n e w s a n d v i e w s

1190 volume 14 | number 11 | november 2008 nature medicine

in such a short time by these embryoid body–derived germ cells. Indeed, in vivo aneuploidy is often associated with male infertility and recurrent pregnancy loss in females, and this situation must be avoided if this technique is to be developed for reliable fertility therapies. Likewise, epigenetic changes may not be cor-rectly executed in vitro, as suggested by inef-ficient embryo development and pregnancy failure.

Finally, given the finding that human ES cells express some of the germ cell markers13, there is an obvious question to be posed—namely, can these strategies be applied to human ES cells? The bet is that we will not have a long wait to find out. Indeed, the human equivalent of the Guan et al.7 study has now been con-firmed and reported14.

1. Seydoux, G. & Braun, R.E. Cell 127, 891–904 (2006).

2. Hofmann, M.C., Hess, R.A., Goldberg, E. & Millan, J.L. Proc. Natl. Acad. Sci. USA 91, 5533–5537 (1994).

3. Feng, L.X. et al. Science 297, 392–395 (2002).4. Kierszenbaum, A.L., Abdullah, M., Ueda, H. & Tres,

L.L. Adv. Exp. Med. Biol. 219, 535–560 (1987).5. Toyooka, Y., Tsunekawa, N., Akasu, R. & Noce, T. Proc.

Natl. Acad. Sci. USA 100, 11457–11462 (2003).6. Geijsen, N. et al. Nature 427, 148–154 (2004).7. Guan, K. et al. Nature 440, 1199–1203 (2006).8. Kanatsu-Shinohara, M. et al. Cell 119, 1001–1012

(2004).9. Hubner, K. et al. Science 300, 1251–1256 (2003).10. Novak, I. et al. Stem Cells 24, 1931–1936 (2006).11. Nayernia, K. et al. Dev. Cell 11, 125–132 (2006).12. Kanatsu-Shinohara, M. et al. Biol. Reprod. 78, 681–

687 (2008).13. Clark, A.T. et al. Hum. Mol. Genet. 13, 727–739

(2004).14. Conrad, S. et al. Nature published online, doi:10.1038/

nature07404 (8 October 2008).

as ES cells because they develop the ability to differentiate into the three major embryonic lineages and to produce teratomas. The devel-opmental potential of these cells was confirmed by injection into blastocysts and detection of a high percentage of chimerism in the offspring. The team concluded that either SSCs are mul-tipotent or, at least, the culture system allows them to revert to an uncommitted, undifferen-tiated state7. It should be noted that an earlier study in SSC culture from neonatal testis8 was able to identify a population of cells similar to that derived from the adult testis. In this latter experimental paradigm, it could be conclu-sively determined that SSCs can dedifferentiate to a pluripotent state12.

Taken together, these reports demonstrate that gametogenesis can be attained in vitro and, conversely, that committed adult germ cells can revert to a pluripotent, ES-like state—both remarkable accomplishments. But this is only the beginning, as the efficiency of the pro-cess is difficult to evaluate but in general seems quite low, and it must be improved for further development for therapeutic application. In addition, a surprising finding is the short time required to recapitulate germ cell differentia-tion in these in vitro gametogenesis models. At one extreme, only 72 h are sufficient to reach the haploid state in vitro11. Thus, more studies are warranted to clarify whether all the steps in the meiotic division (for example, the correct complement of chromosomes and their seg-regation in haploid cells, as well as the occur-rence of recombination) are correctly executed

differentiation (the mouse homolog of DDXR (also known as VASA)) and slightly different culture conditions were used. Although no functional data of the germ cells produced in vitro were provided, Toyooka et al.5 showed that these embryoid body–derived germ cells can be transplanted under the adult testis cap-sule and form tubules containing germ cells at advanced stages of differentiation, including elongated spermatids and spermatozoa. These initial reports did not meet the gold standard of demonstrating that offspring are produced with these in vitro–derived germ cells; however, subsequent publications have shown that this is indeed the case. Notably, Wolfgang Engel and his collaborators were able to produce off-spring by oocyte injection of in vitro–derived sperm-like cells, although their viability was compromised11.

The ES cell–germ cell transition is plastic, as the reverse transition occurs, as well. For example, adult spermatogonial stem cells iso-lated from the testis, under appropriate culture conditions, acquire a phenotype similar to that of undifferentiated ES cells7 (Fig. 1b). Guan et al.7 used transgenic mice expressing EGFP under the control of a germ cell–specific pro-moter (Stra8) to isolate cells similar to SSCs. They then exposed these cells to different cul-ture conditions and followed the appearance of markers of pluripotency7. Several cell lines were derived from these heterogeneous popu-lations of cells, some of them morphologically similar to and expressing many of the mark-ers of ES cells. Functionally, these cells behave

Making eggs: is it now or later?Teresa K Woodruff

Although it has been thought that female mammals develop all the eggs they will ever have by the time they are born, new research suggesting otherwise has now sparked a debate.

