Ultrastructural Characterization of Mouse Embryonic Stem Cell-Derived Oocytes and Granulosa Cells

Abstract

Germ cells are a unique population of cells responsible for transmitting genetic information from one generation to the next. Our understanding of the key mechanisms underlying germ cell development in vivo remains scarce because of insufficient amounts of cell materials available for conducting biological studies. The establishment of in vitro differentiation models that support the generation of germ cells from mouse pluripotent stem cells provides an alternative means for studying reproductive development. The detection and analysis of stem cell-derived germ cells, however, present technical challenges. Methods for determining the developmental stage of germ cells ex vivo, such as gene expression and/or immunochemical analyses are inadequate, frequently necessitating the use of alternative, elaborate methods to prove germ cell identity. We have generated putative oocytes and granulosa cells in vitro from mouse embryonic stem cells and utilized electron microscopy to characterize these cells. Here, we report on the striking ultrastructural similarity of in vitro-generated oocytes and granulosa cells to in vivo oocytes developing within follicles.

Introduction

A fundamental concept in developmental biology is the definition of a robust and reproducible in vitro culture system that supports the induction and differentiation of functional and specific cells from totipotent or pluripotent stem cells. Commitment of mouse embryonic stem cells (ESCs) to differentiate into germ cells under defined culture conditions represents an important milestone in developmental biology [1]. Although the functionality of mouse ESC-derived sperm has been demonstrated [2], that of oocytes has not. Human ESCs can also be used to form primordial germ cells (PGCs), an indication of their potential to contribute to germline development [3 –8]. The potential generation of an unlimited source of germ cells and gametes from human ESCs and other pluripotent cells, for example, induced pluripotent stem cells (iPs), would facilitate the study of reproductive development and provide a new approach to the development of assisted reproduction therapies that do not rely on human samples. A prerequisite to this fascinating approach, however, is the step-by-step characterization of the generated cells of all developmental stages to ensure appropriate recapitulation of germ cell development in vitro. The growth and maturation of oocytes are complex processes of folliculogenesis. A bidirectional mode of communication between the oocyte and the companion granulosa cells (GCs) directs development of the oocyte [9,10]. This communication relies upon the structural integrity of the granulosa–oocyte cell interface [11]. The oocyte and surrounding GCs are structurally and functionally associated via specialized cytoplasmic processes called transzonal projections (TZPs), representing a highly dynamic interaction regulating folliculogenesis in vivo and thereby ensuring appropriate oocyte maturation and ovulation [12,13]. Significant information on the adaptive morphological changes taking place in GCs during the reproductive cycle has been obtained by electron microscopy (EM) in situ as well as in vitro [14 –17]. We used EM to determine the morphological characteristics and cellular interface ultrastructure of in vitro-generated GCs and oocytes. EM techniques can not only provide information on the functionality of the cells under study, but also be performed with minute quantities of a sample. Such techniques are therefore not subject to the experimental restrictions associated with the use of fetal tissues. Further, information garnered using EM techniques in the mouse model would be instrumental toward the establishment of differentiation systems with human pluripotent stem cells.

Materials and Methods

Electron microscopy

For transmission electron microscopy (TEM), cells were fixed with 2.5% glutaraldehyde (Merck) in 0.1 M sodium cacodylate buffer (pH 7.4), postfixed in 1% aqueous osmium tetroxide, dehydrated stepwise in a graded ethanol series, and embedded in Epon 812 (Fluka, Buchs, Switzerland). Ultrathin (50 nm) sections were prepared with an ultramicrotome (EM UC6; Leica), stained first with 1% uranyl acetate and then with 3% lead citrate, and subsequently examined using a Zeiss EM 109 electron microscope (Zeiss). Images were taken on 70 mm films (Maco ORT 25c orthochromatic; Hans O. Mahn & Co., Photo Division).

For scanning electron microscopy (SEM), cells were fixed with 2.5% glutaraldehyde in 0.1 M phosphate buffered saline (pH 7.4) and dehydrated stepwise in a graded ethanol series. Samples were dried with 100% ethanol via CO₂ in a critical-point apparatus (Balzers). Dried samples were mounted onto aluminum stubs with leit-tabs and coated with gold film to a thickness of 40–50 nm in a sputter coater (Leitz). Cells were viewed under a Hitachi S-530 scanning electron microscope operated in a secondary mode at 20 kV.

Cell culture

Mouse ESC lines gcOct4-GFP (XY) [1] and OG2 (ΔPE-Oct4-GFP, XY; XX) [18] were differentiated into oocytes according to 3 protocols: a previously described serum-containing protocol [1], a serum-containing/serum-free differentiation combination protocol, and a serum-free differentiation protocol. Briefly, for the serum-containing differentiation procedure, cells were grown in tissue culture plates in Dulbecco's modified Eagle's medium containing 4.5 g/L glucose and supplemented with 15% fetal calf serum (Hyclone), 2 mM l-glutamine (Invitrogen), 100 μM nonessential amino acids (Invitrogen), 1 μM β-mercaptoethanol (Sigma), 50 μg/mL penicillin and streptomycin (Invitrogen) in the absence of mouse embryonic fibroblast (MEF) feeder cells and leukemia inhibitory factor (LIF) for up to 30 days. The medium was changed every 3 days for 9 days and every other day thereafter. Floating cell aggregates after day 12 of differentiation were collected and processed as published [1]. For the serum-containing/serum-free differentiation combination technique, ESCs were differentiated under serum-containing conditions for 7–12 days. The adherent cell layer was digested with accutase (Invitrogen) and the single-cell suspension was then plated as a suspension culture at a 1:1 ratio in in vitro growth (IVG) media [modified α-MEM supplemented with 3 mg/mL bovine serum albumin (Millipore), 50 μg/mL each penicillin and streptomycin (Invitrogen), 2 mM glutamine (Invitrogen), 0.23 mM pyruvate (Invitrogen), 10 μg/mL insulin, 5 μg/mL transferrin, 5 ng/mL selenium (Invitrogen), 1 ng/mL human recombinant epidermal growth factor (Invitrogen), and 50 ng/mL recombinant mouse stem cell factor (Millipore)] in a humidified incubator at 37°C in 5% O₂ and 5% CO₂. Cells were cultured for 7–15 days, and putative oocytes were then individually picked with a drawn-out glass capillary. For the serum-free differentiation protocol, feeder-free ESCs were plated on gelatin-coated tissue culture plates in N2B27 medium according to the method of Ying et al. [19] and cultured in 5% CO₂ for 48 h. The adherent cell layer was then digested with accutase and the single-cell suspension was replated at a 1:2 ratio in IVG media supplemented with stem cell factor (SCF) and leukemia inhibitory factor (LIF) in suspension plates. After 3 days of culture in 5% CO₂, cell aggregates were transferred to gelatinized 24-well tissue culture plates at a density of 5–6 aggregates/well and further cultivated for 5 days in medium lacking LIF. Cultures were then partially digested in trypsin/ethylenediaminetetraacetic acid 0.05% supplemented with 0.02% DNAse I and replated in oocyte growth media (modified α-MEM supplemented with 3 mg/mL bovine serum albumin [Millipore], 50 μg/mL each penicillin and streptomycin [Invitrogen], 0.23 mM pyruvate [Invitrogen], 5 μg/mL insulin, 5 μg/mL transferrin, 5 ng/mL selenium [Sigma], 1 ng/mL human recombinant epidermal growth factor [Invitrogen], 100 ng/mL recombinant mouse stem cell factor [Millipore], and 5 ng/mL rhFSH [Sigma]) and cultured in a humidified incubator at 37°C in 5% O₂ and 5% CO₂ for additional 12–15 days. Putative oocytes were cultivated in maturation media as previously described to assess their maturation status [20].

