Human Reproduction Update, Vol.10, No.3 pp.193-195, 2004
© European Society of Human Reproduction and Embryology 2004; all rights reserved
Germline stem cells in the postnatal ovary: is the ovary more like a testis?
The Jones Institute for Reproductive Medicine, Eastern Virginia Medical School, 601 Colley Avenue, Norfolk, VA 23507, USA e-mail: gosdenrg{at}evms.edu
 |
Hot Topics—announcement of new series |
|---|
 TOP Hot Topics—announcement of... Germline stem cells in...References |
|---|
B.C.J.M.Fauser
Editor-in-Chief
1I am pleased to announce a new series starting in this issue. ‘Hot topics’ will include occasional editorials and commentaries as appropriate with a forward- looking perspective highlighting recent important research papers or potential revolutionary developments in the field. The article below from one of our Associate Editors Roger Gosden seems an ideal start to the series in that it presents a balanced and constructive assessment, from someone active in the field for several decades, of a paper which has the potential to radically change long-held concepts.
|
|
Germline stem cells in the postnatal ovary: is the ovary more like a testis? |
|---|
|
TOPHot Topics—announcement of...Germline stem cells in...References |
|---|
Roger G. Gosden
‘Men who are orthodox when they are young are in danger of being middle-aged all their lives’ (Walter Lippmann)
2According to worldwide media reports, a group of young researchers at Harvard Medical School recently punctured a supposedly watertight theory of ovarian physiology (Johnson et al., 2004
). In a paper in Nature, they conclude that their data ‘establish the existence of proliferative germ cells that sustain oocyte and follicle production in the postnatal mammalian ovary’. This is indeed a bold break with scientific orthodoxy, and something that Lippmann might have admired. The doctrine was established more than half a century ago (Zuckerman, 1951), and subsequently germline stem cells (GSC) have never been recorded in juvenile or adult ovaries, apart from isolated exceptions in certain species (e.g. Ioannou, 1967). The claim that these cells are generating new oocytes to replace those perishing in atretic follicles is most astounding. If true, it will require a fundamental revision of the underlying mechanisms of the menopause and oocyte ageing. It might even herald breathtaking advances in reproductive technology. In the words of the authors: ‘this work has clinical significance related to the therapeutic expansion of the follicle reserve as a means to postpone normal or premature ovarian failure’. The new paradigm is as revolutionary for ovarian biology as the breakthrough in nuclear transfer was for reproductive cloning a few years ago, and it behoves us to examine the facts. Two branches of evidence were presented to support their hypothesis: (i) inferential, by estimating follicle numbers and death (atresia) rates, and (ii) experimental, by revealing the existence and activity of GSC in the postnatal mouse ovary. I shall now briefly examine them in turn.
|
Johnson et al. counted the numbers of primordial and growing follicles in mouse ovaries using conventional microscopic methods, obtaining results that were broadly consistent with previous studies. The numbers of follicles in single ovaries were somewhat higher than expected (closer to those reported for paired organs), the increase between postnatal days 4 and 8 was atypical (though not statistically significant) and the rate of decline in the immature ovary was somewhat less than anticipated (Jones and Krohn, 1961
). However, these differences were minor compared with the enormous numbers of preantral and antral follicles that they found degenerating or ‘atretic’ in young adult ovaries (>1200), in striking contrast to general experience. According to Byskov (1974), the transit time for atresia is 
4 days for large follicles, and presumably less for smaller ones. In view of the rapid turnover, the authors suspected that new follicles were being formed continuously to avoid exhausting the primordial follicle store and truncating the reproductive lifespan, which is up to a year in this species. These rates of atresia were approximately an order of magnitude higher than reported in other studies, including the same strain (C57BL/6) (Gosden et al., 1983). Scoring atretic follicles is notoriously subjective and, while the data are valuable for comparative studies, reliance should not be placed on them as absolute values for calculating the rate of follicular wastage. Histological protocols can create artefacts that exaggerate the true incidence of follicular atresia, and the use of a harsh fixative in this study (Carnoy’s fluid) might account for the unusual results.
