Genetic ablation of the tumor suppressor PTEN in prostatic epithelial cells (PECs) induces cell senescence. However, unlike oncogene-induced senescence, no hyperproliferation phase and no signs of DNA damage response (DDR) were observed in PTEN-deficient PECs; PTEN loss-induced senescence (PICS) was reported to be a novel type of cellular senescence. Our study reveals that PTEN ablation in prostatic luminal epithelial cells of adult mice stimulates PEC proliferation, followed by a progressive growth arrest with characteristics of cell senescence. Importantly, we also show that proliferating PTEN-deficient PECs undergo replication stress and mount a DDR leading to p53 stabilization, which is however delayed by Mdm2-mediated p53 down-regulation. Thus, even though PTEN-deficiency induces cellular senescence that restrains tumor progression, as it involves replication stress, strategies promoting PTEN loss–induced senescence are at risk for cancer prevention and therapy.
Mutations or deletion of the PTEN locus are common and associated with metastasis and resistance to therapeutic castration in prostate cancer (Cairns et al., 1997; Choucair et al., 2012; Krohn et al., 2012; Costa et al., 2015). Genetic ablation of PTEN or expression of a dominant-negative mutant of PTEN in mouse prostate epithelial cells (PECs) induces prostatic intraepithelial neoplasia (PIN) with full penetrance (Chen et al., 2005; Luchman et al., 2008; Ratnacaram et al., 2008; Papa et al., 2014).
Several studies have shown that the progression of PTEN loss–induced PINs is antagonized by cell senescence in mice (Chen et al., 2005; Alimonti et al., 2010; Di Mitri et al., 2014). Senescence is triggered in response to various stimuli (Yaswen and Campisi, 2007; Courtois-Cox et al., 2008), including the expression of oncogenes in untransformed cells (e.g., RasG12V, E2F1, Raf, Mos, Cdc6, cyclin E, Stat5, and PML; Serrano et al., 1997; Ferbeyre et al., 2000; Michaloglou et al., 2005; Mallette et al., 2007; Courtois-Cox et al., 2008). Oncogene-induced senescence (OIS), by permanently halting cell proliferation and promoting immune surveillance of premalignant lesions, is a barrier against cell transformation (Braig et al., 2005; Chen et al., 2005; Kang et al., 2011). Accordingly, markers of senescence have been observed in premalignant lesions in various human tissues, including the prostate, but not in the corresponding tumors (Chen et al., 2005; Michaloglou et al., 2005; Collado and Serrano, 2010; Vernier et al., 2011). Thus, escaping or avoiding OIS likely represents a critical step toward transformation.
The DNA damage response (DDR) pathway is a central regulator of OIS (Bartkova et al., 2006; Bartek et al., 2007; Mallette et al., 2007; Courtois-Cox et al., 2008). Indeed, the expression of oncogenes has been shown to stimulate cell proliferation, causing replication stress and a robust activation of the DDR pathway (Bartkova et al., 2006; Bartek et al., 2007), whereas inactivation of components of the DDR pathway bypasses OIS (Di Micco et al., 2006; Mallette et al., 2007). Induction of the DDR stabilizes p53 through its phosphorylation by DDR kinases (ATR, ATM, DNA-PK, CHK1, and CHK2; Zhou and Elledge, 2000; Lavin and Gueven, 2006). p53 promotes OIS through transcriptional regulation of an array of genes, including p21, an inhibitor of cell cycle progression (Mirzayans et al., 2012).
PTEN loss–induced senescence (PICS) was also shown to be p53-dependent (Chen et al., 2005), but as no hyperproliferation and DDR activation was observed (Alimonti et al., 2010; Astle et al., 2012), it was concluded that PICS is a new type of senescence (Chen et al., 2005; Courtois-Cox et al., 2008; Astle et al., 2012). Moreover, Di Mitri et al. reported that tumor-infiltrated GR-1–positive myeloid cells antagonize PICS and sustain tumor growth (Di Mitri et al., 2014).
To further characterize PICS in vivo, we analyzed PTEN(i)pe−/− mice in which PTEN is selectively ablated in prostatic luminal cells at adulthood, via the tamoxifen (Tam)-dependent Cre-ERT2 system (Ratnacaram et al., 2008). These mice develop slowly progressing PIN lesions with a highly reproducible kinetics. We took advantage of the strict temporal control of PTEN ablation in this model to characterize the fate of PTEN-deficient PECs. Our study reveals that PTEN ablation stimulates proliferation of PECs during several months, followed by a progressive growth arrest with characteristics of cell senescence. Importantly, we also show that proliferating PTEN-deficient PECs undergo replication stress and mount a DDR that stabilizes p53. However, as p53 is actively down-regulated at early time by Mdm2, cell senescence is delayed by several months.
