Regulation of fetal male germ cell development by members of the TGFβ superfamily

SCR-01018; No. of pages: 7; 4C: 2, 3 Stem Cell Research xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Stem Cell Research journal homepage: www.elsevier.com/locate/scr

Regulation of fetal male germ cell development by members of the TGFβ superfamily Cassy Spiller a,1, Guillaume Burnet a,1, Josephine Bowles a,b,⁎ a b

School of Biomedical Sciences, The University of Queensland, Brisbane, QLD 4072, Australia Institute for Molecular Bioscience, The University of Queensland, Brisbane, QLD 4072, Australia

a r t i c l e

i n f o

Article history: Received 13 December 2016 Received in revised form 4 May 2017 Accepted 15 July 2017 Available online xxxx Keywords: Germ cells TGFβ Activins Nodal Cripto

a b s t r a c t There is now substantial evidence that members of the transforming growth factor-β (TGFβ family) regulate germ cell development in the mouse fetal testis. Correct development of germ cells during fetal life is critical for establishment of effective spermatogenesis and for avoiding the formation of testicular germ cell cancer in later life. Here we consider the evidence for involvement of various TGFβ family members, attempt to reconcile discrepancies and clarify what we believe to be the likely in vivo roles of these factors. © 2017 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction During mammalian fetal development, germ cells migrate to the nascent gonads, colonise them and then respond to signals from somatic cells that make them commit to oogenesis in the fetal ovary or spermatogenesis in the fetal testis (McLaren, 2001) (Fig. 1). In the ovarian environment, retinoic acid (RA) acts directly on germ cells to trigger expression of the critical pre-meiotic gene Stra8 and thereby drive them to embark on meiosis (Bowles et al., 2006; Koubova et al., 2006; MacLean et al., 2007; Bowles et al., 2010, 2016). It is likely that another signaling pathway is also activated in ovarian fetal germ cells because oogenesis (but not meiosis) can proceed in the absence of Stra8 (Dokshin et al., 2013). In the testis, endogenous RA is degraded by the P450 enzyme CYP26B1 and, therefore, germ cells do not express Stra8 and therefore do not initiate meiosis (Bowles et al., 2006; Koubova et al., 2006; MacLean et al., 2007). Instead, they eventually stop proliferating (initiating G1/G0 mitotic arrest during the 12.5 to 14.5 dpc period) and begin to express characteristic molecular markers that evidence their commitment to the male program of germ cell differentiation, spermatogenesis (Adams & McLaren, 2002; Western et al., 2008; Suzuki & Saga, 2008; La Salle et al., 2004).

⁎ Corresponding author at: School of Biomedical Sciences, The University of Queensland, Brisbane, QLD 4072, Australia. E-mail address: [email protected] (J. Bowles). 1 These authors contributed equally to this work.

One signaling molecule produced by the somatic cells of the developing testis is FGF9. This factor has an important role in development of the somatic tissue of the testis because when it is genetically deleted testicular development is compromised, even to the extent of frank male-to-female sex reversal (Colvin et al., 2001; Schmahl et al., 2004). Besides this essential role in somatic testis development, FGF9 also acts directly on testicular germ cells to maintain expression of pluripotency markers, to make germ cells less prone to succumb to meiosis and, eventually, to support the expression of male fate markers Nanos2 and Dnmt3L (Bowles et al., 2010; Barrios et al., 2010; Tian-Zhong et al., 2016). Testicular germ cells express various FGF receptors (Bowles et al., 2010), however direct downstream targets of FGF signaling in these cell types remain unknown. Following the initial and as-yet undefined actions of FGF9, there is evidence that TGFβ signals are the major drivers of the commitment to the male germ cell fate, including initiation of cell cycle arrest as well as the expression of key fate markers. The transforming growth factor beta (TGFβ) superfamily is composed of over 40 cytokines in mammals, including TGFβ isoforms-1,-2 and -3, growth differentiation factors (GDFs), Nodal, Activins and Inhibins, which are considered to make up one major branch of the family and, in the other branch, the bone morphogenic proteins (BMPs). This major subdivision into two families is based on the downstream pathways activated (see below) [reviewed by (Miyazawa et al., 2002)]. Members of this superfamily are involved in a range of cellular processes including control of cell proliferation, differentiation, apoptosis, migration and cell fate specification [reviewed by (Kitisin et al., 2007)]. Here

http://dx.doi.org/10.1016/j.scr.2017.07.016 1873-5061/© 2017 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

2

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx

2. TGFβ signaling

Fig. 1. Regulation of germ cell development during fetal life, in the mouse model. Primordial germ cells begin to colonise the sexually-indifferent gonad at about 10.5 dpc. In the developing ovary, retinoic acid (RA) present in the somatic environment acts directly on germ cells to stimulate expression of the critical pre-meiotic factor Stra8. Germ cells subsequently enter meiosis and express Sycp3, Dmc1, Rec8 and γH2AX. In the developing testis endogenous RA is degraded by Cyp26b1, a P450 enzyme expressed by somatic cells (not shown) and, for this reason, testicular germ cells do not express Stra8 and do not enter meiosis. Sertoli cells produce FGF9 which acts directly on germ cells to stimulate expression of Cripto, the obligate co-receptor for Nodal. The presence of Cripto on the germ cell surface allows auto-upregulation of Nodal which feeds back on germ cells to maintain expression of pluripotency markers, particularly Nanog. When FGF9 is present at low concentrations (from 12.5 dpc) it acts directly on germ cells to induce expression of male fate markers Nanos2, Dnmt3l and P15. Activins and/or TGFβs may act directly on germ cells to stimulate male fate marker expression, to induce mitotic arrest and to maintain the quiescent state.

we focus on members of the TGFβ arm and particularly Nodal, Activins and TGFβ, all of which have been implicated in fetal testicular germ cell development.

TGFβ ligands act as homo- or heterodimers and signal through cell surface serine/threonine kinase receptor complexes composed of two types of receptors [reviewed by (Shi & Massague, 2003)] (Fig. 2). When ligand binds to the extracellular domain of Type II receptors (TGFβR2 for TGFβ, ACVR2A/ACVR2B for Activin and Nodal) the Type II receptor transactivates a Type I receptor (Activin receptor-like kinases (ALKs): ALK5 for TGFβs (also known as TGFβRI) and ALK2, -4 and -7 for Activin and Nodal) by phosphorylation. BMP ligands use a different set of Type I and Type II receptor molecules including ALK2, -3, -6 and BMP receptor type II (BMPR2) in addition to ACVR2A/ACVR2B. Activated receptor complexes in turn phosphorylate intracellular effector proteins of the mothers against decapentaplegic (SMAD) family which subsequently interact with the common mediator SMAD4; the complex then translocates to the nucleus where it acts, together with co-activators and co-repressors, to directly regulate gene transcription. SMAD2 and SMAD3 are the effectors of TGFβ, Activin and Nodal signaling whilst SMAD1, -5 and -8 respond to BMP-activated receptors. Considerable overlap exists in the receptors used by the various ligands as well as in SMAD proteins employed and, therefore, extensive redundancy is to be expected. Although Nodal and Activin share the same receptors, Nodal requires the additional presence of an obligate co-receptor Cripto (also known as TDGF1, a member of the epidermal growth factor-Cripto-FRL1-Cryptic (EGF-CFC) family) to successfully signal through the complex (Shen & Schier, 2000). Cripto is a small GPI-anchored protein that binds, through its CFC domain, to Type I receptor, ALK4, and through its EGF-like domain to Nodal, the overall effect being to greatly potentiate Nodal signaling. Although the presence of Cripto is crucial for Nodal signaling there is evidence that it attenuates Activin signaling: it is reported that when Cripto forms a complex with Type I and Type II receptors it diminishes the strength of Activin signaling by about 50% (Kelber et al., 2008; Gray & Vale, 2012). Hence in the absence of Cripto, Nodal cannot signal at all but Activin signals very strongly and, therefore, genes responsive to high levels of SMAD2/3

Fig. 2. TGFβ signaling pathway components. Each TGFβ morphogen (Activin, Nodal, TGFβ and BMP) signal through different combinations of Type I and Type II receptor kinases. Type I receptors include Alk2, -3, -4, -5, -6 and -7. Type II receptors include Acvr2a, Acvr2b, Tgfβr2 and Bmpr2. Activated receptor complexes phosphorylate specific intracellular SMAD effector proteins (SMAD1, -2, -3, -5 and -8) which subsequently interact with the common mediator SMAD4. The SMAD complex translocates to the nucleus where it acts, together with coactivators and co-repressors, to directly regulate gene transcription. Examples of several gene targets for each pathway are depicted, however this list is by no means exhaustive and is cell-context specific.

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx

signaling are transcriptionally activated. On the other hand, in the presence of Cripto, Nodal can signal but Activin signaling is attenuated making it comparable in strength to Nodal signaling and, as a result, only genes that are responsive to low levels of SMAD2/3 signaling are activated (Kelber et al., 2008; Gray & Vale, 2012). Cripto can also attenuate the effects of TGFβ signaling by binding TGFβ1 and preventing it from interacting with its Type I receptor, ALK5 (Gray et al., 2006). 3. Expression of TGFβ pathway components during testicular germ cell development In order to understand how TGFβ signaling might regulate testicular germ cell development it is critical to consider where and when the various ligands, receptors and co-receptors are expressed. This review focuses on the 12.5 to 15.5 dpc developmental window during which XY germ cells transition from pluripotency to differentiation, and the likely scenario of TGFβ and TGFβ receptor expression and function during this time is summarized graphically in Fig. 3. Although Nodal has been very well-studied, due to its involvement in numerous critical processes during embryogenesis, ES cells and in tumorigenesis, it is only recently that its expression during early development of testicular germ cells has been noted. In retrospect, however, it is perhaps not surprising that Nodal would be active in this system given the similarity in both molecular makeup and totipotent function that germ cells, ES cells and tumorigenic cells share. Very shortly after germ cells colonise the gonad, XY germ cells (but not XX germ cells nor any somatic cell types) begin to express the Nodal co-receptor Cripto (Spiller et al., 2012; Souquet et al., 2012) apparently in response to the signaling molecule FGF9 produced by nascent Sertoli cells (Spiller et al., 2012). Subsequently Nodal expression increases, presumably as a result of auto-upregulation from a baseline level of expression that is potentiated once Cripto is present on the germ cell surface (Spiller et al., 2012). As well as regulating its own expression Nodal induces expression of two Nodal antagonists Lefty1 and Lefty2. Because these factors diffuse more efficiently than Nodal they act to restrict the range of

3

Nodal signaling (Sakuma et al., 2002). Although there is some uncertainty regarding time of onset of Cripto and Nodal expression (Spiller et al., 2012; Souquet et al., 2012) it is clear that both are male-specifically expressed substantially earlier than the male germ cell fate marker gene Nanos2 (Suzuki & Saga, 2008). The expression of Cripto, Nodal, Lefty1 and Lefty2 is transient in the XY germ cells (Spiller et al., 2012), suggesting that XY germ cells are probably exposed to only a brief pulse of Nodal signaling (Spiller et al., 2012; Miles et al., 2013). Our knowledge of the cell type-specific expression of Activin and TGFβs in the developing mouse testis is less clearcut. In contrast to Nodal, germ cells do not express Inhba and Inhbb (genes that encode Activin subunits), at least prior to 15.5 dpc (Miles et al., 2013). These genes are, however, highly expressed by male somatic cells: Inhba is predominantly expressed by the steroidogenic Leydig cells and peritubular myoid cells and Inhbb by the Sertoli, and to a lesser extent, the Leydig cells (Archambeault & Yao, 2010; Yao et al., 2006; Jameson et al., 2012; Jeanes et al., 2005). At the transcript level, all three Tgfb isoforms are detectable in somatic cells of the fetal testis but only Tgfb1, encoding TGFβ1, is expressed by XY germ cells in the period 12.5 to 15.5 dpc with levels increasing over this period (Miles et al., 2013). We consider these results, based on qRT-PCR profiling of sorted germ and somatic cell populations, to be quite reliable and we have used them in formulating the model we offer in this review (Fig. 3). The qRT-PCR study results are, however, not entirely consistent with earlier studies including those based on immunohistochemical evaluations of these ligands. For example, in some studies Tgfb1 transcript and TGFβ1 protein were barely detectable prior to birth (Cupp et al., 1999; Moreno et al., 2010; Memon et al., 2008) and TGFβ3 was expressed by both Sertoli and germ cells from at least 13.5 dpc (Moreno et al., 2010). TGFβ2 was found in quiescent (16.5 dpc) but not proliferating (13.5 dpc) germ cells as well as in Leydig cells (Moreno et al., 2010). It is possible that the TGFβ antibody sensitivity or specificity is not optimal for the detection these ligands and/or that Tgfb1 transcript detected in XY germ cells (Western et al., 2008) is not translated.

