Ovarian stem cells—Potential roles in infertility treatment and fertility preservation

Abstract

One of the principal beliefs in reproductive biology is that women have a finite ovarian reserve, which is fixed from the time they are born. This theory has been questioned recently by the discovery of ovarian stem cells which are purported to have the ability to form new oocytes under specific conditions post-natally. Almost a decade after their discovery, ovarian, or oogonial, stem cells (OSCs) have been isolated in mice and humans but remain the subject of much debate. Studies in mice have shown that these cells can be cultured to a mature oocyte stage in vitro, and when injected into germ-cell depleted ovary they can form follicles and have resulted in the birth of healthy offspring. There are few data from human OSCs but this finding would open the door to novel fertility preservation strategies for women with both age-related and premature ovarian insufficiency (POI). As the number of girls and young women surviving cancer increases worldwide, POI secondary to gonadotoxic treatments, such as chemotherapy, is becoming more common. The ideal fertility preservation approach would prevent delays in commencing life-saving treatment and avoid transplanting malignant cells back into a woman after treatment: OSCs may offer one route to achieving this. This review summarises our current understanding of OSCs and discusses their potential clinical application in infertility treatment and fertility preservation.

Introduction

The dogma that female mammals are born with all of the oocytes they will ever possess has its foundations in a paper from Sir Solomon Zuckerman published in 1951 [1]. Simply put, Zuckerman failed to find any experimental evidence available at that time that he felt was inconsistent with an earlier hypothesis [2] that germ cell production in female mammals ceases prior to birth (reviewed by Zuckerman) [3]. This paper and its main conclusion profoundly affected the subsequent interpretation of experimental and clinical observations relating to ovarian development, function and failure for the next 50 years. A paper published by Jonathan Tilly's laboratory in 2004 reignited this debate by reporting the presence of a population of mitotically active germline stem cells (GSCs) in the mouse ovary which, the authors postulated, maintain oocyte and follicle production in the ovary after birth [4]. The finding of GSCs, or oogonial stem cells (OSCs) as they are now more commonly known, has generated a lively debate in the field over the last decade as it is in direct opposition to the dogma that female mammals have a non-renewable oocyte reserve from birth. This debate has been perceived as representing two clearly opposing viewpoints with no common ground (reviewed by Powell) [5], but there is the possibility that both views can co-exist, with the formation of a population of oocytes at birth that is the main contributor to ovarian function and fertility and subject to little, if any, renewal and the existence of OSCs in adult ovaries that can only be activated under specific circumstances. It is impossible to prove the absence of any given cell in a tissue but the debate cannot be resolved until the presence and function of OSCs within adult ovaries can be unequivocally demonstrated.

Regardless of the physiological significance of these cells what is undeniable are the possible clinical applications of OSCs in infertility and fertility preservation if their potential can be harnessed; this review will address the background to current understanding of OSCs, and provide a speculative discussion of their potential clinical applications. If human OSCs can be grown into fully functional oocytes, can this be harnessed to address the age-related decline in oocyte quality? Could girls and young women about to undergo gonadotoxic therapy, e.g. for cancer, be able to cryopreserve some OSCs within their ovarian cortex prior to commencing treatment? Instead of concentrating on the finite number of primordial follicles within that ovarian tissue, it is conceivable that OSCs could subsequently be retrieved from this tissue and either cultured to form mature oocytes for use in in vitro fertilisation (IVF), or injected back into the woman's ovarian cortex for in vivo development. The number of new follicles that could be generated from OSCs could be much larger than the number of follicles in the stored ovarian tissue, and certainly much larger than the number of mature oocytes that a woman could store using the conventional approach of ovarian stimulation and aspiration of mature oocytes.

Johnson et al. identified cells they considered OSCs whilst investigating follicular atresia in the mouse ovary [4]. They discovered that follicles were dying at a rate such that the ovary would be deplete of oocytes far earlier than is found in vivo. Analysis of the ovary revealed ovoid cells that both immunostained for a germ-cell specific marker (mouse vasa homologue or MVH, a germ-cell specific RNA helicase) and demonstrated incorporation of 5-bromodeoxyuridine (BrdU), indicative of proliferating cells. Furthermore, these cells expressed a meiosis-specific protein (synaptonemal complex protein 3, SCP3) required to initiate meiosis for the production of oocytes. In their final set of experiments, ovarian tissue from wild-type mice was transplanted onto the ovaries of mice which ubiquitously expressed green fluorescent protein (GFP). After 3–4 weeks, the wild-type ovary contained GFP-positive oocytes surrounded by wild-type granulosa cells, persuading the authors that OSCs from the GFP mouse had initiated folliculogenesis in the wild-type mouse and that they had discovered mitotically active OSCs that had the ability to form new oocytes after birth [4].

