Germ cells are totipotent and, in principle, immortal as they are the source for new germ cells in each generation. This very special role requires tight quality control systems. The p53 protein family constitutes one of the most important quality surveillance systems in cells. Whereas p53 has become famous for its role as the guardian of the genome in its function as the most important somatic tumor suppressor, p63 has been nicknamed ‘guardian of the female germ line’. p63 is strongly expressed in resting oocytes and responsible for eliminating those that carry DNA double-strand breaks. The third family member, p73, acts later during oocyte and embryo development by ensuring correct assembly of the spindle assembly checkpoint. In addition to its role in the female germ line, p73 regulates cell-cell contacts between developing sperm cells and supporting somatic cells in the male germ line. Here, we review the involvement of the p53 protein family in the development of germ cells with a focus on quality control in the female germ line and discuss medical implications for cancer patients.
The p53 protein family is the most-studied group of proteins, with the transcription factor p53 (officially known as TP53) itself being the protein with the highest number of publications. Since its discovery in 1979 (Lane and Crawford, 1979; Linzer and Levine, 1979), the role of p53 as the central somatic tumor suppressor has been well-documented (Vousden and Prives, 2009; Joerger and Fersht, 2010; Lane, 1992; Lane and Levine, 2010). In addition, its roles in the regulation of metabolic networks (Vousden and Ryan, 2009) and in the implantation of an embryo in the uterus (Hu et al., 2007) have been discovered and studied. Research in these non-tumor suppressor functions has been sparked by the discovery of its two siblings, p63 and p73 (TP63 and TP73, respectively), in the late 1990s (Kaghad et al., 1997; Yang et al., 1998; Trink et al., 1998; Senoo et al., 1998; Schmale and Bamberger, 1997). Both proteins show a high sequence similarity to p53, reaching 65% in the DNA-binding domain and resulting in a very large overlap of promoter target sequences between all three family members. In addition, all three proteins contain a highly similar oligomerization domain through which they form tetramers (Jeffrey et al., 1995; Lee et al., 1994; Joerger et al., 2009; Coutandin et al., 2009). However, the original sequencing of the p63 and p73 genes also showed significant differences between p53 on the one hand, and p63 and p73 on the other. p63 and p73 exist in several different isoforms that are expressed via at least two different promoters and created through C-terminal splicing events (Yang et al., 1998; Kaghad et al., 1997); both have been shown to have different biological functions (Yang et al., 1999; Mills et al., 1999; Suh et al., 2006). For p53, different isoforms have been also described; however, their biological functions, apart from that of inhibiting wild-type p53 and promoting tumor development, are still controversially debated (Bourdon et al., 2005; Marcel et al., 2011). The p63 and p73 isoforms that, so far, are characterized best all contain a sterile alpha motif (SAM) domain in their C-terminus, which is absent in p53 (Chi et al., 1999).
Studies using knockout-mice revealed that all three proteins serve different biological functions despite their high sequence similarity. The p53-knockout mouse shows some developmental defects, such as craniofacial malformations that include ocular abnormalities and defects in tooth formation, and more than 20% of female embryos and newborn mice die due to defects in neural tube closure (Armstrong et al., 1995). However, the most important effect is that p53-knockout mice develop tumors, in particular thymus T cell lymphomas, as early as six months of age (Donehower et al., 1992; Jacks et al., 1994). The p73-knockout mouse suffers from abnormal hippocampal development, hydrocephalus, chronic infections and inflammation, as well as abnormalities in pheromone sensory pathways (Yang et al., 2000). Inactivation of p63 has even more severe effects, resulting in limb truncations, and the lack of a multi-layered skin and other epithelial structures (Yang et al., 1999; Mills et al., 1999). In stratified epithelial tissue, such as the skin, p63 is highly expressed within the basal layer, which is necessary to build up these multi-layered tissues, including the limb buds required for limb growth (Senoo et al., 2007).