Fig. 3. Proposed roles of Nodal, Activin and TGFβ in regulation of testicular germ cell development in the mouse model. At 12.5 dpc, Nodal/Activin receptors (green) are present on the surface of germ cells, along with the Nodal obligate co-receptor, Cripto (yellow). Nodal (blue) is produced by germ cells and Activin (green) by somatic cells. Both Nodal and Activin proteins are abundant but, because of the presence of Cripto, germ cell-specific signaling via Smad2/3 activation is likely to be weak (thin arrows, small P-SMAD2/3). This low level signaling is probably associated with the maintenance of pluripotency in these cells. TGFβ ligands are also present in the testicular milieu but their effect is likely to be limited to signaling to somatic cells because suitable receptors are not present on germ cells at early timepoints. At 13.5 dpc, levels of Nodal and Cripto begin to diminish and by 14.5 dpc little Cripto protein is available on the germ cell surface; consequently, Activins are free to signal strongly through the Nodal/Activin receptor complexes (wide arrows, large P-SMAD2/3). This signaling is likely to be associated with triggering mitotic arrest and, possibly, male fate marker expression. It is not until at least 15.5 dpc that germ cells become sensitive to direct TGFβ signaling. For simplicity, ligands are depicted as monomers rather than homo- or heterodimers and Activin and TGFβ forms are not specified.

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

4

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx

Irrespective of which cell type produces the TGFβ family ligands, these factors can only directly impact germ cell development if appropriate receptor complexes are expressed by the germ cells themselves. Germ cells and somatic cells in both testis and ovary express genes encoding receptors ALK4 and ACVR2A and -B (Spiller et al., 2012; Miles et al., 2013) and are therefore susceptible to signaling by Activin, however only XY germ cells express Cripto and are therefore potentially Nodal-responsive (Spiller et al., 2012). Analysis by qRT-PCR found only low levels of expression of Alk5 and no appreciable expression of Tgfbr2 in germ cells up until at least 15.5 dpc (Miles et al., 2013) whereas receptors for TGFβ ligands, ALK5 and TGFβRII, were both found to be expressed by germ cells from at least 13.5 dpc, by immunohistochemical evaluation (Moreno et al., 2010). Because of this discrepancy it remains uncertain how responsive germ cells are to signaling by TGFβ isoforms. Once again we have opted to rely on the qRT-PCR results generated by analysis of sorted cell populations (Miles et al., 2013) and, therefore, we propose that TGFβs are unlikely to act directly on testicular germ cells prior to at least 15.5 dpc (Fig. 3). Overall, knowledge of where and when ligands, receptors and co-receptors are expressed suggests that testicular germ cells are sensitive to both Nodal and Activin signaling at early fetal timepoints, although the strength of Activin signaling is likely to be moderated due to the presence of Cripto (Fig. 3). At later timepoints, once Cripto is no longer expressed, testicular germ cells might be subjected to much higher levels of pathway activation, elicited by Activin and, possibly, TGFβs (depending on the availability of TGFβ receptors). 4. Evidence for a role for Nodal in testicular germ cell development Given the important role Nodal signaling plays in various developmental contexts, as well as in ES cell biology and in cancer (Strizzi et al., 2005; Minchiotti, 2005), its sex-specific activation during a critical period in germ cell development is intriguing. The function of endogenous Nodal signaling during early XY germ cell development is still, however, unresolved. Because Cripto is only expressed by germ cells (Spiller et al., 2012; Souquet et al., 2012; Miles et al., 2013), and because Leftys act to limit the range of Nodal signaling (Sakuma et al., 2002) it seems unlikely that Nodal signals to testicular somatic cells. Nonetheless, signaling from germ cells to somatic cells by Nodal remains a theoretical possibility given that Nodal can act in a Cripto-independent manner in some circumstances (Yeo & Whitman, 2001) and that somatic cells do express Activin/Nodal receptors. Here we focus on evidence regarding the role of Nodal in directing testicular germ cell development. 4.1. Does Nodal prevent ectopic meiosis? Our first hypothesis, upon observing activated Nodal signaling in XY germ cells, was that this might be the mechanism by which FGF9 exerts an anti-meiosis effect on testicular germ cells (Bowles et al., 2010). However, we find no evidence that this is true. When we studied a Nodal hypomorphic mouse model, in which Nodal expression in XY germ cells is diminished to 30% of normal levels and there is a clear diminution of P-SMAD2 in the germ cells (approx. 50% of normal levels), there is no upregulation of Stra8, the pre-meiotic marker gene (Spiller et al., 2012). On the other hand, others have found that pharmacological inhibition of the TGFβ branch of signaling (blocking TGFβ, Activin and Nodal functions), using the ALK4, -5 and -7 inhibitor SB431542 (at 20 μM) in fetal testis culture, provokes meiosis marker expression (γH2AX and Sycp3) in ex vivo cultures (Souquet et al., 2012). A similar study (Wu et al., 2013) also found upregulation of meiosis markers (Stra8, Rec8 and Dmc1) upon inhibition of TGFβ signaling using SB431542 (at 40 μM) in fetal testis culture. There are several possible explanations for these apparently discordant results. Firstly, it is possible that TGFβ or Activin, but not Nodal, help ensure testicular germ cells do not enter meiosis – this would

