The role of p63 in germ cell apoptosis in the developing testis

JOURNAL OF CELLULAR PHYSIOLOGY 210:87–98 (2007)

The Role of p63 in Germ Cell Apoptosis in theDeveloping Testis

BETRICE PETRE-LAZAR,1,2,3 GABRIEL LIVERA,1,2,3 STEPHANIE G. MORENO,1,2,3 EMILIE TRAUTMANN,1,2,3

CLOTILDE DUQUENNE,1,2,3 VINCENT HANOUX,1,2,3 RENE HABERT,1,2,3AND HERVE COFFIGNY1,2,3*

1CEA, DSV/DRR/SEGG, Laboratoire de Differenciation et de Radiobiologiedes Gonades, F-92265 Fontenay-aux-Roses, France

2Univ Paris 7 - Denis Diderot, F-92265 Fontenay-aux-Roses, France3INSERM, U566, Unite Gametogenese et Genotoxicite,

F-92265 Fontenay-aux-Roses, France

The fetal and neonatal development of male germ cells (gonocytes) is a poorly understood but crucial process for establishingfertility. In rodents, gonocytes go through two phases of proliferation accompanied by apoptosis and separated by a quiescent periodduring the end of fetal development. P63 is a member of the P53 gene family that yields six isoforms. We detected only the p63protein and no p53 and p73 in the nucleus of the gonocytes of mouse testes. We report for the first time the ontogeny of eachp63 mRNA isoform during testis development. We observed a strong expression of p63g mRNA and protein when gonocytes are inthe quiescent period. In vitro treatment with retinoic acid prevented gonocytes from entering the quiescent period and wascorrelated with a reduced production of p63g isoform mRNA. We investigated the function of p63 by studying the testicularphenotype of P63-null mice. P63 invalidation slightly, but significantly increased the number of gonocytes counted during thequiescent period. As P63-null animals die at birth we used an original organ culture that mimicked neonatal in vivo development tostudy further the testicular development. P63 invalidation resulted in a sharply increased number of gonocytes during the cultureperiod due to a decrease in spontaneous apoptosis with no change in proliferation. P63 invalidation also caused abnormalmorphologies in the germ cells that were also found in P63þ/� adult male mice. Thus, p63 appears as an important regulator of germcell development. J. Cell. Physiol. 210: 87–98, 2007. � 2006 Wiley-Liss, Inc.

The development of the gonocytes is a complex processthat is crucial for establishing fertility in adulthood.Gonocyte proliferation and differentiation must betightly regulated as uncontrolled stem cells are poten-tially malignant. It is thought that human testiculargerm cell tumors originate from in situ carcinomaarising from gonocytes transformed during develop-ment (Skakkebaek et al., 2001). On the other hand, lowlevels of proliferation and differentiation may causesterility. Alterations to gonocyte development havebecome a major health topic as the incidence of testicularcancer in most Western countries has increased inrecent years and sperm production has decreased(Sharpe and Irvine, 2004).

The development of germ cells in mice testes follows awell-known sequence of events. First, the proliferatingprimordial germ cells (PGC) colonize the genital ridge by11.5 days post-coitum (dpc) (Eddy et al., 1981; Ginsburget al., 1990; McLaren, 1995; Yoshimizu et al., 2001). Thegonocytes are now actively proliferating and undergoapoptosis by 12.5–13.5 dpc (Wang et al., 1998; Kasaiet al., 2003). Then, from 14.5 to 19.5 dpc, the gonocytesenter a quiescent period during which mitosis andapoptosis no longer occur (Vergouwen et al., 1991;Nagano et al., 2000). After birth, the gonocytes resumemitosis at the same time as a second wave of apoptosisand start to differentiate into spermatogonia (Moriet al., 1997; Rodriguez et al., 1997; Wang et al., 1998;Boulogne et al., 1999).

The intracellular signaling pathways regulatinggonocyte development are poorly understood. Theprotein p53 is one candidate for regulating the cell cycleand apoptosis during gonocyte development. Thisprotein induces cell cycle arrest and initiates repair orapoptosis (Fei and El-Deiry, 2003; Oren, 2003).Recently, two p53 homologues, p63 and p73, have been

described (Yang et al., 1998; Kaghad et al., 1997). Allthree genes (P53, P63, P73) give rise to complex arrays oftranscripts and proteins by combining a specific N-terminus with a particular C-terminus. The P63 geneencodes six isoforms (Yang et al., 1998), whereas P53and P73 are more complex, with additional N-terminaland C-terminal splice variants (Melino et al., 2002;Bourdon et al., 2005). These genes use alternativepromoters to produce two classes of proteins havingtwo different N-termini: those containing the transacti-vation domain (full-length TA isoforms) and those withno transactivation domain (truncated DN isoforms).In vitro studies have shown that TAp63 and DNp63isoforms have different transcriptional activity, withTAs inducing cell cycle arrest and/or apoptosis, andDNs

� 2006 WILEY-LISS, INC.

Abbreviations: AMH, anti-Mullerian hormone; 3bHSD, 3b hydro-xysteroid dehydrogenase; BrdU, 5-bromo-20-deoxyuridine; dpc,days post-coitum; dpp, days post-partum; ER, estrogen receptor;GCNA1, germ cell nuclear antigen 1; IHC, immunohistochemis-try; ISEL, in situ end labeling; DN, truncated amino-terminaldomain; ND, not detected; PGC, primordial germ cells; RA,retinoic acid; TA, transactivation domain.

Contract grant sponsor: CEA, Universite Paris 7, INSERM;Contract grant sponsor: Electricite de France; Contract grantsponsor: Toxicology Nuclear Environmental program; Contractgrant sponsor: Nuclear Technical and Scientific Institute.

*Correspondence to: Dr. Herve Coffigny, CEA, INSERM U566,Universite Paris 7, DSV/DRR/SEGG/LDRG, Route du Panorama,92265 FONTENAY AUX ROSES, France.E-mail: [email protected]

Received 15 December 2005; Accepted 13 July 2006

DOI: 10.1002/jcp.20829

mostly having an opposite dominant-negative effect onTAs (Yang et al., 1998; Wu et al., 2003). Alternativesplicing also occurs at the 30 end of these genes. For P63,this splicing generates three different proteins, a, b, andg, for each group, TA and DN, that have differentpotencies (Ghioni et al., 2002; Serber et al., 2002). Thelongest are the a isoforms, which specifically contain thesterile a motif (SAM) domain that is found in proteinsinvolved in various processes undergoing protein–protein interactions, such as development, differentia-tion, and apoptosis (Westfall and Pietenpol, 2004).

Recently, TAp63 isoforms have been shown to beinvolved in apoptosis of developing neural cells and inseveral other cell lines (Gressner et al., 2005; Jacobset al., 2005). Cell death is induced by the transactivationof death receptors by p63 and the upregulation ofapoptosome constituents, such as Apaf1 and caspase-9or Bcl2 family members such as Bax or Bim (Calabroet al., 2004; Gressner et al., 2005). Similarly, theexpression of several caspases is stimulated by TAisoforms. p63 also directly upregulates members of theJag/Notch family (Laurikkala et al., 2006). Many ofthese proteins (Apaf1, Bax, Bim, and Notch) have alsobeen reported to be involved in the apoptosis ordifferentiation of postnatal germ cells (Honarpouret al., 2000; Russell et al., 2002; Hayashi et al., 2004;Coultas et al., 2005).

In adult mouse testes, P53 gene invalidationdecreases irradiation-induced apoptosis (Beumeret al., 1998) and causes abnormal growth (seminomaand teratocarcinoma) or differentiation of germ cells(giant-cell degenerative syndrome) (Rotter et al., 1993;Donehower et al., 1995). However, mice lacking P53show no male gonad abnormalities during development(Donehower et al., 1992; Moreno et al., 2002). P73�/�

mice have central nervous system defects and die afterabout 5 weeks due to chronic infections, although notestes abnormalities have been observed in these mice(Yang et al., 2000).

As P53 and P73 do not appear to be essential for testisdevelopment and as p63 expression has been previouslyreported in the developing testis (Nakamuta andKobayashi, 2003, 2004a), we wondered whetherp63 regulates testis development. P63�/� mice die wi-thin a few hours after birth due to dehydration andmaternal neglect (Mills et al., 1999; Yang et al., 1999)and their testicular phenotype has not been previouslyinvestigated.

Our study had two main aims. First, we studied theontogeny of the p63 isoforms in the developing testis andinvestigated whether this is correlated with gonocyteactivity (proliferation/apoptosis versus quiescence).Second, we analyzed the phenotype of the P63 null miceto link the expression pattern to the function of p63 inmale germ cells during testicular development.

