Reviews of Reproduction (2000) 5, 93–98

Molecular mechanisms involved in the differentiation of spermatogenic stem cells Keith A. Sutton Discipline of Biology, School of Biological and Chemical Sciences, University of Newcastle, Callaghan, NSW 2308, Australia

In male mammals, spermatogenesis proceeds for the reproductive lifetime of the animal. The continuation of this process depends upon a pool of spermatogenic stem cells within the testes that undergo asymmetric division to both maintain the stem cell population and give rise to progenitors that will proceed through spermatogenesis to generate mature spermatozoa. Thus, the development of functional spermatozoa may be divided into two distinct stages. The second, the process of spermatogenesis, is dependent upon the first, the successful formation of spermatogenic stem cells. Although spermatogenesis is characterized by marked cellular differentiation, the initial stages of germ line differentiation involve an avoidance of the differentiation signals acting during embryo development. The germ line is set aside early in embryo development and, while the primordial germ cells remain refractory to the differentiation signals affecting the soma, they undergo a number of phenotypic shifts before and after colonizing the genital ridge. Upon colonization of the genital ridge, the somatic tissue of the male genital ridge directs the final differentiation events that result in the formation of spermatogenic stem cells. It is this cell population that provides the basis for the maintenance of spermatogenesis in the adult. The germ line is the one component of the organism that may aspire to immortality and its development and transmission represent the most important tasks undertaken by a multicellular organism. The first identifiable germ line cells are the primordial germ cells (PGCs) which arise early in embryogenesis and maintain the same potential for totipotency as the embryonic stem (ES) cells of the early embryo. PGCs are first identifiable at the epiblast stage of embryo development and can be identified as a discrete cell population positive for alkaline phosphatase, and expressing oct-4, c-kit and MesP-1 (throughout this article italics are used to denote genes (presence of mRNA) and plain text protein). These cells migrate into the extra-embryonic mesoderm until they are reintroduced into the fetus. At this point, the PGCs begin to divide and develop an invasive phenotype that allows them to migrate through the developing embryo and into the genital ridge, the site of the developing gonad. It is only at this stage, when PGCs within a male embryo enter mitotic arrest, that they begin to lose their potential for totipotency. So, while the developing PGCs avoid the differentiation signals that direct the formation of the soma, they must respond to a series of signals that guides their development and migration until they reach the genital ridge (Fig. 1). After colonization of the genital ridge, the invasive phenotype is lost and the further differentiation of the PGCs is determined by signals from the surrounding soma. In the male gonad, an SRY-dependent signal from the soma directs the formation of testes. SRY is the master regulator of mammalian sexual differentiation: even in the absence of a Y chromosome, its presence directs the differentiation of the early bipotential gonad into testes (Koopman et al., 1991). This environment then directs the final step of PGC

differentiation in the male fetus: the formation of spermatogenic stem cells (SSC). Thus, an understanding of the signalling events required for the proper formation of the spermatogenic stem cell involves investigation and understanding of a number of distinct microenvironments within the developing fetus.

Immortality, totipotency and differentiation One problem that arises in discussing the nature of the germ line is nomenclature and the use of the terms ‘immortal’ and ‘totipotent’. Although the germ line is frequently referred to as ‘immortal’, this is only valid at the level of the entire population and, more specifically, only the genome residing within the cells that compose the germ line may be viewed as immortal. A second and more apposite issue arises in relation to the totipotency of the germ line, that is, its ability to generate all the tissues of an organism, including itself. The early germ line may be referred to as totipotent in two distinct senses. The first of these senses involves the normal fertilization process involving the fusion of male and female gametes. Thus, any cell within the male or female PGC population has a potential for totipotency, if a gamete derived from it is involved in the fertilization event. This description and concept of totipotency is distinct from that used when referring to the totipotency of cells in the early embryo, the embryonic stem (ES) cells, that form the basis of gene knockout technology. The ability to obtain ES cell lines from PGCs indicates that, although restricted to one developmental pathway in vivo, manipulation in vitro reveals a potential for totipotency in this population that does not involve the formation of haploid gametes. Such cell lines derived from PGCs are referred to as embryonic germ (EG) cell