PTEN-deficient PECs actively proliferate to generate PINs prior to becoming senescent
To study the consequences of PTEN loss in PECs of adult mice, we analyzed PTEN(i)pe−/− and PTENpe+/+ (control) mice over a period of 12 mo after PTEN ablation (Fig. S1 A). The prostate weight of PTEN(i)pe−/− mice increased during the first 3 mo after PTEN ablation to reach twice that of control mice and remained stable for the following 9 mo (Fig. 1 A). In agreement with previous results (Ratnacaram et al., 2008), the levels of pAKT S473 were enhanced in the prostate of PTEN(i)pe−/− mice, and >75% of the glands in the dorsolateral prostate (DLP) contained PINs between 1 and 12 mo (Fig. S1, B–E). The mitotic index of PECs was approximately four- to fivefold higher in PTEN(i)pe−/− mice than in control mice between 1 and 3 mo, but progressively decreased at a later time (Fig. 1 B). No terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL)-positive apoptotic cells were observed in PECs of PTEN(i)pe−/− mice (Fig. S1 F), but transcript levels of the negative regulators of cell cycle progression Cdkn1a (p21) and Cdkn2a (p16 and p19ARF) were markedly increased at 5 mo and 12 mo (Fig. 1 C). In contrast, transcript levels of p16 and p19ARF were similar in PTEN(i)pe−/− and control mice at 1 mo and slightly increased at 2 mo, whereas those of p21 were increased by 2-fold at 1 mo and by 10-fold at 2 mo (Fig. 1 C). From 2 to 12 mo, >95% PIN cells in DLP of PTEN(i)pe−/− mice displayed nuclear foci of phosphorylated HP1γ (pHP1γ; Fig. 1 D and Fig. S2), indicative of senescence-associated heterochromatin foci (SAHF; Adams, 2007). At 12 mo, most of PIN cells in PTEN(i)pe−/− mice exhibited staining for senescence-associated β-galactosidase (SA-βGal), a well-accepted marker of cell senescence (Dimri et al., 1995; Collado and Serrano, 2006), and ∼20–50% of PIN cells were SA-βGal positive at 2 and 5 mo, respectively (Fig. 1 E). In contrast, at 1 mo, only ∼7% of PEC were pHP1γ-positive in PTEN(i)pe−/− mice, and almost no SA-βGal staining was observed. Transcript levels of several members of the senescence-associated secretory phenotype (SASP; e.g., IL-1α, IL-1β, M-CSF, and TNFα) (Pérez-Mancera et al., 2014) were increased by 4- to 15-fold at 2 and 5 mo in the prostate of PTEN(i)pe−/− mice, whereas their levels were similar in control and PTEN(i)pe−/− mice at 1 mo (Fig. 2). Thus, PTEN deficiency stimulates the proliferation of PECs within 1 mo to form PINs, which become progressively senescent within the following months.
PTEN deficiency in PECs induces replication stress
Because DDR is an inducer of OIS (Bartkova et al., 2006; Di Micco et al., 2006; Mallette et al., 2007), we investigated the levels of phosphohistone H2AX S139 (γH2AX), a faithful marker of DDR activation (Rogakou et al., 1998; Mah et al., 2010). Western blot analysis showed increased levels of γH2AX in DLP extracts of PTEN(i)pe−/− mice at 1 mo (Fig. S3 A), and immunohistological analyses revealed that most PECs in DLP of PTEN(i)pe−/− mice present a nuclear staining of γH2AX, whereas no staining was observed in the prostatic epithelium of age-matched control mice (Fig. 3 A). Although most cells displayed numerous small nuclear foci, ∼5% of the cells exhibited a very intense γH2AX nuclear staining (γH2AX-high), and often, no individual γH2AX foci could be distinguished in these cells (Fig. 3 A). Importantly, γH2AX-high PECs stained positive for the proliferation marker PCNA (Fig. S3 B), and individual γH2AX foci colocalized with foci of PCNA (Fig. 3 B). In contrast, no γH2AX foci were observed in PCNA-positive PECs of age-matched control mice. Therefore, in γH2AX-high PECs, γH2AX foci are localized at the vicinity or at the site of DNA replication, which is characteristic of replication stress (Ward and Chen, 2001). In addition, labeling of DNA-replicating cells with BrdU revealed that all γH2AX-high PECs were BrdU-positive in PTEN(i)pe−/− mice (Fig. 3 C). Moreover, at 1 and 2 mo, PECs of PTEN(i)pe−/− mice displayed numerous nuclear foci of RPA32, a subunit of the single-stranded genomic DNA (ssDNA) complex replication protein A (RPA) that coats stretches of ssDNA in replication stress (Branzei and Foiani, 2005; Byun et al., 2005), although such foci were rarely observed in prostatic epithelial cells of age-matched control mice (Fig. 4 A and not depicted). RPA phosphorylation at various sites occurs at stalled replication forks, and DNA double strand breaks (DSB) produced from stalled DNA replication induce phosphorylation of RPA32 at S4 and S8 (Liaw et al., 2011; Sirbu et al., 2011). Interestingly, at 1 mo, many PEC nuclei of PTEN(i)pe−/− mice contained high levels of S4/S8 phosphorylated RPA32 (Fig. 4 B), indicating that DSB follow replication fork stalling. Numerous nuclear foci of ATR (Ataxia telangiectasia and Rad3-related) and 53BP1 (p53-binding protein 1) were also observed in PECs of PTEN(i)pe−/− mice, whereas none were observed in PECs of age-matched control mice (Fig. 4, C and D), further supporting activation of DDR signaling (Cortez et al., 2001; Lukas et al., 2011). Together these results show that replication stress-mediated DDR signaling is induced in proliferating PTEN-deficient PECs. In agreement with the DDR mediating senescence induction, discrete foci of γH2AX were present in most PIN cells of PTEN(i)pe−/− mice at 5 mo (Fig. S3 C), similar to long term γH2AX foci observed in OIS (Rodier et al., 2011).
