Conditional E2F1 activation in transgenic mice causes ... - Nature

Oncogene (2005) 24, 780–789

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Conditional E2F1 activation in transgenic mice causes testicular atrophy and dysplasia mimicking human CIS Karl Agger1,2,3, Eric Santoni-Rugiu4, Christian Holmberg2, Olle Karlstro¨m2 and Kristian Helin*,1,3 1 Biotech Research & Innovation Centre, 2100 Copenhagen, Denmark; 2Department of Molecular Cell Biology, Institute of Molecular Biology, University of Copenhagen, 1353 Copenhagen, Denmark; 3European Institute of Oncology, Department of Experimental Oncology, 20141 Milan, Italy; 4Department of Pathology, Rigshospitalet, Copenhagen University Hospital, 2100 Copenhagen Denmark

E2F1 is a crucial downstream effector of the retinoblastoma protein (pRB) pathway. To address the consequences of short-term increase in E2F1 activity in adult tissues, we generated transgenic mice expressing the human E2F1 protein fused to the oestrogen receptor (ER) ligand-binding domain. The expression of the ERE2F1 fusion protein, which is inactive in the absence of 4-hydroxy tamoxifen (OHT), was targeted to the testes. We show that short-term activation of E2F1 results in activation of E2F target genes and apoptosis of germ cells. Consistent with our previously published results, the apoptotic response was independent of p53. Persistent E2F1 activation for 3 weeks led to massive apoptosis and severe testicular atrophy with seminiferous tubules containing only Sertoli cells and clusters of undifferentiated spermatogonia. The latter showed high expression of ERE2F1 and excessive mitotic activity, including atypical mitoses. In addition, gonocyte-like dysplastic germ cells, resembling carcinoma in situ (CIS) cells in humans, appeared. Our results show that a relatively short period of deregulated E2F1 activity in testicles can induce premalignant changes. Moreover, we demonstrate the feasibility of tissue-specific expression of conditional ERE2F1 in transgenic mice. Oncogene (2005) 24, 780–789. doi:10.1038/sj.onc.1208248 Published online 8 November 2004 Keywords: neoplasia

E2F; retinoblastoma; testis; apoptosis;

Introduction Some of the best-characterized downstream targets of pRB are the E2F transcription factors. The E2Fs are key regulators of genes required for DNA replication and cell cycle control (Muller and Helin, 2000; Trimarchi and Lees, 2002; Cam and Dynlacht, 2003). *Correspondence: K Helin, Biotech Research & Innovation Centre, Fruebjergvej 3, DK-2100 Copenhagen, Denmark; E-mail: [email protected] Received 29 April 2004; revised 20 August 2004; accepted 5 October 2004; published online 8 November 2004

When associated with pRB and pRB family members, the E2Fs function as active repressors of transcription. The pRB/E2F complex is disrupted when pRB is phosphorylated in late G1 by cyclin-dependent kinases resulting in transcriptional activation of E2F target genes and entry into S phase (Muller and Helin, 2000; Trimarchi and Lees, 2002; Cam and Dynlacht, 2003). The E2F family also plays an important role in the regulation of genes involved in apoptosis, differentiation and development (Muller et al., 2001). The RB1 gene is a prototypic tumour suppressor and the pRB pathway is disrupted in most, if not all, human cancers (Sherr and McCormick, 2002). Mice with one inactivated allele of RB1 develop pituitary tumours with almost 100% penetrance (Clarke et al., 1992; Jacks et al., 1992; Lee et al., 1992; Nikitin and Lee, 1996). Similarly when pRB is targeted by viral oncoproteins in transgenic mice, the mice develop tumours in the target tissue (Symonds et al., 1994). When these mice strains are crossed into an E2F1-deficient background, there is a significant reduction in the tumourigenic potential (Pan et al., 1998; Yamasaki et al., 1998), emphasizing the importance of E2F1 as a downstream effector of the pRB pathway. Seven different E2Fs have been described so far (Trimarchi and Lees, 2002; de Bruin et al., 2003; Di Stefano et al., 2003). E2F1–E2F6 interact with one of the two known DP proteins and this association is necessary for high-affinity binding to DNA. E2F7, the most recently isolated member of the family, contains two DNA-binding domains and associates with the E2F DNA-binding consensus site independently of DP association (de Bruin et al., 2003; Di Stefano et al., 2003). The biological properties of the E2F–DP complex appear to be determined by the E2F moiety. E2F1, E2F2 and E2F3 represent one subgroup, which preferentially binds to pRB and less so to the two pRBrelated proteins p107 or p130 (Muller and Helin, 2000; Trimarchi and Lees, 2002). Acute inactivation of pRb in quiescent mouse fibroblasts is sufficient for re-entry into the cell cycle (Sage et al., 2003), and overexpression of E2F1, 2 or 3 in immortalized quiescent fibroblasts recapitulates this effect, driving the cells into S phase in the absence of serum (Johnson et al., 1993; Lukas et al., 1996; Lomazzi et al., 2002). Mouse embryonic