MATERIALS AND METHODSAnimals

All animal studies (NMRI for p63 expression studies andC57BL/6 for p63 transgenic mice) were conducted in accor-dance with the NIH Guide for Care and Use of LaboratoryAnimals. Animals were housed under controlled photoperiodconditions (lights on 07:00–20:00 h) and were supplied withcommercial feed and tap water ad libitum. Males were cagedwith females overnight and the day after mating was count-ed as 0.5 dpc. Natural birth occurs at 19.5 dpc, which wascounted as 0 days post-partum (dpp). Sex of the mice wasdetermined from the anogenital distance for postnatal mice.Mice were killed by cervical dislocation.

The testes of mice (Charles Rivers, Arbresle, France) werecollected from fetuses between 11.5 and 18.5 dpc and fromnewborn mice at 1 and 3 dpp. For fetuses at 11.5 dpc, gonadmorphology was insufficiently developed to determine the sex.Therefore, we determined the sex by detecting the SRY gene byPCR (primers: F, 50GTCAAGCGCCCCATGAATGCAT30 andR, 50AGTTTGGGTATTTCTCTCTGTG30). We collected fromthese fetuses both gonads and their associated mesonephroitogether because they were tightly attached and because somegerm cells remain in the mesonephric tissue. For older fetuses(from 12.5 dpc onwards), we could morphologically distinguishthe testes from the ovaries as they were rounder and containeda testicular vessel. Neonates were killed by decapitation.

P63�/� embryos were produced by intercrossing heterozy-gous P63Brdm2 (þ/�) mice (C57BL/6 genetic background)purchased from Jackson Laboratories (Bar Harbor, ME, US).The targeted disruption of the murine P63 gene has beendescribed elsewhere (Mills et al., 1999). Briefly, a recombinantallele was generated with a truncated exon 10. If this wastranslated, the resulting transcript would give rise to aninactive protein. P63�/� mice were identified visually by theirlimbless phenotype. At 18.5 dpc, heterozygous P63 invalida-tion did not affect the body weight [1.19� 0.04 and 1.15� 0.03 gin P63þ/þ and P63þ/� mice, respectively (n¼ 25 for each)] butthis was significantly lower in P63�/� mice [0.90� 0.03 g,(n¼ 5)]. P63�/�mice also had short tails [11� 0.1 mm in P63þ/þ

(n¼ 25); 7.17� 0.5 mm in P63�/� (n¼ 5), P< 0.001]. Embryoswere genotyped by PCR using DNA extracted from tail biopsysamples. The PCR protocol was obtained from JacksonLaboratories’ technical support. The primers used to amp-lify the P63 mutant allele were: F, 50GTGTTGGCAAGGA-TTCTGAGACC30 (OIMR1029) and R, 50GGAAGACAATA-GCAGGCATGCTG30 (OIMR1030). Heterozygous intercross-ing only generated a total of 6% P63�/� male fetuses at 16.5and 18.5 dpc. This is probably due to a higher abortion rate ofP63�/� fetuses.

Organotypic culture of the fetal testis

Testes were cultured on Millicell CM filters with a 0.4 mmpore size (Fisher Scientific Labosi, Elancourt, France), aspreviously described (Habert and Brignaschi, 1991; Liveraet al., 2000). XY gonads and adjacent mesonephroi from 11.5-dpc mice were placed on the filter and cultured for 4 days. Thefilter was then floated on culture medium in tissue-culturedishes and incubated at 378C in a humidified atmospherecontaining 95% air/5% CO2. The culture medium was DMEM/HamF12 (Gibco, Cergy Pontoise, France) plus 80 ng/Lgentamicin with no serum, hormones or growth factors. Weused 300 ml of culture medium per well and the culture mediumwas replaced every day. The testes were either cultured in theabsence (control) or presence of 10�6 M all-trans retinoic acid(RA) from Sigma-Aldrich, Saint Quentin Fallavie, France(R-2625).

Testes from P63�/� and P63þ/þ mouse embryos werecollected at 18.5 dpc. For each animal, each testis was cut intotwo pieces and placed on a filter. The pieces were then culturedfor 3 days (corresponding to 2 dpp in vivo) using the sameprocedure as for 11.5 dpc testes.

RT-PCR analysis

Total RNA was extracted from tissue samples usingthe RNeasy kit (Qiagen SA, Courtaboeuf, France) accor-ding to the manufacturer’s protocol. We used random primersand Omniscript reverse transcriptase (Qiagen) to reverse-transcribe 500 ng RNA. No amplicon was obtained in reac-tions in which the reverse transcriptase was omitted. Weused b-actin as an internal control for RNA extraction andcDNA synthesis. The primers for b-actin were: F, 50AAGA-GAGGTATCCTGACCCTG30 and R, 50GGCCATCTCCTGCT-CGAAGT 30 and the annealing temperature was 508C. PCRamplification was stopped before reaching the plateau (22–24 cycles).

We used eight different primers to amplify each of the tenstudied p63 isoforms in separate reactions. All primers werefrom Invitrogen (SARL, Cergy Pontoise, France). The primersequences and the positions to which they bind in the exons

88 PETRE-LAZAR ET AL.

Journal of Cellular Physiology DOI 10.1002/jcp

specific for p63 isoforms are shown in Table 1A and Figure 1.Isoform-specific PCR conditions for each primer pair wereoptimized (Table 1B) and the reaction was stopped before theplateau was reached. The annealing temperature and numberof cycles used are shown in Table 1B. We used either thethymus from 1-dpp mice or a whole 14.5-dpc embryo as apositive control. Each experiment was repeated at least threetimes.

We determined the different levels of the various p63isoforms during testis ontogeny using the same amount ofRNA for all RT-PCR. The intensity of the bands obtained wasdetermined using densitometry with image analysis software(Bio1D). For quantification, we expressed the values obtainedfor the various p63 isoforms with respect to those for b-actin.

All ten p63 PCR amplification products were sequenced byGENOME EXPRESS (Meylan, France) using the primersdescribed above.

Protein extraction and Western blotting

Fetal and neonatal testes were resuspended in homogeniza-tion buffer (20 mM Tris base pH 8.0 (Sigma) containing 150 mMNaCl (Sigma), 0.5 mM EDTA (Sigma), 1% Triton X 100(Sigma), 0.1% sodium dodecyl sulfate (SDS) (Sigma), 10 mMsodium fluoride (NaF) (Sigma), 1 mM sodium orthovanadate(Na3VO4) (Sigma), 10 mM b glycerophosphate (Sigma), and

protease inhibitor cocktail (1 tablet/10 ml extraction solution,Roche)). Samples were placed on ice for 30 min and subject tohomogenization every 5 min. The mixture was then centri-fuged at 48C for 15 min at 11,000g to remove cellular debris,and the protein concentration in the supernatant was measur-ed using the Bradford method. Samples were then frozen at�208C until Western blotting.

We separated 70 mg of total proteins on a 10% polyacryla-mide denaturing gel using Tris/glycine/SDS as running buffer(25 mM Tris base (Sigma), 200 mM glycine (pH 8.3) (Sigma),0.1% SDS (Sigma)), and the separated proteins were blottedonto PVDFmembranes (Amersham Biosciences, Little ChalfontBuckinghamshire, England). Membranes were blocked byincubation for 1 h at room temperature in Tris-buffered saline(TBS) at pH 7.4 containing 0.05% Tween 20 (Sigma) and 5%non-fat dried milk, and then incubated overnight at 48C with amouse monoclonal anti-p63 antibody (4A4, Santa-Cruz Bio-technology) diluted 1:100 in the blocking solution. Themembranes were then incubated for 1 h at room temperaturewith a anti-mouse IgG:HRP-linked antibody diluted 1:2,000 inthe blocking solution. Antibody-protein complexes were visua-lized using the enhanced chemiluminescence’s (ECL) visuali-zation system (Amersham).

