| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Laboratoire de Physiologie et Physiopathologie, CNRS-UMR 7079, Université Paris VI, 75252 Paris cedex 05, France
ABSTRACT
Follicle histogenesis, in which follicles arise from fragmenting ovigerous cords, is a poorly understood mechanism that is strictly dependent upon the presence of germ cells. Our previous studies have shown that severely germ cell-depleted rat ovaries after fetal gamma-irradiation display modifications of follicular endowment and dynamics during the immature period. The primordial follicle stock was absent and the follicles with primary appearance remained quiescent longer than in control ovaries during the neonatal period. The aim of the present work was to analyze the initial steps of follicle histogenesis, and to investigate the etiology of the alterations observed in the development of irradiated ovaries. Just after birth, we observed, in addition to sterile ovigerous cords, the emergence of the first follicles which exhibited several abnormal features as compared to those of control ovaries. Most of the follicles appeared as primary follicles, as they were composed of a layer of cuboidal-shaped granulosa cells surrounding an enlarged oocyte. Interestingly, the granulosa cells of these primary-like follicles did not proliferate and did not express the genes for anti-Müllerian hormone (Amh) or bone morphogenetic protein receptor type II (Bmpr2), both of which are normally expressed from the primary stage onwards. In contrast, the oocytes strongly expressed the gene for growth and differentiation factor 9 (Gdf9), which is normally upregulated from the primary follicle stage onwards, which suggests an uncoupling of granulosa cell development from oocyte development. In addition, irradiated ovaries displayed a higher frequency of follicles that contained 2 or 3 oocytes, which are also referred to as multi-oocyte follicles (MOFs). Examination at the time of follicle histogenesis indicated that MOFs arise from incomplete ovigerous cord breakdown. Taken together, the results of this study indicate that severe perturbations of follicular histogenesis take place following irradiation and massive germ cell depletion during fetal life. In addition to the classically described sterile cords, we have pointed out the differentiation of MOFs and primary-like quiescent follicles, which finally evolve into growing follicles and participate in ovarian function. We propose that these phenotypes are closely correlated to the proportion of granulosa cells to oocytes at the time of neonatal follicle histogenesis.
developmental biology, follicle, follicular development, granulosa cells, ovary
INTRODUCTION
Initial follicular histogenesis is the developmental process that results in the differentiation of the primordial follicles, that constitute the follicular stock, which serves as a reservoir for the reproductive lifespan of the adult. During this process, fetal ovigerous (or ovarian) cords, made up of numerous oocytes and epithelial pregranulosa cells, split off into follicular units, the primordial follicles, which are composed of a single oocyte surrounded by squamous granulosa cells [1–4]. In the rat, follicular histogenesis is initiated in the core of the ovary in the hours following birth and spreads throughout the gonad in the subsequent days. As soon as they are formed, follicles located in the core of the ovary appear as primary follicles that contain cuboidal granulosa cells and initiate their growth [5]. As a result, during the neonatal period, the histo-architecture of the ovary is characterized by the presence of primary follicles located in the center and primordial follicles located at the periphery of the gonad.
The formation of follicular units involves complex remodeling of the pre-existing ovigerous cord basal membrane (BM) and acute synthesis of BM components, which assemble into the newly formed BM that delineates the follicular units [3]. This re-structuring is ensured by the important contribution of epithelial cells, which undergo a decisive step of differentiation. Interestingly, during this process, a wave of oocyte apoptosis takes place. It has been proposed that the massive disappearance of oocytes permits the re-organization of epithelial cells around the surviving oocytes and allows their differentiation into granulosa cells [3, 6, 7].
It has been known for a long time that oocyte plays an important role in ovarian differentiation [8, 9, for review 10]. When the ovary is severely depleted of germ cells during fetal life for either genetic or experimental reasons, the follicular units do not emerge from the sterile ovigerous cords [8–11]. These epithelial structures maintain the morphological and biochemical characteristics of the ovigerous cord [8, 12] and regress progressively from the ovary leaving a streak gonad [8, 13]. When massive but incomplete germ cell depletion was induced by fetal 
-irradiation of rat females at 15.5 days postconception (DPC) destroying more than 90% of oogonia, follicle assembly takes place only in the ovarian regions in which oocytes are present [13–15]. Irradiated females, although fully fertile at the beginning of their reproductive lives, exhibit premature ovarian failure as a consequence of severe germ cell depletion [13].
