PLOS GENETICS RESEARCH ARTICLE The DNA damage response is required for oocyte cyst breakdown and follicle formation in mice Ana Martı́nez-Marchal ID1,2, Yan Huang ID1,2, Maria Teresa Guillot-Ferriols ID1,2, Mònica Ferrer-Roda ID1,2, Anna Guixé1,2, Montserrat Garcia-Caldés2,3, Ignasi Roig ID1,2* 1 Unitat de Citologia i Histologia, Departament de Biologia Cel�lular, Fisiologia i Immunologia, Facultat de Biociències, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Spain, 2 Grup d’Inestabilitat i Integritat del genoma, Institut de Biotecnologia i Biomedicina, Universitat Autònoma de Barcelona, a1111111111 Cerdanyola del Vallès, Spain, 3 Unitat de Biologia Cel�lular i Genètica Mèdica, Facultat de Medicina, a1111111111 Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Spain a1111111111 a1111111111 * ignasi.roig@uab.cat a1111111111 Abstract Mammalian oogonia proliferate without completing cytokinesis, forming cysts. Within these, OPEN ACCESS oocytes differentiate and initiate meiosis, promoting double-strand break (DSBs) formation, Citation: Martı́nez-Marchal A, Huang Y, Guillot- which are repaired by homologous recombination (HR) causing the pairing and synapsis of Ferriols MT, Ferrer-Roda M, Guixé A, Garcia-Caldés the homologs. Errors in these processes activate checkpoint mechanisms, leading to apo- M, et al. (2020) The DNA damage response is required for oocyte cyst breakdown and follicle ptosis. At the end of prophase I, in contrast with what is observed in spermatocytes, oocytes formation in mice. PLoS Genet 16(11): e1009067. accumulate unrepaired DSBs. Simultaneously to the cyst breakdown, there is a massive https://doi.org/10.1371/journal.pgen.1009067 oocyte death, which has been proposed to be necessary to enable the individualization of Editor: Mónica P. Colaiácovo, Harvard Medical the oocytes to form follicles. Based upon all the above-mentioned information, we hypothe- School, UNITED STATES size that the apparently inefficient HR occurring in the oocytes may be a requirement to first Received: February 12, 2020 eliminate most of the oocytes and enable cyst breakdown and follicle formation. To test this Accepted: August 20, 2020 idea, we compared perinatal ovaries from control and mutant mice for the effector kinase of the DNA Damage Response (DDR), CHK2. We found that CHK2 is required to eliminate Published: November 18, 2020 ~50% of the fetal oocyte population. Nevertheless, the number of oocytes and follicles Copyright: © 2020 Martı́nez-Marchal et al. This is found in Chk2-mutant ovaries three days after birth was equivalent to that of the controls. an open access article distributed under the terms of the Creative Commons Attribution License, These data revealed the existence of another mechanism capable of eliminating oocytes. In which permits unrestricted use, distribution, and vitro inhibition of CHK1 rescued the oocyte number in Chk2-/- mice, implying that CHK1 reg- reproduction in any medium, provided the original ulates postnatal oocyte death. Moreover, we found that CHK1 and CHK2 functions are author and source are credited. required for the timely breakdown of the cyst and to form follicles. Thus, we uncovered a Data Availability Statement: All relevant data are novel CHK1 function in regulating the oocyte population in mice. Based upon these data, we within the manuscript and its Supporting Information files. propose that the CHK1- and CHK2-dependent DDR controls the number of oocytes and is required to properly break down oocyte cysts and form follicles in mammals. Funding: This work was supported by the Ministerio de Ciencia e Innovación (BFU2013- 43965-P and BFU2016-80370-P, PID2019- 107082RB-I00, IR; http://www.ciencia.gob.es/ portal/site/MICINN/). AMM was supported by a PIF fellowship from Universitat Autònoma de Barcelona (B16P0048; https://www.uab.cat/). YH was supported by a CSC fellowship from the China PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009067 November 18, 2020 1 / 21 PLOS GENETICS The DDR controls follicle formation in mammals Science Council (201607040048). The funders had no role in study design, data collection and Author summary analysis, decision to publish, or preparation of the manuscript. In mammals, female fertility is limited by the number of germ cells, i.e., the oocytes, females are born with. These oocytes associate with somatic cells to produce follicles that Competing interests: The authors have declared that no competing interests exist. guarantee the production of a fertilizable egg and the synthesis of sex hormones. During fetal development, oocytes initiate meiosis to halve their ploidy. As part of the meiotic program, oocytes self-induce hundreds of DSBs into their genome, which have to be repaired. However, oocytes are not very efficient at repairing DSBs. Also, during fetal development, a great part of the oocytes are eliminated, in a process that has been linked to the formation of follicles. It has been described that the effector kinase of the DDR (CHK2) is responsible for eliminating oocytes with persistent DSBs. Thus, we hypothe- sized that the DDR may be responsible for eliminating oocytes. In agreement with our hypothesis, we found that CHK2 was responsible for eliminating most fetal oocytes and regulating follicle formation in the mouse. Unexpectedly, we found that there is a CHK1- dependent perinatal oocyte death occurring in Chk2-/- mice that alters follicle formation. These data suggest that the CHK2- and CHK1-dependent DDR control the oocyte pool and follicle formation in mammals. Introduction In mice oogenesis, the primordial germ cells migrate to the genital ridges around 10.5 days post-coitus (dpc)[1], become oogonia and proliferate without completing cytokinesis, result- ing in the formation of cysts, i.e., groups of cells joined by the cytoplasm [2–4]. Within these cysts, oocytes differentiate and initiate meiosis around 13.5 dpc. But meiotic progression stops at the end of the first meiotic prophase, around birth [5]. Concomitantly, the majority of the oocytes die and cysts break down, so single oocytes can be surrounded by stromal ovarian cells to form primordial follicles [6]. These follicles containing arrested oocytes represent the pool of germ cells females will use during their entire reproductive lifetime. The massive oocyte elimination occurring during the early stages of oogenesis is not spe- cific to mice, but is common in many other mammalian species [7]. For instance, in humans, only ~10% of the oocytes that initiate meiosis will end up forming a follicle [8]. The reasons behind the effort of eliminating the vast majority of the oocytes produced are still unknown, although it has been speculated that it may help in disassembling cysts to individualize oocytes so follicles can form [9,10]. Additionally, the possible triggers of this massive oocyte death are not completely understood. The activation of the retrotransposable LINE-1 element during fetal development has been shown to trigger oocyte death [11], however, the mechanism behind it is also unknown [12]. Nonetheless, quality-control mechanisms that regulate meiotic prophase progression have been speculated to play a part in this process [7]. Meiosis is the reductional division of the genome that generates haploid cells. At the onset of the first meiotic prophase, SPO11 creates hundreds of double-stranded breaks (DSBs) on the genome, which are repaired by homologous recombination leading to the pairing and syn- apsis of the homologous chromosomes [13]. These processes are tightly regulated to avoid pos- sible deleterious effects originated by errors in recombination or synapsis. Therefore, when errors occur, meiocytes delay their cell-cycle progression, and can even activate programmed cell death [14]. Two mechanisms have been identified that are able to lead mammalian meio- cytes into apoptosis. One dependent on DSB formation and recombination progression, and a second one independent of SPO11-generated DSBs [15–19]. Studies in spermatocytes and oocytes have revealed the central role that the DNA damage-response (DDR) effector kinase, PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009067 November 18, 2020 2 / 21 PLOS GENETICS The DDR controls follicle formation in mammals CHK2, has in the arrest of mammalian meiocytes with persistent recombination intermediates [17,19]. The DDR is a complex signaling network that is responsible for the maintenance of genome integrity. Different DNA lesions activate different DDR pathways. Whereas