© 2000 Journals of Reproduction and Fertility 1359-6004/2000

94

K. A. Sutton

Day after mating

7

8

9.5

10.5

PGC location

Proximal epiblast

Allantois

Hindgut

Genital ridge

Migratory

Quiescent

Phenotype

Proliferative/invasive

12.5

15

Mitotic arrest

Gene expression

oct-4 c-kit

pem MesPI GCNA1

tiar

Fig. 1. Gene expression during primordial germ cell (PGC) differentiation. The location and phenotype of PGCs during mouse fetal development are shown, together with the known patterns of expression of germ line genes. Solid lines denote the developmental period over which genes are expressed (dashed lines indicate where the onset of expression has yet to be determined). Arrowheads denote genes the expression of which precedes PGC origin on day 7 after mating (oct-4) or extends beyond day 15 after mating. Differentiation of PGCs to spermatogenic stem cells (SSCs) occurs after the initiation of sry expression in Sertoli cells at day 10.5 after mating, although it may not be completed until after birth. Expression data are from the following: tiar: Beck et al. (1998); c-kit: Manova and Bachvarova (1991); pem: Pitman et al. (1998); Mesp1: Saga et al. (1996); germ cell nuclear antigen (GCNA1): Wang et al. (1997); oct-4: Yeom et al. (1996).

lines to distinguish them from ES cell lines that are derived from disaggregated blastocysts. The normal totipotent cycle of the germ line and the points at which this compartment may be accessed for manipulation in vitro is outlined (Fig. 2). Throughout this review, the word totipotency will only be used to refer to cells (or cell populations) that may differentiate into any cell population without passing through a gamete stage. The totipotency of the germ line in the normal lifecycle of the organism is implicit in any reference to the male or female germ line. Embryonic germ cells have been derived from PGCs isolated from mouse embryos from up to day 10.5 of gestation; a time at which all other cell lineages within the embryo have undergone a series of lineage commitment and differentiation events. It has not been possible to isolate totipotent cell lines after this time point, although pluripotent mouse EG lines have been established from male embryos at day 12.5 after mating, indicating that PGC differentiation to SSCs is initiated around this time and that these events result in a loss of the totipotent phenotype (Labosky et al., 1994; Stewart et al., 1994). The isolation of human EG cell lines from human fetal testis indicates that this transition from PGC to SSCs, and the corresponding phenotypic shifts from a totipotent potential, through a pluripotent stage, to the production of committed SSCs, which are unable to generate other cell lineages, may be a general feature of mammalian germ line differentiation (Shamblott et al., 1998). However, PGCs must be passaged in vitro under appropriate conditions to generate EG cells; freshly isolated PGCs do not display totipotency (Stewart et al., 1994; M. Pesce, personal

communication). This finding indicates that, although these PGCs have a potential for totipotency, this phenotype is restricted in vivo. These observations pose two questions: why do PGCs retain this potential for totipotency and what is the molecular basis for the control of this property? The retention of the totipotent phenotype is likely to be linked to the requirement for the developing germ line to undergo imprinting (for review, see Surani ,1998). The differentiation of the germ line down either the male or female pathway is triggered only after the PGCs enter the genital ridge. It is in this environment that PGCs commit to either the male or female pathway in response to differentiation signals from the surrounding soma. Thus, PGCs must undergo genome-wide demethylation to erase the inherited imprinting pattern and hence the ability to establish either maternal- or paternal-specific imprinting. It is likely that retention of a totipotent phenotype is related to the epigenetic reprogramming that the germ line must undergo. Therefore, the requirement to erase and re-establish imprinting represents an attribute of PGCs that makes them unique when compared with any other stem cell population within the fetus or adult. Thus, totipotency of this cell population may reflect the erasure of methylation patterns to ensure that their genome remains a blank sheet upon which sexspecific imprinting patterns may be written at a later developmental stage. Analysis of the imprinting pattern in EG lines indicates that EG lines isolated at different stages of PGC development show variation and that, unlike ES cells, EG cell lines do not always maintain a stable imprinting pattern (Labosky et al., 1994). This observation leads to the final