The isoform expressed in these epithelial tissues is ΔNp63α. It has a truncated N-terminal transactivation domain and the longest C-terminus, including the SAM domain (Mills et al., 1999; Yang et al., 1999). A different isoform, TAp63α contains the entire N-terminal transactivation domain and is highly expressed in female germ cells (Suh et al., 2006). Here, p63 serves as a quality control factor, which ensures that oocytes with DNA damage are eliminated before they can be recruited for ovulation (Suh et al., 2006; Livera et al., 2008). The identification of p53-like proteins in germ cells of invertebrates, such as Caenorhabditis elegans (Derry et al., 2001) and Drosophila melanogaster (Ollmann et al., 2000; Brodsky et al., 2000), which are more closely related to p63 than to p53 (Ou et al., 2007), suggests that quality control is the oldest function of the entire p53 protein family (Mateo et al., 2016; Levine et al., 2011). The increasing life span of animals, which results in the development of adult stem cells and renewable tissue, has made tumor suppression an important task. Within the p53 protein family, these tumor suppressor functions are mainly carried out by p73 and p53. In this review, we discuss the involvement of p63 in maintaining the genetic quality of oocytes, the effect of these mechanisms during tumor chemotherapy on the fertility of female patients, and the role of p73 and p53 in this process. p73 is also involved in orchestrating spermatogenesis and we, thus, highlight the involvement of the p53 family in the development of male germ cells. We show that the activity of all three family members is regulated by different mechanisms that include the oligomerization state of p63, and the cellular stability of p73 and p53.
Germ cells are a highly specialized type of cell. Primordial germ cells evolve during embryogenesis from the primitive streak of the blastula and migrate to the developing gonads (Nikolic et al., 2016; McLaren, 2003). During this migration, they begin to multiply and undergo massive epigenetic reprogramming, including the erasure of any imprinting to restore totipotency (Tilgner et al., 2008). In the developing ovary, female primordial germ cells differentiate into oogonia that further multiply mitotically (De Felici and Farini, 2012) before entering meiosis. Immediately prior to and during the process of primordial follicle formation around the time of birth in mice, a large number of germ cells die due to apoptosis (De Felici et al., 2005; Malki et al., 2014). The surviving oocytes are arrested in prophase of meiosis I. Male primordial germ cells enter the mitotic arrest phase after arriving at the genital ridge and remain quiescent in the G0/G1 phase of the cell cycle. After birth, they reinitiate the cell cycle following migration to the basal membrane of the seminiferous tubule where they form tight junctions with Sertoli cells. This tight interaction creates the spermatogonial stem cell niche that supports spermatogonial stem cells and the production of sperm cells throughout the lifetime of males. Since both types of germ cell are potentially immortal – they are passed from generation to generation – they must have very strict quality control systems (Aitken et al., 2011; Levine et al., 2011).
Oocytes are not only different from all somatic cells, they also differ significantly from male germ cells. Only oocytes have extended arrest periods during the cell cycle (dictyate phase) that can last up to a year in mice and several decades in humans (Eppig, 1996). Before mammalian oocytes enter this dictyate arrest around the time of birth, they undergo meiotic recombination by aligning homologous chromosomes during the pachytene stage of meiotic prophase. During this stage, the type II topoisomerase-like DNA transesterase Spo11 promotes hundreds of DNA double-strand breaks, thereby initiating the exchange of genetic material between homologous chromosomes. In the following diplotene phase, chromosomes start to separate but remain attached through chiasmata. At this stage, which occurs around the time of birth − i.e. embryonic day 18.5 (E18.5) to 5 days post birth (P5) in mice − oocytes enter dictyate arrest and remain in this phase until rising levels of lutenizing hormone initiate their re-entry into the cell cycle and recruit oocytes for ovulation (Eppig, 1996).
Currently, it is believed that human females are born with a finite number of oocytes, which are created during embryogenesis from germ stem cells, and that this number declines over the life time (Block, 1952; Richardson et al., 1987). From the approximately seven million primary oocytes created ∼20 weeks after fertilization, only one to two millions are left at the time of birth; and this number continues to decline until a level of ∼25,000 is reached at an age of 37 (Block, 1952, 1953; Johnston and Wallace, 2009) (Fig. 1). From that time, the decline in oocyte number accelerates, and menopause begins when their number drops to below 1000 (Richardson et al., 1987). Cell death in oocytes can be caused by three different mechanisms (Klinger et al., 2015): (i) insufficient amount of growth factor and/or nutrients, or of supporting ovarian somatic cells, (ii) self-sacrifice mechanism by which oocytes donate their cytoplasm to a subset of surviving oocytes and, (iii) DNA damage or chromosomal defects. For the first two mechanisms, cell-cell contacts − as well as a complex interplay between apoptotic pathways and autophagy − are important and have been described in several excellent recent reviews (Streiter et al., 2016; De Felici and Klinger, 2015). The mechanism of the DNA-damage-induced pathway has also been identified, and was found to depend on the p53 protein family (Suh et al., 2006), in particular, on p63 as discussed below.