We stripped the membranes to detect the actin protein usedas reference. Briefly, the membranes were incubated for 15 min

TABLE 1. Nucleotide sequences (50–30) of the primers and conditions used for RT-PCR

A

N-terminal Core primers C-terminal

Name Sequence Name Sequence Name Sequence

P1 (F) TTAGCATGGATTGTATCCGC P3 (R) CCAGCCCCTACAACACAGA P6 (R) ACCACTCCGTGTACCCATAP2 (F) CCAGACTCAATTTAGTGAGC P4 (F) GCGTGCTGGTCCCTTATGAG P7 (R) TCAGGATTTGGCAAGTCTGA

P5 (R) AAAGCAGCAAGTATCGGACA P8 (R) CCATGGATGATTTGGCAAGT

B

RT-PCRname Set of primer T8 annealing

Numberof cycles RT-PCR name Set of primer T8 annealing

Numberof cycles

panp63 P4–P5 55 32 P63g P4–P6 57–55 38TAp63 P1–P3 57 34 TAp63a P1–P8 60–57 42DN p63 P2–P3 57–55 37 DNp63a P2–P8 57–55 42p63a P4–P8 60–57 42 TAp63g P1–P6 57–55 40p63b P4–P7 55–51 42 DNp63g P2–P6 57–55 42

Part A shows the sequences of the primers shown in Figure 1. Part B shows the set of primers and conditions of RT-PCR for each of the p63 isoforms. The set of primersused for the N-terminus: P1–P3 and P2–P3 for the TA and DN group, detect all the three C-terminal variants (a, b, g). The P4–P5 primer pairs for panp63 detect thecore p63, common to all six isoforms. The sets of primers used for C-terminus variants: P4–P8, P4–P7, P4–P6 for the a, b, g types detect the two isoforms (TA and DN).The pair of primers used for TAp63a (P1–P8), DNp63a (P2–P8), TAp63g (P1–P6), DNp63g (P2–P6) recognized only the specified isoform.

Fig. 1. Schematic representation of the p63 isoforms mRNA andthe primer pairs used for RT-PCR. Groups of exons coding forspecific protein domains are indicated. Exons are not to scale. At the50 end (gray boxes), exons 1–3 encode the transactivation domain(TA) for the full-length isoform group, whereas exon 30 encodes thetruncated (DN) isoform group. Exons 4–10 (black boxes) encode theDNA-binding domain (DBD) and the oligomerization domain (OD),which are common to all the transcripts (core). At the 30 end (whiteboxes), alternative splicing occurs for each group. The a-type is

encoded by exons 11–14 with exons 13 and 14 encoding the sterilea motif (SAM) domain. In the b-isoforms, exon 13 is spliced out. Inthe g-isoforms, exons 11–14 are spliced out, but exon 15 isspecifically expressed. The position of the primers (P1–P8) used isshown. Primers P1 and P2 are specific for the N-terminal regionsof the TA and DN isoforms, respectively. Primers binding to theregion common to all isoforms are: P3, P4, and P5. Primers P6, P7,P8 are specific for the C-terminal region of the g, b, and a type,respectively.

p63 AND GAMETOGENESIS 89


in a stripping buffer (62.5 mM Tris-HCl pH 6.8 (Sigma), 2%SDS (Sigma), 100 mM b-mercaptoethanol (Sigma)). Actinexpression was detected as described above using a mousemonoclonal anti-actin (CP01) antibody (Calbiochem MerckEurolab, Fontenay-sous-Bois, France).

Fixation and histological examination

For histology and immunohistochemistry, testes were fixedby incubation for 2 h (fetal and neonatal testes) and for 48 h(adult testes) in Bouin’s fluid. Tissues were dehydrated andwashed in a graded series of ethanol (708, 958, 1008) andbutanol solutions. The tissues were then embedded in paraffinand cut into 5-mm sections. The sections from differentages were mounted together on the same poly-L-lysine-treatedslide.

Immunohistochemistry

After paraffin removal, the sections were rehydrated ingraded alcohol solutions and stained using the avidin-biotin-immunoperoxidase method. Between steps, the sections werewashed in TBS (50 mM Tris base pH 7.4, 9% NaCl). For antigenretrieval, slides were heated in 10 mM of citrate buffer at pH6.0 in a microwave oven (750 W) for 15 min. The slides werethen incubated in 0.3 % H2O2 for 15 min, in 5% BSA in TBS for30 min and with a monoclonal anti-p63 antibody (1:50; 4A4,Santa Cruz Biotechnology) overnight at 48C. The sections werethen incubated for a further 30 min at room temperature in thepresence of biotinylated horse anti-mouse antibody followed byincubation with avidin–biotin–peroxidase complex (VectorLaboratories Inc, Burlingame, CA) for 30 min. We used 3,30-diaminobenzidine (Vector Laboratories) as the chromogen andhematoxylin as the nuclear counterstained. We used similarprotocols for anti-p53 (1:500, CM5, Novocastra, Newcastle,UK) and anti-p73 (1:100, H79, Santa Cruz Biotechnology).

The protocols for detecting germ cell nuclear antigen 1(GCNA1) and 3b-hydroxysteroid-dehydrogenase (3bHSD)have been described elsewhere (Enders and May, 1994; Liveraet al., 2000). We used polyclonal rat anti-GCNA1 antibodies ata dilution of 1:50 (provided by G. Enders) and anti-3bHSDantibodies at a dilution of 1:5,000 (provided by A. Payne).

Control sections in which the primary antibodies wereomitted showed no staining.

Identification and counting of the gonocytes

This procedure has been previously validated for rat (Olaso,1998; Livera et al., 2000) and mouse (Delbes et al., 2004) fetaland neonatal testes. We counted all the germ cells (identifiedby nuclear labeling with GCNA1) in three different sections (atone-third, one-half and two-thirds through the gonad) of testesfrom each 16.5 and 18.5-dpc mouse and cultured testes fromP63þ/þ and P63�/� mice. For each of the three sections, thenumber of gonocytes was divided by the area of the section todetermine mean gonocyte density per unit area. We thenmultiplied this density by the cumulative area of the sectionsfrom the testis to obtain the total number of gonocytes in thewhole testis called the crude count (CC). We used theAbercrombie formula (Abercrombie, 1946) to correct for anydouble counting due to a single cell appearing in two successivesections: TC¼CC�S/(SþD), where TC is the true count, S isthe section thickness (5 mm) and D is the true mean diameter ofthe cell nuclei. D is the mean nuclear diameter measured (DM)on the section, divided by p/4 to correct for overexpression ofsmaller profiles in sections through spherical particles. DM ineach testis was determined by at least 200 random measure-ments using a computerized video micrometer (MicrovisionInstruments, Evry, France).

The cumulative area of the sections from testis wasmeasured by mounting one in 10 testis sections and stainingwith hematoxylin eosin. We measured the area of each sectionusing a computerized video densitometer (Histolab; Microvi-sion Instruments) and added the areas of the sections together.The resulting value was then multiplied by 10.

Treatment with BrdU

We used the Cell Proliferation kit (Amersham Biosciences)according to the manufacturer’s protocol by incubating with1% 5-bromo-20-deoxyuridine (BrdU) solution. Briefly, testeswere cut in two pieces and incubated immediately afterremoval from the fetuses or at the end of the culture periodfor 3 h with BrdU added to the culture medium.

Sertoli cells were distinguished from germ cells using anti-Mullerian hormone (AMH) staining. For AMH immunohisto-chemistry, we used the same protocol as described above, usingan anti-AMH goat antibody diluted 1:500 (Santa CruzBiotechnology). We then incubated the sections for 30 minwith the secondary biotinylated anti-goat antibody (PK-6105,Vector Laboratories Inc). The tissue sections were washed inphosphate-buffered saline (PBS) and BrdU labeling (dilution1:100) was detected using the Cell Proliferation kit (AmershamBiosciences) according to the manufacturer’s protocol. We used3,30-diaminobenzidine for detecting AMH and the VectorAlkaline Phosphatase Substrate kit (SK-5100, Vector Labora-tories Inc) for detecting BrdU. The nuclei were counterstainedwith hematoxylin. AMH was shown as a brown color in thecytoplasm of Sertoli cells. AMH-negative cells in the tubuleswere considered to be gonocytes. Red nuclei were positive forBrdU. We determined the percentage of BrdU-positive cells bycounting 500–1,000 cells per testis.

Detection of apoptosis

Apoptosis was detected using the in situ end labeling (ISEL)method and by immunolabeling for cleaved caspase 3, aspreviously described (Moreno et al., 2001; Delbes et al., 2004).Antibodies against cleaved caspase 3 (Asp 175) was purchasedfrom Cell Signaling and we used the in situ Apoptag S7100 kit(Qbiogene, Illkirch, France) according to the manufacturer’srecommendations for the ISEL method.

We measured the diameters of stained apoptotic cells andcounted as positive only cells �2 mm, as in previous studies(Orth and Boehm, 1990), these were considered to begonocytes. The apoptosis index was obtained in vivo andin vitro by counting at least 500 germ cells per testis.

Statistics

The results are presented as means�SEM. The statisticalsignificance of the difference between two mean values wasevaluated using Student’s t-test.

