Mouse Models of Testicular Germ Cell Tumors

Germ cell tumors arise from anomalies in primordial germ cells (PGCs) (Stevens, 1967), the embryonic precursors of oocytes and sperm. Their normal development follows three steps: migration, proliferation and differentiation into mature germ cells (Mauduit et al., 1999). Abnormalities in these steps can result in sterility, reduced fertility, or in some cases to transformation into ovarian tumors in females or testicular germ cell tumors (TGCTs) in males (Stevens, 1967), both of which present remarkable cellular and tissue heterogeneity reflecting the pluripotent nature of the TGCT stem cell.


Introduction
Germ cell tumors arise from anomalies in primordial germ cells (PGCs) (Stevens, 1967), the embryonic precursors of oocytes and sperm.Their normal development follows three steps: migration, proliferation and differentiation into mature germ cells (Mauduit et al., 1999).Abnormalities in these steps can result in sterility, reduced fertility, or in some cases to transformation into ovarian tumors in females or testicular germ cell tumors (TGCTs) in males (Stevens, 1967), both of which present remarkable cellular and tissue heterogeneity reflecting the pluripotent nature of the TGCT stem cell.
Although representing only 1-2% of all cancers in men, TGCTs are the most common malignancy affecting young men 15-35 years of age (Buetow, 1995).Over the past decade, TGCT incidence has risen ~1.2% per year with about 8,300-8,400 new cases reported in the United States (American Cancer Society, www.cancer.org).TGCT risk varies more than 5fold among ethnic groups and geographic regions (McGlynn et al., 2005;www.cancer.org).In addition, developmental anomalies such as undescended testis are indicators of TGCT risk (Dieckmann & Pichlmeier, 2004).Finally, environmental factors such as pesticides or insecticides strongly influence susceptibility.However, the mechanisms by which genetic and environmental factors affect susceptibility remain elusive.
Genetic factors account for 25% of susceptibility to TGCTs, making these tumors the third most heritable cancer (Gilbert et al., 2011).Family history is a significant risk factor with 8to 10-fold risk for brothers of men with TGCTs and 4-to 6-fold risk for sons (Hutter et al., 1967).The risk increases 75-fold for monozygotic twins (Swerdlow et al., 1999).Genomewide association studies (GWAS) in humans reveal various loci that contribute to susceptibility, but the identity of these genes has not yet been established (Gilbert et al., 2011).Discovery of TGCTs in the 129 family of inbred strains of mice has enabled identification and characterization of specific genes and their interactions (Stevens & Little, 1954;Stevens, 1973).
In this chapter, after discussing the various mouse models for TGCTs, we review the evidence for TGCT genes and their role in tumorigenesis in both humans and mice.We consider the consequence of their mutation, as well as the role of gene interactions, to better understand molecular pathways of PGC transformation and pathogenesis.Finally, we discuss evidence for transgenerational effects that influence TGCT incidence.

Mouse models for human TGCTs
Studies of TGCT development before birth are not feasible in humans.Laboratory mice therefore provide unique opportunities to determine the genetic basis for TGCT susceptibility and for characterizing key components contributing to PGC transformation.
At P6, gonocytes differentiate into Type A1 spermatogonia and begin radial migration in the seminiferous tubule (Nagano et al., 2000).Germ cells undergo either a proliferative period (to maintain the progenitor population throughout the reproductive life of males) or meiosis (to mature germ cells) (Nagano et al., 2000).During these postnatal days, pre-Sertoli cells transform into mature Sertoli cells that support germ cell differentiation and spermatogenesis (Griswold, 1995;Hess et al., 2006).Remarkably, a significant proportion of PGCs remains scattered along the migratory route.These ectopic PGCs are eliminated between E10 and E17 by activating the intrinsic cell death pathway (Stallock et al., 2003).This intrinsic pathway is regulated by BCL2 family members (i.e.BAX, BAK and BCL2), which activate mitochondria leading to secretion of cytochrome C, an initiator of apoptosis (Shimizu et al., 1999;Stallock et al., 2003).Another BAX-dependent apoptotic control point occurs at P10-P13 before meiosis to eliminate germ cells that fail either to migrate in the seminiferous tubules or to repair DNA (de Rooij, 2001).
In the 129 family of inbred strains, 5%-10% of males develop spontaneous TGCTs (Fig. 1) (Stevens & Little, 1954;Stevens, 1967), making these strains highly relevant for learning about TGCT development in humans.Indeed, TGCTs in mice are most similar to pediatric TGCTs in humans, and both species have a similar left-sided preference of the tumors.Bilateral TGCTs are infrequent in mice and usually involve less than 5% of all cases in humans (Dieckmann & Pichlmeier, 2004).The critical period for transformation is between E11.5-E13.5 in mice (Stevens & Little, 1954;Stevens, 1967).Interestingly, syncytial masses of atypical gonocytes are detected in embryonic testis cords from E13 onward in 129 wild-type and mutant males (Stevens, 1967;T. Noguchi & M. Noguchi, 1985;Rivers & Hamilton, 1986;Matin et al., 1999).These atypical cells may result from incomplete cell cycle, mitotic arrest defects, or de-differentiation after their entrance in quiescence (Rivers & Hamilton, 1986).
These so-called embryonal carcinoma (EC) cells escape necrosis, and then proliferate and aggregate together to form clusters, which are correlated with partial or complete germ cell deficiency (Stevens, 1967;T. Noguchi & M. Noguchi, 1985;Rivers & Hamilton, 1986;Matin et al., 1999).Furthermore, EC clusters have been suggested as the precursors of either testicular abnormalities, which have been found in an appreciable frequency (18%) in 129 males, or TGCTs (Rivers & Hamilton, 1986).Similar abnormalities, termed carcinoma in situ, have been found in humans many years before invasive malignancy (Skakkebaek, 1972).Why some germ cells transform into EC clusters while others develop normally is unclear, just as why some abnormal gonocytes die while others persist.