Aside from these data, they noted a discrepancy between the declining numbers of follicles with age, on the one hand, and the rate of follicle disappearance, on the other. Their data reveal that, between 16 and 40 days of age, the numbers of follicles had declined by only 294, a rather lower gradient than in other strains of mice at juvenile ages (Faddy et al., 1983). The authors then applied estimates of the rate of atresia derived from a mathematical model of follicle dynamics: at 21 days of age; our model predicted in mice an egress rate from the primordial follicle store in CBA/Ca mice of 89 follicles per day (48 by death and 41 by growth to the next stage) (Faddy et al., 1987). Johnson et al. reasoned that 2136 follicles (89x24) were lost during the 24 day span starting at 16 days of age. They then drew the conclusion that a difference of 1842 follicles between the number estimated to be lost and the observed decline of only 294 in this time span must be accounted for by replacement of 77 new follicles per day. These arguments are based on dubious assumptions and can be challenged. For one thing, the rate of egress is not constant with age and, for another, follicle dynamics are strain-specific (Faddy et al., 1983, 1987). Prior to 21 days of age, far more than 89 follicles leave the resting pool each day, and fewer afterwards (as a function of the declining reserve). What is more alarming, they applied the rates of disappearance from the CBA/Ca mouse, which is an exceptionally fast-depleting strain, whereas follicles in C57BL/6 ovaries are lost more slowly. The net effect of losing fewer follicles than expected and using a high estimate of atresia will be to overestimate the ‘discrepancy’ in follicle numbers, if any exists at all.
Since it is difficult to infer rates of follicle disappearance using indirect methods, the authors also carried out experimental studies. They treated mice with dimethylbenz[a]anthracene (DMBA) and busulphan, which have selectively sterilizing effects on the ovary. Primordial follicles disappeared rapidly and completely whereas most growing follicles survived for weeks, and some of them even ovulated. These findings are entirely consistent with studies of DMBA reported many years ago by Krarup (1970) and others who concluded that the compound targeted oocytes for selective destruction at the highly sensitive primordial follicle stage. There appears to be no compelling reason for questioning this original interpretation. Nor is there yet a strong basis for believing in the superiority of the explanation of Johnson et al., namely, that primordial follicle disappearance was secondary to the demise of ovarian stem cells after busulphan exposure (by analogy with the testis).
Seeking more direct evidence of postnatal GSC, the authors located naked cells expressing the germ cell marker, Vasa, in the outer layers of the ovary, including the surface epithelium. They reported that the cells were mitotically active judging from morphology and btomo-deoxyuridine BrdU incorporation, which made them candidates for the elusive stem cells. However, it is hard to be sure of cellular character from the limited examples provided. A different interpretation was always attached to these cells in the past, which have often been noted at or near the ovarian surface in immature mice and human fetuses. Under transmission and scanning electron microscopes, they appeared to be supernumerary germ cells, sometimes revealing amoeboid locomotion, and often carrying out a suicidal migration out of the ovary (Motta and Makabe, 1986). The figures published by Johnson et al. are, in fact, wholly consistent with this interpretation. It is likely that some of the cells at prenatal ages are germ cells at various stages of mitosis or meiosis, and, after birth, they could be oocytes from primordial follicles or even naked diplotene cells that had failed to be captured by a layer of granulosa cells. It is not yet possible to completely rule out stem cells, although few, if any, germ cells of any kind are normally found in this location at adult ages.
The expression of genes that are absolutely specific to pre-meiotic or early prophase stages of oogenesis could provide a strong indication that GSC’s can persist in juvenile or adult ovaries. Accordingly, Johnson et al. used RT–PCR and immunocytochemistry to study the expression of Spo11, Dmc1, Scp3 and SCP3 protein, representing genes that are involved at the initiation of meiosis and in the structure of the synaptonemal complex. Surprisingly high levels of transcripts were found, consistent with their hypothesis. Furthermore, we were unable to identify the corresponding proteins in our proteome databases for mature oocytes, although Scp1, Scp2, topoisomerases 1 and 2
and related proteins were present (D.Miller et al., unpublished data). Further studies are desirable to verify the specificity of early meiosis markers and rule out the possibility that the PCR products were long-lived, untranslated messages carried over to the diplotene stage.
The creation of chimaeric ovaries was the most original experiment reported by Johnson et al., and seems to provide prima facie evidence that oocytes are continuously formed in adulthood. The authors grafted small wedges of ovaries from animals transgenic for the ubiquitous fluorescent marker, Green fluorescent protein GFP, to the side of compatible wild-type ovaries. Several weeks later they found some follicles containing labelled oocytes surrounded by unlabelled granulosa cells. If the integrity of follicles is inviolate, this result must imply that newly formed oocytes from the donor had generated new follicles with pre-granulosa cell partners from the host. However, even this intriguing experiment has an alternative explanation, because the mouse ovary is remarkably plastic and small follicles may not be as inflexible as generally assumed. The organ can be disaggregated with proteolytic enzymes into isolated cells or primordial follicles which, after reaggregating the scrambled cells and transplanting them to a recipient animal, can form products that regenerate into follicles that can ovulate (Gosden, 1990). Even more remarkably, after separating somatic and germ cells from mouse and rat ovaries, the suspensions of cells could be experimentally recombined in various combinations to recreate follicles after transplantation to severe combined immunodeficient (SCID) mice. Some of these follicles contained mouse oocytes surrounded by rat granulosa cells, as well as vice versa (Eppig and Wigglesworth, 2000). Thus, it is plausible that, during the trauma of ovarian grafting and subsequent tissue repair, some germ and granulosa cells reshuffle between primordial follicles from the donor and host tissues. Perhaps this even occurs naturally in the intact ovary, because adhesion between cells in small follicles is weak. Thus, the chimaeric follicles described by Johnson et al., although remarkable and interesting, do not necessarily lend support to the existence of a production line for new oocytes.