Teresa K. Woodruff is the Watkins Professor of

Obstetrics and Gynecology in the Department

of Obstetrics and Gynecology, Feinberg School

of Medicine, Northwestern University, 303 East

Superior, Lurie 10-117, Chicago, Illinois 60611-

3015, USA.

e-mail: tkw@northwestern.edu

Although long held as true, the notion that females are born with a finite number of non-replenishable oocytes is somewhat remark-able. How do eggs, made during embryonic life, ‘last’ so long? This traditional narrative of oogenesis was recently challenged; indeed,

the paper by Johnson et al.1 highlighted here suggests the radical notion that the adult ovary must and does have the capacity to create new oocytes. After it was published, the reproductive science community collec-tively asked, “How did we miss this?,” and thus new studies were launched to validate these findings. To date, however, the overall evidence still favors the established dogma. Nonetheless, the field has been energized by this vigorous debate.

The purpose of germ cells, from pollen grains to mouse oocytes, is to produce future progeny and recreate the ‘self ’, not necessar-

ily for the individual but for the species as a whole. Unraveling the mechanisms that drive germ cell development and persistence in mammals is fascinating and challenging. The prevailing dogma of mammalian repro-duction states that females lose the capacity to produce germ cells during fetal develop-ment and are born with a finite number of follicle-enclosed oocytes, only a small num-ber of which will be ovulated after puberty (Fig. 1). Several recent studies from the laboratory of Jonathan Tilly at Harvard led to the hypothesis that adult mice are capa-ble of regenerating oocytes, suggesting the

©20

08 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

n e w s a n d v i e w s

nature medicine volume 14 | number 11 | november 2008 1191

tive interventions offer new options to young patients with cancer19. These include con-ventional hormone stimulation with embryo or egg cryobanking and ovarian tissue cryo-preservation, with promising tissue transplan-tation and follicle maturation technologies also on the horizon (refs. 20–23, and see the related News & Views on this topic on pages 1182–1184). Such strategies are welcome and exciting reproductive interventions that can be offered to those women who will have irretrievable damage to the resident ovarian follicles as a result of the treatment for their cancer.

It is hard to look back on the past four years and not appreciate the number of key discoveries that emerged as a consequence of this discourse. The field has moved forward by affirming what it knew—that making an egg is a biological puzzle that has yet to be entirely solved. New concepts about the essential characteristics of the oocyte, which include its ability to persist over long spans of time, provide the next generation of ques-tions that can be solved by multiple groups bringing varied perspectives and experimen-tal methods to the problem. Reproductive science and medicine have been strength-ened by virtue of the discussion and are now poised to move to the next frontier.

1. Johnson, J., Canning, J., Kaneko, T., Pru, J.K. & Tilly, J.L. Nature 428, 145–150 (2004).

2. Zuckerman, S. Recent Prog. Horm. Res. 6, 63–108 (1951).

3. Johnson, J. et al. Cell 122, 303–315 (2005).4. Lee, H.J. et al. J. Clin. Oncol. 25, 3198–3204

(2007).5. Baker, T.G. Proc. R. Soc. Lond. B 158, 417–433

(1963).6. Gougeon, A. & Chainy, G.B. J. Reprod. Fertil. 81, 433–

442 (1987).7. Byskov, A.G., Faddy, M.J., Lemmen, J.G. & Andersen,

C.Y. Differentiation 73, 438–446 (2005).8. Telfer, E.E. et al. Cell 122, 821–822 (2005).9. Johnson, J. et al. Cell Cycle 4, 1471–1477 (2005).10. Eggan, K., Jurga, S., Gosden, R., Min, I.M. & Wagers,

A.J. Nature 441, 1109–1114 (2006).11. Begum, S., Papaioannou, V.E. & Gosden, R.G. Hum.

Reprod. 23, 2326–2330 (2008).12. Bristol-Gould, S.K. et al. Dev. Biol. 298, 132–148

(2006).13. Bristol-Gould, S.K. et al. Dev. Biol. 298, 149–154

(2006).14. Faddy, M. & Gosden, R. Cell Cycle 6, 1951–1952

(2007).15. Daggag, H. et al. Biol. Reprod. 79, 468–474 (2008).16. Anderson, E.L. et al. Proc. Natl. Acad. Sci. USA 105,

14976–14980 (2008).17. Tilgner, K. et al. Stem Cells published online,

doi:10.1634/stemcells.2008-0289 (18 September 2008).