DDR stabilizes p53 in proliferating PECs of PTEN(i)pe−/− mice.
Because p53 can be stabilized by the DDR (Meek, 2009), we immunodetected p53. Nuclear accumulation of p53 was observed in >90% PECs of PTEN(i)pe−/− mice at 2 mo, whereas almost all PECs displayed nuclear p53 at 5 and 12 mo (Fig. 5 A and Fig. S4 A). No p53-positive nuclei were detected in prostate epithelium of control mice at any time point. Therefore, PTEN ablation induced nuclear accumulation of p53 in PECs within 2 mo and up to 12 mo. In contrast, at 1 mo, only ∼6% of PECs were p53-positive in PTEN(i)pe−/− mice (Fig. 5 A and Fig. S4 A), and most of them were Ki67- and BrdU-positive (Fig. 5 B and Fig. S4 B). This is in agreement with the slightly increased expression of the p53-target gene p21 at 1 mo (Fig. 1 C) and shows that activation of p53 signaling precedes senescence.
Activation of the DDR increases p53 stability and its transcriptional activity through phosphorylation of its serine 15 (Meek, 2009). p53 pS15 nuclear staining was observed in ∼5% PECs of PTEN(i)pe−/− mice at 1 mo, and these cells also stained for the proliferation marker Ki67 (Fig. 5 C). Therefore, p53 is stabilized in proliferating PECs of PTEN(i)pe−/− mice by a DDR-mediated pathway.
p53 stabilization is counteracted by the protein kinase B (AKT)–Mdm2 axis.
To investigate the role of p53 in the control of senescence induced by PTEN loss, we administered Tam to 8-wk-old PSA-CreERT2 mice bearing LoxP-flanked PTEN and p53 alleles to generate PTEN/p53(i)pe−/− mice in which both PTEN and p53 are ablated in PECs at adulthood.
As expected, no nuclear staining for p53 was observed in PECs of PTEN/p53(i)pe−/− mice (Fig. S4, C and D). At 1 and 2 mo, the prostate weights were similar in PTEN/p53(i)pe−/− and in PTEN(i)pe−/− mice (Fig. 6 A). The mitotic index of PECs was similar in PTEN/p53(i)pe−/− and PTEN(i)pe−/− mice at 1 mo, but was higher in PTEN/p53(i)pe−/− mice than in PTEN(i)pe−/− at 2, 5, and 6 mo (Fig. 6 B). At 5 and 6 mo, the prostate weight was threefold higher in PTEN/p53(i)pe−/− than in PTEN(i)pe−/− mice (Fig. 6 A). Whereas PIN lesions in DLP were similar in PTEN/p53(i)pe−/− and in PTEN(i)pe−/− mice at 1 and 2 mo, they were more severe in PTEN/p53(i)pe−/− mice than in PTEN(i)pe−/− mice at 5 and 6 mo with, in some cases, a loss of prostate epithelium architecture (compare Fig. 6 C and Fig. S1 C). Therefore, p53 ablation does not affect early stimulation of PTEN-null PEC proliferation and PIN formation, but impairs their subsequent growth arrest.
Di Mitri et al. reported that GR1-positive myeloid cells infiltrate the prostate of PTENpe−/− mice and secrete the cytokine IL-1RA, an antagonist of IL-1R, which opposes PICS in a paracrine fashion (Di Mitri et al., 2014). FACS analysis of dissociated prostates of PTEN(i)pe−/− mice revealed that GR-1–positive myeloid cells represent ∼2.5–3% of all prostatic cells at 2 and 5 mo after PTEN ablation, but <0.07% at 1 mo in both in PTEN(i)pe−/− and control mice (Fig. 7 A). Moreover, the transcript levels of the marker of the myeloid lineage CD11b, and of Ly6g and Ly6c, the two GR-1 variants, were increased in PTEN(i)pe−/− mice at 2 and 5 mo, but not at 1 mo (Fig. 7, B–D). In agreement with these results, the transcript levels of IL-1rn, coding for IL-1RA protein, were markedly increased at 2 and 5 mo in the prostate PTEN(i)pe−/− mice, whereas they were similar in control and PTEN(i)pe−/− mice at 1 mo (Fig. 7 E). Therefore, IL-1RA–producing GR-1–positive cells cannot account for the delay of PEC senescence entry induced by PTEN ablation.