Origin of spermatogenic stem cells

conceptual challenge presented by germ cells: how do PGCs manage to undergo switches in phenotype and yet retain a potential for totipotency? PGCs undergo a series of phenotypic switches from their moment of origin and could be viewed as undergoing a controlled differentiation program, although they retain a potential for totipotency throughout (Fig. 1). PGCs first develop a migratory phenotype that allows them to enter the base of the allantois. They then lose or have this phenotype suppressed while in this environment and, after being reintroduced into the fetus via the invaginating gut, they must regain this invasive phenotype to complete their migration to the genital ridge (Gomperts et al., 1994; GarciaCastro et al., 1997). During this migratory phase and for a period after colonization of the genital ridge, PGCs undergo rapid proliferation before entering either meiotic arrest in the female or mitotic arrest in the male fetal gonad.

95

Zygote

ES cell lines

Blastula

Gastrulation Soma Germ

SSC cell lines (?)

Origin of the germ line The molecular events that trigger the separation of the germ line from the soma in the early embryo are still largely undefined. In mice, PGCs are first observed on day 7 after mating. The recruitment of epiblast cells into the germ line is dependent upon Bmp4 expression in the extraembryonic ectoderm (Lawson et al., 1999). Bmp4 regulates the formation of the allantois, the recruitment of cells into the PGC population and the initial number of PGCs. This PGC population may be first identified as a result of the restriction of alkaline phosphatase activity, oct-4 expression, MesP1 expression and the initiation of c-kit expression (Chiquoine, 1954; Ginsburg et al., 1990; Manova and Bachvarova, 1991; Saga et al., 1996; Yeom et al., 1996). oct-4 has been proposed as a marker of totipotency given its expression within the totipotent cells of the early embryo (and ES cell lines derived from them) and the PGC lineage, which has the potential for totipotency (for review, see Pesce et al., 1998). The restriction of oct-4 expression to the PGC lineage involves a switch in cis regulatory sequences, with PGCs maintaining the activity of the distal cis regulatory elements and epiblast expression switching to proximal elements (Yeom et al., 1996). Analysis of the regulatory sequences involved in this switch and the control of other early germ cell markers (alkaline phosphatase and MesP-1) may provide the first indications of the nature of the signalling pathways responsible for the recruitment of epiblast cells into the germ line. The expression of various germ line transcripts during PGC development is outlined (Fig. 1). Entry to the germ line appears to be a position-dependent decision, and the committed progenitors then migrate into the allantois, while the somatic tissue of the embryo undergoes lineage commitment events and loss of totipotency (Tam and Zhou, 1996). The retention of the germ line within the extraembryonic tissue is thought to represent one mechanism by which the germ line evades the differentiation signals to which the embryo is exposed and the selection of this site may reflect its relatively undifferentiated state (Dixon, 1994; Downs and Harmann, 1997). In mice, PGCs reside within this location for about a day as a small population of about 50 cells before being reintroduced into the embryo by the invaginating hindgut, which places them within the embryo but at a site outside the developing gonad. By day 9.5, the PGCs are detected within

EG cell lines

Gametogenesis

Fig. 2. Totipotent cycles. The separation of the somatic and primordial germ cell (PGC) lineages occurs during early gastrulation. Three cell populations that provide access to the germ line compartment after manipulation in vitro are embedded within the normal totipotent cycle of the germ line: embryonic stem (ES) cells that are derived from desegregated blastocysts; embryonic germ (EG) cells derived from PGCs that have yet to commit to the male or female lineage; and spermatogenic stem cells (SSCs). EG lines, while derived from PGCs are re-introduced to the cycle by injection into the blastocyst. In contrast to EG and ES cell lines, the ability to culture and manipulate SSC cell lines has yet to be fully demonstrated.

the epithelium of the developing hindgut. At this stage, the PGC population begins to divide rapidly such that, by day 11.5 after mating, the population has increased to approximately 25 000 cells (Tam and Snow, 1981).