RESULTSImmunohistochemistry of p63 in

the developing testis

We localized p63 using an antibody that specificallyrecognizes all the various p63 isoforms. We found thep63 protein specifically in the nucleus of germ cells at allthe studied stages (Fig. 2). We detected large amount ofp63 protein in gonocytes during the quiescent period(15.5 and 18.5 dpc, Fig. 2B,C), whereas the staining wasless intense staining in proliferating fetal (12.5 dpc,Fig. 2A) and neonatal gonocytes (1 dpp, Fig. 2D). As apositive control, we used a prepubertal testis (15 dpp) inwhich p63 was present only in the nucleus of germ cells(Fig. 2E). As previously reported, the staining was mostintense in pachytene spermatocytes but weaker in thespermatogonia (Hamer et al., 2001; Nakamuta andKobayashi, 2003; Kurita et al., 2005).

Expression of the various p63 isoformsin the developing testis

We analyzed the level of p63 isoform mRNA by RT-PCR using the primers described in Table 1 andFigure 1. We detected p63 mRNA at 12.5 dpc



(Fig. 3A,C) using primers targeting a sequence commonto all p63 isoforms (panp63). The level of p63 mRNAincreased fivefold at 15.5 dpc and then graduallydecreased. This age-related change in p63 mRNAproduction is consistent with the p63 protein patternseen by immunohistochemistry. The slow decrease inthe levels of p63 mRNA from 15.5 dpc onwards wasprobably due to a decrease in the gonocytes/Sertoli cellratio resulting from the arrest of the gonocytes while theSertoli cells continue to proliferate.

RT-PCR analysis with primers specific for the N-terminus showed that TA isoforms were produced at allthe stages studied (Fig. 3A,C). At 15.5 and 18.5 dpc, theTA isoform mRNA levels were about twice as high asthose observed at 12.5 dpc and 3 dpp. We detected thetruncated (DN) isoforms mostly during fetal stages, withtheir levels decreasing shortly after birth (1 dpp). Wedetected no obvious developmental changes in the TA/DN ratio.

RT-PCR analysis with primers specific for sequencesencoding the C-terminal end of the p63 isoforms(Fig. 3A,C) showed that mRNA levels for the p63aisoforms decreased strongly, whereas the mRNA levelsfor the p63g isoforms increased about fivefold during thechange from gonocyte proliferation to quiescence (from12.5 to 15.5 dpc). The high levels of p63g mRNA and lowlevels of p63a mRNA continued during the quiescentperiod. After birth, the level of p63g mRNA decreased,whereas the level of p63a mRNA remained unchanged.We detected p63b transcripts in the thymus but never inthe developing testis.

We investigated the change in p63 isoform profile overtime by carrying out a detailed study of the individualp63 a and g isoform expression. We focused on the

transition from the mitotic/apoptotic period of gonocytesactivity to the quiescent period, and the reverse(Fig. 3B,D). When the mitotic and apoptotic activitiesstopped (i.e., from 13.5 to 14.5 dpc), we observed a sharpdecrease in TAp63a mRNA level, whereas TAp63gmRNA level strongly increased. The mRNA levels ofthese isoforms then remained constant at 15.5 (data notshown) and 18.5 dpc (Fig. 3B,D). From the end of thequiescent period to neonatal proliferation and apoptosis(i.e., from 18.5 dpc to 1 dpp), the level of TAp63a mRNAremained low, whereas TAp63g mRNA level signifi-cantly decreased. We observed a similar change overtime for the DNp63 isoforms, with DNp63a mRNA beingpresent at 13.5 dpc and 1 dpp but totally absent from14.5 dpc until 18.5 dpc. By contrast, we detectedDNp63gmRNA during the quiescent period only. The results aresummarized in Table 2.

We also analyzed the production of p63 isoforms byWestern blotting during the course of testis develop-ment (Fig. 4). We detected only the TAp63g isoformbetween 13.5 dpc and 1 dpp, with the anti-p63 antibody(4A4) not detecting any p63a isoforms in the testisduring the same period. As a control, TAp63a wasdetected in the ovary, as previously published (Kuritaet al., 2005). We also obtained the same result with ananti-p63a specific antibody. At 13.5 dpc, the level ofTAp63g protein detected was extremely low but pro-gressively increased between 15.5 and 18.5 dpc. Afterbirth, the level of TAp63g protein decreased slightly.

Correlation between germ cell activitiesand p63 isoforms expression

We investigated the relationship between the switchfrom a to g p63 mRNA isoform and the proliferating/apoptotic status of the fetal gonocytes by preventinggonocytes from entering the quiescent stage with anin vitro treatment of retinoic acid (RA). RA is a knownpotent mitogen and is an apoptotic factor of PGC andfetal gonocytes (Koshimizu et al., 1995; Livera et al.,2000). When 11.5-dpc XY gonads were cultured incontrol medium, they spontaneously differentiated andthe germ cells progressively stopped proliferating andundergoing apoptosis, as observed in vivo, althoughwith a slight delay (Fig. 5A). However, in presence of RA,most germ cells continued proliferating and undergoingapoptosis (Fig. 5A,B). In both conditions, the number ofgerm cells per testis was similar after 4 days of culture(1495�398 in control vs. 1980� 213 in RA-treatedculture) confirming that RA increases both mitosis andapoptosis rates. We detected p63a mRNA in bothconditions, although the level was about 50% higher inRA-treated cultures than in the control cultures, butthis difference was not statistically significant(Fig. 5C,D). Under the control conditions, p63g mRNAwas easily detected by RT-PCR but was undetectable inthe RA-treated cultures (Fig. 5C,E).

Effect of p63 gene invalidation in miceon fetal testis development

First, we checked that the P63 invalidation waseffective in the testis by analyzing the expression ofTAp63g, the major transcript present in 18.5-dpc P63þ/þ

testis (Fig. 3B). We detected no TAp63g mRNA by RT-PCR in P63�/� testes at this stage (Fig. 6A).

At 16.5 and 18.5 dpc, P63�/� testes appeared normalon macroscopic observation and their volumes were notsignificantly smaller than testes from P63þ/þ fetuses

Fig. 2. Location of p63 protein in mouse testis during development.P63 was detected by IHC in the nucleus of the gonocytes (arrows)at all stages during testicular development: at 12.5 dpc (A), 15.5 dpc(B), 18.5 dpc (C), and 1 dpp (D). As a positive control for p63 detection,we used a 15-dpp testis in which the nuclear staining for p63 wasstrong in pachytene spermatocytes (double arrowheads) and weakerin spermatogonia (arrowheads) (E). In the absence of the primaryantibody, no staining was detected at 18.5 dpc (F). Scale barrepresents 10 mm.



[0.690�0.054 mm3 in P63þ/þ vs. 0.550�0.055 mm3 inP63�/� at 16.5 dpc: (n¼ 4 for each); 1.380�0.034 mm3 inP63þ/þ (n¼ 5) vs. 1.273�0.046 mm3 (n¼4) in P63�/� at18.5 dpc]. We observed no gross abnormalities inWolffian derivatives in P63�/� mice: the epididymis,ducts deferens and seminal vesicles appeared to haveformed normally (data not shown), indicating sufficienttestosterone secretion, as the development of thesederivates are androgen-dependent.

We observed no obvious change between P63�/� andP63þ/þ testes at 18.5 dpc in the Sertoli cell population,identified by immunohistochemical analysis of AMH(Fig. 6B,C), in the Leydig cell population, as determinedfrom 3bHSD expression, (Fig. 6D,E), or in the germ cellpopulation, as determined by GCNA1 staining(Fig. 6F,G). The number of gonocytes per testis tendedto increase after P63 invalidation at 16.5 dpc[5,756�415 in P63þ/þ; 6,926� 468 in P63�/�; (n¼4 foreach genotype)] and 18.5 dpc [5,867�495 in P63þ/þ;7,129� 368 in P63�/� (n¼5 for each genotype)]. This

increase in the number of gonocytes in P63�/� testes wasstatistically significant (P¼0.013) when the data from16.5 to 18.5-dpc fetuses testes were pooled. Theproliferative activity of the germ cells was similar inboth genotypes at 16.5 dpc [0 % in P63þ/þ; 0 % in P63�/�

(n¼ 4 for each genotype)] and at 18.5 dpc [0.4�0.2 % inP63þ/þ (n¼ 3); 0.5�0.1 % in P63�/� (n¼ 5)]. Sertoli cellproliferation was unchanged [37.1�1.6 % in P63þ/þ

(n¼ 4); 38.3� 1.2 % in P63�/� (n¼5)]. We also observedno apoptotic germ cells by immunostaining of cleavedcaspase 3 or by the ISEL method in the fetal testes of thetwo genotypes at 16.5 and 18.5 dpc. Thus, p63 invalida-tion did not affect the quiescent period.

Effect of p63 invalidation on p53 and p73 proteinproduction in the developing testis

We investigated whether p63 invalidation modifiesthe expression of other p53 family members byimmunohistochemical analysis of p53 and p73 in

Fig. 3. Detection of p63 isoform mRNA transcripts by RT-PCRduring testicular development in mice. RT-PCR was carried out usingthe primers and conditions presented in Table 1. Part A shows resultsfrom 12.5 dpc to 3 dpp for panp63 (all six p63 isoforms), N-terminal(TA and DN) and C-terminal p63 isoforms (a, b, and g). Part B focusedon the two transition phases of the gonocytes: entering mitotic arrestfrom 13.5 to 14.5 dpc and resuming mitosis from 18.5 dpc to 1 dpp. Inthis part, specific RT-PCR for TAp63a, DNp63a, TAp63g, and DNp63gare shown. The housekeeping gene, b-actin was used as an internalpositive control. As a positive control for p63 (þ) we used mRNA from a

1 dpp mouse thymus (A) and whole embryos at 14.5 dpc (B). Ampliconsizes (in base pairs) of each RT-PCR product are indicated on the right-hand side. Quantitative analysis of the band intensity from threedifferent experiments is shown (C, D). The amounts of p63 isoformswere referenced to that of b-actin and the ratio obtained for each agewas expressed with reference to that obtained at the earliest age atwhich the p63 isoform could be detected (12.5 dpc in C, 13.5 or 14.5 dpcin D). Columns show means�SEM. *P< 0.05 and **P< 0.01 in thepaired statistical comparison between two successive ages usingStudent’s t test [Not detected (ND)].



section from P63þ/þ and P63�/� developing testes(Fig. 7). For both proteins, we used a section from awild-type mouse adult testis as a positive control as thelocation of p53 and p73 in adult testes has previouslybeen determined at this stage (Hamer et al., 2001). p53was repeatedly detected in the spermatogonia of adulttestis (Fig. 7E). In fetuses, we observed p53 in thecytoplasm of the quiescent gonocytes (18.5 dpc) aspreviously reported (Moreno et al., 2002) and stainingwas unchanged after p63 invalidation. As P63�/�micedie within a few hours after birth (Mills et al., 1999;Yang et al., 1999), we removed the testes at 18.5 dpc(i.e., 1 day before birth) and cultured them for 3 days(D3) to investigate the effect of P63 inactivation duringmitosis and apoptosis that normally resumes just afterbirth. Staining for p53 disappeared in gonocytesresuming mitotic and apoptotic activities in bothP63þ/þ and P63�/� (Fig. 7A,C). We found p73 in thecytoplasm of gonocytes from P63þ/þand P63�/� ani-mals at every stage studied, as well as in the cytoplasmof testicular adult germ cells (Fig. 7B,D,F). In conclu-

sion, neither the expression of p53 nor p73 waschanged between P63þ/þ and P63�/� testes.

Effect of p63 invalidation on the development ofgonocytes resuming mitosis and apoptosis

To investigate the effect of p63 invalidation on themitosis and apoptosis that normally resumes after birthwe cultured 18.5 dpc testes for 3 days. On D3 of culture,the number of germ cells was 38% higher in P63�/�

testes than in P63þ/þ testes (Fig. 8A–C). The nucleardiameter of the germ cells [7.89� 0.13 mm in P63þ/þ

(n¼6); 7.91�0.05 mm in P63�/� (n¼ 4)] and the volumeof the testis [0.298� 0.014 mm3 in P63þ/þ (n¼ 6);0.268� 0.020 mm3 in P63�/� (n¼ 4)] were similar inthe two genotypes.

We investigated whether P63 deletion increasedmitotic activity and/or decreased apoptosis. After 3 daysof culture, 37% of the germ cells in P63þ/þ testisincorporated BrdU, indicating that the germ cells hadresumed their mitotic program normally in vitro. Thiswas the same after P63 invalidation (Fig. 8D–F). At endof the culture, the rate of Sertoli cell proliferation wasalso similar in both genotypes [33.8�0.6% in P63þ/þ;33.2�0.7 % in P63�/� (n¼4 for each genotype)].Apoptosis resumed spontaneously in culture, as deter-mined by the presence of cleaved caspase 3 and by ISEL-positive cells (Fig. 8G–L). Both methods of evaluation ofapoptosis showed that invalidation of P63 reduced bythree the rate of apoptosis in germ cells of culturedtestes.

Effect of p63 deletion on germcell differentiation

Although most of the gonocytes in P63�/� testesappeared normal, we observed some abnormal germcells with polymorphic nuclei (Fig. 9A,B) and manymultinucleated germ cells (Fig. 9C) that tended toassociate in clusters after culture [2.3� 0.54% in P63þ/þ

vs. 7.6�0.05% in P63�/�, (n¼3 for each genotype),P<0.01]. After culture, we also observed abnormallylarge mitoses (Fig. 9D) in the middle of the cords. Weobserved no such mitoses in P63þ/þ testes.

As P63�/�mice died at birth, we studied the histologyof adult testis from five P63þ/�mice aged 6–12 months.Adult P63þ/� testis had abnormally large germ cells(Fig. 10) in seminiferous tubules of otherwise almostnormal appearance. Some of these giant cells werelocated close to the basal membrane (Fig. 10A) but weremore frequently found in the lumen of the tubule(Fig. 10B). In one case, we observed partial atrophy ofthe germinal epithelium (Fig. 10C) in a few tubules.

DISCUSSION

This paper highlights an unusual switch in p63isoforms, with the expression of g isoforms beingspecifically increased during the quiescent period, thatis, when gonocytes no longer have apoptotic and mitoticactivities. Furthermore, we have also shown for the firsttime that p63 expression in gonocytes corresponds to anapoptotic function.

We found that p63 is located in the nucleus of thegonocytes throughout fetal development and duringneonatal life. This is consistent with the previouslydescribed distribution of p63 in the developing testis(Nakamuta and Kobayashi, 2003, 2004a). This nuclear

TABLE 2. Expression of individual p63 isoform mRNA duringgonocyte development in mice

Isoform

Phases of gonocytes development

Fetalproliferationand apoptosis

(13.5 dpc)Quiescent period(14.5–18.5 dpc)

Neonatalproliferation andapoptosis (1 dpp)

TAa þþ þ þb � � �g þ þþþþ þþ

DNa þþ � þb � � �g � þþþþ �

Summary of the level of the RT-PCR signal obtained for the various p63 isoforms.The level of expression is given as follows: high (þþþþ), moderate (þþ), low (þ)and not detected (�). The quiescent period (14.5–18.5 dpc) is correlated with p63gisoform expression in contrast to fetal (13.5 dpc) and neonatal (1 dpp)proliferation of gonocytes. P63b was not detected at any of the stages studied.The number of ‘‘þ’’s is with respect to the global signal obtained from RT-PCRthat detected several p63 isoforms (TA, DN, a, g) and RT-PCR that detectedindividual isoforms.

Fig. 4. Detection of p63 isoforms by Western blotting duringtesticular development. p63 expression was studied by Westernblotting with 13.5, 15.5, and 18.5-dpc fetal testes and 1-dpp neonataltestes using an antibody that recognizes all the p63 isoforms.Membranes were stripped to detect actin protein as a reference.Tissue from a P63�/� mouse was used as a negative control. In thedeveloping testis, a single band at about 55 kDa was detected, whichcorresponds to that expected for TAp63g. In the neonatal ovary, usedas a positive control, we detected a strong band of about 77 kDa, whichcorresponds to that expected for TAp63a.



location of p63 contrasts with the cytoplasmic location ofp53 and of p73 in fetal gonocytes. As p63 is the onlyprotein in the p53 family that is expressed in thenucleus, this suggests that p63 may play a major role inthe developing germ cells, although the cytoplasmiclocation of p53 and p73 does not fully exclude theseproteins from playing a regulatory role.

We detected DNp63 mRNA in fetal and neonataltestes, as previously described by Nakamuta et al.(Nakamuta and Kobayashi, 2003, 2004b). Unlike Naka-muta and Kobayashi (2003) who found no TAp63 mRNAin the testis between 13.5 and 15.5 dpc, we detectedTAp63 in the fetal and neonatal testes at all stagestested by both RT-PCR and Western blotting. This isimportant because it goes against the classical ideadeveloped by Nakumuta et al. that the mitotic andapoptotic arrest of gonocytes may be due to changes inthe TA/DN ratio, which is consistent with DNp63 andTAp63 often being, respectively, described as anti-apoptotic and pro-apoptotic proteins in different celltypes (Yang et al., 1998).