RNA isolation

RNA was isolated using the RNeasy Micro Kit (QIAGEN) with on-column DNase digestion as per the manufacturer's instructions. RNA quality was determined using a Nanodrop (ND-1000) spectrophotometer and an Agilent RNA 2100 Bioanalyzer.

For single-cell analysis of oocytes, the Applied Biosystems kit was used for preamplification according to the manufacturer's recommendations.

Reverse transcription–polymerase chain reaction and real-time quantitative reverse transcription–polymerase chain reaction

Two hundred fifty nanograms of RNA was reverse-transcribed to cDNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Real-time quantitative TaqMan reverse transcription–polymerase chain reaction (q-RT-PCR) was performed on 7900ht and ht fast devices (Applied Biosystems) with milliQ H₂O and TaqMan Universal PCR Master Mix (Applied Biosystems). Raw data were analyzed using SDS 2.3 software (Applied Biosystems). Gene expression was normalized to the mouse housekeeping gene Hprt. Quantification of relative gene expression was calculated using the ΔΔCt method. Three technical replicates were used for each real-time q-RT-PCR reaction; a reverse transcriptase blank and a no-template blank served as negative controls. The following primers were used: Oct4: Mm00658129_gH; Blimp1: Mm00476128_m1; Ddx4: Mm00802445_m1; Figla: Mm00488823_m1; Nobox: Mm00453743_m1; Prdm14: Mm01237813_m1; Sohlh1: Mm01338424_g1; Stella: Mm00836373_g1; Sycp3: Mm00488519_m1.

Isolation of germinal vesicle-stage oocytes from ovaries of adult mice

Eight- to 10-week-old CD1 female mice were superovulated by injection of 7.5 IU pregnant mare serum gonadotropin (PMSG) and, 48 h later, by injection of 7.5 IU human chorionic gonadotropin (hCG). Females were sacrificed by cervical dislocation at 17–18 h post-hCG stimulation. Ovaries were dissected and digested for 10 min at 37°C in 0.5 mg/mL collagenase type 1A (Sigma), washed in PBS, and placed in Hepes-buffered CZB medium containing 1 g/L glucose and 0.1% (w/v) polyvinylpyrolidone (ICN). Follicles were released into the medium by mechanical dissociation of the tissue. Cumulus cells were removed mechanically and oocytes with a clearly visible germinal vesicle were collected, washed 3 times in Hepes-buffered CZB medium, and lysed in RNeasy lyses buffer (RLT) to extract RNA for gene expression analysis or further processed for EM. Mice were maintained and used for experiments according to institutional guidelines.

Results

Granulosa cells and the GC–oocyte interface of ESC-derived follicles

The entry of PGCs into meiosis in vivo and in vitro appears to be a cell autonomous process. Subsequent development of PGCs into competent oocytes and ultimate meiotic arrest during folliculogenesis is dependent on 2-way signaling interactions between the oocyte and surrounding GCs [21,22]. Figure 1A–D shows ESC-derived follicle-like structures. As aggregation of PGCs and granulosa-like cells takes place randomly in this system, we obtained spherical aggregates with centrally located single as well as multiple putative germ cells surrounded by various layers of attached presumptive GCs and extracellular matrix (ECM). Light microscopy revealed germ cell clusters surrounded by loosely attached somatic granulosa-like cells (Fig. 1A). SEM analysis demonstrated follicle-like structures, that is, germ cells surrounded by a dense layer of cuboidal cells (Fig. 1B). Figure 1C shows ESC-derived putative GCs in direct proximity to a zona pellucida-like matrix. The granulosa-like cells exhibit a polygonal shape and a smooth surface with few microvilli (Fig. 1C). SEM and TEM analysis revealed a structure specific for GCs, that is, TZPs (Fig. 1C, 1D, arrow) that extend toward the putative oocyte.

FIG. 1.

ESC-derived follicle-like structures. (A) Light microscopy image of ESC-derived follicle-like structures with clusters of granulosa cells loosely surrounding the oocyte. (B) SEM image of a defined follicle structure. Note the layer of densely attached cuboidal cells (asterisk). (C) SEM image of polygonal-shaped granulosa cells around a smooth zona pellucida-like surface (asterisk). The pole of a granulosa cell is facing toward the oocyte (arrow). (D) TEM image of a granulosa cell with an extension stretching towards the oocyte (arrow) (×4,890). ESC, embryonic stem cell; SEM, scanning electron microscopy; TEM, transmission electron microscopy.