explain why no ectopic meiosis was observed in the in vivo study (Spiller et al., 2012). Another possible explanation is that that the dose of SB431542 in the two ex vivo studies was such that off-target effects, such as the inhibition of additional signaling pathways, were induced: the recommended concentration for use of SB431542 as a specific inhibitor of ALK4, -5 and -7 is 10 μM (Inman et al., 2002). Even when used at 10 μM it has been shown that SB431542 induces highly disorganized testis cord structure and reduced Sertoli cell proliferation in ex vivo fetal testis culture (Miles et al., 2013). It remains possible, therefore, that the ectopic meiosis observed in ex vivo studies using SB431542 reflects abnormal somatic development or abnormal production of Sertoli cells products and that this, in turn, permits or initiates ectopic meiosis. Indeed, one of the ex vivo studies demonstrated that the ectopic Stra8 expression achieved after SB431542 treatment actually depends on signaling through RA receptors (Wu et al., 2013). Two additional results relevant to the question of whether Nodal signaling prevents germ cells entering meiosis have been reported. One of the ex vivo studies mentioned above treated XX gonads with recombinant Nodal in an effort to demonstrate its putative meiosis-inhibiting role. Although Stra8 expression was diminished to about 2/3 of its normal level, very high concentrations of Nodal were required to achieve this effect (5 μg/ml). This approach is probably not ideal, in any case, as XX germ cells do not express the Nodal co-receptor, Cripto, at appreciable levels (Spiller et al., 2012; Souquet et al., 2012). It is possible that Nodal, used at this concentration, can signal to germ and/or somatic cells in a non-canonical manner, diminishing the relevance of this experiment to the in vivo situation. The second experiment was conducted in vivo. When Smad4, encoding the common SMAD, SMAD4, was deleted specifically in germ cells (thereby theoretically preventing all TGFβ and BMP signaling) very little testicular germ cell meiosis was observed (Wu et al., 2013). Because this result was at odds with their in vitro inhibitor studies suggesting an anti-meiosis role for Nodal, the authors postulated that Nodal must act through a SMAD-independent pathway to antagonize meiosis onset (Wu et al., 2013), however such a mechanism has not been reported in the literature in any other system to date. As should be clear from the above discussions, we consider that the evidence that Nodal acts endogenously to ensure testicular germ cells avoid meiosis is inadequate. Resolution of this point awaits analysis of a more complete germ cell-specific deletion of Nodal and/or its obligate co-receptor, Cripto, in vivo. 4.2. Does Nodal maintain pluripotency? Using the same Nodal hypomorphic mouse model mentioned above we found that a reduction in Nodal signaling in XY fetal germ cells was associated with diminished levels of expression of the pluripotency-associated markers Oct4, Sox2 and, especially, Nanog (Spiller et al., 2012). In agreement with this, and working with ex vivo cultures, others found that pharmacological block of ALK4/5 and -7 (using SB431542 at the recommended dose of 10 μM) resulted in lowered Nanog expression (Miles et al., 2013). A third study documented an association between elevated Nodal signaling and the maintenance of pluripotency marker expression and also functional pluripotency (propensity to for embryonic germ cell colonies in vitro) (Tian-Zhong et al., 2016). These results all align with evidence that Nodal/Activin signaling maintains pluripotency in human embryonic stem (hES) cell cultures by directly controlling the expression of Nanog (Vallier et al., 2009; Brown et al., 2011). Why or even whether it is necessary for transient Nodal signaling to maintain pluripotency in XY germ cells during the 12.5 to 14.5 dpc window of time is as yet unknown. It is possible that XY germ cells remain relatively pluripotent and proliferative during this period so as to generate a sufficiently large spermatogonial stem cell population and, therefore, ensure high levels of fertility during adult life.

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx

4.3. Does Nodal induce mitotic arrest and the expression of male fate markers? Two studies have reported that ex vivo fetal testis culture in the presence of the TGFβ receptor antagonist SB431542 results in reduced expression of male germ cell fate markers Nanos2 and Dnmt3l, leading authors to suggest that Nodal signaling acts to promote the male-type differentiation of germ cells (Souquet et al., 2012; Wu et al., 2013). A third study that also used SB431542, albeit at a more modest dose, did not find any change in Dnmt3l expression although expression of another male fate marker, MILI, was apparently delayed or diminished as judged by immunofluoresent analysis (Miles et al., 2013). As these ex vivo experiments caused XY germ cells to express Stra8 and enter meiosis (for whatever reason, see above), it is possible that male fate does not progress in these cultures because germ cells have already committed to meiosis. One major shortcoming of this approach, in terms of ascribing a function to Nodal, is the fact that Activin and TGFβ signaling is also inhibited, in both germ cells and the soma, by SB431542. The influence of Nodal signaling on male fate marker expression in testicular germ cells was examined, more specifically, using a Nodal KO model (where Nodal was deleted ubiquitously after 10.5 dpc). This manipulation resulted in diminished levels of Nanos2 expression at 14.5 dpc but the effect was completely rescued by 16.5 dpc (Wu et al., 2013). Puzzlingly, however, there was no evidence that Nodal signaling was actually compromised in this model as both Lefty expression and PSMAD2 levels were reportedly unaffected. These results are at odds with our preliminary study of in vivo Nodal function, using the Nodal hypomorph, where we found that expression of markers of male germ cell differentiation, particularly Dnmt3l, was higher than normal at 14.5 dpc (Spiller et al., 2012). In interpreting this result we reasoned that germ cells might differentiate precociously simply because the length of the pluripotency phase is reduced in the absence of Nodal signaling, as discussed above. It is difficult to reconcile the contrasting results seen in these two in vivo models: it is possible that differing Nodal concentrations in each model (70% knock down (Spiller et al., 2012) versus almost 100% knock down (Wu et al., 2013)) result in different effects, as is the case in ES cells (Vallier et al., 2004). As far as whether Nodal has a direct effect on mitotic arrest, there is as yet little evidence to support or refute the hypothesis. In the Nodal hypomorph model we did observe that P15 expression, which is associated with the onset of mitotic arrest, seemed to be slightly precocious (Spiller et al., 2012). We also found that fewer germ cells were proliferating at 13.5 dpc in the Nodal hypomorphic testis, compared with wildtype. Rather than supporting the theory that Nodal induces mitotic arrest, however, these results indicate that Nodal effectively delays arrest, possibly by virtue of its role in maintaining germ cell pluripotency marker expression. 5. Evidence for a role for Activin and TGFβs in testicular germ cell development Both Activin and TGFβs are considered to be cytostatic, so it is a reasonable hypothesis that these ligands are involved in inducing germ cell mitotic arrest that is characteristic of testicular germ cell development. Despite numerous studies addressing this possibility it is still unclear what role, if any, these ligands play in the regulation of germ cell development during fetal life. It is known that Activin play critical roles in development of the testicular soma and, for this reason, it is difficult to determine whether it impacts directly on germ cell development. For example, Inhba, which encodes Activin A subunits and is expressed mainly by Leydig cells, is required for proper Sertoli cell proliferation and testis cord expansion from 15.5 dpc (Archambeault & Yao, 2010). Moreover, Inhbb null mice, which lack Activin B, do not form the characteristic testicular coelomic vessel and treatment of developing ovaries with Activin B results in ectopic vessel formation (Yao et al., 2006). Similarly, TGFβ ligands seem to impact on development of the testicular

