Making gametes from alternate sources of stem cells: past, present and future

Making gametes in a Petri dish by directed differentiation of human pluripotent embryonic and induced pluripotent stem cells (hES/iPS) is considered one of the most important goals of stem cells research to help infertile couples attain biological parenthood. Research efforts by several groups across the globe have been focused to use ES cells grown in a Petri dish to differentiate into gametes for almost 3–4 decades based on when mouse [1, 2] and human [3] ES cells were first reported; and almost a decade of iPS cells research as they were first reported in 2006 [4]. Since primordial germ cells (PGCs) that arise from the epiblast- stage embryo are the natural precursors to the gametes [5, 6], the first crucial step involves conversion of ES/iPS cells into functional PGC-like cells (PGCLCs) which then spontaneously differentiate into gametes in vitro or when appropriate niche is provided in vivo. This conversion of pluripotent stem cells into PGCLCs remains a big challenge and Hayashi’s group successfully converted mouse ES/iPS cells into PGCLCs [7] whereas specification of human ES/iPS cells into PGCLCs still remains challenging [8, 9]. Efforts are also ongoing to convert primed human ES cells into naïve state to enhance their differentiation ability, as the naïve human ES/iPS cells may be better starting material to make human PGCLCs [10, 11]. This is because whereas mES/iPS cells exist in naïve state, hES/iPS cells are primed in nature being closer to stem cells derived from mouse epiblast state embryo, which do not exhibit any potential to differentiate into the germ cell fate. Research efforts are also directed to convert adult stem cells including spermatogonial stem cells (SSCs) and ovarian stem cells (OSCs) into gametes. Readers may refer to few recent reviews in the field [12‐15].

There exists an additional novel population of pluripotent stem cells termed very small embryonic-like stem cells (VSELs) in all adult organs including testis and ovary, which can also be differentiated into gametes in vitro. Pluripotent VSELs in reproductive tissues were recently reviewed [16] and reasons why they have remained elusive so far were discussed [17]. It is well understood that during early development, PGCs migrate to the gonadal ridge where they differentiate into germ cells and cease to exist thereafter. However, it has been suggested that PGCs migrate not only to the gonads but to all developing organs and survive in few numbers throughout life [18]. There exists a developmental link between PGCs and VSELs in adult tissues and this explains why VSELs in hematopoietic system also express pituitary and sex hormone receptors [19]. Shaikh et al. [20] reported that mouse bone marrow VSELs besides exhibiting the ability to differentiate into 3 germ layers in vitro, in agreement with other reports [18, 21, 22], also differentiate into hematopoietic stem cells and germ cells when conducive culture conditions are provided. When cultured on a Sertoli cell bed and in the presence of Sertoli cells conditioned medium and FSH, VSELs differentiated into germ cells that expressed DAZL, STRA8 and GFRA and transcripts of Gfra, Vasa and Dazl [20]. Similar ability of SSEA1+ pluripotent stem cells from bone marrow to differentiate into male germ cells was recently reported by another group [23]. VSELs have been reported in adult testis [24‐27], ovary [28‐32] and uterus [33, 34]; express receptors for follicle stimulating hormone and survive oncotherapy due to their quiescent nature.

Being developmentally equivalent to PGCs, which are natural precursors to the gametes, VSELs are attractive alternative, pluripotent stem cells in adult gonads to obtain gametes. However, VSELs are not yet widely accepted by the reproductive biologists because of their small size and scarce nature. A recent update on fertility preservation published in September 2017 [35] does not even acknowledge VSELs and this is the main reason for compiling the present review article and compare them with other existing options on which research is ongoing to help infertile couples. Work needs to progress in different directions to ultimately help infertile couples including cancer survivors to achieve biological parenthood. We have earlier discussed that VSELs may be ideal stem cells to differentiate into gametes [36, 37] and that use of VSELs to attain biological parenthood may help avoid legal, ethical and safety issues associated with oncofertility [38]. Year 2016 was remarkable in that significant strides were taken in the field to differentiate stem cells into gametes. An update on various alternative sources of stem cells and progress made using them to obtain sperm and oocytes in vitro is provided in the review.