p63-mediated quality control in oocytes
Surveillance of the genetic quality of oocytes is an important mechanism to ensure that only oocytes without any DNA damage are recruited for ovulation. In mice, quality control is based on two check points, which are triggered by defects in DNA double-strand break repair within days of birth and by defects in chromosomal synapsis within the first two months (Reinholdt and Schimenti, 2005; Di Giacomo et al., 2005). The first check point ensures that only oocytes survive that have repaired the high number of Spo11-induced DNA double-strand breaks by either crossover or noncrossover repair mechanisms (Cole et al., 2012). Activation of this checkpoint leads to a massive loss of oocytes within a few days following birth. Interestingly, this check point appears to remain activated because DNA double-strand breaks that occur at later stages result in elimination of oocytes by apoptosis (Suh et al., 2006). The need for an effective quality control system that ensures the genetic integrity of oocytes is owing to their tetraploid state during the dictyate arrest phase, which makes them particularly vulnerable to DNA damage (Suh et al., 2006; Peters and Levy, 1964). DNA lesions can be caused either by extrinsic factors, such as irradiation or DNA-damaging chemicals, or by intrinsic processes, such as the activation of the long interspersed nuclear element 1 (LINE1) retrotransposon, which becomes activated during epigenetic reprogramming of embryonic germ cells (Malki et al., 2014). The dictyate arrest phase is characterized by high cellular concentration of TAp63α, the oocyte-specific isoform of p63 (Suh et al., 2006; Livera et al., 2008) (Fig. 2). Indeed, expression of TAp63α in mice is not detectable at E16.5, but increases around the time of birth with ∼20% of oocytes showing expression at E18.5 (Kim and Suh, 2014). Subsequently, virtually all oocytes show strong expression of TAp63α at P5 (Suh et al., 2006), and its expression remains high in primary oocytes, but is lost in secondary oocytes that are recruited for ovulation (Suh et al., 2006). Knockout of this oocyte-specific TAp63 isoform in mice does not impair the normal development of oocytes, but makes them insensitive to DNA damage, for example by γ-irradiation (Suh et al., 2006; Livera et al., 2008). These knockout studies have demonstrated that p63 serves as a quality control factor of oocytes in the arrest phase. The high concentration of a pro-apoptotic factor, such as the p53-family member p63, in combination with the long arrest phase of oocytes raised the question how activity of TAp63α is regulated, and how oocytes manage to survive for decades in humans. Indeed, already during the discovery of p63 it had been noticed that the transcriptional activity of the TAp63α isoform on a prototypical p21 promotor is strongly reduced relative to the activity of its shorter isoforms (Yang et al., 1998). The region controlling the transcriptional activity of TAp63α was mapped to approximately the 70 C-terminal amino acids and, in particular, to a stretch of around 15 residues (amino acids 597−610) (Serber et al., 2002). Mutation of this so-called transactivation inhibitory domain (TID) or its deletion restores the transcriptional activity of the mutant to levels of the active isoforms (Straub et al., 2010). Further investigations into the TID-mediated inhibition have revealed an intramolecular mechanism by which TAp63α is locked in a closed and dimer-only conformation, whereas all active forms of the human p53 protein family are tetramers (Deutsch et al., 2011; Luh et al., 2013). Mutational analyses in combination with structural studies using small-angle x-ray scattering have demonstrated that the TIDs of the two TAp63α molecules within the dimer form an antiparallel β-sheet (Coutandin et al., 2016). This two-stranded, anti-parallel β-sheet is extended on both sides by two β-strands formed by a sequence of amino acids within their N-terminal transactivation domain. The complete six-stranded, anti-parallel β-sheet blocks tetramerization by interacting with the central oligomerization domain. This oligomerization domain consists of a dimer of dimers, and has two different interfaces through which two monomers first form dimers, before the two dimers create the full tetramer. (Joerger et al., 2009; Coutandin et al., 2009). Interaction of the inhibitory β-sheet with this second interface blocks tetramerization and locks TAp63α in a dimeric state. In addition, the first ∼36 amino acids of the N-terminal transactivation domain form an α-helix through which the transcription factor interacts with the transcriptional machinery (Burge et al., 2009; Krois et al., 2016; MacPartlin et al., 2005). This helix contains three residues − F19, W23 and L26 − that are conserved in p53 and in which they constitute the interface with its negative regulator E3 ubiquitin-protein ligase mouse double minute 2 homolog (MDM2) (Kussie et al., 1996). In p63, this helix also binds to the central oligomerization domain, thereby stabilizing the compact and closed conformation (Deutsch et al., 2011; Coutandin et al., 2016).