Migration Studies of murine PGCs in vitro and in vivo at different points during the migratory phase, and its termination at the genital ridge, indicate that a number of phenotypic switches are undergone as the cells migrate to and then reach their final destination (Donovan et al., 1986; Gomperts et al., 1994). During this phase, and up to day 12.5 after mating, PGCs undergo rapid division. In the mammalian system, only regulators of survival and proliferation have been identified to date, although the extracellular protein laminin is implicated in the process of PGC migration through the embryo to the genital ridge (Garcia-Castro et al., 1997). Two genes encoding germ line regulators of migration to the genital ridge and PGC proliferation have been identified to date, c-kit and tiar. A third locus, steel, which encodes stem cell factor (SCF), the ligand for the c-Kit receptor, is also required

96

K. A. Sutton

for successful colonization of the genital ridge by PGCs. steel transcripts can be detected along the migratory route of PGCs (Matsui et al., 1990). Expression of c-kit is detectable in mouse PGCs on day 7.5 after mating, while the PGCs are still located in the allantois (Manova and Bachvarova, 1991). Mutation of either c-kit or steel results in a failure of normal PGC migration and proliferation when these cells are reintroduced into the embryo (for review, see Loveland and Schlatt, 1997). The requirement for specific interactions between PGCs and the extracellular matrix is shown by the analysis of integrin β1–/–/wild type chimaeras. Integrin β1–/– cells are capable of contributing to the germ line but these cells fail to migrate correctly and do not contribute significantly to the population of PGCs that colonizes the fetal gonad (Anderson et al., 1999). The requirement for two signals, SCF and laminin, indicates a local microenvironment that both guides the migrating PGCs to the genital ridge and supports their proliferation. The role of the tiar gene product in this process is less clear. Loss of tiar expression results in a phenotype similar to c-kit and steel mutant mice with reduced PGC survival and population of the genital ridge (Beck et al., 1998). Although PGCs can be detected within the genital ridges of tiar –/– fetuses on day 11.5 after mating, their numbers are reduced and, by day 13.5 after mating, PGCs cannot be detected in the fetal gonads. This observation indicates that the defect in the tiar mutant mice primarily affects PGC proliferation and survival once they have reached the genital ridge rather than migration itself. The tiar locus encodes an RNA recognition motif (RRM)– ribonucleoprotein-type RNA-binding protein and, thus, is likely to be involved in the regulation of mRNA stability within the PGCs. Post-transcriptional control of gene expression may be particularly important in PGC development given the retention of totipotency by these cells in conjunction with the requirement for phenotypic shifts. This observation, together with the known importance of post-transcriptional regulation in spermiogenesis, highlights the importance of post-transcriptional regulatory events throughout germ line differentiation. In contrast to the proteins that are expressed in both the fetal and adult germ line, expression of the homeobox protein Pem is restricted to migratory and early post-migratory PGCs (Pitman et al., 1998). Pem protein is first detected on day 8.5 after mating in PGCs associated with the hindgut and expression is high in embryos of both sexes until day 15 after mating. This expression pattern would indicate a role for Pem in the regulation of PGC migration and proliferation, although pem knockout mice are fertile and have no gross defects in PGC migration or proliferation.

Insights from studies in vitro The development of techniques to isolate pure populations of PGCs has facilitated the examination of the growth factor requirements of these cells. Studies have also identified a number of other signalling molecules that may be involved in PGC proliferation and survival; SCF, Gas6, IL-4, leukaemia inhibitory factor (LIF) and basic fibroblast growth factor (bFGF) have all been reported to enhance PGC proliferation (Cooke et al., 1996; Matsubara et al., 1996; Resnick et al., 1998). However, knockout mice with disruptions of the IL-4 receptor and LIF low-affinity receptor do not show any abnormalities in the primordial germ