Despite many reservations about the inferences drawn by the authors from their experiments, Johnson et al. have done biology a service by forcing a review of the foundations of a scientific dogma. It is time for others to engage the pursuit and apply the powerful technologies now available for testing the old theory. Transgenic biology can serve as one of the major tools in this quest, if only in rodent models. By examining expression of GFP reporter constructs for genes specifically involved in early meiosis, the temporal and spatial distribution of GSC can be revealed more precisely. Conditional knockouts can be designed to investigate the impact on surviving follicle numbers of deleting pre-diplotene stages at various ages. And stem cells may be isolated from ovaries using flow cytometry for detailed characterization, a strategy that the Harvard group has already initiated (J.Tilly, personal communication).
Most of all, readers of this journal will be interested in the implications for human biology. Throughout evolution, the mechanisms of reproductive physiology have diverged more than any other life-sustaining system in the body, and extrapolations from one species to another should be made with the utmost caution. However, numerous practical and ethical limitations overlay the study of the human (and even the monkey) ovary, and progress will always be faster with small animal models. Nevertheless, valuable information can be gained from microscopy, as well as from ovarian cDNA libraries and other resources.
According to a recent study, approximately a quarter of the resting follicles sampled in the ovaries of women aged 20–48 years were TdT (terminal deoxynucleotidyl transferase)-mediated dUDP nick-end labelling (TUNEL)-positive (Depalo et al., 2003). If these data are equated with the frequency of apoptosis, they indirectly point to the existence of GSC in our own species, because the reproductive lifespan would otherwise end prematurely if limited by the stock of oocytes endowed at birth. Whatever the explanation for these puzzling results, they should probably not be taken at face value since electron microscopy has revealed that the rate of degeneration among primordial and primary follicles is low, as had always been assumed (de Bruin et al., 2002). More significantly, although GSC (‘oogonia’) are abundant in second trimester fetuses, they are rare after 30 weeks of gestation and virtually extinct at birth (Block, 1953; Baker, 1963; Lintern-Moore et al., 1974; de Bruin et al., 2001). What is even more persuasive, cytological studies of >500 ovaries from infants and young women have never reported finding these cells (Block, 1952; Lintern-Moore et al., 1974; Peters et al., 1976; Gougeon and Chainy, 1987). Of course, GSC would not have to be conspicuously common to be theoretically capable of generating a trickle of new follicles after birth, but it seems highly unlikely that so many expert microscopists have overlooked pre-diplotene stages in the ovary. Besides, pachytene oocytes are especially obvious and have a transit time of several days in the mouse ovary and probably much longer in humans.
In conclusion, on the basis of these new data it is premature to rewrite the life history of the oocyte or to regard the ovary as more like a testis. Whether or not the old theory is correct, we are confronted with the indubitable facts that oocyte quality and follicle numbers decline with age, signalling the end of the reproductive lifespan. We have assumed that ovarian ageing is due to sealing of the fate of stem cells before birth by the decision to die or differentiate into oocytes, as opposed to the tapering of stem cell activity in adult life. Either way, germline stem cells are strange cells indeed.
|
References |
|---|
|
TOPHot Topics—announcement of...Germline stem cells in...References |
|---|
Baker TG (1963) A quantitative and cytological study of germ cells in human ovaries. Proc R Soc Lond B 158,417–433.[Medline]
Block E (1952) Quantitative morphological investigations of the follicular system in women. Variations at different ages. Acta Anat 14,108–123.[Web of Science][Medline]
Block E (1953) A quantitative morphological investigation of the follicular system in newborn female infants. Acta Anat 17,201–206.[Web of Science][Medline]
Byskov AG (1974) Cell kinetic studies of follicular atresia in the mouse ovary. J Reprod Fertil 37,277–285.
de Bruin JP, Nikkels PG, Bruinse HW et al (2001) Morphometry of human ovaries in normal and growth-restricted fetuses. Early Hum Dev 60,179–192.[CrossRef][Web of Science][Medline]
de Bruin JP, Dorland M, Spek ER et al (2002) Ultrastructure of the resting ovarian follicle pool in healthy young women. Biol Reprod 66,1151–1160.