TGCTs in the 129-Ter strain
The Ter allele, also known as Teratoma, has dramatic effects on PGC biology and TGCT susceptibility (Stevens, 1973;T. Noguchi & M. Noguchi, 1985), and was recently identified as a spontaneous mutation in the Dead-end ( Dnd1) gene (Youngren et al., 2005).Ter homozygosity causes severe germ cell deficiency in both sexes, probably mediated through BAX-dependent apoptosis after E8.5 (Stevens, 1973;T. Noguchi & M. Noguchi, 1985;Cook et al., 2009).EC clusters develop from E13 in 129-Ter/Ter male embryos probably due to a defect in G1/S mitotic arrest (Rivers & Hamilton, 1986).The few surviving PGCs successfully migrate to the genital ridges, suggesting that DND1 is not essential for migration (Youngren et al., 2005).Deficiency progresses with age until P11 when PGCs are no longer detectable (T.Noguchi & M. Noguchi, 1985;Cook et al., 2009).As a result, adult mutant males are sterile (Fig. 1) (T.Noguchi & M. Noguchi, 1985).Somatic development of Sertoli and Leydig progenitors is not affecte d i n a c c o r d a n c e w i t h t h e P G C -s p e c i f i c expression of Dnd1 (Weidinger et al., 2003).Interestingly, 17% of 129-Ter/+ males develop tumors, 10% of which are bilateral, but these males are fertile.This rate increases to 94% in homozygotes with 75% of cases having bilateral tumors (T.Noguchi & M. Noguchi, 1985).
DND1 was originally proposed to be a component of the cytidine to uridine RNA editing complex given its similarity with the Apobec complementation factor (A1CF) (Weidinger et al., 2003).More recently, DND1 was shown to block miRNA access to their mRNA targets (Kedde et al., 2007).The reactivated target genes are involved in PGC pluripotency (i.e.Sox2, Nanos1 and Nanog), in cell cycle regulation (i.e.Cyclin-dependent kinase inhibitors (Cdkn) 1b, Cdkn1a) and in PGC survival (i.e.Transformation related protein (Trp53), apoptotic factor Bax and Phosphatase and tensin homolog (Pten)) (Kedde et al., 2007;Cook et al., 2011;R. Zhu et al., 2011).Dnd1 is expressed throughout embryogenesis with an up-regulation between E12.5 and E15.5 in males, the critical period for TGCT development (Youngren et al., 2005;Cook et al., 2009).Thus, loss of DND1 in Ter mutants strongly affects differentiation, survival and entry in quiescence of PGCs and dramatically enhances the TGCT frequency in the 129 strain.

Consomic 129 inbred strains
Chromosome Substitution Strains (CSSs, consomic strains) carry entire chromosomes derived from another inbred strain.Some CSSs were created to study the genetic linkage of the MOLF-derived TGCT modifier genes.MOLF is derived from Mus musculus molossinus, which is genetically distinct from 129 and has no predisposition for TGCTs (Mieno et al., 1989;Matin et al., 1998).We will review results for two consomic strains.

129-M18
The 129-M18 CSS substitutes MOLF-derived chromosome 18 for its homologue in 129 wildtype mice (Anderson et al., 2009a).129-M18 males show complete resistance to develop TGCTs with no homozygous males affected.Four quantitative trait loci (QTLs, Region I-IV) were identified independent of the Dnd1 gene, which is also located on this chromosome.
Region I shows homology with the 10p11 region in humans (Copeland et al., 1993).One candidate gene may be Map3k8, which encodes a mitogen-activated protein kinase (MAPK) that acts downstream of tyrosine kinase-dependent pathway (Patriotis et al., 1994).Region II belongs to a conserved region 5q in humans (Copeland et al., 1993).In mice, this region contains Eif1a gene, which encodes a translational regulator that is functionally related to eIF2α, encoded by Eif2s2 gene, loss of which suppresses TGCT development (Heaney et al., 2009) (Section 3.5).However, the three homologues of Eif1a gene are EIF1AD on chromosome 11, and EIF1AX and EIF1AY on sex chromosomes in humans.Region III contains at least one TGCT enhancer and is conserved with the 18q region in humans (Copeland et al., 1993).An interesting candidate gene is Noxa, which encodes a mitochondria-mediated apoptotic factor (Krishna et al., 2011).Interestingly, Noxa is activated by the TGCT suppressor TRP53 (Michalak et al., 2005;Donehower et al., 1992) (Section 3.3).Region IV contains the F-box only protein 15 gene, one of the few known targets of the stem cell pluripotent factor OCT3/4 (Tokuzawa et al., 2003).
Regions on chromosome 19 are homologous to either 9p, which contains the TGCT modifier doublesex-& Mab3-related transcription factor 1 (Dmrt1) gene (Section 3.4), 9q, 10q, which contains the TGCT modifier Pten (Section 3.2) and 11q in humans (Copeland et al., 1993).An interesting candidate gene is Splicing factor 1 (Sf1), which encodes an RNA binding protein that functions as a pre-mRNA splicing factor (Z. Liu et al., 2001).Interestingly, Sf1 deficiency (heterozygous Sf1 -/ + ) in 129-M19/+ males reduces the incidence of TGCTs (R. Zhu et al., 2010), suggesting that Sf1 may be one of the TGCT enhancer genes on chromosome 19.D19Bwg1357e is a predicted gene down-regulated in the gonads of MOLF strain mice (R. Zhu et al., 2007).This gene has an RNA-binding domain homologous to those in Pum1 and Pum2 genes, which encode two major components of P-bodies, the center of RNA processing (Moore et al., 2003), suggesting a similar role for D19Bwg1357e.
Thus, CSSs are powerful tools for identifying new genes that, alone or in combination and with conventional additive or epistatic effects, confer susceptibility to TGCTs.These candidate genes are involved in RNA biology, epigenetic regulation and intracellular pathways regulating PGC survival, proliferation and pluripotency.

TGCT modifier genes in the 129 strain
Spontaneous and engineered mutations are essential for characterizing molecular pathways involved in PGC development and transformation into TGCTs.In this section, we review the phenotypic traits of 129 mice that carry mutations on TGCT modifier genes.

Kit and Kitlg
In the mouse, mutations at the White-spotting (W) or Steel (Sl) loci cause sterility and severe anemia that lead to in utero or perinatal death in homozygotes (McCoshen & McCallion, 1975;Nocka et al., 1990).The W locus, located on chromosome 5 in mice (region 4q12 in humans), encodes KIT, a tyrosine kinase receptor (Manova et al., 1990).The ligand of KIT, named KITLG, is encoded at the Sl locus on chromosome 10 in mice (region 12q21 in humans) (Flanagan et al., 1991).These two factors are essential for hematopoiesis, melanogenesis and gametogenesis (Bernstein et al., 1991;Besmer et al., 1993).
In the testis, KIT is expressed on the surface of PGCs from E7.5 to E13.5, and then at P5, on differentiating germ cells and interstitial Leydig cells (Manova et al., 1990;Yoshinaga et al., 1991).KITLG is only expressed by the pre-Sertoli and Sertoli cells (Rossi et al., 1991;Tajima et al., 1991).The two forms of KITLG, soluble and membrane-bound, are differentially expressed depending on developmental stages (Matsui et al., 1990;Flanagan et al., 1991;E.J. Huang et al., 1992).The membrane-bound KITLG is predominantly expressed during proliferative periods: between E8 and E14 and just after the birth (P3), and the soluble form when PGCs are quiescent (between E13.5 and P3) (Matsui et al., 1990;Flanagan et al., 1991;E.J. Huang et al., 1992).In addition of their essential role during migration, the interaction of KITLG with its receptor leads to the dimerization of KIT and its auto-phosphorylation activates two major pathways: (1) the PI3K/AKT signaling cascade regulates transcription of mitotic inhibitors such as CDKN1a, CDKN1b and cyclin D1, and enhances the translational factor eIF4E; (2) the MAPK pathway regulates factors involved in pluripotency such as NANOG and SOX2, and in proliferation (Mithraprabhu & Loveland, 2009).By acting on TRP53 activity, KIT/KITLG pathway controls apoptosis of PGCs by regulating the BCL2 components (i.e.BAX, BAK, BCL2) and their cofactors (i.e.PUMA, NOXA) (Pesce et al., 1993;Carson et al., 1994), explaining why ectopic PGCs that lose KIT/KITLG interaction in 129 wild-type mice are eliminated by apoptosis.
Loss of KIT (Kit W and Kit Wv ) or KITLG (Kitlg Sl , Kitlg SlJ and Kitlg Slgb ) leads to massive PGC loss, resulting from high levels of apoptosis beginning on or before E9 (Stevens, 1967;McCoshen & McCallion, 1975;Nocka et al., 1990;Heaney et al., 2008).In these Kit and Kitlg heterozygous mutants, the wild-type allele is sufficient to rescue PGC viability at E13.5 and fertility in adult males.Loss of only membrane-bound KITLG in the Kitlg Sld deletion leads to a mild phenotype characterized by partial PGC deficiency because of a proliferation defect (Fig. 1) (Flanagan et al., 1991;Tajima et al., 1991).At later stages, this mutation adversely affects PGC differentiation resulting in sterility.In Kitlg Sl17H mutants, the membrane-bound form is not functional due to absence of its cytoplasmic tail (Brannan et al., 1992).These mutants present a slight phenotype apparent after birth with anomalies in spermatogenesis, leading to sterility (Fig. 1).Surprisingly, Kitlg Sld and Kitlg Sl17H mutants have normal apoptosis when mice lacking KIT or KITLG have a high apoptotic rate, suggesting that soluble KITLG alone is sufficient to re-establishing normal apoptosis (Flanagan et al., 1991;Brannan et al., 1992).Nonetheless, neither Kitlg Sld nor Kitlg Sl17H mutants are fertile, suggesting that membrane-bound KITLG is necessary for complete PGC development (proliferation and differentiation).Thus, the KIT/KITLG pathway controls the migration, proliferation and survival of PGCs during embryogenesis, and the proliferation, differentiation and the radial development of germ cells after birth.
Although loss of KIT and KITLG have similar effects on PGC development, only Kitlg Sl , Kitlg SlJ and Kitlg Slgb heterozygotes have 2-fold increase in occurrence of TGCT-affected males (Fig. 1) (Stevens, 1967;Heaney et al., 2008), suggesting that KIT is haplosufficient to promote TGCT formation.Interestingly, Kitlg Sld allele has no effect on TGCT susceptibility (Heaney et al., 2008), suggesting that soluble KITLG is sufficient to suppress TGCT formation.The presence of EC clusters and the frequency of testicular abnormalities in Kitlg Sl and Kit W heterozygous males remain to be evaluated.However, given that EC clusters are probably the origin of TGCTs and that some clusters have been described in mice lacking the PI3K binding site on KIT (Kissel et al., 2000), we expect to find EC clusters at least in Kitlg Sl , Kitlg SlJ and Kitlg Slgb heterozygous males.Determining whether soluble KITLG in Kitlg Sld mutants reduces the frequency of these clusters, or whether Kit W mutants also carry EC clusters without transformation into tumors, could highlight the molecular pathway involved in transformation of benign EC masses into malignant tumors.

Pten
Pten encodes a phosphatase that antagonizes both PI3K/AKT and MAPK signaling cascades through its dual phosphatase activities (Myers et al., 1997).Thus, PTEN is a key element of the KIT/KITLG pathway.Interestingly, PTEN regulates its own expression by stabilizing its transcriptional activator TRP53 (Tang & Eng, 2006).
In the testis, PTEN is expressed in PGCs, but not in Sertoli cells (Kimura et al., 2003).Loss of PTEN leads to embryonic death, and heterozygotes have a high tumor incidence (Di Cristofano et al., 1998).PGCs lacking PTEN have defects in mitotic arrest and slow pluripotency loss after E13.5 (Fig. 1), and form EC clusters (Kimura et al., 2003).A high apoptotic rate is observed in the testis after E13.5 but remains insufficient to counterbalance abnormal proliferation.All mutant males develop bilateral TGCTs (Kimura et al., 2003).Pten transcript levels are reduced at least 2-fold in E13.5 gonads in MOLF males compared with 129 males (R. Zhu et al., 2007), suggesting that Pten is one of the genes on chromosome 19 that contribute to the high TGCT frequency.Furthermore, Pten is the only genetic variant that increases TGCT susceptibility to 100% in a mixed genetic background (Kimura et al., 2003).
Adenosine triphosphate-binding domains on PTEN regulate its subcellular localization (Lobo et al., 2009).Defect in these domains results in a predominantly nuclear localization, a DNA repair defect, and an inappropriate regulation of G1/S progression associated with a reduced apoptotic rate (He et al., 2011).PTEN mislocalization also leads to a reduced nuclear TRP53 level and transcriptional activity (He et al., 2011).Thus, PTEN plays an important role in cell growth and tumorigenesis, by regulating apoptosis, pluripotency, chromosome stability, DNA repair and cell cycle arrest (Kimura et al., 2003;Shen et al., 2007;Saal et al., 2008;He et al., 2011).Whether PTEN localization affects TGCT susceptibility is an open question.

Trp53
The Trp53 gene encodes a tumor suppressor expressed in both PGCs and pre-Sertoli cells during embryogenesis, down-regulated after the birth in spermatogonia, and re-expressed in primary spermatocytes at pachytene, suggesting a role for TRP53 in control of meiosis (Almon et al., 1993;Schwartz et al., 1993).TRP53 activity is under the control of the PI3K/AKT pathway and depends of TRP53 phosphorylation (Xu, 2003).TRP53 regulates expression of several genes encoding mitotic regulators (i.e.CDKN1a, cyclin G1), the tumor suppressor PTEN, the pluripotent factor NANOG, and several apoptotic regulators (i.e.BAX, NOXA, PUMA) (Lin et al., 2005;Michalak et al., 2005;Tang & Eng, 2006).TRP53 also activates PUMA, which controls activity of BCL2 components and induces a mitochondria-dependent apoptosis (Chipuk et al., 2004;D. Liu et al., 2010).Thus, TRP53 controls many key cellular pathways including apoptosis, pluripotency, G1/S cell cycle arrest, and meiosis.
DNA damage leads to deficiency of both germ and Sertoli cells after γ-irradiation of fetal testis due to apoptosis and a proliferation defect (Lambrot et al., 2007).PGC deficiency is associated with increased TRP53 activity and induction of BAX, PUMA and CDKN1a expression (Lambrot et al., 2007).In mice, loss of TRP53 phosphorylation leads to constitutive activation of TRP53 and in turn too embryonic lethality (D. Liu et al., 2010).In contrast, Trp53 -/ -mice have normal development but a significantly reduced lifespan due to high predisposition for spontaneous tumors, especially lymphomas (Donehower et al., 1992;Jacks et al., 1994).Disruption of the Trp53 gene in a pure 129 background increases TGCT incidence to 35-50% (Fig. 1) (Rotter et al., 1993;Donehower et al., 1995).Homozygotes have a giant-cell degenerative syndrome characterized by abnormal primary spermatocytes that arrest meiosis at pachytene, and form clusters leading to germ cell deficiency but remain fertile (Rotter et al., 1993).Heterozygotes are fertile with a 2-fold increase of TGCT frequency in the 129 strain (Rotter et al., 1993;Donehower et al., 1995).In a 50/50 mixed background of C57BL/6 and 129, the TGCT incidence is reduced to ~20% (Jacks et al., 1994).On these backgrounds, heterozygotes exhibit apparently normal testicular morphology (Rotter et al., 1993;Jacks et al., 1994;Muller et al., 2000).Thus, the combination of the Trp53 defect and the 129 genetic background results in a synergistic increase of giant-cell syndrome and TGCT penetrance.

Dmrt1
Dmrt1 encodes a male-specific transcriptional factor (Raymond et al., 1999).In the testis, DMRT1 is strongly expressed in pre-Sertoli cells and then in both Sertoli and undifferentiated germ cells from P1 onward (Raymond et al., 1999).DMRT1 disappears in germ cells that enter meiosis, suggesting that DMRT1 regulates initiation of either meiosis, mitotic arrest, or both, in a stage-dependent manner (Raymond et al., 2000;Fahrioglu et al., 2007).During embryogenesis, DMRT1 controls transformation of PGCs into gonocytes by regulating expression of pluripotent factors such as SOX2 and NANOG, and their entrance into a quiescent state by controlling some cell cycle inhibitors such as CDKN2d (Krentz et al., 2009;Murphy et al., 2010).After birth, the DMRT1 control of cell cycle kinases allows mitotic reactivation of male gonocytes.DMRT1 also plays a role in radial migration (Fahrioglu et al., 2007).Finally, DMRT1 acts as a transcriptional gatekeeper that controls the switch from mitosis to meiosis in the undifferentiated spermatogonia (Matson et al., 2010).
Interestingly, DMRT1 also has an indirect control of critical developmental steps by regulating expression of the retinaldehyde dehydrogenases ALDH1A1 and ADH4 (Matson et al., 2010).These enzymes are expressed in Sertoli cells and convert vitamin A-derived retinal into retinoic acid.Retinoic acid regulates cell proliferation, migration and differentiation (Bowles et al., 2006;Koubova et al., 2006).Catabolism of retinoic acid is facilitated by cytochrome P450 enzymes such as CYP26B1, which is highly expressed until E13.5 by mitochondria in Sertoli cells (Li et al., 2009).Moreover, retinoic acid signaling is under the control of NANOS2, which is an RNA binding protein located in P-bodies, and is essential for male PGC development (Tsuda et al., 2003;A. Suzuki & Saga, 2008).PGCs lacking CYP26B1 enter meiosis at E13.5 and have a high apoptotic rate (MacLean et al., 2007).Similarly, NANOS2 loss affects gonocytes that re-enter in the proliferative phase at E15, immediately initiate meiosis, and finally are completely depleted at E18.5 (Tsuda et al., 2003;A. Suzuki & Saga, 2008).DMRT1 loss in the 129 strain increases 10-fold the numbers of PGCs that escape mitotic arrest at E16.5 and return into pluripotent state to form EC clusters but without re-entry into meiosis as observed in Nanos2 -/ -and Cyp26b1 -/ -mutants (Raymond et al., 2000;Krentz et al., 2009).Thus, NANOS2, CYP26B1 and consequently retinoic acid are required in PGCs for quiescence, meiosis and survival, whereas DMRT1 is only required for quiescence and differentiation despite its role on retinoic acid pathway.Furthermore, Dmrt1 -/ -males are sterile and 90% of those develop TGCTs (Fig. 1) (Krentz et al., 2009).Interestingly, heterozygotes are fertile and the incidence of TGCTs is similar to that in 129 wild-type mice, suggesting that DMRT1 is haplo-insufficient for TGCTs.

A y and Eif2s2
The Agouti-yellow (A y ), which is a ~170 kb deletion on chromosome 2 in mice includes the entire coding region of both Eif2s2 and Raly genes as well as a part of the Agouti gene (Michaud et al., 1994).Eif2s2 gene encodes the beta subunit of translation initiation factor eIF2 (Sarre, 1989).Raly encodes an RNA-binding protein that acts in pre-mRNA processing.Agouti encodes a signaling protein involved in the pigment synthesis in melanocytes (Michaud et al., 1994).Interestingly, the A y deletion places the coding region of Agouti under the control of the Raly promoter, resulting in ectopic expression of Agouti (Duhl et al., 1994).As a result, Agouti is expressed in both PGCs and Sertoli cells, while RALY and eIF2s2 are lost in embryonic testes of A y mice (Heaney et al., 2009).Homozygosity for A y results in a pre-implantation lethality, whereas heterozygous A y mice develop obesity, diabetes, yellow coat color traits, and have an increase of the propensity to develop a variety of spontaneous tumors (Duhl et al., 1994).Surprisingly, the A y allele is the only genetic modifier known to suppress TGCT susceptibility in 129 mice and this phenotype results from loss of Eif2s2 (Lam et al., 2004;Heaney et al., 2009).Indeed, loss of one Eif2s2 allele decreases at least 2fold (less for A y allele) the high TGCT susceptibility observed in the 129-M19/M19 males (Fig. 1) (Lam et al., 2004;Heaney et al., 2009).This protective effect is due to a rescue of the G1/S mitotic arrest from E16.5 onward, but without effect on apoptosis, suggesting that Eif2s2 deficiency affects only the mitotic activity but not the survival of PGCs (Heaney et al., 2009).In addition, spermatogenesis is normal in the mutant adult testis although the weight of testes is significantly reduced compared with their control 129-M19/M19 males, suggesting that reduced Eif2s2 impedes but does not repress adult germ cell maturation (Heaney et al., 2009).
By using genetic targeting in mice, several genes have been identified that play a crucial role in PGC development and transformation into TGCTs.These genes act in distinct pathways and understanding their interrelation is a challenge for future research.

From mouse models to humans: Molecular basis of TGCT development
In this section, we compare results for genetic studies of TGCT susceptibility in humans and in mice (Table 1), with an emphasis on gene functions and protein pathways that control development of the PGC stem cell lineage and that modulate susceptibility to TGCTs.

The KIT/KITLG pathway
The KIT/KITLG pathway controls migration, proliferation, survival and differentiation of PGCs during embryogenesis and spermatogenesis.Although both Kit and Kitlg mutations affect development of several stem cell lineages including PGCs, only specific mutations in the Kitlg gene affect TGCT risk in mice (Heaney et al., 2008).In humans, four different GWAS identified allelic variations at the KITLG locus in individuals with TGCTs (Rapley et al., 2009;Turnbull et al., 2010;Kanetsky et al., 2009Kanetsky et al., , 2011)).Somatic mutations of KIT are also reported in men with TGCTs (Looijenga et al., 2003).In addition, deregulated expression of KITLG and KIT was found in TGCT biopsies (Murty et al., 1992).Thus, the KIT/KITLG pathway appears to be crucial for TGCT development both in humans and mice.
Similar evidence supports the hypothesis that the PI3K/AKT and MAPK signaling cascades modulate PGC transformation into TGCTs in humans and in mice:


The SPROUTY 4 (SPRY4) gene encodes an inhibitor of MAPK signaling by inhibiting RAS activation (Leeksma et al., 2002).SPRY4 is associated with TGCT susceptibility in humans (Rapley et al., 2009;Turnbull et al., 2010;Kanetsky et al., 2009Kanetsky et al., , 2011)). TRP53 deficiency is a potent but unusual modifier of TGCT susceptibility in both humans and mice.Although common in many cancers in humans (K.Suzuki & Matsubara, 2011), somatic TRP53 mutations are exceptionally rare in TGCTs in humans (Murty et al., 1994;Peng et al., 1995); somatic Trp53 mutations do not appear to have been surveyed in mice.By contrast, germline TRP53 mutations are the molecular basis for Li-Fraumeni syndrome, which increases susceptibility to various cancers including TGCTs (Malkin et al., 1990), and germline Trp53 mutations also increase susceptibility to many cancers including TGCTs in mice (Rotter et al., 1993;Donehower et al., 1995).Together these observations suggest that TRP53 mutations act in the soma, rather than in the germline, to promote transformation of PGCs.
Various elements of TRP53-mediated apoptosis have been implicated in TGCT development.Indeed, variation within the BAK1 gene is associated with TGCT cases in humans (Rapley et al., 2009;Turnbull et al., 2010).In addition, expression of Cox15, which encodes the mitochondrial cytochrome C oxidase assembly protein that is essential for the cell death program, is altered in testes of 129-M19 mice, which have a dramatically elevated TGCT risk (R. Zhu et al., 2007).
Interestingly, double homozygosity for the KITLG and DMRT1 risk haplotypes increases risk 14-fold in humans (Kanetsky et al., 2011), suggesting that these haplotypes affect TGCTs in a non-additive manner.These interactions could arise either through pathways downstream of KIT that regulate DMRT1 activity, or through membrane-bound KITLG activating an intrinsic pathway in Sertoli cells that modulate the DMRT1 or the retinoic acid pathways.Pre-Sertoli cells are depleted in Kitlg Sl mutants (Tajima et al., 1991) and their development is altered in Dmrt1 -/ -mice (Raymond et al., 2000;Krentz et al., 2009), suggesting that DMRT1 and KITLG are essential to Sertoli cell development and that pre-Sertoli cells might be involved in tumorigenesis, reinforcing the hypothesis of functional relations between KIT/KITLG and DMRT1.Finally, it is possible that at least one element of the KIT/KITLG pathway is a target gene of the transcriptional regulator DMRT1.
Table 1.Candidate modifiers of TGCTs in humans and their orthologues in mice. www.intechopen.com Other tyrosine kinase receptors that are expressed in PGCs may also contribute to TGCT development.A strong candidate is RET, which is activated by the growth factor GDNF receptor GFRA1.Gfra1 gene is located at 19D3 in mice, within a locus that was proposed as candidate TGCT modifier (Matin et al., 1999;Youngren et al., 2003).Furthermore, Ret is a known proto-oncogene (Grieco et al., 1990) and its expression is indirectly under the control of the TGCT modifier DMRT1 (Krentz et al., 2009).Allelic variations were identified at 18q21 in TGCTs in humans (Murty et al., 1994;Crockford et al., 2006).This locus contains the Docking protein 6 (DOK6) gene that plays a role in RET signaling cascade (Crowder et al., 2004).Interestingly, Dok6 gene is located on chromosome 18 in mice, reinforcing the hypothesis of the role of other tyrosine kinase receptors such as GDNF receptors on TGCT development (Anderson et al., 2009a).

Telomerase and TGCTs
By extending the TTAGGG telomeric nucleotide repeats, telomerase counterbalances loss of telomere length that usually occurs during cell division, and thus preserves chromosomal integrity (Venteicher et al., 2009).The active telomerase complex involves TERT (telomerase inverse transcriptase), the RNA component TERC, the ribonucleoprotein dyskerin (encoded by the X-linked DCK1 gene) and several other cofactors (Venteicher et al., 2009).Telomerase is activated by the telomerase Cajal body protein-1 (TCAB1) which is encoded by the WRAP53 gene (Jady et al., 2004).
Mutations in TERT (Marrone et al., 2007), TERC (Vulliamy et al., 2001), DKC1 (Heiss et al., 1998) and more recently WRAP53 (Zhong et al., 2011), which all lead to absence or dysfunction of telomerase, are found in congenital dyskeratosis, a human genetic deficiency characterized by abnormal skin pigmentation, bone marrow failure, and an elevated tumor frequency.Telomere defects are also associated with segmental progeria syndrome, which is characterized by accelerated ageing and is associated with a severe deficiency of adult stem cells in brain, bone marrow and testis (Burtner & Kennedy, 2010).Furthermore, telomerase dysfunction enhances tumor incidence in mice (Blasco et al., 1997).Thus, telomerase seems to play an essential role in stem cell development and cancer formation.
Specific markers of TGCT susceptibility have been recently identified in humans at 12p13, which contains the gene encoding ATF7IP, an enhancer of TERT and TERC transcription, and at 5p15 within TERT (Turnbull et al., 2010).Furthermore, a marker of familial TGCT risk was located 50kb downstream of TERC (Crockford et al., 2006), and amplification of human Xq28 containing DCK1 was found in TGCTs (Skotheim et al., 2001).In mice, the primordial germ cell tumor 1 locus (pgct1), that contains Tert on chromosome 13, has been identified as an enhancer of TGCT susceptibility (Muller et al., 2000).Furthermore, loss of Tert in mice leads to PGC deficiency which increases accross generations, due to both reduced proliferation and increased apoptosis (Lee et al., 1998).This phenotype is more severe in males where PGC deficiency is complete at the sixth generation (Lee et al., 1998).Taking together, these observations highlight the involvement of telomerase and telomere biology in TGCT development.
Interestingly, pgct1 locus interacts with TRP53 to modulate TGCT susceptibility in mice (Muller et al., 2000), and Wrap53 is a natural antisense transcript of Trp53 and regulates the l e v e l s o f T R P 5 3 i n r e s p o n s e t o D N A d a m a ge (Mahmoudi et al., 2009).Furthermore, telomerase dysfunction activates TRP53-dependent apoptosis (Chin et al., 1999).By contrast, progeria-like syndromes have been associated with alterations in TRP53-dependent apoptosis (D. Liu et al., 2010).Given these results and that both congenital dyskeratosis and segmental progeria affect similar stem cell lineages as KIT/KITLG defect (melanogenesis, gametogenesis and hematopoiesis), we propose an interrelation between KIT/KITLG pathway, via TRP53, and telomerase, via TERT and WRAP53, during embryogenesis that contributes to tumorigenesis.

Sex chromosomes and TGCTs
Males with Klinefelter's syndrome (also known as XXY syndrome) have 50-fold greater TGCT risk (Gustavson et al., 1975).Secondly, gain of X chromosomes has been described in TGCTs in humans (Peltomaki et al., 1990;Skotheim et al., 2001).Third, chromosome X from the C57BL/6 strain reduces tumor incidence in 129-Ter/Ter mice (Hammond et al., 2007), suggesting that genes linked to chromosome X both in mice and humans modulate TGCT incidence.In addition to DCK1 at Xq28 in humans, other interesting genes are Sperm protein associated with the nucleus mapped to the X chromosome (SpanX) clusters at Xq27, a locus named Tgct1, which has been linked with bilateral TGCTs and undescended testicular syndrome in humans (Rapley et al., 2000;Kouprina et al., 2004;Crockford et al., 2006).Although the function of SPANX proteins is unknown, evidence is accumulating that suggests their involvement in tumorigenesis.For example, SPANX genes are deregulated in Down's syndrome subjects who have undescended testis and an increased risk of TGCTs (Satge et al., 1997), further suggesting interaction of genes located on chromosome X and chromosome 21 for TGCT susceptibility.SPANX genes have a testis-specific expression that is conserved in rodents and humans, and was also detected in EC clusters and TGCTs (Westbrook et al., 2004;Salemi et al., 2006).
Similarly, a complete loss of chromosome Y in humans (45 XO karyotype) reduces TGCT susceptibility (Soh et al., 1992).Furthermore, the rare gr/gr deletion of the Y chromosome that removes part of the AZFc region (Yq11) is found in men with infertility and lowpenetrance for TGCT susceptibility (Nathanson et al., 2005).This deletion affects DAZ (deleted in azoospermia), BPY2 (Basic protein, Y-linked) and CDY1 (chromodomain protein, Y-linked 1).DAZ encodes an RNA-binding protein that interacts with the P-body component PUM2 in PGCs (Moore et al., 2003).The functions of CDY1 and BPY2 remain unknown.Moreover, aberrant expression of testis-specific protein on Y (TSPY) at Yp11 in humans may contribute to predisposition for TGCTs (Y.F.Lau, 1999), revealing a differential effect of the Y chromosome on TGCT risk.In mice, neither the MOLF-derived nor the C57BL/6-derived chromosome Y significantly affects susceptibility in 129 mice (Hammond et al., 2007).However, by using a sex-reversed mouse model, TGCTs were not found in the absence of the Y chromosome (Anderson et al., 2009b).Furthermore, a possible interaction, which suppress TGCT susceptibility in mice, was described between the Y-linked genes, which have a low-penetrance for TGCTs, and either the Dnd1 gene (chromosome 18) or the genes located on chromosome 19 (perhaps Dmrt1 given its role in sex differentiation and in TGCT susceptibility) (Anderson et al., 2009b).This hypothesis is supported by the fact that Dnd1 has sex-dependent effects on PGC survival and tumor susceptibility.Indeed, Dnd1 is differentially expressed in XX (down-regulation) and XY (up-regulation) gonads (Youngren et al., 2005).Loss of Dnd1 leads to PGC deficiency that increases with age in males, but not www.intechopen.com in females (T.Noguchi & M. Noguchi, 1985).Furthermore, the few mutant germ cells that successfully migrate to the gonad give rise to mature oocytes in females, while in testes, they give rise to TGCTs in 95% of cases (Cook et al., 2009).
Thus, sex chromosomes seem to play a crucial role in TGCT susceptibility.Identification of the candidate genes and of their interrelation remains to be elucidated.