Figure 2A shows a light microscopy image of an ESC-derived follicular structure, with presumptive GCs surrounding the green fluorescent protein (GFP)-positive oocyte-like cell. To analyze the GC–oocyte interface by SEM, in vitro-derived follicles were cracked. Figure 2C and D shows both parts of a cracked aggregate, revealing their morphological similarity to natural follicles. Even though the presumptive oocyte was destroyed during the cracking procedure, a round cavity of 50–60 μm in diameter with residual oocyte cell material is evident (Fig. 2C, D). Multiple layers of GCs surround the cavity. The most central ones exhibiting the typical cuboidal-elongated phenotype of natural polarized cumulus cells, that is, those cells in direct contact to the oocyte (Fig. 2C), whose TZPs orientate toward the oocyte (Fig. 2B). TZPs have been well characterized in many mammals by EM [23,24], and their number and form have been demonstrated to change dynamically during follicular development [12]. To investigate the GC–oocyte interface in more detail, we performed TEM analyses and compared GCs of natural follicles to ESC-derived presumptive GCs. The ultrastructure of TZPs of in vitro-derived GCs appeared to be indistinguishable from the TZPs of their natural counterparts, indicating generation of the interface between the oocyte and GCs, the major control structure for healthy follicle development [12]. ESC-derived GCs of high electron density exhibited long cellular processes containing the same dark fine granulation as the cytoplasm (Fig. 2E). These processes penetrated the matrix, with some reaching a diameter of ∼1 μm, very similar to the branched processes of polarized cumulus cell in natural follicles. The detection of organelles, for example, mitochondria [12,25], within the TZPs of in vitro-derived GCs (Fig. 2F, G) is indicative of recapitulation of folliculogenesis in vitro. Further, the ultrastructural texture of the pale amorphous matrix in direct proximity to the GCs in vitro (Fig. 2E) resembles the zona pellucida of the GC–oocyte interface in vivo (Fig. 2H).

FIG. 2.

Granulosa cell–oocyte interface of ESC-derived follicles. (A) Light microscopy image of an ESC-derived follicle structure. Granulosa cells surround the gcOct4-GFP oocyte. (B) SEM analysis of an ESC-derived follicle-like structure. A long process originates from a somatic granulosa cell toward the oocyte (arrow). (C, D) SEM analysis of both halves of a cracked ESC-derived follicle-like structure. The position of the oocyte is clearly apparent by the 50–60-μm-large, round cavity ( asterisk in C) and residual oocyte cell material in D (asterisk). The granulosa cells surround the oocyte in multiple layers, with the cell layer in direct contact being cuboidal-elongated in shape (C, arrow ). (E) TEM analysis of ESC-derived granulosa cells. Cells exert long processes (TZPs) toward the oocyte (arrow). Note the texture of the pale amorphous matrix (asterisk), which is penetrated by the TZPs (×4,890). (F) TZPs contain organelles such as mitochondria (arrow) and have the same ultrastructural texture as the cell cytoplasm (×4890). (G) TZPs of ESC-derived granulosa cells partly exhibit extreme branching processes (arrow) and reach diameters of almost 1 μm (×12,210). (H) Cellular processes of cumulus cells with branching ends (arrow) of in vivo follicles closely resemble those of in vitro-derived granulosa cells. The ultrastructure of the zona pellucida (asterisk) resembles that of in vitro-derived follicles in E (×4,890). TZPs, transzonal projections.

Ultrastructural characterization of in vitro-generated GC-like cells revealed additional analogies to characteristics of natural GCs described in the literature, for example, the presence of both dark- and light-appearing GCs. Dark GCs possess a dark background nucleoplasm and background cytoplasm (Fig. 3A, C). The electron density is due to the fine granulation of the cytoplasm, which does not correspond to free ribosomes, as free ribosomes would be much larger and stand distinctly out from the granular background (Fig. 3E, F). The cytoplasm of the GCs appears to be partly vacuolated and is populated by mitochondria, ribosomes, and the most prominent organelle of ESC-derived GCs, the rough endoplasmic reticulum. Cisternae of swollen rough endoplasmic reticulum containing granular material wind through the cytoplasm of dark and light in vitro-derived GCs (Fig. 3A, E) as well as pale in vivo GCs (Fig. 3F). The ribosomes of natural dark and light GCs and ESC-derived GCs are either free or associated with rough endoplasmic reticulum (Fig. 3A, B, E, F). The presence of free ribosomes is a characteristic feature of protein-secreting cells, such as the GCs of growing follicles. The mitochondria of both dark in vivo and in vitro-derived GCs appear irregular and elongated, containing pale cristae and an electron-dense matrix (Fig. 3A, B). The mitochondria of light GCs exhibit an oval shape and have tubular–vesicular cristae (Fig. 3E, F). Lipid droplets filled with a transparent gray substance and empty vacuoles frequently surrounded by a membrane (Fig. 3B, D) are distributed throughout the cytoplasm of in vivo GCs. Vacuolization appears to be a prominent feature of ESC-derived GCs (Fig. 3C), manifested mainly by empty white vacuoles and vacuoles containing a residual gray substance (Fig. 3C). Empty vacuoles appear to be an artifact of sample manipulation, created when the gray substance of lipid droplets is washed out during the dehydration step. The nuclei of both in vitro-derived and in vivo GCs are large and have an irregular contour (Fig. 3A–D). The nucleoplasm of ESC-derived GCs, enclosed by the inner nuclear membrane, contains finely dispersed chromatin and clumps of heterochromatin scattered throughout (Fig. 3A, C, E); the nucleoplasm of in vivo GCs is arranged similarly (Fig. 3B, D, F). Taking these data together, the ultrastructural features of in vitro-derived GCs are indistinguishable from natural GCs. Most importantly, to the best of our knowledge, there is no other cell in the body capable of forming unique and typical TZPs. The observed activated rough endoplasmatic reticulum, indicative for metabolically active and steroid-producing cells, further supports their GC-like identity.

FIG. 3.

TEM analyses of ESC-derived dark- and light-colored granulosa cells. (A) Cuboidal-elongated in vitro-derived granulosa cells with electron dense cytoplasm and swollen rough endoplasmatic reticulum. Mitochondria (insert) exhibit pale tubular cristae (×4,890). (B) In vivo-derived dark granulosa cell. Note lipid droplets (asterisk) (×4,890). (C) Dark ESC-derived granulosa cells with numerous vacuoles and droplets containing residual gray translucent substance indicative of lipids (asterisk). Note prominent nucleolus (×4,890). (D) In vivo dark granulosa cell with numerous pseudopodia, branching processes and protrusions of cytoplasmic invaginations, vacuoles (arrowhead), and lipid droplets (asterisk) as detected in in vitro-derived granulosa cell shown in C (×6,950). (E) ESC-derived light granulosa cell with microvilli at the cell surface. The cell exhibits the typical swollen rough endoplasmatic reticulum (asterisk) and oval-shaped tubular mitochondria with dark cristae (insert) (×4,890). (F) In vivo-derived light granulosa cell with swollen rough endoplasmatic reticulum (asterisk) and vesicular–tubular mitochondria (arrow) (×4,890).