Depalo R, Nappi L, Loverro G et al (2003) Evidence of apoptosis in human primordial and primary follicles. Hum Reprod 18,2678–2682.
Eppig JJ and Wigglesworth K (2000) Development of mouse and rat oocytes in chimeric reaggregated ovaries after interspecific exchange of somatic and germ cell components. Biol Reprod 63,1014–1023.
Faddy MJ, Gosden RG and Edwards RG (1983) Ovarian follicle dynamics in mice: a comparative study of three inbred strains and an F1 hybrid. J Endocr 96,23–33.
Faddy MJ, Telfer E and Gosden RG (1987) The kinetics of preantral follicle development in ovaries of CBA/Ca mice during the first 14 weeks of life. Cell Tiss Kinet 20,551–560.[Web of Science][Medline]
Gosden RG (1990) Restitution of fertility in sterilized mice by transferring primordial ovarian follicles. Hum Reprod 5,499–504.
Gosden RG, Laing SC, Felicio LS et al (1983) Imminent oocyte exhaustion and reduced follicular recruitment mark the transition to acyclicity in aging mice. Biol Reprod 28,255–260.[Abstract]
Gougeon A and Chainy GBN (1987) Morphometric studies of small follicles in ovaries of women at different ages. J Reprod Fertil 81,433–442.
Ioannou JM (1967) Oogenesis in adult prosimians. J Embryol Exp Morph 17,139–145.[Web of Science][Medline]
Johnson J, Canning J, Kaneko T et al (2004) Germline stem cells and follicular renewal in the postnatal mammalian ovary. Nature 428,145–150.[CrossRef][Medline]
Jones EC and Krohn PL (1961) The relationships between age, numbers of oocytes and fertility in virgin and multiparous mice. J Endocrinol 21,469–495.[Web of Science][Medline]
Krarup T (1970) Oocyte survival in the mouse ovary after treatment with 9,10-dimethyl-1,2-benzanthracene. J Endocrinol 46,483–495.[Web of Science][Medline]
Lintern-Moore S, Peters H, Moore GPM and Faber M (1974) Follicular development in the infant human ovary. J Reprod Fertil 39,53–64.
Motta PM and Makabe S (1986) Germ cells in the ovarian surface during fetal development in humans. A three-dimensional microanatomical study by scanning and transmission electron microscopy. J Submicrosc Cytol 18,271–290.[Web of Science][Medline]
Peters H, Himelstein-Braw R and Faber M (1976) The normal development of the ovary in childhood. Acta Endocrinol 82,617–630.
Zuckerman S (1951) The number of oocytes in the mouse ovary. Recent Prog Horm Res 6,63–108.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. L Walker, D. C Anderson, J. G Herndon, and L. C Walker Ovarian aging in squirrel monkeys (Saimiri sciureus) Reproduction, November 1, 2009; 138(5): 793 - 799. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Tilly, Y. Niikura, and B. R. Rueda The Current Status of Evidence for and Against Postnatal Oogenesis in Mammals: A Case of Ovarian Optimism Versus Pessimism? Biol Reprod, January 1, 2009; 80(1): 2 - 12. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.P. Jones, L.C. Walker, D. Anderson, A. Lacreuse, S.L. Robson, and K. Hawkes Depletion of Ovarian Follicles with Age in Chimpanzees: Similarities to Humans Biol Reprod, August 1, 2007; 77(2): 247 - 251. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Gargett Review Article: Stem Cells in Human Reproduction Reproductive Sciences, July 1, 2007; 14(5): 405 - 424. [Abstract] [PDF]
|
|
|
|
|
|
J B Kerr, R Duckett, M Myers, K L Britt, T Mladenovska, and J K Findlay Quantification of healthy follicles in the neonatal and adult mouse ovary: evidence for maintenance of primordial follicle supply Reproduction, July 1, 2006; 132(1): 95 - 109. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Ottolenghi, S. Omari, J. E. Garcia-Ortiz, M. Uda, L. Crisponi, A. Forabosco, G. Pilia, and D. Schlessinger Foxl2 is required for commitment to ovary differentiation Hum. Mol. Genet., July 15, 2005; 14(14): 2053 - 2062. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||