18. Hubner, K. et al. Science 300, 1251–1256 (2003).19. Woodruff, T.K. Cancer Treat. Res. 138, 3–11 (2007).20. Agarwal, S.K. & Chang, R.J. Cancer Treat. Res. 138,

15–27 (2007).21. Gosden, R.G. Reprod. Biomed. Online 4 Suppl 1,

64–67 (2002).22. Silber, S.J. et al. Hum. Reprod. 23, 1531–1537

(2008).23. Xu, M., Kreeger, P.K., Shea, L.D. & Woodruff, T.K.

Tissue Eng. 12, 2739–2746 (2006).

mental data collected over the past four years supports the nonrenewable germ cell pool theory. These findings show that circulating adult stem cells (provided through parabio-sis) do not contribute to the ovulated fol-licle pool10, empty germ cell niches are not repopulated by newly formed oocytes11 and markers of stem cells are absent in the post-natal ovary but do appear in the ovarian vasculature where stem cells are known to exist12. Additionally, mathematical models evaluating the kinetics of follicle transitions indicate that the canonical nonreplenish-able follicle pool model describes empirical observations of follicle numbers better than models incorporating neo-oogenesis13,14. Taken together, these papers teach against the idea that germline stem cells initiate neo-oogenesis in the adult animal and support the concept that female germ cells formed before birth are all that will be available from puberty to menopause. Thus, nature has given the female ovary a limited number of nonreplenishable eggs.

However, what if an occasional wander-ing stem cell, or a local one, was to trans-form into a germline stem cell in the adult? If such an event could be documented, that cell would face a high set of hurdles on its way to becoming a functional oocyte, includ-ing progressing from a proliferative cell to a meiotically arrested, follicle-enclosed cell. The hurdles heighten if this oocyte must eventually mature to an egg that can be fertilized and sustain embryogenesis. We know there is something unique about the embryonic gonad that supports oogonia pro-liferation. And we know that there is some-thing equally singular about the adult ovary, which maintains the majority of oocytes in a quiescent pool. Some of the key factors involved in embryonic neo-oogenesis have been identified15,16, and the field is look-ing forward to the next revelation of genes and mechanisms. There is equal interest in re creating oocytes from embryonic stem (ES) cells. Indeed, oocyte-like structures have been formed from ES cells in vitro, again creating a rich vein of investigation17,18. These are the areas where reproductive biology is breaking new ground and creating new paradigms.

The publication of the initial adult oogen-esis paper created a splash in the journal and in the media, suggesting that young women facing life-preserving but sometimes fertility-threatening cancer treatments might have an option for restoring their fertility. However, the ability to repopulate the ovary with a bone marrow transplant or stem cell transfusion now seems unachievable. Despite this reality, a number of existing and emerging reproduc-

existence of adult germline stem cells1,3,4. These papers created a schism within the reproductive science community that is now coming to a resolution.

In humans, fewer than 300,000 of the original pool of about one million healthy oocytes present at birth survive to puberty5,6. This number continues to decline through-out adulthood to the point of extinction around age 50, thus driving menopause. In a series of high-profile papers1,3,4, it was pos-tulated that, at least in mice, the postnatal follicle cohort was dying so quickly that nor-mal adult fertility could not be maintained if it depended solely on the follicle pool present at birth. Indeed, on the basis of this obser-vation, a profoundly provocative statement was made that neo-oogenesis occurs in the adult mammal1. Moreover, it was posited that the source of the germline stem cells is tissue-based cells in the ovarian epithelium or, in a change from the first paper, from the bone marrow1,3. More recently, the group has contended that the ovary contains germ cell niches (resident follicles) that can accept these cells and help them transform into oocytes4. On the basis of these studies, the authors suggest that the adult oocyte popu-lation is renewable and that this process is necessary to supplant an insufficient number of follicles formed during embryonic devel-opment.

The reproductive research community began to think critically about the experi-ments and their interpretation and to design studies to examine the provocative new ideas7–14. Alternative interpretations of the data presented in the original papers and the absence of corroborating evidence of appreciable numbers of stem cells within the adult ovary were the first two hints that the original theories from last century may still stand. Indeed, the preponderance of experi-

Figure 1 Follicles are the functional unit of the ovary. Most of these follicles are present as nongrowing primordial follicles (arrowheads). Once they are activated, the centrally located oocyte and the surrounding granulosa cells begin to grow (arrows) in a coordinated way leading toward ovulation. Scale bar, 50 mm. Figure courtesy of C. Tingen, Northwestern University.

©20

08 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine