundant in the stroma surrounding developing follicles (Fig. 8a,c). In contrast, such elements tended to be sparse in the primordial follicle zone (Fig. 8a,b) and were most frequently localized there to the major stromal partitions between clusters of follicles. DISCUSSION The histological structure of the ovary in G. soricina during fetal and neonatal life did not differ dramatically from that found during comparable stages of development in mammals with more ‘‘typical’’ ovaries, such as those in primates, ruminants, rats, and mice, with most of the surface covered by OSE (Mossman and Duke, 1973; Konishi et al., 1986; Hirshﬁeld and DeSanti, 1995). By the time the female bats reached sexual maturity, the ovary had been transformed into a polarized structure with well-deﬁned regions of follicles at progressive stages of development. The recruitment of follicles for development was consistently observed to occur 1444 KOMAR ET AL. Fig. 3. Ovary from a neonatal Glossophaga soricina stained with hematoxylin. a: Much of the ovary is still occupied by primordial follicles (pf). Some oocytes (arrows) in the rete ovarii, as well as possibly medullary cords or primary follicles, had been activated and have begun to grow. b: Higher power view of part of the rete ovarii (arrowheads). In contrast to the fetal condition (see Fig. 2a,b), many of the rete epithelial cells appeared larger and more active. Scale bars 5 100 mm in a, 50 mm in b. only in a narrow zone along the medullary side of the area occupied by primordial follicles. Similarly, clear evidence of oocyte activation (as indicated by an increase in their size) was observed only in the central part of the ovary in neonatal females. There was no evidence of follicular activation and development within most of the primordial follicle zone closer to the OSE. The current study indicated that polarization of the ovaries in G. soricina occurred between birth and adulthood. The mechanisms whereby this was accomplished remain unclear, but two possibilities seem deserving of consideration. As each ovary grew, its broad mesenteric attachment did as well. This ﬁnding may have displaced the OSE and primordial follicles from the mesenteric side. Alternatively, or in addition, regional differences in follicular atresia may have been involved. In some neonatal animals, the activation (early growth) of follicles on the medullary side of the primordial follicle zone was indeed seen. This pattern of follicular activation has been reported to occur in other species, with these activated follicles subsequently undergoing atresia (Hirshﬁeld and DeSanti, 1995). Fig. 4. Ovarian medulla and adjacent cortex from a neonatal Glossophaga soricina stained with hematoxylin. Multiple oocytes are evident within epithelial-lined channels of the rete ovarii (arrows). Scale bar 5 100 mm. POLARIZED OVARIES OF G. SORICINA 1445 Fig. 6. Portion of ovarian cortex from an adult Glossophaga soricina stained with hematoxylin. An arrowhead shows where the ovarian surface epithelium (OSE) covering the region containing primordial follicles changes to a more ordinary mesothelium (mt). Arrows indicate growing follicles on the medullary side of the primordial follicle zone (pf). Scale bar 5 100 mm. Fig. 5. Ovary obtained from an adult Glossophaga soricina on day 22 of her nonpregnant cycle and stained with hematoxylin and eosin. By this age, the ovary had become polarized, and primordial follicles were restricted to a zone (pf) on its medial side. cl, corpus luteum. The asterisk denotes a probable area of follicular rupture at ovulation. Scale bar 5 100 mm. There is evidence that the rete ovarii might be responsible for at least some of the follicular activation that is observed during ovarian development. Hence, the rete ovarii may contribute to ovarian polarization in G. soricina. In neonatal bats, epithelial cells of the rete ovarii appeared more active histologically than in the fetal specimens. This was suggested by the generally greater size of these cells and their nuclei in the neonatal specimens. Furthermore, activated oocytes—characterized by their increased size compared with oocytes in primordial follicles—were noted in and near elements of the rete ovarii. Previous studies have shown that the rete ovarii and medullary cords may be involved in early follicular development and oocyte activation in other mammals (Byskov, 1975; Byskov et al., 1977; Rajah et al., 1992; Hirshﬁeld and DeSanti, 1995). Presumably in all of these cases, where oocyte activation occurs early in life, it is followed by atresia. Ovulation can take place from anywhere on the ovarian surface in most mammals, but occurred only in the region of the ovary covered by the OSE in G. soricina. This phenomenon is not limited