single-stranded DNA breaks activate the Ataxia Telangiectasia and Rad3-related protein (ATR), which in turn activates the effector kinase CHK1, DSBs activate the ataxia telangiectasia mutated (ATM) kinase, which ultimately activates the effector kinase CHK2. Nonetheless, the activation of either of these two DDR pathways promotes the repair of DNA damage or the elimination of the cells which accumulate excessive damage in order to maintain genome stability [20]. Sev- eral studies have revealed the role of these two pathways in regulating meiotic recombination in mice [17,21–24]. Mammalian gametogenesis is characterized by a marked sexual dimorphism [25]. One of many examples is found in the distinct control of meiotic recombination that occurs in mam- malian oocytes, as compared with spermatocytes [22,26–28]. While spermatocytes repair most DSBs by mid-pachynema, the vast majority of human oocytes still present multiple unrepaired DSBs at the same stage. Based upon the persistence of unrepaired DSBs present in wild-type oocytes at the end of meiotic prophase, we hypothesized that the physiological oocyte death occurring during oogenesis is due to the activation of the CHK2-dependent DDR in those oocytes that accumulate more unrepaired DSBs. Furthermore, due to the co-occurrence of this massive oocyte death and cyst breakdown, we have also hypothesized that the DDR-dependent oocyte elimination contributes to proper cyst breakdown and follicle formation. To test these ideas, the number of oocytes and the cyst breakdown and follicle formation in perinatal female ovaries from control and Chk2-/- mice was analyzed. We found that during fetal development, the oocytes with higher amounts of unrepaired DSBs were eliminated in a CHK2-dependent manner, suggesting a selection process of the fittest oocytes to survive fetal development. We also showed that CHK2 is required to eliminate fetal oocytes as a result of LINE-1 activation, suggesting that the mechanism behind LINE-1-triggered oocyte death depends on DDR. Nev- ertheless, control and Chk2 mutant samples had the same number of oocytes by 4 days post- partum (dpp). Our results also provide evidence that the CHK1 function is required to eliminate Chk2-/- oocytes after birth. Moreover, cyst breakdown and follicle formation were also disrupted when the two effector kinases of DDR were inhibited, proving that the regula- tion of the oocyte population mediated by DDR is required for proper cyst breakdown and fol- licle formation in mammals. Results Chk2 mutant fetal ovaries accumulate oocytes with more unrepaired DSBs than do controls Our model predicts that in the absence of CHK2, ovaries will accumulate oocytes with high numbers of unrepaired DSBs that normally would be eliminated. To test this, the number of γH2AX patches was counted, as a surrogate of unrepaired DSBs [26], in control and Chk2 mutant oocytes at different stages of meiotic prophase. As mentioned earlier and described elsewhere [26,28], wild-type oocytes at pachynema presented multiple unrepaired DSBs. Importantly, Chk2 mutant pachytene-stage oocytes had significantly more patches of γH2AX than did control cells (p<0.0001 T-test, Fig 1, S1 Table and S1 Data). This finding is not due to a delay in the meiotic prophase progression of Chk2 mutant samples (S1 Fig). Thus, these data corroborate our model and suggest that CHK2 eliminates the oocytes with a higher number of unrepaired DSBs at pachynema. Nonetheless, as oocytes completed meiotic prophase, the dif- ference of unrepaired DSBs between the two genotypes decreased. So that, at late-diplonema, PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009067 November 18, 2020 3 / 21 PLOS GENETICS The DDR controls follicle formation in mammals Fig 1. Chk2-/- oocytes present a higher number of unrepaired DSBs at pachynema and early diplonema than do controls. (A) Representative images of control (WT) and Chk2-/- oocytes at pachynema and diplonema. The cells are immunostained against SYCP3 (red) and ɣH2AX (green). The scale bars represent 10 μm and applies to all images. (B) Quantification of the number of γH2AX patches found in WT and Chk2-/- oocytes at pachynema and diplonema. The horizontal lines represent the mean ± standard error of the mean (SEM). � 0.05