One of our most striking findings is the ‘‘mirror-image’’ distribution of the p63a and g isoform mRNAduring fetal gonocyte development, with p63a beingexpressed during the mitotic and apoptotic period andp63g being expressed during the quiescent period. Weobserved a very strong correlation between the increasein p63g isoform levels and the mitotic/apoptotic arrestthroughout the developmental course. Furthermore,when retinoic acid was used to inhibit entry into thequiescence stage, we observed no change from p63a to gisoforms. Finally, p63g decreased in neonates at the endof the quiescent period. As mitotic activity always occurswith apoptotic activity in gonocytes during the fetalperiod, during the neonatal period and in RA-treatedcultures, the expression of specific p63 isoforms may be

linked with either apoptosis or mitosis, or both. How-ever, it is still not known whether normal or RA-inducedchanges in p63a and g isoform production are theconsequence of or the cause of changes in mitotic/apoptotic activity in gonocytes.

Moreover, this switch between p63a and g isoformsshould be interpreted with caution as we detected onlythe TAp63g isoform by Western blotting whereas thisand TAp63a and DNp63 a and g mRNA were alsodetected by RT-PCR. This may due to the differentsensitivity of the detection methods (i.e., RT-PCR ismore sensitive). However, the p53 protein is oftenunstable unless a specific stress is applied to the cell.Similarly, it may be that p63 isoform mRNA is stable butsome p63 isoform proteins are unstable under normalphysiological conditions. A similar discrepancy betweenp63 mRNA and protein levels has already been reportedduring neural development (Jacobs et al., 2005).

Taken together, our results on p63 isoform expressionand their correlation with the proliferation/apoptosisand the quiescent periods of the gonocytes suggest thatthe coordination of p63 isoforms with these processesin germ cells is different to that in somatic cells. Ingonocytes, the level of TAp63g appears to be thedeterminant factor in these processes, rather than thebalance between p63 N-terminal isoform (TA and DN)levels, which does not change significantly duringdevelopment.

The second part of our study used p63 null mice.During fetal development, we found a number ofquiescent gonocytes slightly higher in P63�/� micethan in P63þ/þ mice although no difference wasobserved in proliferation and apoptosis. Thus, altho-ugh the level of p63g isoforms increase when the cellcycle is arrested and apoptosis is stopped, they are notessential for inducing these processes, or they may

Fig. 5. Effect of retinoic acid on testicular p63a and p63g mRNAexpression. The genotypic sex of 11.5-dpc embryos was determinedusing PCR for SRY. XY urogenital tract (gonads and its associatedmesonephroi) were set in organ culture at 11.5 dpc and cultured for 4days in the absence (C) or in the presence of 10�6 M retinoic acid (RA),a known germ cell mitogen and apoptotic factor. At the end of theculture period either BrdU was added to the medium and incubatedfor 3 h and tissues were prepared for histology, or tissues were frozenfor RNA extraction. The percentage of BrdU-incorporating germ cellswas determined before (D0) and after culture (D4) (A). Columnsrepresent the means of five different cultures�SEM (***P< 0.001 in

the unpaired statistical comparison, using Student’s t-test). Thepercentage of cleaved caspase 3 positive cells was determined at D4(B). RT-PCR for p63 a and g isoforms was carried out with threedifferent pools of four to six gonads each (C). P63g was alwaysdetectable in controls whereas it was almost undetectable in the threeRA-treated cultures. A 15.5-dpc testis (T15) was used as a control, as itexpressed both p63 isoforms. Amplicon sizes (in base pairs) of each RT-PCR product are indicated on the right-hand side. Quantitativeanalysis of the band intensity from three different experiments isshown for p63a (D) and p63g isoforms (E).



play a redundant role beside the other members of thep53 family. The invalidation of p63 probably caused anincrease in mitosis and/or a decrease in apoptosisbefore the quiescent period as the number of thegonocytes counted during these periods increasedwithout proliferation or apoptosis changing at thistime. However, we did not study apoptosis and mitosisbefore the quiescent period because the slight increasein the number of fetal gonocytes after p63 invalidationwould have led to expected changes in the percentageof ISEL or BrdU-positive cells being too small to bedetected.

In our organ culture system, in which the mitotic andapoptotic activities are resumed, we also observed asignificant increase in the number of germ cells in P63-null testes. Here, the magnitude of this increase allowedfurther investigation. It was due to decreased apoptosiswith no change in proliferation rates. As we could detectTAp63g proteins only, this isoform is probably involvedin the gonocyte apoptotic process shortly after birth.However, we cannot fully exclude that changes in otherp63 isoforms (detected only as mRNA) may also beresponsible for inducing neonatal apoptosis. The specificfunction of each isoform could only be determined byinvalidating single specific isoforms. Nevertheless,by combining in vivo gene targeting and in vitroapproaches we provide the first evidence that p63 is

involved in spontaneous apoptosis in the germ celllineage. However, the mechanism of action of p63 in thedeveloping testis remains to be determined. There aremany pro-apoptotic factors that are upregulated byTAp63g in different models (Gressner et al., 2005).Among these, the Bcl2 and the Notch families areparticularly interesting as they may also be involved inapoptosis of postnatal germ cells (Dirami et al., 2001;Russell et al., 2002).

Finally, we studied the role of p63 in germ celldifferentiation. As p63 is expressed in PGCs (Nakamutaand Kobayashi, 2004b) and with a specific pattern ingonocytes for each isoform (present study), it seemedsurprising that more defects were not observed in p63-invalidated fetal gonocytes. We observed only milddifferentiation defects in germ cells from fetal P63�/�

testes (frequently aberrant nuclei). This suggests thatp63 isoforms are not strictly required for early fetal germcell differentiation.

After culture, we frequently observed multinucleatedgerm cells in P63�/� testes. These cells were probablyresponsible for the giant mitotic cells often observed inthese P63�/� testes. They closely resembled cellsdescribed in rat neonatal testes after treatment withDBP (Fisher et al., 2003). Such cells are also observed inhuman testicular carcinoma in situ (Cortes et al., 2003),which is thought to originate from the abnormal

Fig. 7. Expression of p53 and p73 proteins in P63þ/þ and P63�/�

developing testes. p53 (A, C, E) and p73 (B, D, F) protein productionwas investigated by immunohistochemistry using specific antibodieson sections from cultured 18.5 dpc P63þ/þ (A, B) and P63�/� (C, D)testes. Wild-type adult testis was used as a positive control (E, F). Astaining was observed for p53 in the nuclei of spermatogonia(arrowheads). The cytoplasm of pachytene spermatocytes (doublearrowheads) was strongly positive for p73. After culture of 18.5 dptestes, p53 was undetectable and only p73 was detected. Scale barrepresents 10 mm and the same magnification was used in everyimage.

Fig. 6. Effect of P63 gene invalidation on fetal testis development.RT-PCR analysis of TAp63g mRNA at 18.5 dpc in P63þ/þ and P63�/�

testes. This transcript is strongly expressed in P63þ/þ testes and istotally absent from P63�/� testis (A). Markers for Sertoli cells [AMH(B, C)], Leydig cells [3bHSD (D, E)] and germ cells [GCNA1 (F, G)]were detected by IHC in P63þ/þ (B, D, F) and P63�/� (C, E, G) testes at18.5 dpc. Scale bar represents 20 mm. The same magnification wasused in every image.



Fig. 8. Effect of P63 gene invalidation on the development of germcells. P63þ/þ and P63�/� testes from 18.5 dpc fetuses (D0) werecultured for three days in organ culture (D3), fixed and processed forhistology. The pictures are for P63þ/þ mice (B, E, H, K) and P63�/�

mice (C, F, I, L) at D3. The total number of germ cells per testis wasdetermined (A) after immunolabeling of the gonocytes using theGCNA1 marker in P63þ/þ and P63�/� testis as illustrated in (B) and in(C), respectively. The percentage of BrdU-positive cells was deter-mined (D) using double IHC against AMH (brown cytoplasm) andBrdU (red nuclei) as shown for the P63þ/þ (E) and P63�/� testes (F).Inside the testicular cords, cells unstained for AMH, a specific markerfor Sertoli cells, were identified as gonocytes (indicated by asterisks).

Gonocytes labeled for BrdU are shown with black arrows andgonocytes unlabeled for BrdU are shown with white arrows. Thepercentage of cleaved caspase 3-positive cells (G) and ISEL-positivecells (J) were determined. Representative IHC for cleaved caspase 3and the ISEL method were carried out on sections from testes fromboth genotypes: P63þ/þ (H, K) and P63�/� (I, L). Stained gonocytes areshown by black arrows and unstained gonocytes by white arrows.Columns represent means�SEM. The number of tests is given inbrackets. **P< 0.01 and ***P< 0.001 in the statistical comparisonbetween P63þ/þ and P63�/� using Student’s t-test. ND, not detectable.Scale bar represents 20 mm. The same magnification was used in everyimage.