cell population and IL-4 receptor mice breed normally in pathogen-free conditions (Kuhn et al., 1991; Ware et al., 1995). It is unclear whether the lack of defects in these knockout mice indicates a degree of degeneracy within the signal environment within the embryo that supports PGC survival and proliferation or artefacts of the PGC culture systems used. Studies on EG lines indicate the key growth factors for the derivation of these cell lines are SCF, LIF and bFGF (Resnick et al., 1992, 1998). One problem that arises when PGCs are grown in vitro to study the involvement of growth factors in their survival and proliferation is the ability of these cultures to give rise to EG cell lines that display totipotency. This ability is not a feature of freshly isolated PGCs, so any culture system designed to study PGCs (rather than EG cell lines) must ensure that the cells retain the PGC phenotype and do not de-differentiate to an EG phenotype. Although knowledge of the signalling pathways involved in PGC proliferation and survival is advancing rapidly as a result of in vitro models, the nature of the chemoattractant that controls the migration of PGCs to the developing gonad in mammals remains elusive. Studies in Drosophila, in which PGCs also arise at a site distant to the developing gonad, have identified the product of a homologue of 3 hydroxy-3-methylglutaryl coenzyme A (HMG–CoA) reductase (a protein involved in cholesterol biosynthesis in mammals) as the enzyme responsible for the synthesis of the chemoattractant molecule (Van Doren et al., 1998). Experiments in vitro have demonstrated the production of a chemoattractant by mouse genital ridges, and this activity can be blocked using antibodies directed against TGFβ1 (Godin et al., 1990; Godin and Wylie, 1991).

The final step: differentiation to spermatogenic stem cells The differentiation of PGCs into spermatogenic stem cells appears to be initiated soon after colonization of the genital ridge. At the end of the migratory–proliferative phase, the gonads of the mouse fetus contain 25 000–30 000 PGCs. Although the broad outlines of this process are known, the signalling events that trigger the cessation of migration and proliferation are, as yet, uncharacterized. It has been suggested, on the basis of studies on PGC migration in mouse embryos, that the aggregation of germ cells within the genital ridge at the end of the migratory phase is an important signalling event triggering a cessation of proliferation (Gomperts et al., 1994). Thus, a germ cell–germ cell signalling event may control and limit PGC proliferation within the genital ridge before or in conjunction with differentiation signals from the surrounding soma. A similar mechanism has been proposed for the regulation of germ cell nuclear antigen (GCNA1) (Wang et al., 1997). PGCs in both male and female mouse embryos become positive for GCNA1 as they colonize the genital ridge at day 10.5 after mating, and its expression is maintained until meiosis I. Between day 11.5 and day 12.5 after mating, the male mouse genital ridge undergoes marked re-organization involving the migration of cells from the adjacent mesonephros and the formation of the sex cords (the precursors of the seminiferous cords). These events do not occur in the female genital ridge (Buehr et al., 1993; for review, see McLaren, 1998). In mice by day 12.5 after mating, PGCs within the male fetal gonad have

Origin of spermatogenic stem cells

entered mitotic arrest and remain quiescent until spermiogenesis is initiated shortly after birth. It is at this stage that the differentiation from PGCs to SSCs occurs in males. This event is poorly defined but involves the loss of positive staining for alkaline phosphatase, Pem and the loss of totipotent potential of the germ cells (Fig. 1). In females, rather than entering mitotic arrest, the germ line enters meiosis and then arrests until the onset of puberty. McLaren and Southee (1997) have suggested that PGCs are preprogrammed to enter meiosis upon reaching the required density. Using dissociation–re-association experiments, McLaren and Southee (1997) demonstrated that the PGCs from male or female genital ridges on day 10.5 and day 11.5 after mating enter meiosis when re-associated with cells from the fetal lung. However, PGCs isolated from male embryos on day 12.5 after mating would not enter meiosis when dissociated from the fetal gonad. When male genital ridges were dissociated and then allowed to re-associate on day 11.5 after mating, the PGCs entered meiosis, indicating that the signal required to prevent PGCs in the male embryo entering meiosis is dependent upon a tightly regulated micro-environment involving specific cell–cell contacts. The microenvironment within the developing male gonad extends the time frame during which EG lines may be generated from the male germ line. However, the failure to observe germ line transmission of EG cells derived from PGCs isolated from male fetal gonads on day 12.5 after mating indicates that these cells have commenced a differentiation programme and have lost their potential for totipotency, although these EG lines are pluripotent (Labosky et al., 1994). In addition to these phenotypic changes, the expression of several markers also changes after the mitotic arrest of the male germ line. While c-kit and oct-4 continue to be expressed by the spermatogenic stem cells of the adult and in the pre-meiotic stages of spermatogenesis, other genes are inactivated. Both alkaline phosphatase activity and a transcript derived from an intronic promoter in the PDGFRα gene cannot be detected within the adult testis but are expressed by the cells of carcinoma in situ (Mosselman et al., 1996). Carcinoma in situ is probably derived from PGCs that have failed to complete their differentiation to spermatogenic stem cells. While Oct-4 has been implicated in the regulation of this intronic promoter, its inactivation as male germ cells progress from PGCs to spermatogenic stem cells indicates that additional factors are important for its regulation, given that Oct-4 is present in both PGCs and spermatogenic stem cells (Kraft et al., 1996). In addition to these two markers, pem expression is also downregulated during this transition. This change in marker expression argues for a male germ line differentiation event driven by the somatic tissue that over-rides early entry into meiosis. The induction of GCNA1 expression across this window (expression is initiated on day 11.5 after mating), before the entry of PGCs into mitotic arrest or meiosis, indicates that the differentiation process is initiated at the time of genital ridge colonization. The induction of GCNA1 in PGCs populating both male and female genital ridges, and the ability of PGCs that develop outside the fetal gonad to produce GCNA1, indicate the involvement of autocrine regulation within the PGC population (Enders and May, 1994; Wang et al., 1997). This finding and observations on germ cell migration and proliferation indicate that facets of