In this closed dimeric state, the transcriptional activity of TAp63α is inhibited by two mechanisms. First, its DNA-binding affinity is reduced approximately twenty-fold (Suh et al., 2006; Deutsch et al., 2011) and, second, the domain responsible for recruiting the transcriptional machinery including the RNA polymerase is buried and inaccessible (Deutsch et al., 2011; Coutandin et al., 2016). Activation of TAp63α requires unfolding of the inhibitory β-sheet, which is achieved by phosphorylation followed by formation of the active and open tetramer (Coutandin et al., 2016; Deutsch et al., 2011) (Fig. 3). However, phosphorylation is not required for stabilization of the tetrameric state but, merely, serves as a trigger to overcome the activation-energy barrier between the closed inactive dimer and the open active tetramer. Once the tetramer is formed, all phosphate groups can be removed without the protein adopting again the closed dimeric state (Deutsch et al., 2011). Further investigations have demonstrated that this closed dimeric state is a kinetically trapped, high-energy conformation, and that activation follows a spring-loaded mechanism (Coutandin et al., 2016).
TAp63α is not the only factor that adopts a pre-activated state. The oocyte contains already all factors (kinases) that are required for the activation of p63 without the need for protein synthesis (Coutandin et al., 2016). This explains why oocytes are far more sensitive to DNA damage than any of the surrounding somatic cells. One of the key questions is which kinase(s) are involved in the activation process. It has been shown that DNA double-strand breaks are most effective in activating TAp63α. Formation of DNA double-strand breaks triggers the recruitment of the Mre11−Rad50−Nbs1 (MRN) complex to the site of the lesion (Lee and Paull, 2005; Marechal and Zou, 2013). This, in turn, activates the ataxia telangiectasia mutated stress kinase (ATM), which phosphorylates and activates the check point kinase 2 (Chk2). Activated Chk2 then phosphorylates TAp63α on Ser582 within a sequence that is located between the SAM and TID domains (Bolcun-Filas et al., 2014). Accordingly, oocytes from Chk2-knockout mice do not die after γ-irradiation treatment despite expressing TAp63α (Bolcun-Filas et al., 2014). In addition, pre-treatment of mouse ovary cultures with Chk2 inhibitor II, followed by γ-irradiation also protects the cells from cell death, demonstrating that phosphorylation through Chk2 is a necessary step in the activation process of TAp63α (Coutandin et al., 2016).
In addition to Chk2, the tyrosine kinase c-Abl has been shown to phosphorylate TAp63α on Tyr110 following treatment with cisplatin (Gonfloni et al., 2009). This phosphorylation appears to stabilize TAp63α, leading to increased cellular levels. In contrast, activation of TAp63α by γ-irradiation, results in a decrease in the cellular levels of TAp63α, suggesting that c-Abl cannot fully activate TAp63α. This interpretation is in agreement with a study using C. elegans, which demonstrated that treatment with the c-Abl inhibitor Imatinib cannot inhibit irradiation-induced oocyte death but, instead renders germ cells even more sensitive to apoptosis induced by DNA double-strand-breaks (Deng et al., 2004).
Transcriptional activation of TAp63α leads to the expression of the BH3-only proteins PUMA and NOXA, which trigger the induction of apoptosis (Kerr et al., 2012). In somatic cells, both proteins are transcriptional targets of p53 (Villunger et al., 2003; Erlacher et al., 2005). Interestingly, in oocytes, TAp63α appears to be the main transcriptional regulator of both proteins. However, it has been suggested that, in the absence of TAp63α and under continuously high levels of DNA damage, p53 initiates apoptosis (Bolcun-Filas et al., 2014). The role of p53 at a time when TAp63α is present at physiological levels needs further investigation.
PUMA and NOXA double-knockout mice are insensitive to DNA damage inflicted by γ-irradiation. Interestingly, the irradiated mice are not only fertile but their offspring are healthy and fertile, suggesting that oocytes are capable of repairing the DNA damage if their elimination by cell death is inhibited (Kerr et al., 2012). The same effect has been seen with Chk2-knockout mice, which remain fertile after irradiation, and whose litter numbers are similar to those of non-irradiated wild-type mice (Bolcun-Filas et al., 2014). No abnormalities have been found in their offspring and no persisting DNA double-stand breaks were detected two month after birth in Chk2−/− Trip13Gt/Gt mice harboring thyroid hormone receptor interactor 13 (TRIP13) alleles that are defective in noncrossover DNA double-strand break repair mechanisms (Bolcun-Filas et al., 2014).