The detailed analysis of follicular populations in irradiated females has revealed that follicular endowment and dynamics are modified during the immature period. The primordial follicle stock is absent as soon as 3 days post-natal (dpn) [13], as well as afterwards, in immature and adult life [13, 14]. During the neonatal period, the growth of the initial follicular waves is delayed, and we have shown that at 3 dpn, in irradiated ovaries, follicles that are categorized as growing follicles due to their morphology are formed by granulosa cells which are not committed to the cell cycle since they are not stained for proliferation marker PCNA; these follicles are considered to be quiescent and remain in this state at least until 6 dpn [13]. This type of quiescent status is abnormal for follicles with primary or secondary morphology.
The aim of this study was to increase understanding of how these follicles form, to determine if they originate from primordial follicles that would interrupt their growth, and to assess the differentiation statuses of both germ cells and granulosa cells. For these purposes, in addition to the histological study, we have analyzed by in situ hybridization and immunohistochemistry the expression of molecular markers of differentiation and maturation of both germ cells and granulosa cells, in control ovaries and ovaries irradiated at 15.5 dpc.
MATERIALS AND METHODS
Animal Handling and Irradiation
Female rats of the Sprague Dawley strain (IFFA-CREDO, France) were mated overnight with males; the following day was considered to be 0 dpc. The animals were maintained on a 12L:12D schedule. Pregnant 15-dpc female rats were exposed to -irradiation using a 60Co source and a total dosage of 1.5 Gy, as previously described [13]. Birth, defined as 0 dpn, generally occurred during the night between 21 and 22 dpc. Control and irradiated females were injected intraperitoneally with 50mg/kg bromodeoxyuridine (BrdU) dissolved in saline. Animal procedures were performed according to the 1986 European Communities Directives and those of the Ministère de l'Agriculture et de la Forêt with approval from the experimental animal committee of the Institut Fédératif de Recherche 83 (Agreement A75-05-24).
Tissue Collection and Processing
Ovaries were collected daily from 0–3 dpn and thereafter at 6, 9, 12, 15, 21, and 28 dpn. For histological examinations of ovarian morphology and follicle counts, ovaries were fixed in Bouin fixative for at least 24 h, dehydrated in a graded series of ethanol, and paraffin-embedded using standard protocols. Five µm-thick sections were stained with hematoxylin-eosin or Tuchmann's blue. For in situ hybridization or immunohistochemistry, ovaries were fixed in 2% paraformaldehyde-PBS (pH 7.2) for 1 h at 4°C, cryoprotected in 18% sucrose in PBS, embedded in Tissue-Tek OCT compound (Miles Inc., Elkhart, IN), cut into 7 µm-thick sections, mounted onto 3-aminopropyltriethoxysilane (Sigma-Aldrich Corp., St. Louis, MO)-treated glass slides, and stored at –20°C.
Identification of Follicular Development Stages and Quantification of Multi-Oocyte Follicles
Follicles were identified according to the stages of follicular development, as described previously [13]. Follicles that contained more than one oocyte enclosed within a granulosa cell layer or layers were counted at 9, 15, 21 and 28 dpn for every fifth ovarian section. Atretic follicles were identified by the presence of pycnotic granulosa cells. Follicle numbers were analyzed using the SSCP software and are expressed as the mean ± SEM. Statistical significance was determined by one-way ANOVA, followed by Tukey post-hoc multiple comparison test. The differences were considered significant for P < 0.05.
In Situ Hybridization
The cDNAs used for the synthesis of the different riboprobes were obtained by RT-PCR, subcloned into the pGEMTeasy or pGEM3Z vector, and verified by DNA sequencing. The gene, together with their GenBank accession numbers and positions (nucleotides, nt) are: Ybx2, AF073954, nt 680-1326; Gdf9, X81899, nt 3–546; Fst, M31586, nt 196–635; Amhr2, NM_030998, nt 65–686; Foxl2, NM_012020, nt 215-1083; Lama1, NM_008480, nt 7101–7453; Mmp14, NM_031056, nt 278-1139; Amh, S988336, nt 1546–2198; Cyp19a1, M33986, nt 797-1487; Inhba, M37482.1, nt 1040–1483; Inha NM_012590.1, nt 256–765; Bmpr2 17–1297, nt 17–1297. Sense and antisense riboprobes were generated by in vitro transcription with digoxigenin-labeled deoxy-UTP (Roche, Mannheim, Germany) and the appropriate T7 or SP6 polymerase (Roche). In situ hybridization was performed as previously described [13].