Apoptosis and neoplasia in ER-E2F1 transgenic mice K Agger et al

781

fibroblasts lacking E2F1–E2F3 fail to proliferate and show a dramatic decrease in the expression of several E2F target genes, suggesting that these three E2Fs are required for transactivation of E2F target genes essential for S phase entry (Wu et al., 2001). In contrast, E2F4 and E2F5 appear mainly to work as active repressors of transcription, and mice lacking either of the two proteins show specific differentiation defects (Lindeman et al., 1998; Humbert et al., 2000; Rempel et al., 2000). Overexpression of these proteins does not induce S phase in the absence of serum (Lukas et al., 1996). E2F6 has repressor properties but is unique among the DP-binding E2Fs, since it does not bind to any of the pocket proteins and it does not have a transactivation domain (Morkel et al., 1997; Cartwright et al., 1998; Gaubatz et al., 1998; Trimarchi et al., 1998). E2F7, like E2F6, represses transcription from E2F target genes (de Bruin et al., 2003; Di Stefano et al., 2003). Since disruption of the pRB pathway ultimately leads to increased E2F activity, we asked whether overexpression of E2F1 in vivo could mimic loss of pRB. Several studies have shown that overexpression of E2F1 in transgenic mice can induce apoptosis and have oncogenic effects. E2F1 overexpression in the skin of transgenic mice is sufficient to induce skin tumours, when either coexpressed with H-RAS or in the absence of p53 (Pierce et al., 1998a, b). Moreover, overexpression of E2F1 in the mouse liver is sufficient to induce liver tumours within 12 months of age (Conner et al., 2000). Previously, we showed that overexpression of E2F1 in mouse testes induces p53-independent apoptosis and testicular atrophy (Holmberg et al., 1998). However, these studies employed constitutive overexpression of E2F1. As somatic mutation is the typical source of oncogene activation, we sought to generate transgenic mice with conditional tissue-specific E2F1 activity. This is a particularly important strategy in the analysis of E2F1’s oncogenic potential, since the apoptotic effect of E2F1 overexpression may interfere with proper development. Conditional E2F1 overexpression has previously been achieved in tissue culture models by means of fusion to the ligand-binding domain of the human oestrogen receptor (ER) resulting in the chimeric protein, ER-E2F1 (Vigo et al., 1999). The ER domain has been mutated in a way that renders it unresponsive to endogenous oestrogens, yet it is readily activated by the oestrogen analogue 4-hydroxy tamoxifen (OHT) (Littlewood et al., 1995). Upon exposure to OHT, ER-E2F1 is released from an inactive complex in the cytoplasm and translocates to the nucleus to activate target genes (Vigo et al., 1999). Here, we show that the ER-E2F1 system is functional in transgenic mice, and that induction of E2F1 activity in the testis for 48 h resulted in massive apoptosis of germ cells independently of p53. We also show that induction of E2F1 activity for 21 days resulted in severe atrophy of seminiferous tubules. This recapitulates our previous results from constitutive E2F1 overexpression in testicles (Holmberg et al., 1998). It is important to note, however, that the activation of ER-E2F1 for 21 days led

to the appearance of undifferentiated spermatogonia with aberrant mitotic figures and large gonocyte-like dysplastic germ cells with features of human carcinoma in situ (CIS). Thus, activation of ER-E2F1 in these transgenic mice led to morphological changes that mimic well premalignant changes of human testes.

Results Generation of ER-E2F1 transgenic mice To obtain tissue-specific expression of the ER-E2F1 protein in transgenic mice, we used the promoter from the hydroxy-methyl-glutaryl-CoA-reductase (HMG) gene. This promoter has been reported to direct the expression of transgenes to a wide range of different tissues when integrated at low copy number (Gautier et al., 1989; Holmberg et al., 1998). However, when integrated at high copy number, it is reportedly expressed in testicles only (Mehtali et al., 1990). A schematic drawing of the transgenic construct is shown in Figure 1a. This construct was microinjected into the pronucleus of fertilized eggs, which were implanted into pseudopregnant females to generate transgenic pups. Pups born after implantations were screened for the transgene by Southern blotting using genomic DNA from tail biopsies. We obtained five stably transmitting lines, two of those with approximately 10 tandem copies of the transgene (Figure 1b). Neither of the lines displayed any obvious phenotypic difference from wild-type controls (data not shown). Characterization of ER-E2F1 transgenic mice After establishment of the transgenic lines, the expression pattern of the transgene was determined by Northern blotting (data not shown). Four of the lines