Geijsen et al. [39] isolated SSEA1+/OCT4+ PGCs from mouse ES cells derived embryoid bodies. Treating these PGCs with retinoic acid resulted in the derivation of continuously growing embryonic germ cell lines that also showed erasure of methylation marks. These cells underwent further differentiation into functional haploid male gametes which when injected into oocytes resulted in the development of blastocysts. Nayernia’s group reported differentiation of mouse ES cells into sperm resulting in the birth of fertile offspring [40]. The mES cells differentiated into SSCs, which underwent meiosis in vitro, generated haploid sperm which were capable of fertilizing mouse oocytes after ICSI and resulted in the birth of live pups. However, a follow up publication showing hES cells also exhibit similar ability to produce sperm in vitro [41] was later retracted. Later research efforts were focused to first convert mES cells into PGCLCs and Hayashi et al. [7] transplanted mES cells derived PGCLCs into testicular environment that led to the formation of haploid sperm. But none of these studies demonstrated functionality of the in vitro derived gametes. Equizabel et al. [42] reported complete differentiation of human iPS cells into post-meiotic germ cells. iPS cells derived from various sources (skin, cord blood) on a human foreskin fibroblast feeder support successfully differentiated into germ cells. They discussed that iPS cells are epigenetically better predisposed to differentiate into germ cells as they may harbor some predisposing epigenetic memory of the original somatic cell in agreement with earlier reports [43]. Zhou et al. [44] first successfully reproduced complete maturation of mES cells derived PGCLCs that were further cultured on a testicular somatic cells support with sequential exposure to morphogens and sex hormones could reproduce meiosis resulting in haploid spermatid, which also resulted in viable and fertile offspring after ICSI. Highlight of the study was that they could track various events during differentiation in a very nice manner. Meiosis was tracked in vitro by studying sequential expression of specific markers. Initially chromosomal synapsis and DNA double stranded breaks and their resolution by homologous recombination repair monitored by studying expression of SPO11 and RAD51. Expression pattern of phosphorylated H2A histone family member X recapitulated meiosis progression as it was broadly distributed throughout the nucleus on D8 reflecting an association with double stranded breaks in DNA and later focal appearance on sex chromosomes suggested completion of synapsis. The nucleus showed expression of SYCP1 and SYCP3. Later an up-regulation of meiotic markers Dmc1, Stra8 and Sycp3 was observed by D10 followed by up-regulation of transcripts specific for haploid cells including Prm1, haprin and acrosin. However, success rate to obtain pups using gametes derived from mES cells of 1.89–6.67% is much lower compared to 9.46% in controls. Medrano et al. [45] converted human fibroblasts directly into haploid cells by over expression of 6 germ-line related factors. Whole genome analysis of spermatogonial cells obtained by differentiation of mES cells revealed aberrant methylation pattern, which could restrain their spermatogenic potential [46]. Saitou and Miyauchi [15] suggested that the epigenetic status of the differentiated gametes and embryos obtained by Zhou et al. [44] and Medrano et al. [45] need to be further investigated. The inefficient outcome and emerging data highlight safety issues associated with in vitro differentiated gametes from ES/iPS cells.