Immunofluorescence and Immunohistochemistry
For the immunohistochemical analysis, sections from at least three control and three irradiated ovaries were used. All of the antibodies used were diluted in PBS plus 0.05% BSA. For the negative controls (not shown), the primary antibody was omitted. Fluorescence-labeled sections were mounted in para-phenylenediamine-containing medium, and double-labeled sections for in situ hybridization and immunocytochemistry were mounted in glycerol-gelatin (Sigma-Aldrich). The slides were observed under an epifluorescence microscope (Carl Zeiss). For immunodetection of collagen IV after the detection of the Y box protein 2 (Ybx2) transcript by in situ hybridization, the sections were washed in PBS, incubated with anti-collagen IV antibody (diluted 1/100) overnight at 4°C (AbCys, Paris, France), washed with PBS, and incubated with an Alexa 488-conjugated anti-rabbit IgG secondary antibody (Molecular Probes, Eugene, OR). Immunofluorescence detection of BrdU after Gdf9 transcript detection by in situ hybridization was performed on sections that were subjected to citrate buffer antigen retrieval (microwaved for 5 min) and post-fixed in 4% paraformaldehyde in PBS for 30 min at 4°C. The tissues were incubated with the anti-BrdU antibody (diluted 1/100; Roche) overnight at 4°C, washed with PBS, and incubated with an Alexa 488-conjugated anti-mouse IgG secondary antibody (Molecular Probes).
DEAD box polypeptide 4 (DDX4; also known as mouse Vasa homologue) and KIT were immunodetected sequentially. Sections were incubated with either anti-DDX4 (diluted 1/1000; kindly provided by T. Noce) or anti-KIT (diluted 1/150, C19; Santa Cruz Biotechnology, Santa Cruz, CA) primary antibody overnight at 4°C. Alexa 488-conjugated anti-rabbit serum was used as the first secondary antibody (Molecular Probes). After washing and blocking in PBS plus 10% BSA, the sections were incubated with the second primary antibody, either anti-KIT or anti-DDX4, overnight at 4°C. After washing in PBS, the second primary antibody was detected using an Alexa 566-conjugated anti-rabbit IgG secondary antibody (Molecular Probes). Labeling was performed in both orders for DDX4 and KIT labeling, and the labeling of serial sections gave similar results (data not shown).
Double immunolabeling was performed for keratin 8 (KRT8, clone LE-41, kindly provided by Dr. F.B. Lane) and DDX4. Frozen tissue sections were delipidized in chloroform, rehydrated in PBS, and endogenous peroxidase activity was blocked with 3% hydrogen peroxide for 5 min. Sections were incubated with anti-KRT8 primary antibody (diluted 1/5) for 1 h at room temperature, and thereafter for 1 h with anti-mouse biotinylated antibody (diluted 1/300, RPN 1001; Amersham Biosciences, Piscataway, NJ) and for 30 min with a peroxidase-conjugated streptavidin-horseradish complex (LSAB+ Kit; DAKO Corp., Carpinteria, CA). The reaction product was developed with 3,3'-diaminobenzidine tetrahydrochloride (DAKO). After washing in PBS, subsequent immunofluorescence was performed with the anti-DDX4 primary antibody (diluted 1/1000) incubated overnight at 4°C. After washing, the slides were incubated with an Alexa 488-conjugated anti-mouse IgG secondary antibody (Molecular Probes).
Semi-Thin Sections
Ovaries were recovered at 1 dpn from control and in utero-irradiated Sprague Dawley female rats. The ovaries were dissected from the ovarian capsule, fixed for 24 h at 4°C in 2% paraformaldehyde, 0.5% glutaraldehyde in 0.1M cacodylate buffer (pH 7.2), and post-fixed for 2 h at room temperature in 2% osmium tetroxide (OsO4). The tissues were then dehydrated through a graded series of ethanol and embedded in Epoxy Embedding Resin (Sigma). One micron-thick sections were stained with methylene blue and safranin red.