to G. soricina, but has also been noted in equids (mares, donkeys, zebras), armadillos (pp. 25, 362, and 381 in Mossman and Duke, 1973), and the cuis (Galea musteloides)—a cavid rodent (Weir and Rowlands, 1974). In these animals, only limited portions of the ovaries are covered by OSE, and these are the usual sites of ovulation. There are other species of micro- and megachiropteran bats that have a restricted distribution of OSE and primordial follicles (Rasweiler and Badwaik, 2000), raising the possibility that ovulation may be limited to a speciﬁc region of the ovary in these species as well. Interestingly, however, studies of two species of bats belonging to the same family (Phyllostomidae) as G. soricina—Desmodus rotundus (common vampire bat) and Carollia perspicillata (shorttailed fruit bat)—have established that their ovaries exhibit the more generalized pattern of ovarian organization common to most mammals (de Bonilla and Rasweiler, 1974; Quintero and Rasweiler, 1974). In these two species of bats, the OSE is much more extensive, and ovulation can apparently take place from most of the ovarian surface (Rasweiler, unpublished data). It has been unclear why the extent of the ovary covered by the OSE varies so markedly between mammalian species. Upon reﬂection, however, it seems reasonable that follicular rupture should occur at such a surface, because ovulation and internal fertilization are the culmination of a highly coordinated and carefully timed sequence of events. Therefore, any attempt by follicles to rupture at or near the hilus might delay the process, and/or decrease the likelihood of successful conceptions. Mechanisms to direct ovulation away from the hilus would seem to be particularly important when attachment to the broad ligament or mesovarium is broad, the female normally releases only a single oocyte during each cycle, and the species has limited breeding seasons. 1446 KOMAR ET AL. Fig. 7. Parallel sections of an ovary from a captive-bred adult Glossophaga soricina collected on day 1 post coitum (p.c.), stained with Masson’s trichrome procedure. a: The black arrow illustrates the point of ovulation through the primordial follicle zone (pf). The white arrow denotes a tangential section through the neck portion of an older, regressing CL (cl) from the previous cycle, also depicted in b. This image does not show the full width of the neck (which was narrower than the body of the CL), but indicates that follicle rupture presumably had occurred through the overlying ovarian surface epithelium (OSE). b: The box denotes early growing follicles on the medullary side of the primordial follicle zone. b, bursa; rf, newly ruptured follicle; mt, mesothelium. Scale bar 5 150 mm. G. soricina clearly exhibits the ﬁrst two characteristics and at least some populations breed seasonally, with a maximum female reproductive potential of 2 young/year (Fleming et al., 1972; Willig, 1985; Zortea, 2003). That these characteristics are shared to a large extent by a variety of unrelated species with polarized ovaries raises the possibility that the direction of ovulation toward the OSE may be a primitive mammalian reproductive characteristic that has been retained to promote successful oocyte escape from the ovary. These considerations may explain why there is evidence that the OSE also inﬂuences ovulation in some of the more commonly studied mammals, such as rabbits, sheep, and pigs. In several species, the OSE has been identiﬁed as a source of proteolytic enzymes involved in stigma formation (Cajander and Bjersing, 1975; Colgin and Murdoch, 1997; reviewed by Murdoch and McDonnel, 2002). Raw- Fig. 8. Relative vascularity of ovarian zones in adult Glossophaga soricina as revealed by the immunolocalization of factor VIII-related antigen in vascular endothelial cells. Endothelial cells containing factor VIII-related antigen have been labeled by the brown reaction product. Tissues were counterstained with hematoxylin. a,c: Blood vessels lined by reactive endothelial cells (examples identiﬁed by arrowheads) were more common around growing follicles and in major stromal partitions between clusters of primordial follicles. b: Relatively few such vessels were seen in among the primordial follicles. pf, primordial follicle zone. Scale bars 5 100 mm in a, 50 mm in b, 200 mm in c. son and Espey (1977) investigated the role of the OSE in ovulation by removing this epithelial layer from one ovary in rabbits after mating, but before ovulation, leaving the second ovary as an intact control. In that study, ovulation rates were reduced 47% and 45%, if the ovaries were denuded of the OSE 2 or 5 hr p.c., respectively (Rawson and Espey, 1977). Removal of the OSE in other species such as frogs, sheep, and pigs, also