ESC-derived oocytes

We have developed a suspension culture system that supports the generation of ESC-derived oocytes of different follicular stages. Independent of the differentiation media used to initiate differentiation of ESCs, we obtained putative oocytes of various developmental stages at different yields. Serum-free conditions appeared best to promote germ cell differentiation, and XY-ESCs yielded a higher percentage of oocytes than XX-ESCs. Expression of specific oocyte marker genes was assessed by qRT-PCR in groups of 3 in vitro-derived oocytes and compared with superovulated germinal vesicle-stage oocytes and ESCs. As expected, the generated oocytes (Fig. 4A) exhibited a gene expression pattern very similar to that of control samples (Fig. 4B). Morphologically, in vitro-generated oocytes (Figs. 4A and 5A) and in vitro-matured oocytes generated from 18.5 days postcoitum (dpc) germ cells under the same culture conditions (Fig. 5C) exhibited a very similar shape, size, and microvillous cell surface (Fig. 5C, D). Interestingly, the surface topography of some in vitro-derived oocytes analyzed by SEM resembled exactly that of unfertilized oocytes in vivo, with the typical sparse and relatively uniform microvillous area and a clear mosaic surface containing a polar microvillous-free region (Fig. 5B–D). TEM analyses of generated oocytes revealed large, slightly eccentric nuclei with a characteristic pale background nucleoplasm (paler than the ooplasm) (Fig. 5E). Free ribosomes appeared individually or arranged in clusters of 5–10 (Fig. 5E, insert). Elongated staples of fibrils not associated with ribosomes were found throughout the cytoplasm (Fig. 5E, insert, F). Oocyte-specific cortical granules containing a characteristic electron-dense center and a pale outer ring (Fig. 5F, arrow) were mostly distributed throughout the cytoplasm and appeared only sporadically at the cell periphery (Fig. 5F and inserts), a distinct feature of immature oocytes. Further, we detected round as well as elongated mitochondria with a dense matrix and pale, irregularly arranged cristae and intermitochondrial vacuoles, all features of natural oocytes. Noteworthy is that the majority of ESC-derived oocytes showed signs of apoptosis and autophagy, which might be indicative of atresia (Fig. 5E, H). The oocyte shown in Fig. 5E lost its spherical shape. Numerous lysosomes, clear vesicles, and autophagosomes as well as dark cytoplasmic structures were detected. The ESC-derived oocyte shown in Fig. 5H contains numerous apoptotic bodies and secondary lysosomes within the dark condensed cytoplasm. Mitochondria and other organelles cannot be identified. The zona pellucida and most of the microvilli at the oocyte surface appear to have been lost. In comparison, oocytes derived from natural 18.5-dpc PGCs grown for 13 days in the same culture medium as in vitro-derived oocytes (Fig. 5G) also exhibit dark-stained mitochondria without cristae, numerous vacuoles, and secondary lysosomes; other organelles cannot be identified. Microvilli either have been lost or have retracted from the zona pellucida. These observations indicate that our culture conditions support the maturation of PGCs into oogonia, but further optimization of these conditions is required for the generation of functional oocytes.

FIG. 4.

Gene expression of oocyte markers in a pool of 3 in vitro-derived oocytes. (A) Light microscopy image of one of the analyzed oocytes. (B) Real-time quantitative reverse transcription–polymerase chain reaction of oocyte markers in a pool of 3 ESC-derived oocytes compared with XX-ESCs.

FIG. 5.

Ultrastructural analysis of in vitro-derived oocytes. (A) Light microscopy image of an in vitro-derived oocyte. (B) SEM image of an in vitro-derived oocyte with a polar microvillous-free surface (asterisk). (C) SEM image of fetal in vivo oocytes cultivated under the same culture conditions as the in vitro-derived oocytes. Note the polar microvillous-free region (asterisk). (D) SEM image of an in vitro-derived oocyte with a polar microvillous-free surface (asterisk). Note the similarity of the shape and surface structure to in vivo oocytes shown in C. (E) TEM image showing an overview of an in vitro-derived oocyte. Numerous lysosomes, clear vesicles, lipid droplets (arrow), and autophagosomes (arrowhead) are distributed throughout the cytoplasm. n: nucleus (×1,100). Insert: Elongated mitochondria with tubular–vesicular cristae and vacuoles are arranged between staples of fibrils (×6,950). (F) TEM image of oocyte-specific cortical granules (arrow) with the typical electron-dense center and pale outer ring of in vitro-derived oocytes. Cortical granules appear only sporadically in the cell periphery (inserts). Note vacuolated mitochondria (arrowhead) (×6,950). (G) TEM image of oocytes derived from natural 18.5-day postcoitum primordial germ cells grown for 13 days in the same culture medium as in vitro-derived oocytes. The dark-stained mitochondria contain only few cristae, numerous vacuoles, autophagosomes (arrowhead), and secondary lysosomes. Other organelles are not identifiable. Microvilli are lost or retracted from the zona pellucida (arrow) (×6,950). (H) TEM image of in vitro-derived oocyte showing numerous apoptotic bodies (arrow) and secondary lysosomes in the dark, condensed cytoplasm. Mitochondria or other organelles are not identifiable. The zona pellucida and most of the microvilli at the oocyte surface are lost (×6,950).