RESULTS
Follicle Ontogeny in Fetal Irradiated Ovaries
We first compared the time course of follicular histogenesis in irradiated and control ovaries. In both cases, there was no evidence of follicle differentiation before birth. Despite massive depletion of the germ cells, fetal irradiated ovaries displayed the same histo-architecture as control ovaries (Fig 1, A–D). They contained fibronectin-negative ovigerous epithelial cords (Fig 1, B and D), which were formed by preganulosa cells and oocytes that had not yet attained the dictyate stage (Fig. 1, A and C). From 0 dpn onwards, the initiation of folliculogenesis was evidenced, as reported previously [3], by the progressive deposition of a continuous basal membrane in the narrow clefts that insinuate into remodeling ovigerous cords (arrowheads, Fig. 1, F and H). As a consequence, the follicles gradually differentiated and appeared as epithelial units delineated by a collagen IV labeled BM and contained a unique oocyte, as revealed by the detection of Ybx2 transcripts (arrows, Fig. 1, E–H and K–N). In control ovaries, follicle histogenesis gave rise to primary follicles in the core of the ovary (Fig. 1, I, K, and L) and to a pool of primordial follicles in the periphery of the ovary (Fig. 1, K and L). In irradiated ovaries, splitting of the ovigerous cords occurred in areas that contained the few surviving oocytes and did not display a specific localization (Fig. 1, G, H, J, M, and N). As soon as they were formed, the first follicles that differentiated in the irradiated ovaries appeared as primary-like follicles (Fig. 1J), which resembled the primary follicles in the centers of the control ovaries (Fig. 1I). These follicles were formed by a ring of cuboidal granulosa cells and a single oocyte, the size of which (35–50 µm) was similar to that of oocytes in the control primary follicles. Fetal -irradiation did not alter the chronology of follicular histogenesis but modified significantly the characteristics of follicle supply. Indeed, although some primordial follicles were occasionally observed (for data from 9 dpn, see [13]) the primordial follicle stock present at the periphery of the control ovary (Fig. 1, K and L) was not observed in the irradiated ovary (Fig. 1, M and N).
|
To evaluate the cellular relationships during the time course of follicle histogenesis in control and irradiated ovaries, we compared the distributions of germ cells, and epithelial pregranulosa and granulosa cells. Germ cells were identified by staining for the DDX4 protein [16], and epithelial cells were identified by their staining for KRT8 [17]. At the periphery of the 0-dpn control ovaries, DDX4-positive oocytes were grouped together in clusters surrounded by a layer of epithelial cells that expressed KRT8 (Fig. 2A, arrow). In contrast, in the core of the ovary near the hilum, the oocytes were more distant from each other and each one was encircled by epithelial pregranulosa cells (Fig. 2A, arrowhead). In irradiated ovaries, the partially sterilized ovigerous cords were thin, and the scarce isolated oocytes were surrounded by numerous KRT8-positive epithelial cells (Fig. 2B, arrowheads). Estimation of the relative proportion of oocytes and epithelial cells performed on semi-thin sections of 1-dpn ovaries confirmed that in the control ovaries, the oocyte-to-epithelial cell ratio was significantly higher at the periphery (approximately 2–3 epithelial cells per oocyte per section; Fig. 2C) than in the core of the ovary (4–5 epithelial cells per oocyte per section; Fig. 2D). In the irradiated ovaries, irrespective of the region analyzed, oocytes, regardless of whether they were isolated or arranged into small clusters, were surrounded by numerous epithelial cells (>5–6 epithelial cells per oocyte per section; Fig. 2E). Therefore, oocyte-to-epithelial cell ratio in irradiated ovaries was either similar to or lower than that observed in the core of the control ovaries.
|
In our comparative study of collagen IV-labeled BM deposition and Ybx2 transcript-labeled oocyte distribution, partially fragmented ovigerous cords were observed more frequently in irradiated ovaries than in control ovaries. In these structures, the clefts between the epithelial cells did not join (Fig. 1, M and N, open arrows) and consequently, there was no individualization of the follicular units. These structures, which contained several oocytes, could be observed throughout the development of the ovary (Fig. 1, M and N, Fig. 3, B–E). Progressively, these structures appeared as multi-oocyte follicles (MOFs), which were delineated by a continuous BM that separated the granulosa cells from the external mesenchymal tissue. Interestingly, we occasionally observed incompletely separated follicles that were joined by a bridge of epithelial cells (Fig. 3E). From 15 dpn, the MOFs contained 2 to 4 oocytes (Fig. 3, F–H) and were present mainly at the preantral and small antral stages of development in irradiated ovaries (Fig. 3 F–H). It is worth noting that there were many more MOFs in the irradiated than in the control ovaries (at 21 dpn: 2.83 ± 0.54 in irradiated ovaries, n = 7 vs. 0.15 ± 0.14 per ovary in the control, n = 6; P < 0.0001). From 28 dpn onward, the MOFs were either healthy or atretic and diminished in number.