In unstressed cells, p53 is maintained at low level by proteasomal degradation through mouse double-minute 2 (Mdm2)-mediated polyubiquitylation (Rodriguez et al., 2000). Stabilization of p53 involves the dissociation of p53 from Mdm2 through modifications of p53, including phosphorylation by DDR kinases ATR or ATM. Once stabilized, p53 translocates to the nucleus, whereas Mdm2 is submitted to proteasomal degradation. At 1 mo, high levels of Mdm2 were present in the nuclei of pAKT S473–positive PECs of PTEN(i)pe−/− mice (Fig. 8 A). Moreover, Mdm2-positive cells were p53-negative, and p53 was only detected in Mdm2-low cells (Fig. 8 B). To demonstrate that Mdm2 was involved in p53 degradation at early time, PTEN(i)pe−/− mice were treated with Nutlin-3, an inhibitor of Mdm2–p53 interaction (Vassilev et al., 2004) 1 mo after PTEN ablation. Our results show that Nutlin-3a treatment increases p53 protein levels in prostatic epithelial cells and enhances the transcript levels of p53 target gene p21 (Fig. 8, C–E). Thus, low p53 levels at early time result from Mdm2-mediated p53 degradation.
Because AKT can phosphorylate Mdm2 at several serine residues (S166 and S186 in human, equivalent to S163 and S183, in mouse), inducing its nuclear translocation and accumulation (Feng et al., 2004; Wei et al., 2013), and increasing both its association and E3 ligase activity toward p53 (Ashcroft et al., 2002; Mayo et al., 2002; Ogawara et al., 2002), AKT might promote Mdm2-mediated p53 proteasomal degradation in PECs of PTEN(i)pe−/− mice. Importantly, most nuclei of PECs were Mdm2 pS163–positive at 1 mo, whereas Mdm2 pS163 was not immunodetected in normal prostate epithelium of PTEN(i)pe−/− mice and of control mice (Fig. 9 A and not depicted). The level of Mdm2 pS163 decreased, however, between 1 and 2 mo, while p53 increased and was detected in most cells at 5 mo, whereas Mdm2 pS163 was detected no more (Fig. 9 A and Fig. 5 A). Moreover, because Casein Kinase I (CKI) family members have been shown to phosphorylate N-terminal sites of p53 and to decrease p53 affinity to Mdm2, and because phosphorylation of Mdm2 by CKIδ leads to Mdm2 degradation and p53 expression (Knippschild et al., 2014), we determined their transcript levels in PTEN-deficient PINs at various time points. Interestingly, the transcript levels of CKIα, -δ, -γ, and -ε were induced 3 and 5 mo after gene invalidation, and at 5 mo, the transcript levels of CKIε correlated with p21 levels (Fig. 9, B and C; and Fig. S5).
Thus, as DNA damage has been shown to induce the interaction between Mdm2 and CKI (Inuzuka et al., 2010), it is likely that 3 mo after PTEN ablation, CKI-mediated Mdm2 degradation, and/or decreased binding of Mdm2 to p53 overcomes early Mdm2 activation and contributes to p53 stabilization.
We show here that PTEN ablation in PECs of adult mice induces PINs formation, the progression of which is antagonized by a cell senescence barrier, in agreement with previous studies (Chen et al., 2005; Di Mitri et al., 2014). Because PTEN knock-down in cultured cells was shown to induce a rapid senescence state in the absence of proliferation and DDR, in contrast to OIS (Alimonti et al., 2010; Astle et al., 2012), it was proposed that PICS is a new type of senescence. Moreover, it was shown that PICS is antagonized by IL-1RA secreted by infiltrated GR1-positive myeloid cells, allowing proliferation of a subset of PTEN-deficient PECs and, thereby, PIN formation (Alimonti et al., 2010; Di Mitri et al., 2014).
We demonstrate here that GR1-positive cells and IL-1RN are not present in the prostate 1 mo after PTEN ablation in PTEN(i)pe−/− mice and, thus, cannot contribute to the high proliferation rate and the absence of senescence markers at this stage. Therefore, we conclude that the first temporal consequence of PTEN ablation in mouse PECs is to increase their rate of proliferation, allowing PIN formation in most prostate ducts within 1 mo. Signs of senescence are observed only later, after several weeks of active proliferation.
Unlike previous studies (Alimonti et al., 2010; Astle et al., 2012), PTEN ablation in PECs induced γH2AX nuclear foci, clearly reflecting activation of the DDR. Our results are in agreement with a previous study showing that PTEN knock-down in human-immortalized fibroblasts induces cell senescence with accumulation of γH2AX foci (Kim et al., 2007). Importantly, proliferating PECs in PTEN(i)pe−/− mice displayed a very strong γH2AX staining, with numerous intense foci, and presented characteristics of replication stress, including accumulation of foci of RPA32 and RPA32 pS4/S8, as well as ATR, the DDR kinase activated by replication stress and a major effector of replication stress–mediated DDR (Mazouzi et al., 2014).