Figure 1 Generation of ER-E2F1 transgenic mice. (a) The transgenic construct used in these studies. The promoter from the HMG gene was used to drive expression of the ER-E2F1 fusion protein. (b) Southern blot using genomic DNA from tail biopsies identifying founder mice. In total, 10 mg DNA was digested with BamH1 blotted to nylon membrane and hybridized to a 32P-labelled E2F1 cDNA probe. The copy number of the transgene was determined by comparing the signal obtained from hybridization to wild-type genomic DNA spiked with a known amount of transgenic DNA construct Oncogene


782

expressed the transgene exclusively in testes, whereas one expressed the transgene in testes and thymus (data not shown). The expression of the transgene in the testes and thymus in this line was confirmed by Western blotting using the monoclonal antibody KH95 that is specific for human E2F1 (Figure 2a). The steady-state level of the ER-E2F1 transcript was measured by realtime PCR using cDNA prepared from transgenic testes (Figure 2b). These data indicate that the ER-E2F1 mRNA is highly abundant in transgenic testes. Limitation of transgene expression to testes and thymus likely reflects an effect of the relatively high copy number of the integrated transgene (B10 copies, see Figure 1b) (Mehtali et al., 1990). The data reported below has been obtained using this transgenic line. Importantly, however, activaton of E2F1 led to similar effects in the other transgenic lines (data not shown). We made immmunohistochemical staining to further characterize the expression of the ER-E2F1 fusion protein. Testes sections from transgenic mice and wild-type littermates were stained to visualize expression of the ER-E2F1 protein. It is evident from Figure 2c that the protein was expressed strongly in the meiotic and maturing germ cells indicated with grey arrows (secondary spermatocytes, spermatids and spermatozoa, in order of differentiation), while the less-differentiated mitotic cells (i.e. spermatogonia and primary spermatocytes; located more in the periphery of the seminiferous tubules) stained less intensely as indicated by black arrows (Figure 2c). The observed staining was specific for the transgenic mice, since no staining was observed in the germ cells of wild-type littermates (Figure 2d). ER-E2F1 is functional in transgenic mice Tamoxifen dissolved in peanut oil was injected intraperitoneally to induce the activity of the transgene. The main by-product of tamoxifen metabolized in the liver is OHT (Poon et al., 1995). The activity of the transgene was induced by injection of 2 mg of tamoxifen per animal. In the absence of inducer, the ER-E2F1 protein is expressed but localized to the cytoplasm, while 24 h of tamoxifen treatment led to its localization in the nucleus (Figure 3a, arrows). Immunohistochemical stainings of NIH3T3 cells expressing ER-E2F1 with and without

Figure 2 Expression of ER-E2F1 in transgenic mice. (a) Western blotting using protein lysates from transgenic (Tg) and wild-type (WT) testes as well as from NIH3T3 cells transfected with pHMGERE2F1 (Control). The blots were probed with a monoclonal antibody (KH95) specific for human E2F1. The ER-E2F1 fusion protein has an apparent molecular weight of 90 kDa. (b) Real-time RT–PCR on RNA prepared from Tg and WT testes. The reaction was performed using primers specific for human E2F1 and was normalized to beta actin. The expression level is shown in arbitrary units. (c,d) Immunohistochemical analyses of tissue sections from Tg and WT mice. Sections were stained with KH95 antibody specific for human E2F1. Positive reaction is seen as a red–brown precipitate. Black arrows indicate the less-differentiated cells of the basal germinal epitelium expressing low levels of the transgene. Grey arrows with black indicate the more mature cells expressing the transgene to a higher level Oncogene

OHT was performed to confirm the functionality of the system in mouse cells (Figure 3b compare OHT with þ OHT). To test whether the ER-E2F1 protein was functional, we determined whether three known E2F target genes, Ccne1, APAF1 and p73, were upregulated by tamoxifen treatment using real-time RT–PCR. Ccne1 is one of the