5

soma although the exact roles are difficult to discern due, presumably, to redundancy (Memon et al., 2008). The strongest evidence that Activin regulates XY germ cell development comes from analysis of the Inhba-null mouse model (Mendis et al., 2011). In this line the number of testicular germ cells is abnormally high at 15.5 dpc and at birth, and they abnormally express the proliferation marker PCNA at 17.5 dpc, suggesting that germ cell arrest is compromised in the absence of Activin A. This study also showed that, in ex vivo culture, Activin A is able to repress germ cell proliferation (Mendis et al., 2011). However, despite these compelling findings it is difficult to be confident that Activin A acts directly on germ cells to induce mitotic arrest; as mentioned above Activin A is critical for correct Sertoli cell development and so an indirect mechanism could be involved. In fact, in the same mouse line it was reported that Sertoli cell proliferation was decreased from at least 13.5 dpc resulting in half the normal number of Sertoli cells being present in the testis at birth (Mendis et al., 2011). There is also some in vivo evidence that TGFβ isoforms are involved in triggering XY germ cell mitotic arrest. When Tgfb1 or Tgfb2 were deleted, contrasting phenotypes were observed at birth, the former resulting in diminished germ cell numbers and the later in increased germ cell numbers (Memon et al., 2008). Presumably this indicates a possible role for TGFβ2, at least, in triggering mitotic arrest. In another study deletion of the TGFβ-specific Type II receptor TGFβRII, specifically in germ cells, resulted in an increase in the number of cycling gonocytes thereby demonstrating a direct anti-proliferative effect of TGFβ on germ cells (Moreno et al., 2010). Although this seems like straightforward evidence that TGFβ directly triggers mitotic arrest other interpretations are also possible. Expression studies indicate TGFβRII is not expressed by germ cells in the critical 12.5 to 15.5 dpc period (Miles et al., 2013) making the hypothesis that TGFβ acts directly on germ cells problematical. The TGFβRII receptor is highly expressed by somatic cells however (Miles et al., 2013). The TNAP-Cre driver line, used for the ‘germ cellspecific’ deletion of the receptor, is known to be problematic in terms of cell-specific expression [reviewed by Hammond and Matin, 2009] and so it is possible that TGFβRII expression in somatic cells has been affected in this mouse model, and that TGFβ ligands can therefore no longer signal to somatic cells and, thus, that any effect on germ cells is likely to be indirect. The potential for Activin and/or TGFβs to initiate mitotic arrest critical for the correct development of testicular germ cells has also been studied in ex vivo culture systems. Treatment of gonads with SB431542 prevented their mitotic arrest, as judged by Ki67 and P27KIP1 expression (Miles et al., 2013). Tellingly, the same effect was not observed when an ALK5-specific inhibitor was used instead of SB431542: this result indicates that it is Activin/Nodal signaling rather than TGFβ signaling that is critical for germ cell mitotic arrest. Unfortunately, although this result is quite compelling, it is not possible to conclude that Activin/Nodal is acting directly on germ cells to bring about mitotic arrest, especially given that the development of the gonadal soma is abnormal under these conditions (Miles et al., 2013). Seemingly contrary to the conclusion reached in the above study, ex vivo culture of 13.5 dpc testis in TGFβ2 for 24 h substantially decreased the number of BrdU-positive gonocytes, suggesting that TGFβ2 might be capable of inducing quiescence in vivo (Moreno et al., 2010). Puzzlingly, another study, in which 13.5 dpc testes were treated with TGFβ2 for 48 h, found no effect on germ cell numbers (Sarraj et al., 2013). The problem with the theory that TGFβ2 induces mitotic arrest in germ cells of the fetal testis is that suitable receptors for TGFβ ligands are probably not expressed by germ cells at the appropriate time, in vivo. Interestingly, TGFβ2 protein is detectable in quiescent germ cells (Moreno et al., 2010) suggesting that, regardless of what triggers the arrest, a germ cell-specific autocrine mechanism might maintain it. Consistent with this idea TGFBR2 (also known as betaglycan), a nonsignaling accessory receptor that acts to increase the binding of TGFβs, especially TGFβ2, to type II receptors is not expressed in testicular