The investigations described above have revealed the central role of TAp63α as the ‘guardian of the maternal germ line’, and this role of p63 and its activation mechanism have opened new possibilities of how to preserve the fertility of female cancer patients.
Many drugs used in tumor therapies inflict DNA damage, with some of them also causing DNA double-strand breaks. Since these chemotherapeutic drugs are not cancer cell specific, they also create DNA lesions in oocytes, which can activate cell death pathways. As mentioned above, oocytes are particularly sensitive to DNA damage. Indeed, it has been shown that four to ten DNA double-strand breaks are sufficient to kill them (Suh et al., 2006), whereas somatic cells survive much higher levels of DNA damage. Chemotherapy-induced cell death in oocytes can, thus, result in elimination of the entire oocyte pool, which not only results in infertility but also triggers dramatic changes in the hormonal system of the treated patients (Jeruss and Woodruff, 2009; Johnston and Wallace, 2009; Peate et al., 2009). In the worst case, ovarian failure results in early menopause independent of the age of the patient (Johnston and Wallace, 2009), and the likelihood of ovarian failure increases with the age of the patient. In one study, treatment of female breast cancer patients with the typical regimen of cyclophosphamide, methotrexate and 5-fluorouracil (CMF) resulted in amenorrhea in 61% of patients below the age of 40 and 95% of those above the age of 40 (Maltaris et al., 2008; Goldhirsch et al., 1990). One particularly important aspect of preserving the oocyte reservoir is that of children treated for childhood cancers. Improved therapies have dramatically increased the five-year survival rate for many childhood cancers. However, the flipside of this success is that preserving fertility has become a more-pressing medical problem, particularly when considering that, currently, one in 250 adults is thought to be a survivor of childhood cancer in industrial societies (Maltaris et al., 2009; Blatt, 1999). In one particular study, which investigated the survival of childhood cancer, 8% of all female patients suffered from premature menopause, compared with 0.8% of a sibling control − which is close to the reported average of 0.9% of the general population (Johnston and Wallace, 2009; Sklar et al., 2006). This number increased to 30−40% in patients who were treated with irradiation in the pelvis, in combination with alkylating-agent chemotherapy (Byrne et al., 1992; Sklar et al., 2006). Currently, measures of fertility preservation include cryopreservation of oocytes or ovarian tissue and hormonal treatment to establish a pre-puberty-like state during the time of chemotherapy treatment. However, all these measures have side effects, delay the start of chemotherapy treatments or can only provide an option for later pregnancies based on in vitro fertilization. A better option would be to either reestablish the oocyte pool after the end of the treatment or to protect the oocytes more efficiently. The first option might be feasible on the basis of recent findings that oocytes can be produced from germ stem cells even in adult mice (White et al., 2012; Zou et al., 2009). This observation opens the possibility that reactivation of these stem cells replenishes the oocyte reserve; however, in order to become a viable option, far more research is required into how these stem cells are affected by chemotherapy and their possible activation afterwards. For the latter option to become feasible, a better understanding of the mechanisms that trigger apoptotic pathways in damaged oocytes is necessary, including a detailed structural description of the activation mechanism of TAp63α. Because TAp63α has, thus far, mainly been found in oocytes, the development of an inhibitor to prevent its activation seems to be an option. It is worth noting that, besides germ cells, the only other known location where TAp63α is expressed is in certain stem cells within the sheath of hair follicles in the skin (Su et al., 2009). However, it is not known which oligomeric state this population of TAp63α adopts and what its particular function is. First experiments with Chk2 inhibitors have already shown that the concept of inhibiting the activation pathway of TAp63α is a successful one (Rinaldi et al., 2017; Coutandin et al., 2016). Another proposed approach is the use of imatinib to inhibit the kinase c-Abl (Gonfloni et al., 2009); this has been suggested to be protective during treatment with cisplatin but not with doxorubicin (Morgan et al., 2013). However, the fertoprotective effect has been challenged in another study (Kerr et al., 2012), which also suggested that treatment with imatinib is harmful because this drug also inhibits the receptor tyrosine kinase c-Kit (Tuveson et al., 2001; Krystal et al., 2000). The survival of oocytes depends on the function of this protein (Carlsson et al., 2006).