97

PGC development may be regulated in a cell-autonomous fashion rather than being directed by the soma. In this scenario, the role of the soma may be presented as controlling when, rather than if, the germ line enters meiosis. Evidence for the role of somatic signals in regulating the number of PGCs within the genital idge and developing gonad comes from experiments using rat fetal gonads which demonstrated the induction of apoptosis in germ cells after addition of TGFβ1 and TGFβ2 (Olaso et al., 1998). Such observations indicate that there is some somatic control over germ line survival in the male gonad. The complete set of receptors for the TGFβ superfamily expressed by the fetal testis is yet to be fully characterized, so it is unclear whether this observation reflects a direct effect of TGFβ1 and TGFβ2 on the PGC population or an indirect effect via somatic cells. The defects in spermatogenesis observed in BMP8b and desert hedgehog (Dhh) knockout mice indicate that these signalling pathways are likely to be critical for the correct formation of the spermatogenic stem cell population in the fetus and its maintenance in the adult (Bitgood et al., 1996; Zhao et al., 1996). Cell transfer experiments indicate that the differentiation from PGC to spermatogenic stem cell may not be completed until the onset of spermatogenic initiation (Brinster and Avarbock, 1994). Since it is possible that these observations reflect differences in cell survival during isolation, the development of simplified separation techniques on the basis of α6 and β1 integrin expression by SSCs should allow the time of SSC establishment to be more accurately defined (Takashi et al., 1999). It is interesting to note that the male germline is in mitotic arrest over the period when the transition from PGC to SSC occurs, indicating that differentiation is occurring in the absence of proliferation.

Conclusions Although understanding of the cell biology of the germ line is considerable, remarkably little is known about the molecular and biochemical processes involved in the control of PGC origin and differentiation. Indeed, to date, the spermatogenic stem cell is totally uncharacterized at this level. Successful completion of this differentiation step is essential if spermatogenesis is to be established in the adult testis, and improper or incomplete differentiation may result in the formation of carcinoma in situ. The conjunction of improved isolation techniques, in vitro culture systems and the ability to assay for biological end points using repopulation studies indicates that the stage is now set for a rapid advancement in our understanding of both the generation and maintenance of the spermatogenic stem cell population. The author thanks Shaun Roman and John Aitken for critical reading of the manuscript.

References Key references are indicated by asterisks. *Anderson R, Fassler R, Georges-Labouesse E, Hynes RO, Bader BL, Kreidberg JA, Schaible K, Heasman J and Wylie C (1999) Mouse primordial germ cells lacking β1 integrins enter the germline but fail to migrate normally to the gonads Development 126 1655–1664