Discussion

The first systematic study of the culture and growth of mouse oocytes in vitro was published by John Eppig in 1977 [26]. Eppig and coworkers demonstrated that mouse oocytes from primordial follicles could be grown and matured in vitro to produce live offspring after fertilization [27,28]. These data marked a major breakthrough in the field of reproductive biology. Based on this success, various in vitro culture systems for female gametes of different mammalian species were established, such as rat [29], hamster [30], cat [31], pig [32], sheep [33], goat [34], cow [35,36], and human [37]. All these maturation systems, including the mouse model, are still under investigation in an effort to broaden our understanding of the complex mechanisms involved in germ cell development and maturation. In this study, we performed comparative ultrastructural analyses of follicle-like structures and bona fide oocytes from mouse ESCs with their natural counterparts. With our data, we supplement the genetic and immunological data obtained by standard analyses with a description of the unique ultrastructural properties of oocytes and their supporting GCs during gametogenesis in vivo. Specifically, we performed SEM analysis on in vitro-derived follicle-like structures and compared their morphology with that of natural mammalian follicles described by Makabe and coworkers in 2006 [13]. Striking morphological similarities indicate that our culture system supports the derivation of germ cells of different developmental follicular stages. During the early growth of follicles, the shape of GCs progressively changes from flattened to cuboidal and to columnar [38] [39]. Well-defined in vitro-derived follicles (Fig. 1B), covered with a layer of densely attached polyhedral cells, form the stratum granulosum, indicating that the follicles are in the late phase of growth. Cracked in vitro-derived follicles, such as the one shown in Fig. 2C and D, exhibit the same morphological features as cracked mammalian follicles, with a residual macerated ooplasm as previously described [13]. We utilized TEM analyses to assess differences in the dynamic structure of TZPs, which form the morphological basis of the oocyte–GC interface and play a crucial role in the structural integrity of developing follicles and maturing oocytes [12]. Many of our in vitro-derived GCs developed microvilli, cytoplasmic invaginations, pseudopodia, and cytoplasmic protrusions (not shown here), which have been associated with cell expansion and are considered indirect signs of GC luteinization both in vitro and in vivo [40]. Further, we observed morphologically different subpopulations of ESC-derived GCs, that is, those exhibiting differences in the darkness of background cytoplasm and nucleoplasm as well as in content and distribution of lipid droplets and rough endoplasmic reticulum. This heterogeneity of a given GC population has also been reported by Nottola and coworkers [16] for in vivo follicular aspirates cultured in vitro. Further, cuboidal GCs with a dark background cytoplasm and nucleoplasm have been found in primordial polyovular and unilaminar follicles of the golden hamster in vivo [41]. We also detected these light and dark GCs in natural mouse follicles (Fig. 3B, D, F). Both dark and light GCs are frequently found in developing endocrine tissues and reflect a change in the complex secretory cycle [41]. The fine dispersion of chromatin and the peripheral patches of heterochromatin, which we found in the nuclei of ESC-derived GCs (Fig. 3A, C, E), are features typical of metabolically active cells [40,42]. Mitochondria of ESC-derived GCs vary in size and shape and have mostly tubular–vesicular cristae (Fig. 3A, E). These types of mitochondria are frequently found in luteinizing GCs in vivo [43] and in human GCs of follicles containing a fertilizable oocyte [16]. In vitro-derived dark GCs contain mitochondria with a dense matrix and pale cristae (Fig. 3A, insert), whereas light GCs contain mitochondria with a pale matrix and dark cristae (Fig. 3E, insert). Makino et al. [44] found an association between mitochondria of human GCs exhibiting a pale matrix with poor fertilization outcome in vitro. Mitochondria with longitudinally oriented cristae have been found in human ovarian carcinoma and in the lungs, liver, and ureter of different mammals [16]. Researchers have hypothesized that this morphological transformation of mitochondria could be due to negative environmental conditions, such as high oxygen concentration, presence of cytostatic drugs or antibiotics and aging [45 –47]. We did not observe such mitochondria in our culture-derived GCs, indicating that we successfully generated functional GCs in vitro. A most prominent feature of in vitro-derived GCs is a rich network of swollen rough endoplasmic reticulum and a well-developed Golgi apparatus. These features indicate that there is ongoing luteinization [40,43,48,49]. Isolated rough endoplasmic reticulum and ribosomes of luteinizing GCs are responsible for synthesizing the enzymes involved in steroid production [50]. The presence of numerous lipid droplets within the cytoplasm of in vitro-derived GCs (Fig. 3C), an early sign of luteinization [50], correlates with active steroidogenesis [16]. Taken together, in vitro-derived GCs exhibit ultrastructural characteristics typical for metabolically active and steroid-producing cells.

The ovarian follicle represents a morphological and functional unit wherein both somatic and germ cells play a pivotal role in follicle maturation and formation of fully competent, fertilizable oocytes [51]. Under appropriate culture conditions, in vitro-matured cumulus-enclosed oocytes of mouse, rat, cow, and sheep have been shown to possess a similar developmental capacity to mature oocytes in vivo [52 –54]. SEM analysis clearly demonstrated that ESC-derived cuboidal-shaped GCs closely surround the oocyte (Fig. 2C, D) and extend long TZPs toward the oocyte (Fig. 1D). In vivo, TZPs are most abundant in preantral follicles. By TEM analysis, we demonstrated that ESC-derived cumulus cells extend typical TZPs, some of which contain mitochondria, an observation consistent with literature [12,25]. Albertini et al. [12] demonstrated the active bidirectional movement of mitochondria within the TZPs by vital staining of intact mouse follicles with MitoTracker^TM, a cell-permeant probe that contains a thiol-reactive chloromethyl moiety for labeling mitochondria. TZPs are an essential requirement for healthy follicle development [12] and the ultrastructure of TZPs of our in vitro-derived GCs demonstrates a profound conformity to that of natural GCs. This data demonstrate the establishment of an interface, that is, the major control site between the oocyte and GCs in vitro. TEM analysis showed that TZPs penetrate a mucous substance, which appears as a gray homogeneous material (Fig. 2E, F). Cumulus cells synthesize and secrete hyaluronic acid as well as produce a hyaluronan-rich ECM matrix [55,56]. The ultrastructure of the ECM matrix of the in vivo cumulus–oocyte complex (Fig. 2H) closely resembles that of the matrix penetrated by the TZPs of the in vitro-derived GCs (Fig. 2E).

Comparative morphological analysis of ESC-derived oocytes and in vivo oocytes revealed remarkable similarities (Fig. 5A–H). The oocyte surface varies greatly from one stage of the cell cycle to the other and even within a given stage, suggesting a close association between the surface characteristics and the maturation status of the oocyte [57]. A mature, unfertilized mouse oocyte is characterized by small blebs and sparse, relatively uniform microvilli [58] as well as a clear mosaic surface, that is, a microvillous membrane with a smooth, microvillous-free polar region. This region is relatively free of organelles [59] and has fewer CGs [60]. SEM analysis of in vitro-derived oocytes from late cultures revealed a mosaic surface topography typical for mature unfertilized oocytes (Fig. 5D).

TEM analysis demonstrated a background nucleoplasm that was paler than the background cytoplasm (Fig. 5E). ESC-derived oocytes exhibited slight vacuolization of ooplasm, a feature typical for immature and aged oocytes [61]. Mature oocytes showed spontaneous or vesicle-fused vacuolization, with the vacuoles usually filled with fluid [62]. Culture-derived oocytes had numerous mitochondria of varied shapes (Fig. 5F). These mitochondria exhibited an electron-dense matrix and pale, irregularly arranged cristae, a typical feature of fetal mouse oocytes [63]. Some mitochondria showed a vacuolated matrix, which has been described for oocytes of polyovular and unilaminar follicles in the golden hamster [41]. Round, oval, and elongated mitochondria are found in mouse growing oocytes [63] and contain large intermitochondrial vacuoles, as shown in Fig. 5F (arrowhead). Numerous fibrillar arrays were detected among the mitochondria of our in vitro-derived oocytes (Fig. 5E, insert). This arrangement is seen in mature oocytes and can only be detected in glutaraldehyde-fixed samples [25]. In the mouse, these fibrils form polysomes and a hexagonal crystalline pattern. The fibrillar bundle seen in Fig. 5E (insert) appeared long, without any pattern, and was not associated with ribosomes as described by Zamboni [25]. Future work should address whether this nonassociation with ribosomes can be attributed to a smaller size, different metabolic activity, or a thinner and atypical zona pellucida of the oocyte.