RNA biology, epigenetic regulation and TGCTs
The translational complex eIF2 is composed of the alpha (eIF2s1), beta (eIF2s2) and gamma (eIF2s3) subunits (Sarre, 1989).Eif2s2 is a TGCT modifier in mice (Heaney et al., 2009), but remains to be confirmed in humans.Eif2s3 is located on both Y (Eif2s3y) and X (Eif2s3x) chromosomes in mice (in humans, no Y homologues have been found) (Ehrmann et al., 1998).Given the role of sex chromosomes discussed above, Eif2s3 should be considered as a candidate for TGCT susceptibility.EIF4E is another translational factor which acts downstream of PI3K/AKT pathway.Interestingly, eIF4E is a potent oncogene (Wendel et al., 2004), regulated by the poly(A)-binding protein and its cofactor PAIP2 (Yanagiya et al., 2010).Paip2 is located on mouse chromosome 18 near the Ter locus (Asada et al., 1994).Finally, Eif3a encodes another translational factor located on chromosome 19 in mice within a region involved in TGCT susceptibility (Matin et al., 1999;Youngren et al., 2003), but its direct role in TGCTs remains to be tested in mice and humans.
Translation is under the control of P-bodies, which regulate many mRNA processes (i.e.post-transcriptional regulation, degradation, storage, transport, and stabilization) in cells under normal physiological conditions or in response to stress (Seydoux & Braun, 2006;Parker & Sheth, 2007).P-bodies are essential for male PGC development and probably also for tumorigenesis given the involvement of its major components (Hayashi et al., 2008; K.M. Nelson & Weiss, 2008).DDX3, which encodes a helicase of P-body, is located on both Y (DDX3y) and X (DDX3x) chromosomes in mice and humans, and has been identified as a tumor suppressor (Chao et al., 2006).Given the role of sex chromosomes and the close location of Ddx3y and Eif2s3y in mice, DDX3 should also be considered as a candidate for TGCT susceptibility.Interestingly, another DDX gene (DDX23) is located at 12q13 in humans, a locus which has been associated with TGCTs (Crockford et al., 2006).In addition, PGCs lacking DICER1, another helicase, are depleted around E13.5 in mice (Hayashi et al., 2008), suggesting an essential role of Dicer1 in PGC development.Dicer1 has been also identified as a potential tumor suppressor in mice and probably also in humans (Su et al., 2010).Nanos1, which encodes an RNA binding protein, is located on chromosome 19 in mice, within a region involved in TGCT susceptibility (Matin et al., 1999;Youngren et al., 2003).Finally, allelic variations within the X-fragile mental retardation (FMR) genes (FMR1, FMR1nb and FMR2) at Xq27 in humans have been identified in TGCTs (Crockford et al., 2006).These genes encode RNA binding proteins that are involved in translation regulation through an interaction with the DICER and Argonaute proteins (Jin et al., 2004).The number of CGG trinucleotide repeats in the 5'-untranslated region of FMR1 and their DNA methylation pattern determine the severity of FMR-related pathogenicity (Fu et al., 1991;McConkie-Rosell et al., 1993).Furthermore, loss of Fmr1 alters both proliferation and differentiation of several stem cell lineages including PGCs (Castren et al., 2005), suggesting a role of FMR1 in TGCT susceptibility.
Additional evidence shows an association between DNA methylation and TGCT susceptibility.First, DNA methylation in male PGCs is largely erased at E10.5-E11.5 and partially reestablished at E13-E14 (Sasaki & Matsui, 2008).This epigenetic reprogramming completely overlaps the critical period for TGCT formation (E11.5-E13.5)(Stevens & Little, 1954;Stevens, 1967).Secondly, the DNA methyltransferase DNMT3L has been recently identified as a novel marker of TGCTs in humans (Minami et al., 2010), in accordance with the fact that TGCTs have a distinct epigenetic profile from other cancers (Ushida et al., 2011).Third, other genetic markers for TGCTs have been identified: (1) in humans, at 2p23 which contains DNMT3A, and at 12q13 which contains gene encoding the histone deacetylase HDAC7 (Crockford et al., 2006); (2) in mice, on a region of chromosome 18 that contains genes encoding two methyl-binding proteins MBD1 and MBD2 (Anderson et al., 2009a), and on chromosome 19 at a region that contains the remodeling chromatin factor Smarca2 (Matin et al., 1999;Youngren et al., 2003).Other studies correlate up-regulation of DNMT3A with demethylation of specific loci in TGCTs in humans (Ishii et al., 2007), and inhibition of both DNMTs and HDACs with prevention of cancer formation (W.G.Zhu & Otterson, 2003).
Another TGCT marker has been identified at 4q24 in humans near the CENP-E gene that encodes a centromeric-associated protein required for establishing and maintaining of the mitotic checkpoint (Abrieu et al., 2000;Rapley et al., 2009).CENP-E is described in mice as tumor-suppressing or -promoting factor depending on the context (Weaver et al., 2007).Interestingly, TRP53 has been associated with DNA demethylation (Ashur-Fabian et al., 2010) and several TRP53-target genes show aberrant methylation pattern in TGCTs in humans (Christoph et al., 2007).

Transgenerational genetic effects on TGCT susceptibility
Growing evidence suggests an alternative molecular basis of inheritance that complements conventional Mendelian inheritance with a similar strength, frequency and persistence across multiple generations (V.R. Nelson & Nadeau, 2010).These alternatives could involve four different forms: (1) environmental factors that create an epigenetic state that persists across generations (transgenerational environmental effects); (2) ancestral genetic factors that are sufficient to initiate epigenetic inheritance (transgenerational genetic effects); (3) environmental factors that have an epigenetic effect only in genetically predisposed individuals (transgenerational gene-environmental interactions); and (4) genetic variants in parents and others in offspring that interact to create an epigenetic state in offspring (transgenerational gene-gene interactions) (V.R. Nelson & Nadeau, 2010).The evidence for transgenerational genetic effects, as well as the implication inheritance of TGCT susceptibility will be discussed in this section.

Environmental effects
Pesticides and insecticides were intensively used in agricultural industries during previous decades.These chemicals, which are now found in foods and water, contain endocrine disruptors that act primarily through nuclear hormone receptors such as estrogen and androgen receptors (Danchin et al., 2011).In utero, perinatal or neonatal exposures affect male reproduction leading to infertility with an increase of testicular abnormalities and germ cell cancers in humans (A.Giwercman & Y.L. Giwercman, 2011).These effects can be reversible, permanent or even transgenerational, and involve alterations of DNA methylation (Anway et al., 2005;Danchin et al., 2011;Kalfa et al., 2011).
Other natural nutrients could also modulate predisposition to disease such as TGCTs.As discussed above, vitamin A plays an essential role in PGC development and might also influence TGCT susceptibility.Similarly, folate which is a methyl-donor nutrient affects DNA methylation and is essential for PGC development (Danchin et al., 2011).Determining the effects of folate and vitamins on TGCT susceptibility, and whether supply of these nutrients can reverse TGCT development is a challenge for future research both in rodents and humans.A related challenge is understanding relation between genetic modifiers and environmental factors on TGCT susceptibility.

Parent-of-origin effects
While environmental factors could affect reproductive health across generations in a parentof-origin manner (Anway et al., 2005;Kalfa et al., 2011), two examples reveal similar parental-dependent transgenerational effect of TGCT modifier genes in mice.
Loss of KITLG (Kitlg Sl , Kitlg SlJ and Kitlg Slgb ) in mice causes 2-fold increase of TGCT susceptibility in the 129 strain (Stevens, 1967;Heaney et al., 2008).Surprisingly, among the progeny of the Kitlg Slgb / + males, wild-type sons never develop TGCTs (5% expected), whereas wild-type sons of the reciprocal crosses (maternal Kitlg Slgb allele) are affected at the expected rate (Fig. 2) (Heaney et al., 2008).Thus, absence of one Kitlg allele in the male germline leads to an epigenetic change that affects TGCT susceptibility in the predisposed 129 strain.Whether this protective effect persists across generations is currently being tested (E.Leung & J.H. Nadeau).Determining the molecular basis of this epigenetic modification, and whether other Kitlg mutant males (i.e.Kitlg Sl and Kitlg SlJ ) have similar effects remain to be determined.