Fig. 9. Effect of P63 invalidation on germ cell differentiation duringtestis development. Germ cells were detected by IHC as GCNA1-positive cells (A) or AMH-negative cells (B, C, D). In 18.5-dpc (D0)P63�/� testes, occasional abnormal germ cells (indicated by asterisk)with unusual nuclei and chromatin distribution were seen (A, B). Inboth cases, the lower parts show an enlarged view of the nucleus and

highlight the normal nuclear structure of the germ cell (a) comparedto the abnormal germ cell nucleus (b). In 18.5-dpc P63�/� testescultured for 3 days, multinucleated germ cells organized in clusters(C) and abnormally large mitotic cells covering most of the center ofthe seminiferous cords (D) were frequently observed. Scale barrepresents 20 mm.



differentiation of fetal gonocytes (Rajpert-De Meytset al., 1998; Rorth et al., 2000), possibly after exposureto estrogens or xenoestrogens (Skakkebaek, 2002;Sharpe and Irvine, 2004). Diethylstilbestrol, an estro-gen receptor agonist has also been reported to perturbp63 expression in the Mullerian duct (Kurita et al.,2004). Taken together, these data suggest that estro-genic compounds may act on gonocyte developmentthrough a p63-dependent mechanism. Finally, the giantgerm cells observed in the seminiferous tubules of adultP63þ/� mice closely resemble those described in thetestes of Bcl2/Bcl-x-overexpressing mice and in Bax orp53-null mice (Rotter et al., 1993; Knudson et al., 1995;Rodriguez et al., 1997). This is particularly interestingbecause TAp63a and g, the only isoforms expressed inadult testes (Kurita et al., 2005), upregulate Bax anddownregulate Bcl2 in various models (Katoh et al., 2000;Gressner et al., 2005). All these germ cell abnormalitiesmay either reflect incomplete germ cell differentiationor be a consequence of the defect in the apoptoticpathway.

In conclusion, we have shown that: (i) p63 mRNAisoforms are developmentally regulated and changewith the apoptotic and mitotic activities of gonocyteswith TAp63g being the major protein isoform in thedeveloping testis and (ii) p63 is physiologically involvedin the establishment of the germ cell lineage and acts, atleast in part, by increasing germ cell apoptosis. Finally,our study suggests that p63 may play an important rolein preventing testicular lesions as apoptosis provides amechanism for removing incorrectly differentiatedgonocytes, which are thought to give rise to germ celltumors.

ACKNOWLEDGMENTS

We thank Dr B. Dutrillaux for his constant help, V.Neuville for animals care, Dr A. Mills for helpfuldiscussions, J.B. Lahaye for technical assistance withPCR experiments and Pr V. Rouiller-Fabre for criticalreading of the manuscript. We thank Dr G. Enders andDr A. Payne for generously providing the anti-GCNA1and the anti-3bHSD antibodies, respectively. We alsothank Alex Edelman & Associates for editing theEnglish manuscript. This work was supported in partby the CEA, Universite Paris 7, INSERM and by grantsfrom Electricite de France and the Toxicology NuclearEnvironmental program. B. Petre-Lazar held a fellow-ship from the Nuclear Technical and Scientific Institute,CEA Saclay, France.

LITERATURE CITED

Abercrombie M. 1946. Estimation of nuclear population from microtome sections.Anat Rec 94:238–248.

Beumer TL, Roepers-Gajadien HL, Gademan IS, van Buul PP, Gil-Gomez G,Rutgers DH, de Rooij DG. 1998. The role of the tumor suppressor p53 inspermatogenesis. Cell Death Differ 5:669–677.

Boulogne B, Olaso R, Levacher C, Durand P, Habert R. 1999. Apoptosis andmitosis in gonocytes of the rat testis during foetal and neonatal development.Int J Androl 22:356–365.

Bourdon JC, Fernandes K, Murray-Zmijewski F, Liu G, Diot A, Xirodimas DP,Saville MK, Lane DP. 2005. p53 isoforms can regulate p53 transcriptionalactivity. Genes Dev 19:2122–2137.

Calabro V, Mansueto G, Santoro R, Gentilella A, Pollice A, Ghioni P, Guerrini L,La Mantia G. 2004. Inhibition of p63 transcriptional activity by p14ARF:Functional and physical link between human ARF tumor suppressor and amember of the p53 family. Mol Cell Biol 24:8529–8540.

Cortes D, Thorup J, Visfeldt J. 2003. Multinucleated spermatogonia incryptorchid boys: A possible association with an increased risk of testicularmalignancy later in life? Apmis 111:25–30; discussion 31.

Coultas L, Bouillet P, Loveland KL, Meachem S, Perlman H, Adams JM, StrasserA. 2005. Concomitant loss of proapoptotic BH3-only Bcl-2 antagonists Bik andBim arrests spermatogenesis. Embo J 24:3963–3973.

Delbes G, Levacher C, Pairault C, Racine C, Duquenne C, Krust A, Habert R.2004. Estrogen receptor {beta}-mediated inhibition of male germ cell linedevelopment in mice by endogenous estrogens during perinatal life. Endocri-nology 145:3395–3403.

Dirami G, Ravindranath N, Achi MV, Dym M. 2001. Expression of Notchpathway components in spermatogonia and Sertoli cells of neonatal mice. JAndrol 22:944–952.

Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA, Jr., ButelJS, Bradley A. 1992. Mice deficient for p53 are developmentally normal butsusceptible to spontaneous tumours. Nature 356:215–221.

Donehower LA, Harvey M, Vogel H, McArthur MJ, Montgomery CA, Jr., ParkSH, Thompson T, Ford RJ, Bradley A. 1995. Effects of genetic background ontumorigenesis in p53-deficient mice. Mol Carcinog 14:16–22.

Eddy EM, Clark JM, Gong D, Fenderson BA. 1981. Origin and migration ofprimordial germ cells in mammals. Gamete Res 4:333–362.

Enders GC, May JJ, 2nd. 1994. Developmentally regulated expression of a mousegerm cell nuclear antigen examined from embryonic day 11 to adult in male andfemale mice. Dev Biol 163:331–340.

Fei P, El-Deiry WS. 2003. P53 and radiation responses. Oncogene 22:5774–5783.Fisher JS, Macpherson S, Marchetti N, Sharpe RM. 2003. Human ‘testicular

dysgenesis syndrome’: A possible model using in-utero exposure of the rat todibutyl phthalate. Hum Reprod 18:1383–1394.

Ghioni P, Bolognese F, Duijf PH, Van Bokhoven H, Mantovani R, Guerrini L.2002. Complex transcriptional effects of p63 isoforms: Identification of novelactivation and repression domains. Mol Cell Biol 22:8659–8668.

Ginsburg M, Snow MHL, McLaren A. 1990. Primordial germ cells in the mouseembryo during gastrulation. Development 110:521–528.

Gressner O, Schilling T, Lorenz K, Schulze Schleithoff E, Koch A, Schulze-Bergkamen H, Maria Lena A, Candi E, Terrinoni A, Valeria Catani M, Oren M,Melino G, Krammer PH, Stremmel W, Muller M. 2005. TAp63alpha inducesapoptosis by activating signaling via death receptors and mitochondria. Embo J24:2458–2471.

Habert R, Brignaschi P. 1991. Developmental changes in vitro testosteroneproduction by dispersed Leydig cells during fetal life in rats. Arch Androl27:65–71.

Hamer G, Gademan IS, Kal HB, de Rooij DG. 2001. Role for c-Abl and p73 in theradiation response of male germ cells. Oncogene 20:4298–4304.

Hayashi T, Yoshinaga A, Ohno R, Ishii N, Kamata S, Yamada T. 2004. Expressionof the p63 and Notch signaling systems in rat testes during postnataldevelopment: Comparison with their expression levels in the epididymis andvas deferens. J Androl 25:692–698.

Honarpour N, Du C, Richardson JA, Hammer RE, Wang X, Herz J. 2000. AdultApaf-1-deficient mice exhibit male infertility. Dev Biol 218:248–258.

Jacobs WB, Govoni G, Ho D, Atwal JK, Barnabe-Heider F, Keyes WM, Mills AA,Miller FD, Kaplan DR. 2005. p63 is an essential proapoptotic protein duringneural development. Neuron 48:743–756.

Kaghad M, Bonnet H, Yang A, Creancier L, Biscan JC, Valent A, Minty A, ChalonP, Lelias JM, Dumont X, Ferrara P, McKeon F, Caput D. 1997. Monoallelically

Fig. 10. Effect of inactivation of one allele of the P63 gene on the differentiation of germ cells in the adulttestis. Hematoxylin-eosin staining of testicular sections from adult P63þ/�mice at 6 months of age (A) and12 months of age (B, C). Abnormally large germ cells were seen occasionally near the basementmembrane (white arrows) and more frequently in the lumen of the seminiferous tubules (arrowheads). Inone case, partial atrophy of the germinal epithelium (black arrows) was observed with some seminiferoustubules remaining normal. Scale bar represents 50 mm. The same magnification was used in every image.



expressed gene related to p53 at 1p36, a region frequently deleted inneuroblastoma and other human cancers. Cell 90:809–819.