|
Analysis of Oocyte Maturation
The previous observations of Beaumont [18] on meiosis progression between 16.5 and 19.5 dpc indicate that each meiotic stage lasts longer in 15.5-dpc irradiated ovaries than in control ovaries. This results in a much larger variety of meiotic stages for a given age during this period of time. We investigated the meiosis statuses of the oocytes in control and irradiated ovaries after birth. In the control ovaries at 0 dpn, numerous oocytes exhibited the characteristics of the diplotene stage, i.e., the chromatin was homogenous and the nucleolus was clearly visible (Fig. 4A, arrowheads). In the irradiated ovaries, most oocytes displayed a spumescent and granular nucleus, which is characteristic of the pachytene stage (Fig. 4B, open arrows). However, diplotene oocytes were occasionally observed in the irradiated ovaries (Fig. 4B, arrowhead).
|
To assess more precisely the maturation stage of the germ cells in the irradiated ovaries, we compared the expression of different molecular markers for germ cell lineage differentiation in control and irradiated ovaries (Fig. 4, C–H). These markers included DDX4, which is expressed by germ cells at all developmental stages [16], KIT, which is expressed by oocytes from the diplotene stage onward [19], and Gdf9, which is a growth factor synthesized by the oocytes within growing follicles [20]. Double-immunolabeling confirmed that in control ovaries, the vast majority of germ cells (DDX4-positive cells) localized in the central part of the ovary had attained the diplotene stage, as illustrated by KIT staining (Fig. 4, C and E). In contrast, in the irradiated ovaries, KIT immunolabeling was faintly discernable in the DDX4-positive cells (Fig. 4, D and F). These results suggest that in irradiated ovaries at birth, much like during fetal life [18], some oocytes display delayed meiotic progression. Interestingly, certain oocytes expressed Gdf9, similar to the oocytes present in the growing follicles in the centers of the control ovaries (Fig. 4, G and H). They, thus, have initiated growth as the most mature oocytes of control ovaries. Later in development, while a gradient of oocyte maturation was observed in control ovaries, with small dormant oocytes weakly expressing Gdf9 at the periphery and large growing oocytes displaying a high level of Gdf9 expression in the center of the ovary (Fig. 5A), no such gradient was observed in the irradiated ovaries. All the oocytes were strongly stained for Gdf9 transcripts (Fig. 5C) and their sizes were similar to those of oocytes in the growing follicles present in the cores of the control ovaries (Fig. 5, B and D). Therefore, their development during the neonatal period was similar to that of the most mature oocytes in the control ovaries.
|
Since GDF9 is involved in granulosa cell proliferation [20], and since we have previously demonstrated that the follicles in irradiated ovaries are quiescent at 3 dpn and 6 dpn [13], we compared Gdf9 expression using in situ hybridization (Fig. 5, E–G) and cell proliferation using immunodetection of BrdU incorporation (Fig. 5, F and H). In the 6-dpn control ovaries, the growing follicles contained an oocyte that highly expressed Gdf9 and several proliferating BrdU-positive granulosa cells (Fig. 5, E and F, arrowheads). In 6-dpn irradiated ovaries, the majority of the follicles contained a large oocyte that strongly expressed Gdf9 surrounded by cuboidal granulosa cells that were negative for BrdU (Fig. 5, G and H, arrows). Growing follicles with proliferating granulosa cells were rarely observed (Fig. 5, G and H, arrowheads). These data suggest that in irradiated ovaries, oocytes mature independently of the quiescent status of the granulosa cells.
Analysis of Granulosa Cell Differentiation
To analyze further the processes of follicle differentiation and maturation, we compared the expression patterns of several differentiation markers of the granulosa cell lineage in control and irradiated ovaries during the fetal and neonatal periods (Fig. 6). As a first step, we studied factors that are characteristic of the granulosa cell lineage, the expression of which is initiated in pregranulosa cells of fetal ovigerous cords, such as follistatin (Fst) [21], AMH-receptor type II (Amhr2) [22], and Forkhead box L2 (Foxl2) [23]. At 17 dpc, transcripts encoding these proteins were detected in the ovigerous cords of both control and irradiated ovaries (Fig 6, C–H). The expression pattern was similar to that of KRT8 staining of the epithelial pregranulosa cells of fetal ovigerous cords [17] (Fig 6, A and B). The apparent weaker staining in the control compared to the irradiated gonads results from the difference in the number of unstained germ cells, i.e., a very high number in the control ovaries and low number in the irradiated ovaries. Interestingly, from birth onward, in the irradiated ovaries, as previously described for control ovaries at the onset of follicle histogenesis [3], epithelial cells that differentiated into granulosa cells displayed upregulation of genes that encode the laminin 