The formation of cortical granules begins during oocyte maturation [25] and the distribution of cortical granules in the mouse oocyte cortex changes dynamically during meiotic maturation [60]. Cortical granules are round or elliptical in shape, measure 0.2–0.5 μm in diameter, and consist of a highly dense matrix surrounded by a single smooth membrane [25]. As described by Ducibella et al. [60], in immature oocytes, cortical granules are distributed asymmetrically throughout the cortex, whereas in mature oocytes, they are localized mainly to the cortex periphery and are fewer in number. TEM analysis of ESC-derived oocytes showed numerous cortical granules distributed over the entire cortex (Fig. 5F), with only a few granules at the oocyte periphery (Fig. 5F, insert), which is indicative for immature oocytes. The distribution of cortical granules and their competence to undergo cortical reaction are important indicators of oocyte cytoplasmic maturation and provide valuable information about the fertilizable lifespan of the oocyte [60]. Overall, these ultrastructural data suggest that the analyzed ESC-derived oocytes resembled immature and growing wild-type oocytes. We carried out some IVF experiments with oocytes derived under the tested differentiation conditions and at later culture time points, but were not able to achieve fertilization. The immature nature and the low yields of oocytes impeded the appropriate experimental setup for oocyte activation and subsequent IVF and ICSI. For the ultimate proof of functionality, our differentiation conditions need to be further improved to obtain higher yields of oocytes.

In vivo, only few follicles ovulate; the vast majority degenerate and die [64] during atresia, a process that can occur at any stage of follicular development [65]. This process assures adequate oocyte quality, prevents the survival of oocytes with crossover chromosome defects, and limits the number of oocytes within the growing follicle [66]. TEM analysis of in vitro-derived oocytes revealed signs of apoptosis and autophagy (Fig. 5E, H), which are indicative of atresia. As shown in Fig. 5E, ESC-derived oocytes had lost their spherical shape, a structural feature that was used by Escobar et al. to select for atretic rat oocytes. These oocytes showed clear vesicles, numerous lysosomes, and autophagosomes as well as dark cytoplasmic structures, all signs of autophagy [67]. Signs of apoptosis, such as clumps of compact chromatin, apoptotic bodies, and membrane blebbing, were not observed, consistent with the results of Devine and coworkers [65]. Ribbon-like fibril bundles were found around the nucleus, a finding previously reported by Ortiz et al. [68] in atretic oocytes of rat antral follicles. In comparison, in vitro-derived oocytes, such as the one shown in Fig. 5H, exhibited signs of apoptosis. Numerous apoptotic bodies and secondary lysosomes were distributed within the breaking, condensed cytoplasm. Mitochondria could not be identified and the zona pellucida, organelles, and most of the oocyte surface microvilli were lost. Escobar et al. [67] proposed that oocyte cell death probably starts with autophagic degradation of cytoplasmic components, such as mitochondria. It therefore appears that the in vitro-derived oocyte shown in Fig. 5E was analyzed at an initial phase of cell death, whereas the oocyte shown in Fig. 5H was isolated at an advanced stage of apoptotic cell death. Natural mouse oocytes derived from 18.5-dpc PGCs (Fig. 5G) exhibited similar features when grown under the same culture conditions, indicative of ongoing atresia. All oocytes shown in Fig. 5 were freely floating in the culture media at the time of collection, that is, had already been extruded from the follicle-like structures. At this time, it is not known whether the autophagic and/or apoptotic signs shown by our in vitro-derived oocytes are a consequence of the culture conditions or are caused by apoptotic-inducing stimuli present in the cellular milieu.

Although the derivation and culture of female germ cells from pluripotent stem cells in vitro may lead to ultrastructural alterations in oocytes, this study demonstrates that a simple in vitro culture system can create the essential components required for oocyte development: oocytes, an ECM-based interface, and granulosa-like cells. Future efforts to optimize our in vitro differentiation conditions should bring us closer to our ultimate goal of generating fully functional oocytes and demonstrate whether all key events of germ cell development in vivo can be truthfully recapitulated in vitro.

Footnotes

Acknowledgments

The authors thank J. Lange, M. Sinn, C. Ortmeier, and G. Verberk for providing technical assistance, L. Gentile and T. Esteves for providing helpful discussions, and the Department of Zoology (University of Osnabrück) for granting our group access to their electron microscope. This work was supported by the Max Planck Society and the DFG grant for Research Unit Germ Cell Potential (FOR 1041).

Author Disclosure Statement

The authors declare that no competing financial interests exist.

References

Hubner

, Fuhrmann

, Christenson

LK.

, Kehler

, Reinbold

, De La Fuente

, Wood

, Strauss

JF.

3rd , Boiani

, Scholer

. 2003. Derivation of oocytes from mouse embryonic stem cells. Science, 300:1251–1256.

Nayernia

, Nolte

, Michelmann

HW.

, Lee

JH.

, Rathsack

, Drusenheimer

, Dev

, Wulf

, Ehrmann

IE.

, Elliott

DJ.

, Okpanyi

, Zechner

, Haaf

, Meinhardt

, Engel

. 2006. In vitro-differentiated embryonic stem cells give rise to male gametes that can generate offspring mice. Dev Cell, 11:125–132.

Bucay

, Yebra

, Cirulli

, Afrikanova

, Kaido

, Hayek

, Montgomery

. 2009. A novel approach for the derivation of putative primordial germ cells and sertoli cells from human embryonic stem cells. Stem Cells, 27:68–77.

Chen

HF.

, Kuo

HC.

, Chien

CL.

, Shun

CT.

, Yao

YL.

, Ip

PL.

, Chuang

CY.

, Wang

CC.

, Yang

YS.

, Ho

. 2007. Derivation, characterization and differentiation of human embryonic stem cells: comparing serum-containing versus serum-free media and evidence of germ cell differentiation. Hum Reprod, 22:567–577.

Clark

AT.

, Bodnar

MS.

, Fox

, Rodriquez

RT.

, Abeyta

MJ.

, Firpo

MT.

, Pera

. 2004. Spontaneous differentiation of germ cells from human embryonic stem cells in vitro. Hum Mol Genet, 13:727–739.

Kee

, Gonsalves

JM.

, Clark

AT.

, Pera

. 2006. Bone morphogenetic proteins induce germ cell differentiation from human embryonic stem cells. Stem Cells Dev, 15:831–837.