98

K. A. Sutton

Beck ARP, Miller IJ, Anderson P and Streuli M (1998) RNA-binding protein TIAR is essential for primordial germ cell development Proceedings National Academy of Sciences USA 95 2331–2336 Bitgood MJ, Shen L and McMahon AP (1996) Sertoli cell signaling by desert hedgehog regulates the male germline Current Biology 6 298–304 Brinster RL and Avarbock MR (1994) Germline transmission of donor haplotype following spermatogonial transplantation development Proceedings National Academy of Sciences USA 91 11 303–11 307 Buehr M, Gu S and McLaren A (1993) Mesonephric contribution to testis differentiation in the fetal mouse Development 117 273–281 Chiquoine AD (1954) The identification, origin and migration of the primordial germ cells in the mouse embryo Anatomical Record 118 135–146 Cooke JE, Heasman J and Wylie CC (1996) The role of interleukin-4 in the regulation of mouse primordial germ cell numbers Developmental Biology 174 12–21 Dixon KE (1994) Evolutionary aspects of primordial germ cell formation. In Ciba Foundation Symposium 182: Germline Development pp 92–110 Eds J Marsh and J Goode. John Wiley and Sons Ltd, Chichester Donovan PJ, Stott D, Cairns LA, Heasman J and Wylie CC (1986) Migratory and postmigratory mouse primordial germ cells behave differently in culture Cell 44 831–838 Downs KM and Harmann C (1997) Developmental potency of the murine allantois Development 124 2768–2780 Enders GC and May JJ (1994) Developmentally regulated expression of a mouse germ cell nuclear antigen examined from embryonic day 11 to adult in male and female mice Developmental Biology 163 331–340 Garcia-Castro MI, Anderson R, Heasman J and Wylie C (1997) Interactions between germ cells and extracellular matrix glycoproteins during migration and gonad assembly in the mouse embryo Journal of Cell Biology 138 471–480 Ginsburg M, Snow MH and McLaren A (1990) Primordial germ cells in the mouse embryo during gastrulation Development 110 521–528 Godin I and Wylie CC (1991) TGFβ1 inhibits proliferation and has a chemotropic effect on mouse primordial germ cells Development 113 1451–1457 Godin I, Wylie CC and Heasman J (1990) Genital ridges exert long-range effects on mouse primordial germ cell numbers and direction of migration in culture Development 108 357–363 *Gomperts M, Garcia-Castro M, Wylie C and Heasman J (1994) Interactions between primordial germ cells play a role in their migration in mouse embryos Development 120 135–141 Koopman P, Gubbay J, Vivian N, Goodfellow P and Lovell-Badge R (1991) Male development of chromosomally female mice transgenic for Sry Nature 351 117–121 Kraft HJ, Mosselman S, Smits HA, Hohenstein P, Piek E, Chen Q, Artzt K and van Zoelen EJ (1996) Oct-4 regulates alternative splicing of plateletderived growth factor alpha gene promoter in human embryonal carcinoma cells Journal of Biological Chemistry 271 12 873–12 878 Kuhn R, Rajewsky K and Muller W (1991) Generation and analysis of IL-4 knockout mice Science 254 707–710 *Labosky PA, Barlow DP and Hogan BLM (1994) Mouse embryonic germ (EG) cell lines: transmission through the germline and differences in the methylation imprint of insulin-like growth factor 2 receptor (Igf2r) gene compared with embryonic stem (ES) cell lines Development 120 3197–3204 *Lawson AL, Dunn NR, Roelen BAJ, Zeinstra LMZ, Davis AM, Wright CVE, Korving JPWFM and Hogan BLM (1999) Bmp4 is required for the generation of primordial germ cells in the mouse embryo Genes and Development 13 424–436 Loveland KL and Schlatt (1997) Stem cell factor and c-kit in the mammalian testis: lessons originating from mother nature’s gene knockouts Journal of Endocrinology 153 337–344 McLaren A (1998) Gonad development: assembling the mammlian testis Current Biology 8 R175–177 McLaren A and Southee D (1997) Entry of mouse embryonic germ cells into meiosis Developmental Biology 187 107–113