However, scepticism surrounded the idea of OSCs amongst reproductive biologists [6], [7]. A key finding supporting claims that adult mouse ovaries retain the capacity for oogenesis came in a paper that reported that OSCs had been isolated and cultured from neonatal and adult mouse ovaries [8]. These cells, termed female germline stem cells (FGSC), were initially identified using the same criteria used by Johnson et al. [4] i.e., expression of MVH and BrdU incorporation. By employing a cell-sorting approach using an antibody against Ddx4 (DEAD box polypeptide 4, another name for MVH), the authors reported the ability to isolate and purify OSCs in mice. Furthermore, by transplanting GFP-positive OSCs into the ovaries of infertile mice, they were able to produce live GFP-positive offspring.

The main findings of this key study were developed further by White et al. [9] who not only managed to isolate human OSCs using DDX4 (the human orthologue of MVH, or VASA), but they were able to isolate, culture and form early follicle-like structures after injection of both mouse and human OSCs into ovarian tissue which was xenotransplanted into NOD-SCID (non-obese diabetic – severe combined immunodeficiency) mice to provide a suitable environment for early folliculogenesis (Fig. 1) [9].

Interestingly, and from an entirely separate line of evidence, the case for post-natal neo-oogenesis has been bolstered by a recent analysis of the accumulation of microsatellite mutations in the germline in female mice. There was a positive correlation between thus-determined oocyte ‘depth’ and mouse age, i.e. oocytes in older mice were found to have undergone more mitotic divisions than those in younger mice [10]. Although this may be in part explained by the “production-line” hypothesis, whereby oocytes are ovulated in the order in which they were formed [11], that explanation does not appear adequate for the degree of ‘depth’ identified, and the alternative explanation of post-natal de novo oogenesis could not be ruled out experimentally by the authors [12].

Of note, OSCs are not the only stem cells to have been shown to produce oocyte-like cells under the right conditions. Hayashi et al. recently published evidence that both embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) can be induced into primordial germ cell-like cells (PGCLCs) which, upon combination with gonadal somatic cells and transplantation into mice, can generate germinal vesicle-stage oocytes [13]. These oocytes can then be matured in vitro and, after IVF, can produce offspring which are themselves fertile. This study highlights the need to consider the appropriate stage of somatic cell support and so far has only been carried out in mice, but it offers an alternative model of germ cell development.

The isolation of OSCs has been independently performed by only a small number of other groups to date [14], [15], [16] although full corroboration is as yet lacking. Many criticisms have been levelled at the experimental techniques utilised and the interpretation of the findings in these OSC experiments [7], [17]. For example, mathematical modelling has been used to challenge Johnson et al.’s follicular atresia rate finding [18], although mathematical data supporting the theory of post-natal folliculogenesis in the mouse has also been produced [19]. Critics have been sceptical of the use of Ddx4/DDX4 as a cell surface marker in cell sorting as it has an intracytoplasmic localisation in oocytes [20]. However, it has been proposed that Ddx4/DDX4 has a transmembrane-spanning domain in OSCs, before becoming intracellular in more mature oocytes [9], although experimental evidence for this is still lacking. The argument for the existence of OSCs has been further strengthened by subsequent work reporting the isolation of OSCs in mice using alternative markers for cell sorting [14], [21]. Wu's lab showed improved purification efficiency by employing interferon-inducible transmembrane protein 3 (Ifitm3, or fragilis), a widely accepted early germ cell-specific surface marker [21], whilst Pacchiarotti et al. used transgenic mice that expressed GFP under the control of another germline-specific marker, Oct-4 [14], [22].

It has been questioned why these cells have gone undiscovered for so long [6]; however, it seems that OSCs are exceedingly rare, with White et al., reporting that they constitute a mere 0.014% of all cells in mouse ovaries [9]. They also seem to become rarer with increasing age, declining from 1 to 2% of cells in the neonatal mouse ovary to 0.05% in the adult [13]. The variation between values reported by different groups is likely to reflect the very different isolation methods used. This also perhaps explains the timespan between the initial purported discovery of these cells and their isolation from ovarian cortex. Furthermore, there are concerns about in vitro transformation given that the length of time it takes for these cells to establish in culture (10–12 weeks in the mouse and 4–8 weeks in humans) [9] is much greater than for the equivalent cells in the male [23]. Still, live offspring have been produced from such OSC cultures in mice [8] and early follicle-like structures have been generated in human tissue [9]. Despite these on-going disputes, the discovery of OSCs offers exciting new potential strategies for clinical application.

The age-related decline in female fertility is a central feature of human reproduction [24]. Whilst not the only aspect of concern, it is within the ovary and oocyte that most of this decline resides and will be of greater significance as pregnancy at advanced age becomes more prevalent. Currently the main strategy reproductive medicine has to offer is the use of oocyte donation: whilst undoubtedly successful, it is of huge consequence both for individual couples and for society [25]. ‘Social’ oocyte storage is increasingly prevalent, but even with multiple ovarian stimulation cycles the number of oocytes will be limited, and the financial costs high. Might OSCs have a role here? The isolation of OSCs from older women indicates the possibility, and the fact that they are capable of proliferation (i.e. they are pre-meiotic) means that the age-related aneuploidy risk might not exist, or at least be much less of an issue. OSCs also provide an unparalleled opportunity for the improved understanding of the basis of such clinically relevant aspects of human oocyte biology. Options for both the treatment and prevention of age-related fertility loss therefore exist at least in theory, but are critically dependant on understanding of whether, and why, neo-oogenesis ceases at the menopause.