Our group has reported VSELs in adult human [24] and mouse [25‐27] testis. Other groups have suggested possible presence of pluripotent stem cells as a sub-population among SSCs in adult human testis as well. Lim et al. [47] reported that SSEA-4 (pluripotent marker) and GFRA (marker for spermatogonial stem cells) do not co-express in testicular cells; Izadyar et al. [48] reported a sub-population among SSCs with pluripotent characteristics; Stimpfel et al. [49] reported pluripotent markers in frozen testicular biopsies of azoospermic men and Virant-Klun et al. [50] reported SSEA-4 positive stem cells in testicular tissue biopsies of azoospermic men by FACS analysis. For almost a decade between 2004 and 2014, lot of excitement was generated when different groups described ES-like colonies on culture of testicular tissue and it was proposed that testicular stem cells are the only stem cells in the body that can spontaneously reprogram to pluripotent state (discussed in details in [16]). Rather than reprogramming of SSCs, it is the VSELs that possibly grow as ES-like colonies as discussed earlier [51].These VSELs were found to survive oncotherapy in mouse [25‐27] and human [52] testis. Kurkure et al. [52] have reported presence of VSELs along with Sertoli cells in the azoospermic testicular tissue biopsies of survivors of different kinds of childhood cancers subjected to various regimens of oncotherapy. The underlying reason for quiescence of VSELs lies in their epigenetic status and developmental origin. Being an overlapping population of PGCs, VSELs exhibit biallelic expression of imprinted genes including high levels of H19 and no IGF2 [53]. Every time a VSELs undergoes asymmetric cell division to self-renew and give rise to a SSC, genomic imprinting with increased (monoallelic) expression of IGF2 and reduced levels of H19 occurs in the SSC. As a result SSCs have greater ability to divide and undergo clonal expansion and do not survive chemotherapy. Patel and Bhartiya [27] proposed that VSELs are the most primitive stem cells in the testes that undergo asymmetric cell division to self-renew and give rise to the SSCs that further differentiate into germ cells, undergo meiosis to produce sperm. Anand et al. [25, 26] showed that VSELs and GFRA positive stem cells exist as two distinct populations of stem cells in the testis. VSELs are smaller in size, with a surface phenotype of LIN-CD45-SCA-1+ in mouse testis with nuclear OCT-4 whereas SSCs express cytoplasmic OCT-4. VSELs can be isolated as SSEA-1 positive stem cells from mouse testis and as SSEA-4 positive stem cells from human testicular tissue. Thus, the endogenous VSELs that survive oncotherapy could be targeted and strategies developed to differentiate them into sperm.

VSELs that survive chemotherapy were reported to spontaneously differentiate into sperm [54] since they are developmentally equivalent to the PGCs – natural precursors to gametes as discussed above. No growth factors were added exogenously to support differentiation/meiosis. Culture of seminiferous tubular cells from busulphan (25 mg/Kg) treated mouse was carried out using Sertoli cells conditioned medium and in the presence of FSH. Sertoli cells attached to the bottom of the culture dish to form a feeder layer whereas small sized, spherical VSELs, that survived chemotherapy were present singly at the start of culture. Over next few days, stem cells were observed to undergo asymmetric, symmetric cell divisions and clonal expansion. Three weeks culture resulted in the spontaneous differentiation of sperm and various stages of spermiogenesis were also documented. However, as the initial cells were not purified, it was possible that few mature stem/germ cells including SSCs existed in the initial cultures that could have differentiated into sperm. The approaches of in vitro differentiation of mouse VSELs by our group [54] and mES cells by Zhou et al. [44] were compared and discussed [55].

To address the above mentioned limitation of possible contamination with mature germ cells, Shaikh et al. [22] documented that VSELs purified from mouse bone marrow also differentiate into male germ cells that expressed DAZL, STRA8 and GFRA when cultured on a Sertoli cells bed and in the presence of Sertoli cells conditioned medium. The differentiated germ cells also expressed transcripts of Gfra, Vasa and Dazl. There was no chance of contaminating germ cells being present in the bone marrow. In the same month, April 2017 an article published by Nayernia’s group [23] also reported that SSEA-1positive, pluripotent VSELs from mouse bone marrow have the ability to differentiate into male germ cells that express markers specific for primordial germ cells, spermatogonial stem cells and spermatogonia including Mvh, fragilis, Dppa3, Stra8, DAZL, Piwil2, b1, and a6-integrins as well as meiotic-specific marker SYCP3. They had treated the cells with retinoic acid and cultured for longer time whereas our group [22] only added FSH and used Sertoli cells conditioned medium.

Galdon et al. [56] published an exhaustive systemic review describing the research efforts over last 50 years to achieve spermatogenesis in vitro. Reuter et al. [57] also reviewed progress on spermatogenesis using male germ cells in vitro using testicular tissue fragments, single cell suspensions or three-dimensional culture environments. Sato et al. in 2011 published two articles wherein they successfully cultured gonocytes/primitive spermatogonia in neonatal mouse testis over 2 months positioned at gas-liquid interphase [58]. Spermatids/ sperm obtained in vitro resulted in fertile offspring. They also developed a technology termed in vitro transplantation whereby they transplanted germ stem cells into removed tubules of genetically infertile mice (W/W^w) in vitro [59]. The transplanted cells colonized, proliferated and differentiated into spermatid and sperm. The detailed protocols were later published describing differentiation of germline stem cells into sperm in an organ culture – a process that took about 6 weeks in vitro [60]. Elhija et al. [61] could produce sperm in vitro starting with immature mouse testis whereas Hileihel et al. [62] used juvenile Rhesus monkey testicular cells and obtained spermatids in vitro. However, lot more needs to be done to bring this to the clinics.