Manova K and Bachvarova RF (1991) Expression of c-kit encoded at the W locus of mice in developing embryonic germ cells and presumptive melanoblasts Developmental Biology 146 312–324 Matsubara N, Takahashi Y, Nishina Y, Mukouyama Y, Yanagisawa M, Watanabe T, Nakano T, Nomura K, Arita H, Nishimune Y, Obinata M, and Matsui Y (1996) A receptor tyrosine kinase, Sky, and its ligand Gas 6 are expressed in gonads and support primordial germ cell growth or survival in culture Developmental Biology 180 499–510 Matsui Y, Zsebo KM and Hogan BLM (1990) Embryonic expression of a haematopoietic growth factor encoded by the the Sl locus and the ligand for the c-kit Nature 347 667–669 Mosselman S, Looijenga LHJ, Gillis AJM, van Roijen MA, Kraft HJ, van Zoelen EJJ and Oosterhuis JW (1996) Aberrant platelet-derived growth factor α-receptor transcript as a diagnostic marker for early human germ cell tumours of the adult testis Proceedings National Academy of Sciences USA 93 2884–2888 Olaso R, Pairault C, Boulogne B, Durand P and Habert R (1998) Transforming growth factor b1 and b2 reduce the number of gonocytes by increasing apoptosis Endocrinology 139 733–740 Pesce M, Gross MK and Scholer HR (1998) In line with our ancestors: Oct-4 and the mammalian germ Bioessays 20 722–732 Pitman JL, Lin TP, Kleeman JE, Erickson GF, MacLeod C (1998) Normal reproductive and macrophage function in Pem homeobox gene deficient mice Developmental Biology 202 196–214 Resnick JL, Bixler LS, Cheng L and Donovan PJ (1992) Long-term proliferation of mouse primordial germ cells in culture Nature 359 550–551 Resnick JL, Ortiz M, Keller JR and Donovan PJ (1998) Role of fibroblast growth factors and their receptors in mouse primordial cell growth Biology of Reproduction 59 1224–1229 Saga Y, Hata N, Kobayashi S, Magnuson T, Seldin MF and Taketo MM (1996) MesP1: a novel basic helix–loop–helix protein expressed in the nascent mesodermal cells during mouse gastrulation Development 122 2769–2778 *Shamblott MJ, Axelman, Wang S, Bugg EM, Littlefield JW, Donovan PJ, Blumenthal PD, Huggins GR and Gearhart JD (1998) Derivation of pluripotent stem cells from cultured human primordial germ cells Proceedings National Academy of Sciences USA 95 13 726–13 731 Stewart CL, Gadi I and Bhatt H (1994) Stem cells from primordial germ cells can re-enter the germ line Developmental Biology 161 626–628 Surani MA (1998) Imprinting and the initiation of gene silencing in the germ line Cell 93 309–312 Takashi S, Avarbock MR and Brinster RL (1999) β1- and α6-integrin are surface markers on mouse spermatogonial stem cells Proceedings National Academy of Sciences USA 96 5504–5509 Tam PPL and Snow MHL (1981) Proliferation and migration of primordial germ cells during compensatory growth in mouse embryos Journal of Embryology and Experimental Morphology 64 133–147 Tam PPL and Zhou SX (1996) The allocation of epiblast cells to ectodermal and germ-line lineages is influenced by the position of the cells in the gastrulating mouse embryo Developmental Biology 178 124–132 Van Doren M, Broihier HT, Moore LA and Lehman R (1998) HMG-CoA reductase guides migrating primordial germ cells Nature 396 466–469 Wang D, Ikeda Y, Parker KL and Enders GC (1997) Germ cell nuclear antigen (GCNA1) expression does not require a gonadal environment or steroidogenic factor 1: examination of GCNA1 in ectopic germ cells and in Ftz-F1 null mice Molecular Reproduction and Development 48 154–158 Ware CB, Horowitz MC, Renshaw BR et al. (1995) targeted disruption of the low-affinity leulemia inhibitory factor receptor gene causes placental, skeletal, neural and metabolic defects and results in perinatal death Development 121 1283–1299 Yeom YI, Fuhrmann, Ovitt CE, Brehm A, Ohbo K, Gross M, Hubner K and Scholer HR (1996) Germline regulatory element of oct-4 specific for the totipotent cycle of embryonal cells Development 122 881–894 Zhao GQ, Deng K, Labosky PA, Liaw L and Hogan BLM (1996) The gene encoding bone morphogenetic protein 8b is required for the initiation and maintenance of spermatogenesis Genes and Development 10 1657–1669