Kasai S, Chuma S, Motoyama N, Nakatsuji N. 2003. Haploinsufficiency of Bcl-xleads to male-specific defects in fetal germ cells: Differential regulation of germcell apoptosis between the sexes. Dev Biol 264:202–216.

Katoh I, Aisaki KI, Kurata SI, Ikawa S, Ikawa Y. 2000. p51A (TAp63gamma), ap53 homolog, accumulates in response to DNA damage for cell regulation.Oncogene 19:3126–3130.

Knudson C, Tung K, Tourtellotte W, Brown G, Korsmeyer S. 1995. Bax-deficientmice with lymphoid hyperplasia and male germ cell death. Science 270:96–99.

Koshimizu U, Watanabe M, Nakatuji N. 1995. Retinoic acid is a potent growthactivator of Mouse primordial germ cells in vitro. Dev Biol 168:683–685.

Kurita T, Mills AA, Cunha GR. 2004. Roles of p63 in the diethylstilbestrol-induced cervicovaginal adenosis. Development 131:1639–1649.

Kurita T, Cunha GR, Robboy SJ, Mills AA, Medina RT. 2005. Differentialexpression of p63 isoforms in female reproductive organs. Mech Dev 122:1043–1055.

Laurikkala J, Mikkola ML, James M, Tummers M, Mills AA, Thesleff I. 2006. p63regulates multiple signaling pathways required for ectodermal organogenesisand differentiation. Development 133:1553–1563.

Livera G, Rouiller-Fabre V, Durand P, Habert R. 2000. Multiple effects ofretinoids on the development of Sertoli, germ and Leydig cells of fetal andneonatal rat testis in culture. Biol Reprod 62:1303–1314.

McLaren A. 1995. Primordial germ cells in mammals. In: Zagris N, editor.Organization of the early vertebrate embryo. New York: Plenum Press. pp 1–9.

Melino G, De Laurenzi V, Vousden KH. 2002. p73: Friend or foe in tumorigenesis.Nat Rev Cancer 2:605–615.

Mills AA, Zheng B, Wang XJ, Vogel H, Roop DR, Bradley A. 1999. p63 is a p53homologue required for limb and epidermal morphogenesis. Nature 398:708–713.

Moreno SG, Dutrillaux B, Coffigny H. 2001. Status of p53, p21, mdm2, pRbproteins, and DNA methylation in gonocytes of control and gamma-irradiatedrats during testicular development. Biol Reprod 64:1422–1431.

Moreno SG, Dutrillaux B, Coffigny H. 2002. Study of the gonocyte cell cycle inirradiated TP53 knockout mouse foetuses and newborns. Int J Radiat Biol 78:703–709.

Mori C, Nakamura N, Dix DJ, Fujioka M, Nakagawa S, Shiota K, Eddy EM.1997. Morphological analysis of germ cell apoptosis during postnatal testisdevelopment in normal and Hsp 70-2 knockout mice. Dev Dyn 208:125–136.

Nagano R, Tabata S, Nakanishi Y, Ohsako S, Kurohmaru M, Hayashi Y. 2000.Reproliferation and relocation of mouse male germ cells (gonocytes) duringprespermatogenesis. Anat Rec 258:210–220.

Nakamuta N, Kobayashi S. 2003. Expression of p63 in the testis of mouseembryos. J Vet Med Sci 65:853–856.

Nakamuta N, Kobayashi S. 2004a. Developmental expression of p63 in the mousetestis. J Vet Med Sci 66:681–687.

Nakamuta N, Kobayashi S. 2004b. Expression of p63 in the mouse primordialgerm cells. J Vet Med Sci 66:1365–1370.

Olaso R. 1998. Role du Transforming Growth Factor b dans le controle de laspermatogenese fœtale et neonatale chez le rat [Doctorat]: Paris 7.

Oren M. 2003. Decision making by p53: Life, death and cancer. Cell Death Differ10:431–442.

Orth JM, Boehm R. 1990. Functional coupling of neonatal rat Sertoli cells andgonocytes in coculture. Endocrinology 127:2812–2820.

Rajpert-De Meyts E, Jorgensen N, Brondum-Nielsen K, Muller J, SkakkebaekNE. 1998. Developmental arrest of germ cells in the pathogenesis of germ cellneoplasia. Apmis 106:198–204; discussion 204–206.

Rodriguez I, Ody C, Araki K, Garcia I, Vassali P. 1997. An early and massivewave of germinal cell apoptosis is required for the development of functionalspermatogenesis. EMBO J 16:2262–2270.

Rorth M, Rajpert-De Meyts E, Andersson L, Dieckmann KP, Fossa SD, GrigorKM, Hendry WF, Herr HW, Looijenga LH, Oosterhuis JW, Skakkebaek NE.2000. Carcinoma in situ in the testis. Scand J Urol Nephrol Suppl 205:166–186.

Rotter V, Schwartz D, Almon E, Goldfinger N, Kapon A, Meshorer A, DonehowerLA, Levine AJ. 1993. Mice with reduced levels of p53 protein exhibit thetesticular giant-cell degenerative syndrome. Proc Natl Acad Sci 90:9075–9079.

Russell LD, Chiarini-Garcia H, Korsmeyer SJ, Knudson CM. 2002. Bax-dependent spermatogonia apoptosis is required for testicular developmentand spermatogenesis. Biol Reprod 66:950–958.

Serber Z, Lai HC, Yang A, Ou HD, Sigal MS, Kelly AE, Darimont BD, Duijf PH,Van Bokhoven H, McKeon F, Dotsch V. 2002. A C-terminal inhibitory domaincontrols the activity of p63 by an intramolecular mechanism. Mol Cell Biol 22:8601–8611.

Sharpe RM, Irvine DS. 2004. How strong is the evidence of a link betweenenvironmental chemicals and adverse effects on human reproductive health?Bmj 328:447–451.

Skakkebaek NE. 2002. Endocrine disrupters and testicular dysgenesis syndrome.Horm Res 57:43.

Skakkebaek NE, Rajpert-De Meyts E, Main KM. 2001. Testicular dysgenesissyndrome: An increasingly common developmental disorder with environ-mental aspects. Hum Reprod 16:972–978.

Vergouwen RP, Jacobs SG, Huiskamp R, Davids JA, de Rooij DG. 1991.Proliferative activity of gonocytes, Sertoli cells and interstitial cells duringtesticular development in mice. J Reprod Fertil 93:233–243.

Wang RA, Nakane PK, Koji T. 1998. Autonomous cell death of mouse male germcells during fetal and postnatal period. Biol Reprod 58:1250–1256.

Westfall MD, Pietenpol JA. 2004. p63: Molecular complexity in development andcancer. Carcinogenesis 25:857–864.

Wu G, Nomoto S, Hoque MO, Dracheva T, Osada M, Lee CC, Dong SM, Guo Z,Benoit N, Cohen Y, Rechthand P, Califano J, Moon CS, Ratovitski E, Jen J,Sidransky D, Trink B. 2003. DeltaNp63alpha and TAp63alpha regulatetranscription of genes with distinct biological functions in cancer anddevelopment. Cancer Res 63:2351–2357.

Yang A, Kaghad M, Wang Y, Gillett E, Fleming MD, Dotsch V, Andrews NC,Caput D, McKeon F. 1998. p63, a p53 homolog at 3q27-29, encodes multipleproducts with transactivating, death-inducing, and dominant-negative activ-ities. Mol Cell 2:305–316.

Yang A, Schweitzer R, Sun D, Kaghad M, Walker N, Bronson RT, Tabin C,Sharpe A, Caput D, Crum C, McKeon F. 1999. p63 is essential for regenerativeproliferation in limb, craniofacial and epithelial development. Nature 398:714–718.

Yang A, Walker N, Bronson R, Kaghad M, Oosterwegel M, Bonnin J, Vagner C,Bonnet H, Dikkes P, Sharpe A, McKeon F, Caput D. 2000. p73-deficient micehave neurological, pheromonal and inflammatory defects but lack spontaneoustumours. Nature 404:99–103.

Yoshimizu T, Obinata M, Matsui Y. 2001. Stage-specific tissue and cell interactionsplay key roles in mouse germ cell specification. Development 128:481–490.



The role of p63 in germ cell apoptosis in the developing testis

Documents