With the increasing survival rate of cancer patients, questions of the quality of life following the end of the treatment become also increasingly important. For female patients, preserving their fertility and normal hormonal levels is of central importance. One potential therapeutic approach could be inhibition of the TAp63α-activating kinase Chk2 as shown in recent proof-of-principle studies (Rinaldi et al., 2017; Coutandin et al., 2016).
p53 and p73 in quality control in oocytes
The other family members, p73 and, in particular, p53 also might play a role in the quality control of oocytes. Although oocytes in TAp63−/− mouse ovaries survive 5 days after irradiation, all primary oocytes are eliminated after 7 days (Bolcun-Filas et al., 2014). Additional knockout of p53 rescued the double-knockout oocytes to a similar level as those in Chk2−/− mice, resulting in long-term survival of oocytes. Chk2 also phosphorylates p53, explaining why elimination of this kinase leads to a complete rescue by blocking activation of both TAp63α and p53 (Rinaldi et al., 2017). This points to a timeline of events, in which TAp63α is directly responsible for the elimination of compromised oocytes following DNA damage and in which p53 potentially acts at later time points. This might also reconcile additional findings that p53 is not necessary for oocyte death induced by doxorubicin or γ-irradiation (Perez et al., 1997; Kerr et al., 2012; Suh et al., 2006) because of the different time scales of these studies. Another study proposed a model that connects all three family members (Kim et al., 2013) and identified TAp63α as the master regulator of cisplatin-induced oocyte cell death, with TAp63α controlling the expression of c-Abl, TAp73 and potentially of p53. c-Abl-mediated phosphorylation of TAp73 leads to the expression of the apoptosis regulator Bax, thereby initiating apoptosis. However, in this study, several results could not be explained, including the apparent toxicity of cisplatin, not only to primary oocytes but also to somatic ovarian cells (Hutt et al., 2013).
p73 also has a role in ensuring correct chromosomal segregation (Tomasini et al., 2009). Male and female TAp73−/− mice are both infertile despite showing normal mating behavior, i.e. they have no defects in pheromonal pathways as described for the p73-knockout mouse (Yang et al., 2000) and female mice show normal cyclicity (Tomasini et al., 2008; Tomasini et al., 2009). In females, this infertility was shown to be caused by ovulated oocytes being trapped under the bursa and not being able to migrate towards the fallopian tube (Tomasini et al., 2008). Investigation of in vitro matured germinal vesicles and ovulated oocytes revealed high levels of spindle abnormalities, such as multipolar spindles, spindle relaxation and scattering. Furthermore, in vitro fertilization of TAp73−/− oocytes resulted in high numbers of embryos with multinucleated blastomeres and in blastocysts with an abnormal cell number. These results suggested a link between TAp73 and the spindle assembly checkpoint, which regulates the correct attachment of sister chromatids both to the mitotic and meiotic spindle (Tomasini et al., 2009). Indeed, in oocytes from TAp73−/− mice, several components of the spindle assembly checkpoint were found to be mislocalized, including the mitotic kinases Bub1 and BubR1 (Tomasini et al., 2009). The same study also demonstrated a direct physical interaction between TAp73α and Bub1, Bub3 and BubR1. The interaction with Bub1 was mapped to the C-terminus of TAp73α (Tomasini et al., 2009). Although TAp73α contains a domain with a high sequence similarity to the TID of p63 at its C-terminus, it does not form closed dimers but is a constitutively open tetramer (Luh et al., 2013) (Fig. 4). This conformation makes interaction with the C-terminal domains possible without any major conformational rearrangements. Interestingly, in female mice, the concentration of TAp73 decreases in oocytes with increasing age (Guglielmino et al., 2011). This observation suggests that problems with chromosomal segregation, which could result in a higher number of oocytes with aneuploidy, are more likely in older female mice and could also contribute to the well-known increase of the risk of trisomies in humans with increasing age of the mother (Guglielmino et al., 2011; Gaulden, 1992).