disrupted ovulation (reviewed by Murdoch and McDonnel, 2002). These ﬁndings illustrated that the OSE was involved in the ovulatory process and that compromising its integrity reduced ovulation efﬁciency. Because of the polarization of its ovaries, G. soricina should provide a unique model to further deﬁne the possible role(s) of the OSE in controlling follicular development and rupture. It is noteworthy that over 90% of ovarian cancers are believed to be derived from the OSE (reviewed by Auersperg et al., 2001). Therefore, more thorough elucidation of the function(s) of this epithelial layer may enhance POLARIZED OVARIES OF G. SORICINA our understanding of the causes of ovarian cancer. The identiﬁcation of additional, function-related markers for the OSE might also ultimately lead to better detection and treatment methods. The ovaries of G. soricina are of further interest because initial follicular activation and growth were limited to a very narrow area along the medullary side of the primordial follicle zone and away from the OSE. The basis for this ﬁnding remains speculative, but several reasonable possibilities come to mind. As the primordial follicle zone was found to be relatively undervascularized (present study; ovarian cortex; Herrmann and Spanel-Borowski, 1998), follicular activation along its medial edge could be stimulated by a higher oxygen tension and/or exposure to blood-borne factors due to the presence of more vascular elements in the medullary region of the ovary. Some factor(s) required for follicular activation might also be produced in the latter region by the vascular endothelial cells. It is now accepted that endothelial cells in different regions can vary substantially in their functional and synthetic capabilities. Finally, we must consider the possibility that the OSE may play a critical role in maintaining primordial follicles in a quiescent state. During ovarian development, the OSE may induce the formation of a distinct zone or microenvironment that promotes maintenance of primordial follicles and inhibits follicular activation. Alternatively, the OSE may continue to produce a factor that inhibits follicular activation. In summary, the current study better deﬁned several unique characteristics of the ovaries in G. soricina. Their polarized structure may be exploited to answer important questions regarding ovarian development and function. These would include the investigation of local factors involved in the long-term maintenance of primordial follicles in a quiescent state, recruitment of select follicles for growth and development, and how the OSE inﬂuences the process of ovulation. Finally, this animal model has the advantage of being readily maintained, bred, and worked with at low cost in modest facilities. ACKNOWLEDGMENTS The authors thank the personnel in the Department of Life Sciences, University of the West Indies, St. Augustine, Trinidad (particularly Dr. Indira OmahMaharaj) for generously providing assistance and the use of Departmental facilities during the course of our ﬁeldwork. We also thank Simeon Williams, Dr. Karen Sears, Dr. Scott Weatherbee, and Dr. Lee Niswander for ﬁeld assistance. We thank the Wildlife Section-Forestry Division, Ministry of Agriculture, Land and Marine Resources of the Republic of Trinidad and Tobago for providing required collecting and export permits. R.R.B. was funded by the NSF foundation, and J.J.R. was funded by a grant from the Population Council. LITERATURE CITED Alveraz J, Willig MR, Jones K Jr, Webster WD. 1991. Glossophaga soricina. Mamm Species 379:1–7. Auersperg N, Wong AST, Choi K-C, Kang SK, Leung PCK. 2001. Ovarian surface epithelium: biology, endocrinology, and pathology. Endocr Rev 22:255–288. 1447 Badwaik NK, Rasweiler JJ IV. 2002. Pregnancy. In: Crichton EG, Krutzsch PH, editors. Reproductive biology of bats. London: Academic Press. p 221–294. Byskov A-G. 1975. The role of rete ovarii in meiosis and follicle formation in the cat, mink, and ferret. J Reprod Fertil 45:201–209. Byskov AG, Skakkebaek NE, Stafanger G, Peters H. 1977. Inﬂuence of ovarian surface epithelium and rete ovarii on follicle formation. J Anat 123:77–86. Cajander S, Bjersing L. 1975. Fine structural demonstration of acid phosphatase in rabbit germinal epithelium prior to induced ovulation. Cell Tissue Res 164:279–289. Colgin DC, Murdoch WJ. 1977. Evidence for a role of the ovarian surface epithelium in the ovulatory mechanism of the sheep: secretion of urokinase-type plasminogen activator. Anim Reprod Sci 47:197–204. Cretekos CJ, Weatherbee SD, Chen C-H, Badwaik NK, Niswander L, Behringer RR, Rasweiler JJ IV. 2005. Embryonic staging system for the short-tailed fruit bat, Carollia perspicillata, a model organism for the mammalian