Perspective in infertility: the ovarian stem cells

Both hypogonadotropism or hypergonadotropism concur to ovary failure in approximately 20 % of the infertile women and the clomiphene citrate (CC) has been widely adopted in the past for generic ovulatory dysfunctions. Recently, the metformin has been shown to improve the rate of ovulation in particular in the polycystic ovary syndrome unresponsive to CC and/or to gonadotropins [2].

The intrauterine semen insemination is recommended for the infertility due to cervical factors or to the male ejaculatory dysfunction. The procedure is simple since the donor sperm is inseminated in parallel with induction of the ovulation, but the success rate is not so high.

This procedure was initially developed in 1978 by Steptoe and co-workers for treating the tubal dependent infertility but is currently used for other types of infertility [3]. The procedure includes four sequential steps as follows:

Preliminary observation on OSCs dates on 1951 by Zuckerman et al. [5]. They asserted a long-held dogma according to which in postnatal mammalian ovaries of most species no renewable germinal OSCs are available and, thus, supporting the hypothesis that during the lifetime a numerically fixed pool of oocytes are committed for the fertility. This original pool would account in woman for approximately 10⁶ oocytes in puberty, but the number declines with aging until exhaustion at menopause.

This assumption was subsequently refuted by Tilly and co-workers [6], who postulated the presence of mitotically active OSCs in both juvenile and adult murine ovaries capable to warranty both oocyte and follicle availability after birth. In their experiments, they observed a discordance between the rate of follicle depletion and the reproductive lifespan and found normal histological aspect in ovaries from mice receiving Busulfan showing healthy maturing follicles and corpora lutea. Thus, they concluded that oocytes underwent normal differentiation in these ovaries and grafted ovarian fragments from adult wild type mice in ovaries of green fluorescent protein (GFP)-expressing mice with the aim to explore the occurrence of GFP-positive oocytes in wild-type grafts. The granulosa cells surrounding GFP-positive oocytes within the transplanted fragment were found to be GFP-negative thus suggesting that transgenic OSCs migrated into the graft and formed new follicles in adult mice. Immediate criticisms were raised to these observations and focused not only the methods used to distinguish healthy from unhealthy follicles, but also the computation system used as well as the Busulfan treatment which is a well-known cytotoxic agent for spermatogonial cells, but with unknown effects on OSCs [5].

Other investigators confuted the Tilly’s hypothesis and simulated the dynamics of murine follicle progression during the mouse lifespan by employing two mathematical models, namely the ‘stem cell’ and the ‘fixed pool’ model and argued that the physiological decline in follicle populations supported the theory of ‘fixed pool population’ of oocytes as lifetime ovarian reserve [7]. However, the debate continued when Tilly and colleagues speculated the existence of a putative reservoir of OSCs in the BM of adult mice. In fact, they primarily described the expression of germ cell line markers including Oct4, MVH (mouse vasa homologue), DazI, Stella and Fragilis in BM from adult female mice and then transplanted their BM into adult females pre-sterilized with both Cyclophosphamide and Busulfan, and detected a significant generation of new oocyte-containing follicles and corpora lutea [8]. This result was confirmed a few years later by transplanting BM from GFP-expressing into wild type mice previously sterilized with Cyclophosphamide and Busulfan. The offspring produced by treated females derived from the recipient germline compartment, suggested that the transplantation rescued fertility by either protecting existing oocytes from cytotoxic agents, or reestablishing in the host the somatic cell population, necessary to reinstitute oogenesis [9]. Thus, these preliminary observations gave rise to further studies aimed at OSC isolation and transfer into sterile animals to recover their fertility.

The first attempt to isolate and culture germinal OSCs in mammals was completed by Zou and co-workers who purified neonatal and adult female OSCs from mice through a cell-sorting approach [10]. They first performed a two-step enzymatic digestion of murine ovarian tissue followed by immunomagnetic isolation of cells expressing the germinal marker MVH, also known as Ddx-4 (DEAD box polypeptide 4). This protein is a 79.3 kDa peptide with the extracellular COOH-domain and a cytoplasmatic tale that is uniquely expressed by germinal ovarian and spermatogonial cells [11]. After culturing OSCs from neonatal and adult mice for more than 15 and 6 months respectively, the Authors infected the isolated Ddx-4⁺ cells with the MSCV-PGK-GFP virus and transplanted them into ovaries of infertile female mice. They ultimately observed that the animals produced GFP⁺/Ddx-4⁺ oocytes.

Therefore, Tilly and collaborators validated a FACS-based protocol to purify OSCs from both murine and human ovarian cortex (Fig. 1) as putative location [12]. The investigation was based on the immunological detection of a putative cell surface variant of Ddx-4, commonly considered cytoplasmic, through a rabbit polyclonal antibody against the COOH-terminus of the protein. Once purified by sorting, the cells showed a germline gene expression pattern and were thus established in culture, engineered to express GFP and injected into biopsies of adult human ovarian cortical tissue. The fragments were subsequently xenografted into immunodeficient female mice and new follicles containing GFP-positive oocytes were detected 1–2 weeks later. Major skepticism arose with respect to the choice of using Ddx-4 as cell surface marker used for the OSC sorting since its localization apparently occurs in cytoplasm rather than on plasma membrane [13]. On the other hand, to strengthen the criticism, Zhang and colleagues [14] generated a fluorescent mouse model for both in vivo and in vitro tracing of Ddx-4 expressing OSCs, and in the absence of fluorescence in ovaries they stated that OSCs did not enter in mitosis, nor contributed to the oocyte renewal. Table 2 lists an incomplete panel of markers differentially expressed by OSCs and oocytes.