1 subunit (lama1) (Fig. 6I), matrix metallopeptidase 14 (Mmp14) (Fig. 6J), and urokinase-type plasminogen activator (data not shown). In contrast, regarding the expression of functional markers of follicular maturation, noticeable differences were observed between the control and irradiated ovaries. In the controls, the expression of inhibin subunit (Inha) transcripts detected in granulosa cells of primordial follicles (Fig. 6K) displayed upregulation at the moment of transition into the primary follicle stage (Fig. 6K) [24]. In irradiated ovaries, Inha expression, which was negative in sterile ovigerous cords (Fig. 6L), was quite similar in newly formed primary-like follicles and in the primordial follicles in the controls (Fig. 6, K and L). We also analyzed the expression of markers characteristic of entry into the growing stage in immature animals from the primary follicle stage onwards, such as anti-Müllerian hormone (Amh), inhibin ßA subunit (Inhba), and aromatase (Cyp19a1) [13, 24–26]. Whereas, at 3 dpn, the granulosa cells of all the growing follicles expressed Amh in the control ovaries, those from the primary-like or secondary-like follicles of irradiated ovaries did not express Amh (Fig. 6, M and N). The AMH transcript and protein were detected in irradiated ovaries from 6 dpn onward (data not shown). In addition, whereas the granulosa cells of growing follicles expressed Inhba and Cyp19a1 in control ovaries at 6 dpn, no staining was detected in follicles of comparable size in the irradiated ovaries (Fig. 6, O–R and [13]) until 9 dpn. As we observed that the granulosa cells in primary-like follicles did not proliferate although the oocytes strongly expressed Gdf9, which is known to be involved in granulosa cell proliferation [20], we hypothesized that granulosa cells are not sensitive to the GDF9 signal. To test this hypothesis, we tracked by in situ hybridization the expression of bone morphogenetic protein receptor type II (Bmpr2), which is one of the type II receptors involved in the GDF9 signaling pathway in granulosa cells [27]. Bmpr2 transcripts were expressed by granulosa cells from the primary stage onward in control ovaries (Fig. 6S), but not in the primary-like or secondary-like follicles of 6-dpn irradiated ovaries (Fig. 6T). Taken together, these data indicate that there is a delay in the maturation of the granulosa cells of primary follicles in irradiated ovaries compared to control ovaries.
|
DISCUSSION
In this study, histological analyses, as well as the comparison of germ and epithelial cell maturation markers in neonatal period have shown that in irradiated ovaries, follicle histogenesis immediately gives rise to follicles that are classified as primary or secondary, instead of the primordial follicle pool. In normal ovarian development, similar follicles, characterized by a large oocyte and a ring of cuboidal epithelial cells that form one layer or more, have been described in the centers of 3-dpn control ovaries, just after follicle histogenesis [5]. These follicles enter growth immediately after differentiation and constitute the initial follicular waves that are destined to disappear during the wave of atresia that begins at 18 dpn [5, 28]. The follicles appear in a region of the ovary in which the oocyte-to-epithelial cell ratio is lower than in the periphery. Irradiation, which deprives to a large extent the fetal ovigerous cords of their germ cells, leads to a pattern of follicle histogenesis that appears to be similar to that observed in the centers of control ovaries. Indeed, in irradiated ovaries, oocytes are systematically surrounded by an abnormally high number of epithelial cells, irrespective of the region of the ovary. However, the fate of these primary-like follicles observed in irradiated ovaries differs from that of the first follicular waves in the controls, since, as we have described previously, they do not immediately enter growth, and are consequently saved from the massive wave of atresia that begins at the end of the third week [13].