Gene-gene interactions
The intercross of Trp53 -/ + and Kit Wv -/ + m i c e y i e l d s { Kit Wv -/ + ; Trp53 -/ + } double heterozygotes which are intercrossed to generate double homozygotes (Jordan et al., 1999).Surprisingly, {Kit Wv -/ -; Trp53 -/ -} males are fertile due to a rescue of PGCs at E13.5 (Jordan et al., 1999), confirming that the high rate of apoptosis observed in Kit Wv -/ -embryonic gonads depends on KIT/KITLG-TRP53 interactions.Loss of both TRP53 alleles is necessary to rescue fertility because both {Kit Wv -/ -; Trp53 + / + } and {Kit Wv -/ -; Trp53 -/ + } males are sterile (Jordan et al., 1999).Interestingly, only 10% of normal {Kit Wv -/ -; Trp53 -/ -} germ cells are present in the adult testes, compared to 34% after birth.This deficiency in the adult testes correlates with presence of testicular abnormalities (germ cell radial migratory defect, meiotic failure and apoptosis).As discussed above, these testicular abnormalities could be the origin of TGCTs.Unfortunately, this study was conducted on a mixed background and the TGCT susceptibility of these offspring remains to be investigated in the 129 strain.However, intercrosses of Trp53 -/ + and Kitlg SlJ -/ + mice give the double heterozygous {Kitlg SlJ -/ + ; Trp53 -/ + } males that have a surprising 4-fold reduced TGCT frequency (7%; 26% expected) (Lam et al., 2004), suggesting that Kitlg and Trp53 genes can interact to counterbalance their effects on TGCT susceptibility.The effect of this gene interaction on PGC development remains to be investigated.
The intercross of C57BL/6-Bax -/ -and 129-Ter/Ter mice yields double homozygotes {Ter/Ter; Bax -/ -} that show at least 50% of PGCs rescued at E13.5 (Cook et al., 2009), demonstrating that the PGC deficiency in 129-Ter/Ter males is due to BAX-mediated apoptosis.However, rescued PGCs are completely lost in adult, suggesting that mechanisms affecting PGC development after E13 are BAX-independent in 129-Ter/Ter mice.This hypothesis is supported by the observation that loss of BAX does not affect the TGCT frequency, which remains elevated (91%; 94% expected) (Cook et al., 2009).Surprisingly, the {Ter/Ter; Bax -/ + } males have a 2-fold decrease of tumor risk and {Ter/Ter; Bax + / + } males do not develop TGCTs (Cook et al., 2009).This protective phenotype is interesting but we cannot determine whether it results from a genetic background effect (C57BL/6 vs 129) or from a genetic interaction (Bax vs Dnd1).Backcrosses of the double homozygotes {Ter/Ter; Bax -/ -} to pure C57BL/6 background mice suppress the TGCT susceptibility (Cook et al., 2011), favoring a background effect on TGCT susceptibility.
The double {Ter/Ter; Sf1 -/ + } mutants in the 129 strain have a 2-fold reduction in risk due to a reduced frequency of bilateral tumors (R. Zhu et al., 2010).Despite this protective effect on TGCT formation, all mutant males are sterile due to a complete germ cell deficiency (R. Zhu et al., 2010).This phenotype is similar as those observed in the {Ter/Ter; Bax -/ + } males suggesting a common pathway involving both SF1 and BAX on PGC deficiency in 129-Ter/Ter mice.
The interaction of Ter allele with the four risk alleles Kitlg SlJ , mutated Trp53, M19 and A y in the respective double heterozygous males increased TGCT susceptibility at least 2-fold in the 129 strain by enhancing the frequency of bilateral tumors (Lam et al., 2007).Surprisingly, all {Ter/ + ; + / + } males of these four intercrosses also have a 2-fold increase in TGCT frequency, suggesting transgenerational epistasis that acts only in the presence of the Ter allele in the offspring generation.The mechanism underlying interactions with the Ter allele remains elusive, although the new role of DND1 in the micro-RNA biology (Kedde et al., 2007) may be one ways by which this transgenerational effect acts.

Conclusions
Mouse models of TGCTs have made major contributions to stem cell biology, developmental biology of the PGC lineage, and genetic and epigenetic studies of TGCT susceptibility.
The germ cell lineage, which has been termed the 'mother of all stem cells', carries DNA and other molecular features that together constitute our genetic and epigenetic heritage (Donovan, 1998).Controlling differentiation and proliferation of PGCs, which belong to the only lineage of cells that naturally show totipotency is therefore essential for maintaining their integrity.Many factors have been shown to affect molecular mechanisms regulating pluripotency versus differentiation, proliferation versus death, and migration versus stasis during normal PGC development (Pesce et al., 1993;Lin et al., 2005;Shen et al., 2007;Heaney et al., 2009;Murphy et al., 2010;Cook et al., 2011).In parallel, insights are beginning to emerge about the ways in which anomalies in these factors and processes lead to transformation (Rotter et al., 1993;Kimura et al., 2003;Youngren et al., 2005;Heaney et al., 2008;Krentz et al., 2009).Despite their fundamental relevance to understanding important aspects of human biology, these studies are obviously difficult to conduct in humans.Thus, studies of PGCs and their transformed derivatives in mice will likely remain relevant to our understanding of the genetics and developmental origins of TGCTs in humans.
Despite being a common cancer in young men (Buetow, 1995), with heritable TGCT risk among the highest of all cancers (Gilbert et al., 2011), the genetic control of inherited susceptibility has proven to be remarkably elusive.However, with the availability of complete panels of genetic markers, high-throughput assays, and increasingly rigorous analytical methods, both linkage analysis and GWAS have begun to yield results.The first breakthrough involved the gr/gr deletion on the X chromosome (Nathanson et al., 2005), followed more recently with GWAS evidence for KITLG, SPRY4, DMRT1, TERT, BAK1 and ATF7IP as strong candidate susceptibility genes (Rapley et al., 2009;Turnbull et al., 2010;Kanetsky et al., 2009Kanetsky et al., , 2011)).The recent evidence for involvement of KITLG in humans was anticipated by corresponding evidence in mice involving mutations in the Kitlg gene (Heaney et al., 2008).The fact that some but not all Kitlg mutants affect susceptibility implies that allele-specific tests and structure-function studies will both be important to understand the ways in which KITLG variants affect TGCT susceptibility in humans.
More recently, heritable epigenetic changes have been shown to strongly influence TGCT risk in mouse models.In particular, TGCT modifier genes in the parental generation were shown to interact with the Dnd1 modifier to increase both the number of affected males and the proportion of bilateral cases (Lam et al., 2007).Similar evidence for transgenerational genetic effects has also been reported for Kitlg mutants (Heaney et al., 2008).These and related discoveries suggest that heritable epigenetic changes might be as important as conventional genetic effects in controlling inherited TGCT risk, and might account for the substantial difference in risk between sons and brothers of cases (Hutter et al., 1967).
Identifying the nature of these epigenetic factors, characterizing their molecular mechanisms, and testing their contribution to TGCT susceptibility in humans and in mice remain major challenges.

Fig. 1 .
Fig. 1.Male germ cell development in mouse models for TGCTs in the 129 strain.

Fig. 2 .
Fig. 2. Example of parental transgenerational effects in a mouse model for TGCTs.Similarly, wild-type sons of females with the A y allele, which acts as a TGCT suppressor due to the deletion of Eif2s2, have a 65% of risk being affected (80% expected)(Heaney et al., 2009; J.D. Heaney & J.H. Nadeau, unpublished), revealing another example of a transgenerational effect on TGCTs transmitted in this case through the maternal germ-lineage.