p73 is essential for spermatogenesis
In contrast to the female germ line, spermatogonial germ cells remain active during the entire reproductive life time. These germ cells, called spermatogonia type A, are located on the basement membrane of seminiferous tubules; they divide mitotically to maintain the stem cell population and, after further cell divisions, produce spermatogonia type B cells, which are committed to finally become sperm cells (Oakberg, 1956; Hess and Renato de Franca, 2008; Cheng and Mruk, 2010). After further mitotic cell divisions, these cells enter meiosis I and are now called spermatocytes type I and, in response to doubling their DNA in S-phase enter prophase that lasts 24 days and includes recombination of the genetic material. Each type I spermatocyte creates two spermatocyte type II cells that, during meiosis II, create haploid spermatids and finally mature sperm cells. This entire process is accompanied by migration of the cells from the basal membrane of the seminiferous tubule through the blood–testis barrier (BTB) towards the lumen (Cheng et al., 2010). Of crucial importance in the entire maturation process are Sertoli cells, large somatic cells that reach from the basal membrane into the lumen. Sertoli cells provide multiple layers of docking sites for developing sperm cells, and are essential for their structural and metabolic support, each cell supporting 30–50 developing germ cells in deep cytoplasmic pockets (Fig. 5) (Walker, 2009; Cheng et al., 2010). In addition, Sertoli cells form the BTB and generate junctions, including tight junctions, gap junctions and adherens junctions, as well as desmosome-like structures, to separate the stem cell compartment of mitotically dividing spermatogonia from the apical development compartment that contains meiotic and postmeiotic cells (Cheng and Mruk, 2010; Xia et al., 2005).
Histochemical analysis of male TAp73−/− mice revealed a drastic loss of developing germ cells and mature spermatozoa that resulted in almost ‘empty’ seminiferous tubules (Holembowski et al., 2014; Inoue et al., 2014) (Fig. 5) but, at the same time, detected apoptotic sperm cells in the lumen. However, the stem cell compartment with spermatogonia appeared to contain the normal number of cells. Similarly, the number of meiotic spermatocytes was relatively unaffected, whereas round and elongated spermatids and spermatozoa were lacking. In addition, electron microscopy images revealed that the germ epithelium is disorganized, with Sertoli cells found with thin cytoplasmic pockets and a disorganized BTB. Comparison of transcription programs of wild-type and TAp73−/− germ cells revealed that proteins involved in the regulation of cell–cell contacts and in cell migration, such as the serine proteases inhibitors SPINK2 and SPINT1, and several metalloproteases of the ADAM and MMP families are downregulated (with the exception of MMP13), whereas mRNA levels of the metalloprotease inhibitors TIMP1 and TIMP4 were increased (Inoue et al., 2014; Holembowski et al., 2014). Expression of TAp73 is specific to germ cells and Sertoli cells of TAp73−/− mice do not show any of the aberrant gene expression patterns found to be deregulated in germ cells. These results indicate that the male infertility in TAp73−/− mice is mainly caused by defective cell-cell adhesions between developing germ cells and Sertoli cells, and not by deregulation of a genetic quality control program − although increased staining of γH2AX, which is indicative of DNA double-strand breaks, was observed in testis of TAp73−/− mice (Inoue et al., 2014). Furthermore, interaction with the kinase c-Abl and increased tyrosine phosphorylation has been reported following X-ray irradiation of wild-type mouse testis with a dose of 4 Gy (Hamer et al., 2001). However, involvement of p73 in genetic quality control during spermatogenesis appears to be unlikely (see also below). The interpretation that TAp73 regulates cell-cell contacts and is not part of a quality control mechanism in male germ cells is also consistent with the observation that some of the male TAp73−/− mice are only subfertile and not sterile because some of the germ cells are able to fully develop.
Investigations of the role of p73 in spermatogenesis have revealed a function of this p53 family member that is different from the role of p63 in oocytes. While knockout of p63 does not affect the development of oocytes but, instead, compromises the surveillance of genetic quality in these cells, knockout of p73 leads to changes in cell-cell contacts within the seminiferous tubules. This results in the loss of nurturing contacts between Sertoli cells and the developing sperm cells, blocking their further development.
p63 and p53 in quality control of male germ cells
Testis of humans and great apes express a unique p63 isoform that contains an additional 37 amino acids at its N-terminus (Beyer et al., 2011). This isoform, called GTAp63α, is created by insertion of the LTR region of the human endogenous retrovirus 9 upstream of the TAp63 gene (Beyer et al., 2011). GTAp63α is expressed in spermatogonia and developing spermatozoa but not in mature spermatides (Beyer et al., 2011). Detection of DNA damage leads to cleavage of the C-terminus of GTAp63α by caspases, resulting in the removal of its TID and activation of the protein (Beyer et al., 2011). It has been proposed that GTAp63α gene expression, which is driven by the long terminal repeat (LTR) of human endogenous retrovirus (ERV9), is important to enable longer reproductive periods in hominids (Beyer et al., 2011).
Although staining of mouse testis for TAp63 was originally reported to be negative (Suh et al., 2006), a very recent publication suggests that TAp63 does play a role in quality control in mouse spermatogenesis (Marcet-Ortega et al., 2017). Mice having a Trip13 mutation arrest spermatocytes in the pachytene stage of meiosis due to unrepaired DNA double-strand breaks. Crossing these mice with mice that are defective in one of the three p53 family members revealed that deficiency in either p53 or TAp63 enables spermatocytes to further progress in meiosis. In contrast, p73 deficiency has no effect (Marcet-Ortega et al., 2017). Staining of p53 protein showed only very low or no detectable levels in testis of wild-type mice; however, p53 levels are elevated during the pachytene stage of prophase I in mice carrying a Trip13 mutation. Similarly, focal staining of TAp63 was reported in wild-type mice at around the pachytene stage, which increased and was detectable at earlier stages in Trip13 mutant mice (Marcet-Ortega et al., 2017). Interestingly, this pachytene cell-cycle arrest in spermatocytes of Trip13 mutant mice can be prevented by eliminating the DSB-responsive kinase ATM or ATM-effector kinase Chk2 (Pacheco et al., 2015), suggesting that genetic quality control in mouse spermatogenesis functions similarly to that in oocytes, i.e. on the basis of an ATM−Chk2−TAp63/p53 axis. However, one important difference between the quality control in oocytes and spermatocytes is that spermatocytes are not arrested in prophase and show only transient expression of TAp63 around the pachytene stage.
Finally, a role for p53 in the quality control within spermotogenia has been reported by Yin et al. (1998). Both sexes of p53-knockout mice are fertile, although male p53−/− mice and male mice of strains with reduced p53 expression show multinucleated giant spermatocytes that are probably due to defects in meiosis (Rotter et al., 1993). p53 is also involved in the removal of damaged male germ cells. Although no staining for p53 has been observed in spermatogonia of wild-type mice, γ-irradiation of 4–5 Gy results in p53 protein stabilization (Beumer et al., 1998), consistent with the regulation of p53 by its cellular concentration (Fig. 4) (Livingstone et al., 1992). Interestingly, the dose of 4–5 Gy is 10-fold higher than that necessary to eliminate all oocytes from the mouse ovary (Suh et al., 2006). p53−/− mice also show a 50% increase of differentiating spermatogonia compared to wild type, and an increased number of giant spermatogonia following irradiation with 5 Gy (Beumer et al., 1998).
Conclusions and future perspectives
During evolution, quality control of germ cells was probably the first task of the p53 protein family, with p63 being its most-ancient member. However, isoform-specific mouse knockout studies have revealed that additional developmental tasks were acquired during evolution. One important aspect of these more-resent functions is the protection of stem cells and their surrounding cells. Interestingly, female and the male germ lines have developed different mechanisms to protect their genomic integrity, which also reflects the different strategies used by both types of germ cells. The female germ line contains quiescent stem cells and a limited number of arrested germ cells, whereas male germs cells are actively dividing resulting in the mass production of sperm cells. These differences in quality surveillance are also reflected in the fact that point mutations that are responsible for human genetic diseases disproportionately arise from the male germ line, whereas chromosomal aberrations are mostly caused by non-disjunction during the development of female germ cells. Understanding the molecular details of germ cell generation and the surveillance of their genetic integrity will undoubtedly be important to treat infertility. This is particularly relevant for cancer patients, as infertility and premature ovarian failure − which result in premature menopause − are becoming increasingly prevalent owing to higher survival rates of patients. To identify the exact molecular mechanisms by which chemotherapeutic drugs lead to apoptosis of germ cells is crucial to preserve germ cell pools during the treatment.
Funding was provided by the Deutsche Forschungsgemeinschaft [grant numbers DO 545/8-1 and EXC 115], the Centre for Biomolecular Magnetic Resonance (BMRZ), and the Cluster of Excellence Frankfurt (Macromolecular Complexes). M.T. was supported by a Fellowship from the Der Fonds der Chemischen Industrie, D.C. was supported by a Boehringer Ingelheim Fonds PhD Fellowship.
The authors declare no competing or financial interests.