Hans Scholer’s group was the first to suggest that pluripotent stem cells could differentiate into gametes [75]. Although the functionality was not proved, the authors showed formation of oocytes and follicle-like structures in vitro starting with both female and male mES cells. Embryonic stem cells were coaxed to undergo spontaneous differentiation to form embryoid bodies from which PGCs were isolated, expanded in vitro and differentiated. Later Hayashi et al. [76] used female mES/iPS cells and induced them into PGCLCs, which were then aggregated with ovarian somatic cells and on transplantation under the ovarian bursa, gave rise to GV stage oocytes that successfully contributed to the formation of fertile offspring after IVM and IVF. However, the process to obtain pups from in vitro PGCLCs derived oocytes was less efficient (approximately 3.9%) compared to in vivo PGC derived (approximately 12.7%) oocytes. Almost half of the zygotes that were obtained from the PGCLCs derived oocytes had 3PN with 2 maternal and 1 paternal chromosomes. They failed to extrude polar body. Later Hayashi and Saitou [77] published their protocols and discussed the existing limitations that need to be overcome for the robust generation of mature gametes or for application of the culture system to other species, including humans and livestock [9, 78].

The differentiation of PGCLCs and PGCs to obtain oocytes in vitro was reviewed [14, 79] and highlighted low competence of in vitro produced oocytes starting with PGCLCs. Two major success stories were reported in 2016. Complete in vitro generation of fertile oocytes starting with PGCs isolated from female fetal gonads collected on 12.5 dpc was reported [80]. Various stages during oogenesis like meiosis, oocyte growth and genomic imprinting were reproduced in vitro. It led to the formation of fertile offspring – up to7 from a cultured gonad. The frequency of development of normal fertile pups from 2-cell stage embryos starting with PGCs to derive oocytes was comparable to that observed after natural delivery in mice. Hikabe et al. [81] could reconstitute whole process of oogenesis in vitro from mES and iPS cells (derived from fetal and adult tail tip fibroblasts) to generate eggs which were further used to derive pluripotent stem cell lines. However, they reported 3.5% success rate using mES/iPS cells derived gametes compared to 61.7% using in vivo generated gametes. These results clearly suggest that the in vivo developed PGCs and stem cell derived PGCLCs in vitro are not identical and was discussed by Ge et al. [82]. Converting ES/iPS cells into PGCLCs involves extensive epigenetic changes [83] which may be difficult to attain in vitro. Hayashi et al. [84] published their protocols to produce functional primordial germ cells in vitro, which further differentiate into oocytes and healthy pups.

VSELs are located in the ovary surface epithelium along with OSCs [85].Virant-Klun’s group has initially reported in vitro differentiation of human ovarian small sized (2–4 um), pluripotent stem cells, obtained by gentle scraping of ovary surface epithelium (OSE), into oocyte-like structures [28, 29]. Later, Parte et al. [30] detected similar small sized VSELs along with slightly bigger OSCs among the ovary surface epithelial cells smears which on 3 weeks culture, spontaneously differentiated into oocyte-like structures. The ovary surface epithelial cells attach at the bottom of the culture dish to form a feeder layer whereas VSELs are observed to undergo proliferation/ differentiation into oocyte-like structures. Later, Parte et al. [86] observed that various processes like formation of germ cell nests, Balbiani body-like structures and cytoplasmic streaming extensively described during fetal ovary development, were indeed well recapitulated during in vitro oogenesis in adult OSE cultures along with characteristic expression of stem/germ cell/oocyte markers. Sriraman et al. [31] reported similar stem cells in mouse OSE cells collected by enzymatic digestion. These stem cells survived chemotherapy and could differentiate in vitro into oocyte-like structures. VSELs were recently reported by another group in adult mouse ovaries and their differentiation into oocyte-like structures in vitro [32]. However, none of these studies have tested the functional competence of these oocyte-like structures that differentiate from the VSELs.