Tilgner

, Atkinson

SP.

, Golebiewska

, Stojkovic

, Lako

, Armstrong

. 2008. Isolation of primordial germ cells from differentiating human embryonic stem cells. Stem Cells, 26:3075–3085.

West

FD.

, Machacek

DW.

, Boyd

NL.

, Pandiyan

, Robbins

KR.

, Stice

. 2008. Enrichment and differentiation of human germ-like cells mediated by feeder cells and basic fibroblast growth factor signaling. Stem Cells, 26:2768–2776.

Eppig

. 2001. Oocyte control of ovarian follicular development and function in mammals. Reproduction, 122:829–838.

10.

Eppig

JJ.

, Wigglesworth

, Pendola

. 2002. The mammalian oocyte orchestrates the rate of ovarian follicular development. Proc Natl Acad Sci U S A, 99:2890–2894.

11.

Senbon

, Hirao

, Miyano

. 2003. Interactions between the oocyte and surrounding somatic cells in follicular development: lessons from in vitro culture. J Reprod Dev, 49:259–269.

12.

Albertini

DF.

, Combelles

CM.

, Benecchi

, Carabatsos

. 2001. Cellular basis for paracrine regulation of ovarian follicle development. Reproduction, 121:647–653.

13.

Makabe

, Naguro

, Stallone

. 2006. Oocyte-follicle cell interactions during ovarian follicle development, as seen by high resolution scanning and transmission electron microscopy in humans. Microsc Res Tech, 69:436–449.

14.

Amsterdam

, Keren-Tal

, Aharoni

, Dantes

, Land-Bracha

, Rimon

, Sasson

, Hirsh

. 2003. Steroidogenesis and apoptosis in the mammalian ovary. Steroids, 68:861–867.

15.

Dhar

, P Dockery

WS O.

, Turner

, Lenton

EA.

, Cooke

. 1996. The human ovarian granulosa cell: a stereological approach. J Anat, 188,Pt 3:671–676.

16.

Nottola

SA.

, Heyn

, Camboni

, Correr

, Macchiarelli

. 2006. Ultrastructural characteristics of human granulosa cells in a coculture system for in vitro fertilization. Microsc Res Tech, 69:508–516.

17.

Motta

PM.

, Nottola

SA.

, Familiari

, Makabe

, Stallone

, Macchiarelli

. 2003. Morphodynamics of the follicular-luteal complex during early ovarian development and reproductive life. Int Rev Cytol, 223:177–288.

18.

Szabo

PE.

, Hubner

, Scholer

, Mann

. 2002. Allele-specific expression of imprinted genes in mouse migratory primordial germ cells. Mech Dev, 115:157–160.

19.

Ying

QL.

, Stavridis

, Griffiths

, Li

, Smith

. 2003. Conversion of embryonic stem cells into neuroectodermal precursors in adherent monoculture. Nat Biotechnol, 21:183–186.

20.

Eppig

JJ.

, O'Brien

. 1996. Development in vitro of mouse oocytes from primordial follicles. Biol Reprod, 54:197–207.

21.

Adams

IR.

, McLaren

. 2002. Sexually dimorphic development of mouse primordial germ cells: switching from oogenesis to spermatogenesis. Development, 129:1155–1164.

22.

Gosden

RG.

, Mullan

, Picton

HM.

, Yin

, Tan

. 2002. Current perspective on primordial follicle cryopreservation and culture for reproductive medicine. Hum Reprod Update, 8:105–110.

23.

Anderson

, Albertini

. 1976. Gap junctions between the oocyte and companion follicle cells in the mammalian ovary. J Cell Biol, 71:680–686.

24.

Hertig

AT.

, Adams

. 1967. Studies on the human oocyte and its follicle. I. Ultrastructural and histochemical observations on the primordial follicle stage. J Cell Biol, 34:647–675.

25.

Zamboni

1970. Ultrastructure of mammalian oocytes and ova. Biol Reprod, 2,Suppl 2:44–63.

26.

Eppig

. 1977. Mouse oocyte development in vitro with various culture systems. Dev Biol, 60:371–388.

27.

Eppig

JJ.

, Schroeder

. 1989. Capacity of mouse oocytes from preantral follicles to undergo embryogenesis and development to live young after growth, maturation, and fertilization in vitro. Biol Reprod, 41:268–276.

28.

O'Brien

MJ.

, Pendola

JK.

, Eppig

. 2003. A revised protocol for in vitro development of mouse oocytes from primordial follicles dramatically improves their developmental competence. Biol Reprod, 68:1682–1686.

29.

Daniel

SA.

, Armstrong

DT.

, Gore-Langton

. 1989. Growth and development of rat oocytes in vitro. Gamete Res, 24:109–121.

30.

Roy

SK.

, Greenwald

. 1989. Hormonal requirements for the growth and differentiation of hamster preantral follicles in long-term culture. J Reprod Fertil, 87:103–114.

31.

Jewgenow

1998. Role of media, protein and energy supplements on maintenance of morphology and DNA-synthesis of small preantral domestic cat follicles during short-term culture. Theriogenology, 49:1567–1577.

32.

Hirao

, Nagai

, Kubo

, Miyano

, Miyake

, Kato

. 1994. In vitro growth and maturation of pig oocytes. J Reprod Fertil, 100:333–339.

33.

Cecconi

, Barboni

, Coccia

, Mattioli

. 1999. In vitro development of sheep preantral follicles. Biol Reprod, 60:594–601.

34.

Huanmin

, Yong

. 2000. In vitro development of caprine ovarian preantral follicles. Theriogenology, 54:641–650.

35.

Harada

, Miyano

, Matsumura

, Osaki

, Miyake

, Kato

. 1997. Bovine oocytes from early antral follicles grow to meiotic competence in vitro: effect of FSH and hypoxanthine. Theriogenology, 48:743–755.

36.

Gutierrez

CG.

, Ralph

JH.

, Telfer

EE.

, Wilmut

, Webb

. 2000. Growth and antrum formation of bovine preantral follicles in long-term culture in vitro. Biol Reprod, 62:1322–1328.

37.

Roy

SK.

, Treacy

. 1993. Isolation and long-term culture of human preantral follicles. Fertil Steril, 59:783–790.

38.

Van Blerkom

, Peitro

. 1979. The Cellular Basis of Mammalian Reproduction. Urban & Schwarzenberg: München.

39.

Da Silva-Buttkus

, Jayasooriya

GS.

, Mora

JM.

, Mobberley

, Ryder

TA.