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

6

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx

germ cells until after birth (Sarraj et al., 2010). A function for TGFβ signaling in post-differentiation testicular germ cells would be consistent with results from the TGFβRII germ-cell specific deletion where it was found that adults have a reduced pool of spermatogonial stem cells and ultimately become sterile (Moreno et al., 2010). 6. Concluding remarks Germ cell development is largely determined by signaling from the somatic cell environment (McLaren, 2001) (Fig. 1). In this work we have reflected on the available data regarding likely role(s) of TGFβ family factors, Nodal, Activin and the TGFβ isoforms in direct regulation of germ cell development within the environment of the fetal testis. We identify four main complications that hinder our ability to definitively assign functions to TGFβ molecules in this system. Firstly, the different TGFβ ligands share receptors and intracellular signaling pathways and, therefore, the potential to act redundantly, complicating the interpretation of both in vitro and in vivo studies. Secondly, because TGFβ signaling is critical for normal development of the testicular soma it is challenging to differentiate direct effects on germ cells from indirect effects via somatic cells. Thirdly, it is difficult to interpret ex vivo studies in which the use of exogenous compounds could possibly cause off-target effects. Finally, even the results of in vivo studies in this field are subject to interpretation, largely because of the need to use imperfect conditional targeting approaches. Despite the difficulties mentioned above and numerous caveats inherent to the various studies, we are able to draw some tentative conclusions. Based on our analysis of available evidence we suggest that FGF9, produced by Sertoli cells of the nascent testis, induces Cripto expression on the surface of XY germ cells and that this allows the autoupregulation of Nodal (Fig. 3). Although testicular germ cells are presumably bathed in Activin produced by Sertoli, Leydig and, potentially, other somatic cells at this time, the presence of Cripto should theoretically keep any signaling they induce at modest levels. The effect of Nodal/Activin signaling at early timepoints (12.5 dpc) appears to be to maintain the expression of pluripotency genes, an effect potentially mediated by direct activation of Nanog expression. Such a function mirrors what is known in human ES cells (Vallier et al., 2009; Brown et al., 2011; Xu et al., 2008), human iPS cells (Takahashi et al., 2007) and mouse epiblast stem cells (Takahashi et al., 2007) - in vitro cell types with significant similarity to germ cells. Whether Nodal/Activin signaling is necessary to prevent XY germ cells from entering meiosis remains contentious. Considering that RA, the key Stra8 inducing factor, is not normally present in the fetal testis there does not appear to be a need for this function. Resolution of this point awaits analysis of appropriate in vivo mouse models. There is good evidence that Activin induces mitotic arrest and male germ cell fate, although it is not yet certain that this is a direct effect on germ cells. However, because Activin receptors are expressed by germ cells at the appropriate time, and because Activin is clearly being produced by somatic cells, it seems likely that this is the case. It is also possible that TGFβs induce these features of male germ cell differentiation, although receptor availability data suggests that the impact of TGF ligands on germ cells might be seen only later in fetal life, potentially involving the maintenance of germ cell quiescence. Germ cell development in the fetal testis is critical for successful spermatogenesis and, given that this is such an important and fundamental process, it makes sense that substantial redundancy would exist in the system. TGFβ signaling is incredibly complex (Hata & Chen, 2016) and, in the arena of mouse fetal germ cell development, we so far know relatively little. An additional complication, not discussed here, is the potential for BMP signaling to effectively antagonize TGFβ signaling by limiting availability of the common SMAD, SMAD4. Future studies on this topic will require greater sophistication and subtlety with attention, in particular, to whether effects are direct or indirect, whether the concentration of ligands and inhibitors used

in ex vivo studies are appropriate and whether conclusions drawn in in vivo experiments are sound. Timing and duration of exposure to the various ligands is also something to consider because, as we have learned from stem cell studies (Chen et al., 2013; Lee et al., 2011) the exact state of germ cells when treated with a particular ligand or inhibitor is likely critical to their response. In our system we hypothesise that low levels of P-SMAD2/3 activation (Fig. 3), expected when Nodal or Activin signal in the presence of Cripto, are associated with maintenance of pluripotency and higher levels, as might be expected in germ cells once Cripto is no longer present, are associated with cessation of proliferation and differentiation. Such a scenario would mirror the situation that is emerging in cancer biology where low levels of Nodal signaling are tumorigenic (pluripotent) and high sustained levels of Activin and TGFβ are cytostatic (Kelber et al., 2008; Gray & Vale, 2012; Kelber et al., 2009). An improved understanding of the subtleties of TGFβ signaling in the context of fetal germ cell development is critical to our ongoing efforts to understand the etiology of testicular germ cell cancer. It is hypothesized that this type of cancer is seeded during fetal life when the regulation of germ cell maturation, which includes the suppression of pluripotency as well as mitotic arrest and the commitment to spermatogenesis, goes awry (Skakkebaek et al., 1987; Rijlaarsdam & Looijenga, 2014).