ATR can phosphorylate p53 S15 directly or via activation of CHK1 (Mazouzi et al., 2014), and our results show that actively proliferating cells exhibit a nuclear staining for p53 pS15 and that p53 mainly accumulates in Ki67-positive and BrdU-positive PECs. Therefore, p53 is stabilized by replication stress–mediated DDR activation in PTEN-deficient PECs, and replication stress–mediated p53 activation contributes to trigger senescence, as in OIS (Bartkova et al., 2006; Bartek et al., 2007).
In addition, we show that Mdm2 accumulates in the nuclei of AKT pS473–positive PECs at 1 mo and that nuclear Mdm2 is phosphorylated at S163. It is well documented that Mdm2 can be phosphorylated at S163 by AKT, resulting in its translocation in the nucleus, while it increases its E3-ligase activity toward p53 and thus down-regulates p53 (Mayo and Donner, 2001; Mayo et al., 2002; Ogawara et al., 2002; Feng et al., 2004; Fenouille et al., 2011). It has been shown that PTEN ablation in mouse cells and tissue results in a dramatic decrease of p53 expression (Freeman et al., 2003), and pharmacological inhibition of PI3K prevents p53 activation by DNA damage (Bar et al., 2005). Moreover, activation of AKT by Her2 in breast cancer cells increases Mdm2-mediated p53 degradation via AKT-mediated Mdm2 phosphorylation on S163 (Zhou et al., 2001). Importantly, we show that early p53 degradation in PTEN-deficient PECs is mediated by Mdm2 and provide evidence that CKI contributes to p53 stabilization at later time. We thus propose that PTEN loss has two simultaneous and opposed consequences. Besides initiating a senescence program through replication stress mediated activation of the DDR and p53 stabilization, AKT activation induced by PTEN ablation contributes to delay senescence via enhanced Mdm2-mediated p53 degradation.
An important characteristic of our mouse model is the possibility to strictly control the time of floxed genes ablation by induction of the CreERT2 recombinase activity. In other models, PTEN ablation occurs in the undifferentiated prostate of young mice (Chen et al., 2005; Alimonti et al., 2010), because of the early activity of the Pb-4 promoter (in the prostate bud of the newborn) that drives the expression of a constitutively active Cre recombinase (Wu et al., 2001). As we induced Cre-mediated recombination selectively in luminal prostatic epithelial cells after puberty, PTEN was ablated in a fully developed prostate and well-differentiated PECs. It was shown that PIN development is accelerated when PTEN is ablated before puberty (Luchman et al., 2008). Thus, the lower proliferation rate of PECs in our model, which reaches a maximum of ∼10%, although it reaches ∼40% in previously studied models (Chen et al., 2005; Alimonti et al., 2010), might contribute to the slow establishment of cell senescence, as replication stress might be less intense.
Our results diverge from several published studies on PICS. The discrepancy between our observations made in vivo in epithelial cells and in vitro results obtained in nonepithelial cells (Alimonti et al., 2010; Astle et al., 2012) might result from the obvious growth conditions differences. Indeed, in vitro cell culture conditions induce cellular stress, because partial O2 pressure is much higher in culture medium than in tissues, and because high levels of growth factors push the cells to actively proliferate (Serrano and Blasco, 2001). Thus, cells in culture might induce pathways to cope with such stress conditions. In any event, our data highlight that in vitro data on senescence should not be extrapolated to in vivo situations.
As we demonstrate here that PTEN loss–induced cell senescence in prostate epithelial cells is mediated at least in part by replication stress after a phase of enhanced proliferation in vivo, this type of senescence resembles other types of senescence, in particular OIS, in contrast to previous conclusions drawn from others (Alimonti et al., 2010). Thus, approaches for cancer prevention and therapy based on PICS induction are at risk, because replication stress induced by PTEN loss might result in an accumulation of mutations, including of p53, before p53 is stabilized and senescence initiated, and/or in senescent cells, leading to senescence escape and formation of adenocarcinoma. Supporting this idea, p53 mutations cooccur with PTEN mutations in a high proportion of advanced prostate tumors (Chen et al., 2005; Lotan et al., 2011; Markert et al., 2011).
In contrast, selective inhibitors of AKT or of Mdm2-p53 interaction might enforce senescence and prevent or delay prostate cancer progression. However, the safest approach remains probably the elimination of senescent cells.
Materials and methods
Mice were maintained in a temperature- and humidity-controlled animal facility with a 12 h light/dark cycle. Breeding and maintenance of mice were performed in the accredited IGBMC/ICS animal house (C67-2018-37), in compliance with French and EU regulations on the use of laboratory animals for research, under the supervision of D. Metzger who holds animal experimentation authorizations from the French Ministry of Agriculture and Fisheries (Nos. 67-209 and A 67-227). All animal experiments were approved by the Ethical committee Com’Eth (Comité d’Ethique pour l’Expérimentation Animale, Strasbourg, France). Animals were euthanized with carbon dioxide and cervical dislocation, and tissues were immediately collected, weighed and frozen in liquid nitrogen, or processed for biochemical and histological analysis.