During fetal development in irradiated ovaries, epithelial pregranulosa cells display the same pattern of expression of granulosa cell lineage markers as those in control ovaries. In the hours following birth, as follicle histogenesis is initiated, epithelial cells that differentiate into granulosa cells demonstrate upregulation of lama1and Mmp14 expression, as previously described for control ovaries [3]. These regulatory events and the downregulation of LIM homeobox protein 9 (Lhx9) expression [12], which occur concomitantly with the commitment of epithelial cells to the granulosa cell differentiation pathway [12] and the basal membrane remodeling associated with follicle histogenesis [3], take place in both control and irradiated ovaries. Interestingly, the further differentiation of granulosa cells is delayed in irradiated ovaries compared to control ovaries. The expression of biochemical markers that characteristically appear at the primary stage in neonatal ovaries, such as Amh, Inhba, and Cyp19a1 [13, 24–27, 29], is not observed in primary-like follicles of irradiated ovaries during the neonatal period. As for Inha, the level of transcript expression in newly formed primary-like follicles of irradiated ovaries is quite similar to that of primordial follicles in control ovaries. Taken together, these data (Table 1) show that although the morphological aspects of the neo-formed follicles in irradiated ovaries correspond to those of primary or secondary follicles, the biochemical characteristics of granulosa cells correspond more closely to those of primordial follicles
|
Dysgenic ovaries, which are the result of fetal depletion of germ cells following irradiation or treatment with anti-mitotics, such as busulphan, have long been used as experimental models to assess the management of the primordial follicle stock [14, 15, 30]. Observations made in immature and young adult animals have lead to contradictory interpretations that variously suggest acceleration or slowing of the release of follicles from the primordial follicle stock in response to an experimental reduction in the size of the follicular reserve [14, 30]. Although the process of follicle histogenesis is the limiting step in the acquisition of ovarian functionality, none of these studies have analyzed the timing and modalities of follicle assembly in ovaries depleted of more than 90% of their oocyte pool. Based on our present observations, we suggest that in irradiated ovaries, even at birth, there is no real follicle stock, and that the primary follicles observed during the first week of life do not result from a transition from primordial to primary follicles. It seems likely that this phenomenon also occurs in ovaries treated with busulphan during fetal life, as their phenotypes are very similar to those of in utero-irradiated ovaries [30]. Thus, it is inappropriate to utilize these experimental models to elucidate the mechanisms that regulate the recruitment of follicles from the follicle reserve and their entry into the growing pool [30]. Our own observations have shown that in irradiated ovaries, the very restricted number of primordial follicles does not decrease in a statistically significant fashion before puberty, and that primary-like and secondary-like follicles may be considered as the oocyte reserve [13].
Multi-oocyte follicles, previously called poly-ovular follicles, have been described in several species, such as rabbits, humans, rats, and some mouse strains [31–35]. The hypothetical etiologies of MOFs differ according to the age at which they are observed. In infant humans and in young immature rats, MOFs are supposed to originate from the persistent remains of pairs or groups of germ cells from the ovigerous cords that have failed to become separated into mono-oocyte follicles [32, 33]. In adults, they are believed to form from the fusion of adjacent follicles [36]. Our present study of follicular formation and fate in irradiated ovaries suggests that MOFs may well originate from incompletely separated follicular units. Thus, we propose a model for follicle histogenesis in partially sterilized ovaries (Fig. 7). When ovigerous cords are completely sterilized, they do not split off and they retain fetal characteristics. When very few oocytes are present within an ovigerous cord, the process of fragmentation fails to occur or to reach completion, so that several oocytes remain in a single epithelial structure, thereby forming a MOF. When the density of oocytes is sufficient for follicle histogenesis to occur, the relatively high number of epithelial cells results in the differentiation of follicles, which must be classified as primary or secondary follicles. This scheme raises questions regarding the cellular interactions between the oocyte and the surrounding epithelial cells. Considering that the oocyte is indispensable for follicle histogenesis and that the epithelial cells of ovigerous cords that differentiate into granulosa cells play a major role during this process [3, 37], the oocyte may well send a morphogenic signal to epithelial cells to induce their reorganization and BM remodeling. The in utero-irradiated ovary model suggests the existence of a gradient of sensitivity to this signal. Indeed, it appears that epithelial cells need to be close to the oocyte to perceive this signal. In addition, they also must receive inductive information from the adjacent mesenchymal tissue to acquire a characteristic polarized cell phenotype and to complete follicle histogenesis [3]. In massively sterilized ovigerous cords, an imbalance between the different cellular categories, namely oocytes, epithelial cells, and mesenchymal cells, could explain the formation of multi-oocyte and primary-like follicles.
|
Since irradiation mainly destroys cells that are in mitosis, surviving germ cells are considered to be those that have entered the mitotic arrest that precedes meiosis [18]. Nevertheless, the observation of an increased mitotic index of germ cells 24 h after irradiation has been interpreted as a regenerating mechanism that may allow oogonia to re-enter the mitotic cycle and proceed beyond the deleterious effects of the irradiation [18]. In line with this hypothesis is the observation that during the days following irradiation, each stage of meiosis prophase is longer in irradiated ovaries than in control ovaries, resulting in a larger variety of meiosis stages for any given age [18]. For instance, at 19.5 dpc, while the predominant oocyte stage is the zygotene in control ovaries, in irradiated ovaries, leptotene, zygotene, and pachytene oocytes appear in similar numbers [15, 18]. Our histological observations and the comparison of DDX4 and KIT staining suggest that this type of time-lag in meiotic progression may still be present at 0 dpn. Nevertheless, since by this time certain oocytes have already attained the dictyate stage and this delay is observed for only a small proportion of the oocytes, it seems unlikely this time-lag is responsible for the anomalies of follicle histogenesis.