Acknowledgements This work was funded by the Queensland Cancer Council (APP1080080) and the National Health and Medical Research Council (APP1109405) of Australia. We thank Humphrey Yao for helpful comments on the manuscript. References McLaren, A., 2001. Mammalian germ cells: birth, sex, and immortality. Cell Struct. Funct. 26, 119–122. Bowles, J., Knight, D., Smith, C., Wilhelm, D., Richman, J., Mamiya, S., Yashiro, K., Chawengsaksophak, K., Wilson, M.J., Rossant, J., Hamada, H., Koopman, P., 2006. Retinoid signaling determines germ cell fate in mice. Science 312, 596–600. Koubova, J., Menke, D.B., Zhou, Q., Capel, B., Griswold, M.D., Page, D.C., 2006. Retinoic acid regulates sex-specific timing of meiotic initiation in mice. Proc. Natl. Acad. Sci. U. S. A. 103, 2474–2479. MacLean, G., Li, H., Metzger, D., Chambon, P., Petkovich, M., 2007. Apoptotic extinction of germ cells in testes of Cyp26b1 knockout mice. Endocrinology 148, 4560–4567. Bowles, J., Feng, C.W., Spiller, C., Davidson, T.L., Jackson, A., Koopman, P., 2010. FGF9 suppresses meiosis and promotes male germ cell fate in mice. Dev. Cell 19, 440–449. Bowles, J., Feng, C.W., Miles, K., Ineson, J., Spiller, C., Koopman, P., 2016. ALDH1A1 provides a source of meiosis-inducing retinoic acid in mouse fetal ovaries. Nat. Commun. 7, 10845. Dokshin, G.A., Baltus, A.E., Eppig, J.J., Page, D.C., 2013. Oocyte differentiation is genetically dissociable from meiosis in mice. Nat. Genet. 45, 877–883. Adams, I.R., McLaren, A., 2002. Sexually dimorphic development of mouse primordial germ cells: switching from oogenesis to spermatogenesis. Development 129, 1155–1164. Western, P.S., Miles, D.C., van den Bergen, J.A., Burton, M., Sinclair, A.H., 2008. Dynamic regulation of mitotic arrest in fetal male germ cells. Stem Cells 26, 339–347. Suzuki, A., Saga, Y., 2008. Nanos2 suppresses meiosis and promotes male germ cell differentiation. Genes Dev. 22, 430–435. La Salle, S., Mertineit, C., Taketo, T., Moens, P.B., Bestor, T.H., Trasler, J.M., 2004. Windows for sex-specific methylation marked by DNA methyltransferase expression profiles in mouse germ cells. Dev. Biol. 268, 403–415. Colvin, J.S., Green, R.P., Schmahl, J., Capel, B., Ornitz, D.M., 2001. Male-to-female sex reversal in mice lacking fibroblast growth factor 9. Cell 104, 875–889. Schmahl, J., Kim, Y., Colvin, J.S., Ornitz, D.M., Capel, B., 2004. Fgf9 induces proliferation and nuclear localization of FGFR2 in Sertoli precursors during male sex determination. Development 131, 3627–3636. Barrios, F., Filipponi, D., Pellegrini, M., Paronetto, M.P., Di Siena, S., Geremia, R., Rossi, P., De Felici, M., Jannini, E.A., Dolci, S., 2010. Opposing effects of retinoic acid and FGF9 on Nanos2 expression and meiotic entry of mouse germ cells. J. Cell Sci. 123, 871–880. Tian-Zhong, M., Bi, C., Ying, Z., Xia, J., Cai-Ling, P., Yun-Shan, Z., Mei-Wen, H., Yan-Ru, N., 2016. Critical role of Emx2 in the pluripotency - differentiation transition in male gonocytes via regulation of FGF9/NODAL pathway. Reproduction 151, 673–681. Miyazawa, K., Shinozaki, M., Hara, T., Furuya, T., Miyazono, K., 2002. Two major Smad pathways in TGF-beta superfamily signalling. Genes Cells 7, 1191–1204. Kitisin, K., Saha, T., Blake, T., Golestaneh, N., Deng, M., Kim, C., Tang, Y., Shetty, K., Mishra, B., Mishra, L., 2007. Tgf-Beta signaling in development. Sci. STKE 2007 (cm1). Shi, Y., Massague, J., 2003. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell 113, 685–700.

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016

C. Spiller et al. / Stem Cell Research xxx (2017) xxx–xxx Shen, M.M., Schier, A.F., 2000. The EGF-CFC gene family in vertebrate development. Trends Genet. 16, 303–309. Kelber, J.A., Shani, G., Booker, E.C., Vale, W.W., Gray, P.C., 2008. Cripto is a noncompetitive activin antagonist that forms analogous signaling complexes with activin and nodal. J. Biol. Chem. 283, 4490–4500. Gray, P.C., Vale, W., 2012. Cripto/GRP78 modulation of the TGF-beta pathway in development and oncogenesis. FEBS Lett. 586, 1836–1845. Gray, P.C., Shani, G., Aung, K., Kelber, J., Vale, W., 2006. Cripto binds transforming growth factor beta (TGF-beta) and inhibits TGF-beta signaling. Mol. Cell. Biol. 26, 9268–9278. Spiller, C.M., Feng, C.W., Jackson, A., Gillis, A.J., Rolland, A.D., Looijenga, L.H., Koopman, P., Bowles, J., 2012. Endogenous Nodal signaling regulates germ cell potency during mammalian testis development. Development 139, 4123–4132. Souquet, B., Tourpin, S., Messiaen, S., Moison, D., Habert, R., Livera, G., 2012. Nodal signaling regulates the entry into meiosis in fetal germ cells. Endocrinology 153, 2466–2473. Sakuma, R., Ohnishi Yi, Y., Meno, C., Fujii, H., Juan, H., Takeuchi, J., Ogura, T., Li, E., Miyazono, K., Hamada, H., 2002. Inhibition of Nodal signalling by Lefty mediated through interaction with common receptors and efficient diffusion. Genes Cells 7, 401–412. Miles, D.C., Wakeling, S.I., Stringer, J.M., van den Bergen, J.A., Wilhelm, D., Sinclair, A.H., Western, P.S., 2013. Signaling through the TGF beta-activin receptors ALK4/5/7 regulates testis formation and male germ cell development. PLoS One 8, e54606. Archambeault, D.R., Yao, H.H., 2010. Activin A, a product of fetal Leydig cells, is a unique paracrine regulator of Sertoli cell proliferation and fetal testis cord expansion. Proc. Natl. Acad. Sci. U. S. A. 107, 10526–10531. Yao, H.H., Aardema, J., Holthusen, K., 2006. Sexually dimorphic regulation of inhibin beta B in establishing gonadal vasculature in mice. Biol. Reprod. 74, 978–983. Jameson, S.A., Natarajan, A., Cool, J., DeFalco, T., Maatouk, D.M., Mork, L., Munger, S.C., Capel, B., 2012. Temporal transcriptional profiling of somatic and germ cells reveals biased lineage priming of sexual fate in the fetal mouse gonad. PLoS Genet. 8, e1002575. Jeanes, A., Wilhelm, D., Wilson, M.J., Bowles, J., McClive, P.J., Sinclair, A.H., Koopman, P., 2005. Evaluation of candidate markers for the peritubular myoid cell lineage in the developing mouse testis. Reproduction 130, 509–516. Cupp, A.S., Kim, G., Skinner, M.K., 1999. Expression and action of transforming growth factor beta (TGFbeta1, TGFbeta2, and TGFbeta3) during embryonic rat testis development. Biol. Reprod. 60, 1304–1313. Moreno, S.G., Attali, M., Allemand, I., Messiaen, S., Fouchet, P., Coffigny, H., Romeo, P.H., Habert, R., 2010. TGFbeta signaling in male germ cells regulates gonocyte quiescence and fertility in mice. Dev. Biol. 342, 74–84. Memon, M.A., Anway, M.D., Covert, T.R., Uzumcu, M., Skinner, M.K., 2008. Transforming growth factor beta (TGFbeta1, TGFbeta2 and TGFbeta3) null-mutant phenotypes in embryonic gonadal development. Mol. Cell. Endocrinol. 294, 70–80. Strizzi, L., Bianco, C., Normanno, N., Salomon, D., 2005. Cripto-1: a multifunctional modulator during embryogenesis and oncogenesis. Oncogene 24, 5731–5741. Minchiotti, G., 2005. Nodal-dependant Cripto signaling in ES cells: from stem cells to tumor biology. Oncogene 24, 5668–5675. Yeo, C., Whitman, M., 2001. Nodal signals to Smads through Cripto-dependent and Cripto-independent mechanisms. Mol. Cell 7, 949–957.