, Baithun

, Stark

, Franks

, Hardy

. 2008. Effect of cell shape and packing density on granulosa cell proliferation and formation of multiple layers during early follicle development in the ovary. J Cell Sci, 121:3890–3900.

40.

Suzuki

, Kitai

, Tojo

, Seki

, Oba

, Fujiwara

, Iizuka

. 1981. Ultrastructure and some biologic properties of human oocytes and granulosa cells cultured in vitro. Fertil Steril, 35:142–148.

41.

Weakley

. 1966. Electron microscopy of the oocyte and granulosa cells in the developing ovarian follicles of the golden hamster (Mesocricetus auratus) J Anat, 100:503–534.

42.

Rotmensch

, Dor

, Furman

, Rudak

, Mashiach

, Amsterdam

. 1986. Ultrastructural characterization of human granulosa cells in stimulated cycles: correlation with oocyte fertilizability. Fertil Steril, 45:671–679.

43.

Amsterdam

, Rotmensch

. 1987. Structure-function relationships during granulosa cell differentiation. Endocr Rev, 8:309–337.

44.

Makino

, Ozaki

, Matsubara

, Sato

, Ikuta

, Nishizawa

, Suzumori

. 2005. Role of apoptosis controlled by cytochrome c released from mitochondria for luteal function in human granulosa cells. Am J Reprod Immunol, 53:144–152.

45.

Barastegui

CA.

, Ruano-Gil

, Ros

, Juvells

, Bargallo

. 1984. Atypical cristae (paracrystalline inclusions) in mitochondria of epithelial cell of the rat ureter. II. Image analysis. J Submicrosc Cytol, 16:727–733.

46.

Romert

, Matthiessen

. 1986. Tetracycline-induced changes in hepatocytes of mini-pigs and mini-pig foetuses as revealed by electron microscopy. Acta Pathol Microbiol Immunol Scand A, 94:125–131.

47.

Skinnider

LF.

, Ghadially

. 1976. Chloramphenicol-induced mitochondrial and ultrastructural changes in hemopoietic cells. Arch Pathol Lab Med, 100:601–605.

48.

Bjersing

1967. On the ultrastructure of granulosa lutein cells in porcine corpus luteum. With special reference to endoplasmic reticulum and steroid hormone synthesis. Z Zellforsch Mikrosk Anat, 82:187–211.

49.

Schmidt

CL.

, Kendall

JZ.

, Dandekar

PV.

, Quigley

MM.

, Schmidt

. 1984. Characterization of long-term monolayer cultures of human granulosa cells from follicles of different size and exposed in vivo to clomiphene citrate and hCG. J Reprod Fertil, 71:279–287.

50.

Crisp

TM.

, Dessouky

DA.

, Denys

. 1970. The fine structure of the human corpus luteum of early pregnancy and during the progestational phase of the mestrual cycle. Am J Anat, 127:37–69.

51.

Canipari

2000. Oocyte—granulosa cell interactions. Hum Reprod Update, 6:279–289.

52.

Schroeder

AC.

, Eppig

. 1984. The developmental capacity of mouse oocytes that matured spontaneously in vitro is normal. Dev Biol, 102:493–497.

53.

Sirard

MA.

, Parrish

JJ.

, Ware

CB.

, Leibfried-Rutledge

ML.

, First

. 1988. The culture of bovine oocytes to obtain developmentally competent embryos. Biol Reprod, 39:546–552.

54.

Vanderhyden

BC.

, Armstrong

. 1989. Role of cumulus cells and serum on the in vitro maturation, fertilization, and subsequent development of rat oocytes. Biol Reprod, 40:720–728.

55.

Buccione

, Schroeder

AC.

, Eppig

. 1990. Interactions between somatic cells and germ cells throughout mammalian oogenesis. Biol Reprod, 43:543–547.

56.

Rodgers

RJ.

, Irving-Rodgers

HF.

, Russell

. 2003. Extracellular matrix of the developing ovarian follicle. Reproduction, 126:415–424.

57.

Suzuki

, Jeong

BS.

, Yang

. 2000. Dynamic changes of cumulus-oocyte cell communication during in vitro maturation of porcine oocytes. Biol Reprod, 63:723–729.

58.

Jackowski

, Dumont

. 1979. Surface alterations of the mouse zona pellucida and ovum following in vivo fertilization: correlation with the cell cycle. Biol Reprod, 20:150–161.

59.

Eager

DD.

, Johnson

MH.

, Thurley

. 1976. Ultrastructural studies on the surface membrane of the mouse egg. J Cell Sci, 22:345–353.

60.

Ducibella

, Anderson

, Albertini

DF.

, Aalberg

, Rangarajan

. 1988. Quantitative studies of changes in cortical granule number and distribution in the mouse oocyte during meiotic maturation. Dev Biol, 130:184–197.

61.

Nottola

SA.

, Macchiarelli

, Coticchio

, Bianchi

, Cecconi

, De Santis

, Scaravelli

, Flamigni

, Borini

. 2007. Ultrastructure of human mature oocytes after slow cooling cryopreservation using different sucrose concentrations. Hum Reprod, 22:1123–1133.

62.

Van Blerkom

1990. Occurrence and developmental consequences of aberrant cellular organization in meiotically mature human oocytes after exogenous ovarian hyperstimulation. J Electron Microsc Tech, 16:324–346.

63.

Odor

DL.

, Blandau

. 1969. Ultrastructural studies on fetal and early postnatal mouse ovaries. II. Cytodifferentiation. Am J Anat, 125:177–215.

64.

Hirshfield

. 1991. Development of follicles in the mammalian ovary. Int Rev Cytol, 124:43–101.

65.

Devine

PJ.

, Payne

CM.

, McCuskey

MK.

, Hoyer

. 2000. Ultrastructural evaluation of oocytes during atresia in rat ovarian follicles. Biol Reprod, 63:1245–1252.

66.

Lobascio

AM.

, Klinger

FG.

, Scaldaferri

ML.

, Farini

, De Felici

. 2007. Analysis of programmed cell death in mouse fetal oocytes. Reproduction, 134:241–252.

67.

Escobar

ML.

, Echeverria

OM.

, Ortiz

, Vazquez-Nin

. 2008. Combined apoptosis and autophagy, the process that eliminates the oocytes of atretic follicles in immature rats. Apoptosis, 13:1253–1266.

68.

Ortiz

, Echeverria

OM.

, Salgado

, Escobar

ML.

, Vazquez-Nin

. 2006. Fine structural and cytochemical analysis of the processes of cell death of oocytes in atretic follicles in new born and prepubertal rats. Apoptosis, 11:25–37.