Generation of mouse cohorts
PTEN(i)pe−/− mice were generated as described (Ratnacaram et al., 2008). In brief, mice carrying one copy of the PSA-Cre-ERT2 transgene, expressing the Tam-inducible Cre-ERT2 recombinase in prostate epithelium under the control of the human PSA promoter, were crossed with mice carrying LoxP-flanked (floxed) alleles of PTEN (L2 allele; a gift from T. Mak, Campbell Family Institute for Breast Cancer Research, Toronto, Canada, and A Suzuki, Akita University School of Medicine, Akita, Japan) to generate PSA-Cre-ERT2(tg/0)/PTENL2/L2 (tg, transgenic) and PSA-Cre-ERT2(0/0)/PTENL2/L2 mice. PTEN/p53(i)pe−/− mice were generated by intercrossing PSA-Cre-ERT2(tg/0)/PTENL2/L2 mice with p53L2/L2 mice carrying floxed alleles of p53 (a gift of A. Berns, The Netherlands Cancer Institute, Amsterdam, Netherlands; Jonkers et al., 2001) to generate PSA-Cre-ERT2(tg/0)/PTENL2/L2/p53L2/L2 and PSA-Cre-ERT2(0/0)/PTENL2/L2/p53L2/L2 mice. Gene ablation was induced by intraperitoneal injection of Tam performed daily for 5 d (1 mg/mouse) to 8- to 10-wk-old mice, as described (Fig. S1 A; Metzger et al., 2005), to generate mutant PTEN(i)pe−/− and PTEN/p53(i)pe−/− mice. Respective control mice (PTENpe+/+ and PTEN/p53pe+/+) did not bear the PSA-Cre-ERT2 transgene (pe, prostate epithelium; (i), induced).
PSA-Cre-ERT2, PTENL2/L2, and p53L2/L2 mice were backcrossed on C57BL/6 mice for more than eight generations before intercrossing. Mice were genotyped by PCR performed on genomic DNA isolated from ear biopsies, using the DirectPCR extraction kit (102-T; Viagen) and primers as described (Jonkers et al., 2001; Ratnacaram et al., 2008).
Treatment of mice
Control and PTEN(i)pe−/− mice were administered i.p. with 5-bromo-2'-deoxyuridine (BrdU, Sigma-Aldrich) at 50 mg/kg in 100 µl of sterile saline solution (0.9% NaCl), 3 h before sacrifice. 1 mo after gene ablation, PTENpe−/− mice were treated (i.p.) with Nutlin-3 (daily 100 mg/kg; AdooQ Biosciences) or vehicle for 2 d and sacrificed 4 h after the last administration.
Prostate tissue samples were immediately fixed in ice-cold 4% formaldehyde supplemented with 1 tablet/10 ml of PhoSTOP (04 906 837 001; Roche). Prostate samples were embedded in paraffin, and 5-µm serial sections were cut. For histopathological analyses, paraffin sections were stained with H&E. For immunofluorescence staining, sections were processed as previously described (Ratnacaram et al., 2008), except that sections were incubated over night with primary antibodies diluted 1:200, unless indicated. Primary antibody used for immunofluorescence were directed against AKT pS473 (4060; Cell Signaling Technology), γH2AX (05-636; EMD Millipore; 1:600), p53 (CM5; Vector Labs), p53 pS15 (12571; Cell Signaling Technology), pHP1γ (2600; Cell Signaling Technology), RPA32 (GTX113004; Tebu-Bio), RPA32 pS4/S8 (A300-245A; Bethyl Laboratories), ATR (2790; Cell Signaling Technology), Ki67 (M7248; Dako), BrdU (600-401-C29; Rockland), ATM pS1966 (200-301-400; Rockland), Mdm2 (sc-965; Santa Cruz Biotechnology), Mdm2 pS166 (human)/pS163 (mouse; 3521; Cell Signaling Technology), PCNA (ab2426; AbCam), 53BP1 (NB100-305; Novus Biologicals). Secondary antibodies (CY3 AffiniPure goat anti–rabbit IgG [H+L], CY3 AffiniPure goat anti–mouse IgG [H+L], CY5 AffiniPure goat anti–mouse IgG [H+L], and CY5 AffiniPure donkey anti–rabbit IgG [H+L]) were from Jackson Immunoresearch.
10-μm frozen prostate sections were cut, fixed in 2% formaldehyde and 0.2% glutaraldehyde, stained in 100 mM K3Fe(CN)6, 100 mM K4Fe(CN)6, 2 mM MgCl2, 150 mM NaCl, citric acid phosphate buffer (0.2 M Na2HPO4 and 0.1 M Citric acid), and 1 mg/ml X-gal (Roche) for 6 h and counterstained in hematoxylin (Dimri et al., 1995).