The survival rates from childhood cancer have increased significantly over the last 15–20 years [26], resulting in many young women suffering from premature ovarian insufficiency (POI) after being exposed to gonadotoxic treatment. At present, approaches available to women who want to attempt to preserve their fertility prior to gonadotoxic therapy include having oocytes or embryos cryopreserved after undergoing ovarian stimulation (and subsequent IVF with their partner's or donor's sperm in the case of embryo cryopreservation). Unfortunately, both of these approaches involve hormonal medication and lead to delays in commencing treatment for the patient's disease. The alternative option, which is the only approach appropriate for pre-pubertal girls [27], [28], is cryopreservation of ovarian tissue with subsequent re-transplantation into the woman after cessation of her treatment. The first livebirth from this method was in 2004 when a woman with successfully treated Hodgkin's lymphoma underwent orthotopic autotransplantation of cryopreserved ovarian tissue [29] and there have been a number of livebirths reported since [30]. There have, however, been concerns surrounding this approach with certain diagnoses as there is a risk of reintroducing malignant cells within the transplant into the patient, particularly with haematological malignancies [27], [31], [32]. Consequently, an approach that avoids a delay in commencing life-saving treatment and employs a method to prevent transplanting cancer cells back into a patient would be preferable. OSCs may offer the potential to provide this approach.

To date, human OSCs have only been grown to early follicle-like structures in a xenotransplantation model [9], which is unacceptable in clinical use. However, the development of a multi-step culture system in cows and humans that supports folliculogenesis and oocyte growth from the primordial follicle stage means that there is the potential to produce mature oocytes from OSCs completely in vitro, if the correct somatic cell support is available [33], [34] (Fig. 2). This serum-free culture model supports oocyte development in a shorter time frame compared with that seen in vivo, by culturing primordial follicles in small ovarian cortical strips to the pre-antral stage, before these pre-antral follicles are dissected out and cultured individually in activin-supplemented media [33], [34]. Once the follicles reach the antral stage, the oocyte–granulosa cell complexes (OGCs) can be removed and placed on membranes for a final period of development prior to undergoing in vitro maturation (IVM) [35], [36].

Oocyte development and maturation is a multifaceted process and, as such, livebirths have only been achieved with follicles/oocytes grown from the primordial stage in mice [37], [38]. In particular, there are concerns that in vitro oocyte culture and IVM may interfere with the complex genome imprinting stages and epigenetic mechanisms that are required for the development of a fully competent oocyte and subsequent embryo [39]. Animal studies have so far been reassuring; human studies will be a very considerable challenge.

With regard to human OSCs, if they can be transplanted into human ovarian cortex and the ensuing primordial follicles cultured in vitro in a serum-free multi-step system, any resultant mature, competent oocytes could be used in IVF. This opens the door to a novel fertility preservation approach that would be of use in those female patients who, for reasons of age or urgency for treatment, cannot have mature oocytes removed prior to gonadotoxic treatments that may render them infertile. Additionally, the expansion of OSC number in culture may allow for a considerably greater number of oocytes to be available to the women.

Another possible strategy is the injection of isolated OSCs into a patient's ovaries, where they, theoretically, could undergo neo-oogenesis in vivo and generate an entire population of hormone-secreting follicles as well as allowing a more physiological oocyte maturation. This may have the additional benefit of reversing the climateric symptoms and general health consequences associated with the menopause, at least temporarily. As with in vitro folliculogenesis, this remains a distant prospect at present.

Although an increasing number of women with POI have an underlying iatrogenic cause [40], the aetiology of POI is heterogeneous, with the majority of cases being idiopathic [41]. Unless a family is aware of a hereditary genetic cause, women with non-iatrogenic POI often present too late for current fertility preservation methods to be feasible. At present, these women can either hope to spontaneously conceive (with an approximately 5% chance of spontaneously conceiving following diagnosis [41]), or opt to use donor oocytes and IVF. Hypothetically, OSCs could also be employed in the management of these women as it may be possible to isolate OSCs from these women's ovaries, despite their menopausal state. Indeed, OSCs have already been isolated in aged mouse ovaries and have been shown to undergo oogenesis when transplanted into a young mouse [42]. Again, these OSCs could be cultured as above and used in IVF.

Of note, White et al. have shown that OSCs can be isolated from cryopreserved, as well as fresh, human ovarian tissue [9]. This is encouraging from a fertility preservation point-of-view as it means that ovarian cortical tissue removed from girls or women can be safely stored until OSCs are required, thus allowing for technical advances in their isolation.