7

Wu, Q., Kanata, K., Saba, R., Deng, C.X., Hamada, H., Saga, Y., 2013. Nodal/activin signaling promotes male germ cell fate and suppresses female programming in somatic cells. Development 140, 291–300. Inman, G.J., Nicolas, F.J., Callahan, J.F., Harling, J.D., Gaster, L.M., Reith, A.D., Laping, N.J., Hill, C.S., 2002. SB-431542 is a potent and specific inhibitor of transforming growth factorbeta superfamily type I activin receptor-like kinase (ALK) receptors ALK4, ALK5, and ALK7. Mol. Pharmacol. 62, 65–74. Vallier, L., Mendjan, S., Brown, S., Chng, Z., Teo, A., Smithers, L.E., Trotter, M.W., Cho, C.H., Martinez, A., Rugg-Gunn, P., Brons, G., Pedersen, R.A., 2009. Activin/Nodal signalling maintains pluripotency by controlling Nanog expression. Development 136, 1339–1349. Brown, S., Teo, A., Pauklin, S., Hannan, N., Cho, C.H., Lim, B., Vardy, L., Dunn, N.R., Trotter, M., Pedersen, R., Vallier, L., 2011. Activin/Nodal signaling controls divergent transcriptional networks in human embryonic stem cells and in endoderm progenitors. Stem Cells 29, 1176–1185. Vallier, L., Reynolds, D., Pedersen, R.A., 2004. Nodal inhibits differentiation of human embryonic stem cells along the neuroectodermal default pathway. Dev. Biol. 275, 403–421. Mendis, S.H., Meachem, S.J., Sarraj, M.A., Loveland, K.L., 2011. Activin a balances Sertoli and germ cell proliferation in the fetal mouse testis. Biol. Reprod. 84, 379–391. Hammond, S.S., Matin, A., 2009. Tools for the genetic analysis of germ cells. Genesis 47, 617–627. Sarraj, M.A., Escalona, R.M., Western, P., Findlay, J.K., Stenvers, K.L., 2013. Effects of TGFbeta2 on wild-type and Tgfbr3 knockout mouse fetal testis. Biol. Reprod. 88, 66. Sarraj, M.A., Escalona, R.M., Umbers, A., Chua, H.K., Small, C., Griswold, M., Loveland, K., Findlay, J.K., Stenvers, K.L., 2010. Fetal testis dysgenesis and compromised Leydig cell function in Tgfbr3 (beta glycan) knockout mice. Biol. Reprod. 82, 153–162. Xu, R.H., Sampsell-Barron, T.L., Gu, F., Root, S., Peck, R.M., Pan, G., Yu, J., AntosiewiczBourget, J., Tian, S., Stewart, R., Thomson, J.A., 2008. NANOG is a direct target of TGFbeta/activin-mediated SMAD signaling in human ESCs. Cell Stem Cell 3, 196–206. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., Yamanaka, S., 2007. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861–872. Hata, A., Chen, Y.G., 2016. TGF-beta signaling from receptors to Smads. Cold Spring Harb. Perspect. Biol. 8. Chen, A.E., Borowiak, M., Sherwood, R.I., Kweudjeu, A., Melton, D.A., 2013. Functional evaluation of ES cell-derived endodermal populations reveals differences between Nodal and Activin A-guided differentiation. Development 140, 675–686. Lee, K.L., Lim, S.K., Orlov, Y.L., Yit le, Y., Yang, H., Ang, L.T., Poellinger, L., Lim, B., 2011. Graded Nodal/Activin signaling titrates conversion of quantitative phospho-Smad2 levels into qualitative embryonic stem cell fate decisions. PLoS Genet. 7, e1002130. Kelber, J.A., Panopoulos, A.D., Shani, G., Booker, E.C., Belmonte, J.C., Vale, W.W., Gray, P.C., 2009. Blockade of Cripto binding to cell surface GRP78 inhibits oncogenic Cripto signaling via MAPK/PI3K and Smad2/3 pathways. Oncogene 28, 2324–2336. Skakkebaek, N.E., Berthelsen, J.G., Giwercman, A., Muller, J., 1987. Carcinoma-in-situ of the testis: possible origin from gonocytes and precursor of all types of germ cell tumours except spermatocytoma. Int. J. Androl. 10, 19–28. Rijlaarsdam, M.A., Looijenga, L.H., 2014. An oncofetal and developmental perspective on testicular germ cell cancer. Semin. Cancer Biol. 29, 59–74.

Please cite this article as: Spiller, C., et al., Regulation of fetal male germ cell development by members of the TGFβ superfamily, Stem Cell Res. (2017), http://dx.doi.org/10.1016/j.scr.2017.07.016