10-μm paraffin-embedded prostate sections were cut and stained with TUNEL kit (11684817910; In Situ Cell Death Detection kit; Roche) following manufacturer’s instructions.
Microscope image acquisition
Microscopic analyses were performed at 21°C using a Leica Microsystems DM4000B with the following objectives: 10×, N Plan, dry, numerical aperture (NA): 0.25; 20×, HC Plan Apo dry, NA: 0.7; 40×, N Plan, dry, NA: 0.65; 63×, HCX Plan Apo, oil immersion, NA: 1.32–0.6; and 100, HCX PL Fluotar, oil immersion, NA: 1.30–0.6. Images were acquired with a 12-bit Photometrics Cool Snap FX camera by the Micro-Manager software. Pictures were edited with the Fiji software (Schindelin et al., 2012). The Cell Counter Fiji plugin was used for quantification of pathological features and histological markers.
RNA extraction and analysis
RNA was isolated from DLP and anterior prostate samples, and quantitative PCR of retrotranscribed RNAs were performed as described (Gali Ramamoorthy et al., 2015). Primer sequences were as follows: p16 forward: 5′-GAACTCTTTCGGTCGTACCC-3′, reverse: 5′-CAGTTCGAATCTGCACCGTA-3′; p19ARF forward: 5′-GCTCTGGCTTTCGTGAACAT-3′, reverse: 5′-GTGAACGTTGCCCATCATC-3′; p21 forward: 5′-TCTTCTGCTGTGGGTCAGGAG-3′, reverse: 5′-GAGGGCTAAGGCCGAAGATG-3′; IL-1α forward: 5′-AGACGGCTGAGTTTCAGTGAG-3′, reverse: 5′-TAAGGTGCTGATCTGGGTTG-3′; IL-1β forward: 5′-ACGACAAAATACCTGTGGCC-3′, reverse: 5′-TGGGTATTGCTTGGGATCCA-3′; M-CSF forward: 5′-GACCCAGGATGAGGACAGAC-3′, reverse: 5′-TTCCCATATGTCTCCTTCCA-3′; TNFα forward: 5′-CACTTGGTGGTTTGCTACGA-3′, reverse: 5′-CCCCAAAGGGATGAGAAGTT-3′; IL-1rn forward: 5′-TGAGCTGGTTGTTTCTCAGG-3′, reverse: 5′-GAAAAGACCCTGCAAGATGC-3′; Ly6c forward: 5′-ATAGCACTCGTAGCACTGCA-3′, reverse: 5′-ACCTTGTCTGAGAGGAACCC-3′; Ly6g forward: 5′-TGTGCTCATCCTTCTTGTGG-3′, reverse: 5′-AGGGGCAGGTAGTTGTGTTG-3′; CD11b forward: 5′-CAAAGAACAACACACGCAGG-3′, reverse: 5′-GGCTCCCCAACCAGTGTATA-3′; CKIε forward: 5′-TGAAGCATGGAGTTGCGTGT-3′, reverse: 5′-TGTTGGCACCCAGGTAGATG-3′; CKIα1 forward: 5′-TCCGCATCCTTTTCAGGACC-3′, reverse: 5′-AGAAACCTGTGGGGGTTTGG-3′; CKIδ forward: 5′-TACCTCACACACGGCCAACA-3′, reverse: 5′-GCAGCCGCATACTCACTTTC-3′; CKIγ1 forward: 5′-GGGTGGTTGAGAGAAGCAGT-3′, reverse: 5′-CCTGTCCTCCTAAATCAGCGG 3′; CKIγ2 forward: 5′-GCTATCAAGCTGGAGCCCAT-3′; reverse: 5′-TAGTAGACCTGAGGGACGCC 3′; CKIγ3 forward: 5′-TGGCAAGGCTTAAAGGCTGA-3′; reverse: 5′-TGGGAAGTTCTCACACAACACT-3′; 18S forward: 5′- TCGTCTTCGAAACTCCGACT-3′, reverse: 5′-CGCGGTTCTATTTTGTTGGT-3′.
Preparation and analysis of prostate tissue protein extracts
Prostates were crushed in ice-cold modified radioimmunoprecipitation assay buffer (50 mM Tris, pH 7.5, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 5 mM EDTA, 5% glycerol, supplemented with protease inhibitor cocktail [05892970001; Sigma-Aldrich], PhoSTOP [PHOSS-RO; Sigma-Aldrich] and 1,4-dithiothreitol 10 mM) with a potter. Protein extracts (25 µg) were electrophoresed on 8–15% SDS-polyacrylamide gels and electroblotted to Hybond nitrocellulose membranes (Amersham Biosciences). Proteins were detected using primary antibodies directed against γH2AX (05-636; EMD Millipore; 1:1,000), AKT (4691; Cell Signaling Technology; 1:1,000), AKT pS473 (4060; Cell Signaling Technology; 1:1,000), and tubulin (TUB-2A2; IGBMC; 1:10,000). Membranes were probed with secondary HRP-conjugated secondary antibodies (Amersham Biosciences; 1:10,000), which were revealed using an enhanced chemiluminescence detection system (Pierce).