783

best-characterized E2F-regulated genes, and is strongly upregulated by E2F1 activation in tissue culture systems (DeGregori et al., 1995; Vigo et al., 1999; Muller et al., 2001). APAF1 and p73 are proapototic genes upregu-

lated by E2F1 (Irwin et al., 2000; Moroni et al., 2001). Real-time RT–PCR was performed on RNA from testes of transgenic mice injected with tamoxifen as well as controls. The expression levels of the target genes were normalized to beta-actin. Tamoxifen treatment for 48 h led to a 3–4-fold induction of the target genes when compared WT mice treated with tamoxifen (Figure 3c). These data demonstrate that a potent burst of ectopic E2F1 activity is readily obtained by tamoxifen delivery to ER-E2F1 transgenic mice. Short-term activation of E2F1 in vivo induces apoptosis Overexpression of E2F1 in tissue culture leads to apoptosis within 24 h (Hickman et al., 2002). To test whether increased E2F1 activity in vivo results in apoptosis, we applied the TUNEL technique to testes tissue sections (Gavrieli et al., 1992). As shown in Figure 4, injection of tamoxifen 48 h prior to killing of the animals led to a strong burst in TUNEL-positive cells in the seminiferous tubules of transgenic mice (Figure 4b), whereas no staining was detected in mockinjected mice (Figure 4a) or wild-type mice injected with tamoxifen (Figure 4c). Apoptosis was confined to the spermatogonia and the primary spermatocytes located in the periphery of the tubules. As judged by the immunohistochemical staining for ER-E2F1 (Figure 2b), these cells appeared to express ER-E2F1 less strongly than the differentiating cells located more centrally in the tubules (compare cells indicated with grey arrows with cells indicated with black arrows). As the latter cells are either undergoing meiotic division or have completed doing so, we reason that the cells engaged in meiotic processes are protected from the direct apoptotic signals elicited by elevated E2F1 activity. Induction of p53-independent apoptosis Overexpression of E2F1 can lead to p53-dependent or -independent apoptosis (Hickman et al., 2002). Previously, we have shown that E2F1 induces p53-independent apoptosis in the testes of transgenic mice resulting in severe atrophy and infertility (Holmberg et al., 1998). However, as that study was based on continuous overexpression, we wished to test whether

Figure 3 The ER-E2F1 fusion protein is functional in vivo. (a) Immunohistochemical stainings of tissue sections from Tg mice using the KH95 antibody. Positive reaction is seen as a red–brown precipitate. The ER-E2F1 fusion protein is present in the cytoplasm of cells from Tg mice injected with the control (oil). Cells with cytoplasmic staining are indicated with arrows in the upper picture. (b) Immunohistochemical stainings of NIH3T3 cells expressing HAtagged ER-E2F1. HAER-E2F1-expressing NIH3T3 cells were grown on coverslips and induced for 1 h by addition of 600 mM OHT to the media. The cells were fixed with 2% paraformaldehyde in and probed with the HA-specific antibody 12CA5. (c) Real-time RT–PCR using RNA from transgenic (TG) and wild-type (WT) mice treated with tamoxifen for 48 h. The reactions were carried out using primer sets against Ccne1, p73 and APAF1 and the results were normalized to beta actin Oncogene


784

Figure 4 Short-term activation of E2F1 in vivo induces apoptosis independent of p53. (a) Testes section of a transgenic mouse injected with control (oil) showed no TUNEL-positive cells 48 h after injection. (b) Transgenic mice injected with tamoxifen 48 h prior to analysis, contain numerous TUNEL-positive cells. (c) TUNEL staining performed on wild-type mice injected with tamoxifen for 48 h; no TUNEL-positive cells are observed. (d) Apoptosis persists when ER-E2F1 is activated in p53 null background showing that E2F1 can induce p53-independent apoptosis in testes

Table 1 Averagea number of TUNEL-labelled cells/testicular tubule 48 h after injection

Wild-type mice ER-E2F1 mice ER-E2F1; P53/ mice

Mock injected

Tamoxifen injected

0 0 0

0 20.8 21.4

a

Values are an average of 10 seminiferous tubules

the rapid apoptotic response after activation of ERE2F1 was delayed or diminished in the absence of p53. Thus, we generated ER-E2F1; p53/ compound mice and detected apoptotic cells in the absence or presence of OHT. As in the wild-type background, increased E2F1 activity in p53-deficient mice resulted in apoptosis of germ cells (Figure 4d), whereas no apoptosis was observed in mock-injected ER-E2F1; p53/ compound mice (data not shown). Indeed, the response appeared indistinguishable from that observed in the wild-type background (47.4% atrophic tubules in wildtype background compared to 52.2% in p53/ background). In addition, we have calculated the average number of apoptotic cells per tubule in mice treated and untreated with tamoxifen (Table 1), which also confirms the notion that there are no measurable differences in the apoptotic response between the two genotypes. These results show that short-term activation of E2F1 in vivo results in rapid p53-independent apoptosis. Oncogene