Interestingly, our results show that, in irradiated ovaries, oocytes initiate growth and highly express Gdf9, whereas granulosa cells display delayed differentiation and proliferation. Therefore, there is a discrepancy between the differentiation status of the oocyte and that of granulosa cells in irradiated ovaries. The uncoupling of oocyte and granulosa cell maturation has already been described in other experimental models. In mice Foxl2-deficient that display ovarian failure [38, 39], primordial follicle formation is impaired, possibly as a result of a basal membrane deposition defect [39]. Epithelial pregranulosa cells fail to proliferate, although oocytes enter growth and express high levels of Gdf9 [38, 39]. In mice that lack connexin 43, which is a constitutive protein of the gap junctions between granulosa cells, follicles do not grow beyond the one-layer stage and have the appearance of primary follicles [40, 41]. Oocytes, which are ten-fold less numerous than in control ovaries, continue to grow, albeit more slowly than in controls, despite the blockage of folliculogenesis [40, 41]. In addition, Gdf9-knockout mice in which folliculogenesis is blocked at the primary stage level, display abnormally large oocytes [42]. In both Gdf9- and connexin 43-deficient mice, the etiology of the phenotype is closely linked to oocyte-granulosa cell dialogue. In irradiated ovaries, the absence of granulosa cell proliferation may well result from the unresponsiveness of granulosa cells to the GDF9 signaling pathway, since we have observed that Bmpr2, which is one of the type II receptors for GDF9 [20, 27], is not expressed in granulosa cells of primary-like follicles during the neonatal period. On the other hand, the immediate entry into growth of the oocytes may well be a consequence of the unbalanced oocyte-epithelial ratio. Indeed, during normal ovarian development, it has been proposed that oocyte growth occurs in a second phase of follicular development after flattened granulosa cells become cuboidal in shape [43]. The presence of cuboidal cells would thus appear to be a prerequisite for oocyte growth [43]. In irradiated ovaries, the unbalanced oocyte-epithelial ratio leads to the presence of surnumerous granulosa cells, which are cuboidal in shape, as they surround surviving oocytes.
Our results indicate that primary-like follicles of irradiated ovaries are different from the primary follicles that appear at 0 dpn in the cores of control ovaries and constitute the initial follicular waves. Our previous studies comparing the development of control ovaries and ovaries depleted in germ cells either in utero or during the neonatal period have offered new insight in the characteristics of initial follicular waves and their role in ovarian physiology [13, 44]. In the present work, we have shown that an unbalanced oocyte-epithelial cell ratio at the time of follicle histogenesis is probably responsible for the differentiation of follicles that are characterized by primary follicle morphology and primordial-follicle maturation status. Factors, which in control ovaries govern the immediate growth and maturation of the initial follicular waves, are absent or inoperative in irradiated ovaries. We consider that further examination of follicle histogenesis, comparing control and severely germ cell-depleted ovaries, will certainly furnish new information on the processes and mechanisms [43, 45] involved in primordial follicle assembly as well as on the primordial to primary follicle transition.
ACKNOWLEDGMENTS
We thank Drs. T. Noce and F.B. Lane for the generous gifts of DDX4 and keratin antibodies, respectively. We also thank Dr. R.S. Viger for his critical reading of the manuscript.
FOOTNOTES
2 Correspondence: S. Magre, Laboratoire de Physiologie et Physiopathologie, UMR 7079, Université Paris VI, cc 256, 7 quai Saint-Bernard, 75005 Paris, France. FAX: 331 4427 2650; solange.magre{at}snv.jussieu.fr 
3 Current address: Ontogeny-Reproduction Research Unit, Room T1–49, Centre Hospitalier Universitaire de Quebec (CHUQ)- Laval Research University Centre (CHUL), 2705 Laurier Boulevard, Ste-Foy, Quebec, Canada G1V 4G2.
4 Current address: Gene Regulation Section, Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892.
1 Supported by Electricité de France and the Ministère de l'Education Nationale et de la Recherche Scientifique et Technique (MENRST, France). S.M.G. is a recipient of a fellowship from Organon (Akzo Nobel, France).
Received: 30 December 2005.
First decision: 17 January 2006.
Accepted: 17 July 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  C. Wang and S. K. Roy Development of Primordial Follicles in the Hamster: Role of Estradiol-17{beta} Endocrinology, April 1, 2007; 148(4): 1707 - 1716. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