Another group of investigators described the presence of both very small embryonic-like stem cells (VSELs) and germinal OSCs within the ovarian surface epithelium (OSE) of adult mammals. To obtain these cells from OSE, the samples were enzymatically digested or mechanically scraped and the putative OSCs were analyzed for the expression of OCT-4A, SSEA-4(stage specific embryonic antigen-4), Fragilis, CD133 and others. These cells were then maintained in continuing cultures for 21 days and underwent spontaneous differentiation into oocyte-like structures suggesting that to identify OSCs from ovaries it is preferable to use stemness markers rather than Ddx-4 [15]. Based on these observations, the same Authors set up an immunomagnetic sorting of OSE stem cells from sheep ovaries using SSEA-4 as major surface marker. Therefore, the cells were typed by immunocytochemistry for specific markers of both staminal and germ line profiles as well as by immunofluorescence and RT-PCR, and then cultured for 3 weeks. The SSEA-4⁺ OSE stem cells from those animals underwent dynamic changes resembling spontaneous differentiation into oocyte-like structures [16].

In subsequent studies, Zou and co-workers [17] further tried to isolate OSCs by magnetic sorting through three germinal markers, namely CD9, Stpb-c (short-type pituitary gland and brain-cadherin) and Fragilis. However, the isolation of OSCs efficiently turned out only by targeting the protein Fragilis, since CD9 was detected on membranes of both germ and somatic cells, including those of granulosa, whereas Stpb-c was inapplicable since the appropriate primary antibody was actually unavailable.

Based on these different procedures, in an attempt to investigate the occurrence of OSCs in ovaries from adult women, we tried to isolate these cells from freshly collected ovarian cortical tissue, using a Ddx-4 antibody-based magnetic sorting, as described by Woods and Tilly [18]. Fig. 2 depicts the major steps of this procedure.

To verify the enrichment in OSCs of the cell suspension obtained by immunomagnetic selection, we evaluated their phenotype by flow cytometry using both FITC-conjugated anti-Ddx-4 and PE-conjugated anti-human OCT4A as ovary lineage specific marker. The flow cytometry analysis is illustrated in Fig. 3. The Ddx4⁺ cell population was increased up to 24 % (lower section: right) of the original 2 % (left) value as detected in the initial cortical ovarian suspension, with a remarkable fluorescence intensity suggesting the high expression of Ddx4 molecule. Ddx4 was also detected by confocal microscopy revealing high Ddx4 fluorescence intensity that supported the flow cytometry data (Fig. 4). Therefore, the evidence of Ddx4 on OCT4A⁺ cells from ovarian cortex by both flow cytometry and fluorescence microscopy provided further evidence to the existence of OSCs in the human ovary and that their sorting is reliable by appropriate combined separation methodology.

Once isolated from the OSE, OSCs have been investigated from several Authors for their potential application in infertility treatment. Indeed, Hayashi and coworkers induced murine female embryonic stem cells to differentiate into primordial germ cell–like cells (PGCLCs) that were utilized to reconstitute ovaries. After transplantation under the ovarian bursa of adult mice, the reconstituted ovaries were then isolated and histologically analyzed, and the PGCLCs were revealed to contribute to the oocyte formation. Once reaching metaphase II, these cells were thus fertilized and generated two-cell embryos that, transferred into foster mothers, developed pups able to reach adulthood [19].

If working on animal models was quite easy, direct investigation of OSCs in humans revealed more troubles both for ethical issues and for the scarcity of OSCs in adult OSE. Among the groups exploring this topic, Stimpfel and co-workers successfully characterized and differentiated in vitro stem cells from the adult human ovarian cortex [20]. The isolated cells expressed pluripotency markers as alkaline phosphatase, SSEA-4 and OCT4 together with Ddx4 as prototypic ovarian germ line marker. These cells exhibited a high degree of plasticity since they, once adequately stimulated, differentiated in various somatic cells of mesoderm, ectoderm and endoderm layers, without forming teratomas in immunodeficient mice. However, this may also depend on the cell exiguity as well as on their slow proliferation.

At present, an intriguing theory postulates that the impaired bioenergetics capacity in oocytes correlates with the decline of quality in both eggs and embryos in particular in advanced age of fertile women since recurrent morphologic alterations are frequently observed in the oocyte mitochondria of adult women. This concern led to define a clinical protocol using the oocyte cytoplasm extracted from a young donor that was transferred into the eggs of the recipient woman. The procedure, however, had short life because of concerns regarding the consequent mithocondrial heteroplasmy, whose result was that children generated by this protocol possessed three different sources of genetic material, namely the biological parents and the oocyte donor, for the presence of maternally inherited DNA in mithocondria. Therefore, according to the U.S. Food and Drug Administration (FDA), this procedure was considered a genetic manipulation of human germ cells and was no longer applied [21].

The discovery that OSCs are present in fertile women ovaries has driven to an emerging modified version of ooplasmic transfer, called AUGMENT (autologous germline mitochondrial energy transfer) in which autologous germline-derived cytoplasmic extract or purified mitochondria can be used for the bioenergetics reinforcement of oocytes. Another potential approach to overcome the defective egg energy includes the improvement of both biological and chemical compounds bioavailability, with the purpose to enrich the mitochondrial number or the efficacy of ATP production in oocytes [21, 22].

These studies emphasize the increasing interest in human OSCs for future application in treatment of human infertility that undoubtedly needs further investigation in this specific topic.