order Chiroptera, based upon timed pregnancies in captive-bred animals. Dev Dyn 233:721– 738. de Bonilla H, Rasweiler JJ IV. 1974. Breeding activity, preimplantation development, and oviduct histology of the short-tailed fruit bat, Carollia, in captivity. Anat Rec 179:385–403. Evarts JJ, Rasweiler JJ IV, Behringer RR, Hennighausen L, Robinson GW. 2004. A morphological and immunohistochemical comparison of mammary tissues from the short-tailed fruit bat (Carollia perspicillata) and the mouse. Biol Reprod 70:1573– 1579. Fleming TH, Hooper ET, Wilson DE. 1972. Three Central American bat communities: structure, reproductive cycles and movement patterns. Ecology 53:555–569. Hamlett GW. 1934. Uterine bleeding in a bat Glossophaga soricina. Anat Rec 60:9–17. Hamlett GW. 1935. Notes on the embryology of a phyllostomatid bat. Am J Anat 59:327. Herrmann G, Spanel-Borowski K. 1998. A sparsely vascularised zone in the cortex of the bovine ovary. Anat Histol Embryol 27: 143–146. Hirshﬁeld AN, DeSanti AM. 1995. Patterns of ovarian cell proliferation in rats during the embryonic period and the ﬁrst three weeks postpartum. Biol Reprod 53:1208–1221. Ingram DL. 1962. Atresia. In: Zuckerman S, editor. The ovary. Vol. 1. London: Academic Press. p 247–273. Konishi I, Fujii S, Okamura H, Parmley T, Mori T. 1986. Development of interstitial cells and ovigerous cords in the human fetal ovary: an ultrastructural study. J Anat 148:121–135. Mossman HW, Duke KL. 1973. Comparative morphology of the mammalian ovary. 1st ed. Madison, WI: The University of Wisconsin Press. Murdoch WJ, McDonnel AC. 2002. Roles of the ovarian surface epithelium in ovulation and carcinogenesis. Reproduction 123:743– 750. Peters H, McNatty KP. 1980. The ovary. London: Granada Publishing. Quintero F, Rasweiler JJ IV. 1974. Ovulation and early embryonic development in the captive vampire bat, Desmodus rotundus. J Reprod Fertil 41:265–273. Rajah R, Glaser EM, Hirshﬁeld AN. 1992. The changing architecture of the neonatal rat ovary during histogenesis. Dev Dyn 194:177–192. Ramirez N, Sobrevila C, De Enrech NX, Ruiz-Zapta T. 1984. Floral biology and breeding system of Bauhinia benthamiana (Leguminosae), a bat-pollinated tree in Venezuelan ‘‘Llanos’’. Am J Bot 71:273–280. Rasweiler JJ IV. 1972. Reproduction in the long-tongued bat, Glossophaga soricina. I. Preimplantation development and histology of the oviduct. J Reprod Fertil 31:249–262. Rasweiler JJ IV. 1973. Care and management of the long-tongued bat, Glossophaga soricina (Chiroptera: Phyllostomatidae), in the laboratory, with observations on estivation induced by food deprivation. J Mammal 54:391–404. 1448 KOMAR ET AL. Rasweiler JJ IV. 1974. Reproduction in the long-tongued bat, Glossophaga soricina. II. Implantation and early embryonic development. Am J Anat 139:1–35. Rasweiler JJ IV. 1975. Maintaining and breeding neotropical frugivorous, nectarivorous and pollenivorous bats. In: Duplaix-Hall N. editor. Small mammals in captivity. International Zoological Yearbook 15:18–30. Regent’s Park, London: The Zoological Society of London. Rasweiler JJ IV. 1977. The care and management of bats as laboratory animals. In: Wimsatt WA, editor. Biology of bats. Vol. III. New York: Academic Press. p 519–617. Rasweiler JJ IV. 1979a. Bats as models in studies on folliculogenesis, menstruation, early pregnancy, and sperm survival. In: Alexander NA, editor. Animal models for research on contraception and fertility. Hagerstown, MD: Harper and Row Publishers. p 437–446. Rasweiler JJ IV. 1979b. Differential transport of embryos and degenerating ova by the oviducts of the long-tongued bat, Glosophaga soricina. J Reprod Fertil 55:329–334. Rasweiler JJ IV. 1993. Pregnancy in Chiroptera. J Exp Zool 266:495–513. Rasweiler JJ IV, Badwaik NK. 2000. Anatomy and physiology of the female reproductive tract. In: Crichton EG, Krutzsch PH, editors. Reproductive biology of bats. London: Academic Press. p 157– 219. Rasweiler JJ IV, de Bonilla H. 1972. Laboratory maintenance methods for some nectarivorous and frugivorous phyllostomatid bats. Lab Anim Sci 22:658–663. Rawson JMR, Espey LL. 1977. Concentration of electron dense granules in the rabbit ovarian surface epithelium during ovulation. Biol Reprod 17:561–566. Silva SSP, Peracchi AL. 1999. Visits of bats to ﬂowers of Lafoensia glyptocarpa Koehne (Lythraceae). Rev Brasil Biol 59:19–22. Weir BJ, Rowlands AL. 1974. Functional anatomy of the hystricomorph ovary. Symp Zool Soc Lond 34:303–332. Willig MR. 1985. Reproductive patterns of bats from Caatingas and Cerrado biomes in northeast Brazil. J Mammal 66:668– 681. Zortea M. 2003. Reproductive patterns and feeding habits of three nectarivorous bats (Phyllostomidae: Glossophaginae) from the Brazilian Cerrado. Brazil J Biol 63:159–168.