Fresh prostates were chopped with a scalpel blade immediately after dissection, and dissociated for 4 h in 1 ml/50 mg of prostate tissue of dissociation buffer (RPMI-1640 cell culture medium [Thermo Fisher Scientific] supplemented with 10% heat-inactivated FBS, Hepes, pH 7.2, 10 mM, 200 µg/ml hyaluronidase [H3506; Sigma-Aldrich], 2.5 mg/ml collagenase [C0130; Sigma-Aldrich], and 25 mM CaCl2) at 37°C with gentle shaking. After centrifugation (400 g; 5 min), pellets were washed twice with 10 ml of PBS and resuspended in 2 ml of prewarmed 0.05% trypsin (Thermo Fisher Scientific) supplemented with 5 mM EDTA. After a 5 min incubation at 37°C with gentle shaking, trypsin was inactivated by addition of RPMI-1640 supplemented with 10% heat-inactivated FBS, Hepes, pH 7.2, 10 mM, and 300 µg/ml DNase I (DN25; Sigma-Aldrich). Remaining cell clumps were physically dissociated by 10 passages through an 18-G needle and 10 passages through in a 20-G needle. After filtration on a 40-µm cell strainer, cell suspensions were centrifuged at 400 g for 5 min, washed twice in ice-cold PBS, and resuspended in FACS buffer (PBS supplemented with 5 mM EDTA and 1% heat-inactivated FBS).
Cells were incubated with anti-CD16/32 antibodies (BD PharMingen; 1:50) for 15 min on ice to block nonspecific binding sites to Fc receptors, stained with antibodies directed against Epcam (PE-Cy7; Biolegend), CD45 (PerCP-Cy5.5; eBiosciences), CD11b (BV421; BD biosciences), and GR1 RB6-8C5 (FITC; eBiosciences; 1:50 each) for 15 min on ice, and analyzed on a BD LSR II Flow Cytometer (IGBMC; cytometry service) with FlowJo software. The presence of GR1-positive cells was determined by gating CD11bhighGR1high cells in EpcamlowCD45high cells. The proportion of CD11bhighGR1high cells was determined as the number of CD11bhighGR1high relative to the total number of single cells of dissociated prostate cells that were analyzed.
Statistical analysis was performed with the one-way ANOVA test with Prism (GraphPad).
Online supplemental material
Fig. S1 shows the generation and characterization of PTEN(i)pe−/− mice. Fig. S2 shows pHP1γ immunodetection in PEC of PTEN(i)pe−/− mice. Fig. S3 shows DDR marker immunodetection of in PEC of PTEN(i)pe−/− mice. Fig. S4 shows p53 expression PEC of PTEN(i)pe−/− mice and efficient ablation of p53 in PTEN/p53(i)pe−/− mice. Fig. S5 shows CKIα1, -δ, -γ1, -γ2 and -γ3 transcript levels in the prostate of PTEN(i)pe−/−, and control mice at various time points after gene ablation.
We thank T. Mak and A. Suzuki for floxed PTEN mice and A. Berns for floxed p53 mice; the staff of the mouse, histopathology, cell sorting, and imaging facilities from Institut de Génétique et de Biologie Moléculaire et Cellulaire and Institut Clinique de la Souris for excellent assistance; P. Kessler and G. Laverny for helpful discussions; and E. Weiss and B. Keyes for critical reading of the manuscript.
This work was supported by funds from the Centre National de la Recherche Scientifique, the Institut National de la Santé et de la Recherche Médicale, the Université de Strasbourg, the Fondation ARC pour la Recherche sur le Cancer, the Ligue Contre le Cancer, Alsace Contre le Cancer, the Association pour la Recherche sur les Tumeurs de la Prostate, the Centre d’Ingénierie Moléculaire Européen, and by French state funds through the Agence Nationale de la Recherche grant ANR-10-LABX-0030-INRT under the frame program Investissements d’Avenir labeled ANR-10-IDEX-0002-02. M. Parisotto was supported by the Association pour la Recherche à l’IGBMC (ARI), E. Grelet by the Ministère de l’enseignement supérieur et de la recherché, R. El Bizri by the Al Bizri Foundation, Y. Dai. by the China Scholarship Council, J. Terzic by the Ecole de l’Institut National de la Santé et de la Recherche Médicale Liliane Bettencourt, and D. Eckert by the Ligue Contre le Cancer.
The authors declare no competing financial interests.
Author contributions: D. Metzger conceived the study together with M. Parisotto; M. Parisotto, E. Grelet, R. El Bizri, Y. Dai, J. Terzic, D. Eckert, L. Gargowitsch, and J.-M. Bonert performed experiments; D. Metzger and M. Parisotto analyzed data and wrote the manuscript.
E. Grelet and R. El Bizri contributed equally to this paper.
M. Parisotto’s current address is Université de Montréal, Montréal, Québec, Canada.