Similar to mouse testis, VSELs survive chemotherapy in mouse ovaries also [31]. VSELs were quantitated by flow cytometry as small sized cells (3–5 μm) with a surface phenotype of LIN−/CD45−/SCA-1+ in normal adult (0.02 ± 0.008%) and chemoablated (0.03 ± 0.017%) ovaries and further treating chemoablated mice with FSH resulted in further increase of VSELs (0.08 ± 0.03%). Table 2 is a compilation of various studies reporting beneficial effects of transplanting mesenchymal cells in chemoablated ovaries. A human baby has also been born by transplanting autologus, bone marrow derived mesenchymal cells directing into the non-functional ovaries of a patient with POF [87].

Besides VSELs, OSCs also exist in the adult ovary (similar to VSELs and SSCs in the testis) and have been extensively used to generate primordial follicles using innovative strategies. In fact OSCs were detected and published initially in a landmark paper by Tilly’s group in 2004 [88]. Pluripotent SSEA4 positive VSELs were reported later in adult ovary [28, 29]. Indeed two populations of stem cells exist in ovary surface epithelium and VSELs with nuclear OCT-4 differentiate and give rise to OSCs with cytoplasmic OCT-4 [85, 89].

Zou et al. [90] established a neonatal mouse FGSC line, cultured for more than 15 months, FGSCs from adult mouse were maintained for more than 6 months. These stem cells were GFP labeled and on transplantation in ovaries of infertile mice underwent oogenesis and GFP positive offspring. Later, Tilly’s group successfully expanded DDX-4 positive stem cells in vitro for months from both mouse and human ovarian cortex [91]. Injecting GFP positive human OSCs into human ovarian cortical biopsies leads to assembly of primordial follicles when transplanted in immuno-deficient female mice –resulted in complete assembly of primordial follicles. These protocols were later published [92].

‘Eggbert’ was born by culturing primordial follicles isolated from newborn mice in vitro using a two-step culture method. The pup however, developed health problems including obesity and neurological abnormalities [93]. Later the protocol was further modified and another 59 pups were born to provide company to ‘Eggbert’ [94]. But the process remained inefficient. Of 1160 2-cells embryos transferred, only 66 (5.7%) developed to term and 7 pups (10.6%) died at birth. The remaining 59 pups (27 females, 32 males) survived to adulthood. By comparison, of 437 transferred two-cell stage embryos derived from in vivo-grown oocytes, 76 (17.4%) developed to term and 4 (5.3%) died at birth. The remaining 72 pups (35 females, 37 males) survived to adulthood. Shen et al. [95] transplanted fetal 12.5 dpc ovaries under the kidney capsule that initiated oocyte growth from the pre-meiotic germ cells. Subsequently, the primary and early secondary follicles generated in the ovarian grafts were isolated and cultured for 16 days in vitro. The mature oocytes ovulated from these follicles were able to fertilize in vitro to produce live offspring. Offspring born by IVF were normal and able to successfully mate with both females and males. The patterns of the methylated sites of the in vitro mature oocytes were similar to those of normal mice.

Another alternative is to develop artificial ovary that implies 3D culture of preantral, immature follicles in a scaffold which could be a source of oocytes upon transplantation [96]. This approach could negate introduction of malignant cells. Woodruff’s group developed a two-step follicle culture strategy that recapitulated the dynamic human follicle growth environment in vitro [97]. Follicles developed from the preantral to antral stage and produced meiotically competent metaphase II oocytes after in vitro maturation. Later Kniazeva et al. [98] reported live births on transplanting primordial to primary follicles from young ovaries encapsulated in fibrin beads. The group recently developed a bioprosthetic ovary using microporus scaffolds to restore ovarian function in bilaterally ovariectomized mice. Follicles matured, ovulated, secreted hormones, formed corpus luteum that secreted progesterone to support pregnancy and resulted in birth of fertile pups [99].