Prolonged activation of E2F1 induces testicular atrophy and dysplasia mimicking human CIS To explore the effect of increased E2F1 activity over a longer period, we injected tamoxifen every second day for 3 weeks. This resulted in significantly atrophic seminiferous tubules (Figure 5b). The tubules in the transgenic mice were reduced in number and often in size (Figure 5b,d and f) due to severe hypoplasia of spermatogenetic cells resulting in total absence of the more mature cells (spermatids and spermatozoa). The peritubular interstitial tissue was oedematous and contained groups of Leydig cells with normal appearance. Some tubules contained only Sertoli cells but in the majority of tubules, clusters of remaining aberrantly maturing cells comprising undifferentiated spermatogonia and rare spermatocytes were observed. Interestingly, these clusters were characterized by high expression of ER-E2F1 (Figure 5d) and high mitotic activity (Figure 5f). Some of the mitotic figures might represent a reactive attempt to regenerate the injured structures and provide mature spermatogenic forms by the remaining cells. However, in many of the transgenic tubules the mitotic figures were atypical (tri-/multipolar; with chromosomal bridges or loose fragments, etc.) (Figure 5f, arrows), which is considered to be a feature of cells with malignant potential. In addition, the intratubular cell clusters contained variable amounts of gonocyte-like, large dysplastic cells with abundant clear cytoplasm and one or more large, polymorphic, often bizarre nuclei with prominent nucleoli (Figure 5f,


785

Figure 5 E2F1 activation results in atrophy and apoptosis in vivo. (a) Low-power photomicrograph of a section of normal testis in 5– 6-week-old wild-type mice after 21 days of exposure to tamoxifen. The seminiferous tubules have uniform size and cellularity. (b) Lowpower photomicrograph of a section of atrophic testis in age-matched transgenic mice after 4 weeks of tamoxifen treatment. Note the atrophy of seminiferous tubules, which are diminished in number and often in size due to reduced cellularity and are separated by oedematous stroma. (c) Immunostaining for transgene expression is negative in seminiferous tubules of control mice. (d) Immunohistochemical staining for transgene expression in transgenic testes showing strong expression in the clusters of spermatogonia and spermatocytes. (e) The seminiferous tubules in wild-type mice display normal, complete spermatogenesis and are separated from one another by a slight amount of interstitial connective tissue containing small groups of Leydig cells. (f) Section of atrophic seminiferous tubules in transgenic mice. Marked apoptosis causes severe hypoplasia of spermatogenetic cells with total absence of the more mature cells (spermatids and spermatozoa). Aberrant maturation of the remaining cells results in clusters of undifferentiated spermatogonia and rare spermatocytes. These clusters are characterized by high mitotic activity, including atypical mitoses (arrows), and appearance of large, dysplastic CIS-like cells with one or more large, polymorphic, bizarre nuclei and prominent nucleoli (arrowhead). No invasive GCT component is seen at this stage. Oedematous interstitial spaces containing aggregates of normal Leydig cells surround the atrophic tubules

Oncogene


786

grey arrows). Taken together, these features resembled well those of human CIS (Mostofi et al., 1994; Cheville et al., 1998). The number of tubules affected by CIS-like changes varied in different areas and was proportional to the degree of tubular atrophy, consistent with the fact that the latter is a major predisposing factor for the development of human CIS. Thus, we demonstrate here that persistent ectopic E2F1 activity over a 3-week period suffices to drive apoptotic and mitotic programmes and to induce dysplastic changes similar to those observed in the development of human testicular cancer.

Discussion We generated a system, which enables conditional activation of E2F1 in mouse testes. Thus, we were able to study the direct consequences of increased E2F1 activity in normally developed tissue. We found that 48 h of increased E2F1 activity results in a potent apoptotic response of germ cells. The apoptotic response was also present in p53 null mice. These data are in agreement with previous results from our group and demonstrate that E2F1 can trigger a p53-independent apoptotic response in the testes. Interestingly, the p53 pathway seems to be intact in testicular cancer and treatment of adult male germ cell tumours (GCT) with the p53 inducer cisplatin has a highly curing effect (Chaganti and Houldsworth, 2000). One could speculate that in testes another checkpoint is responding to inactivation of the pRB pathway and that this other apoptotic checkpoint is mutated in GCT leaving the p53 pathway intact. It is well established that increased E2F1 can lead to 53-dependent and -independent apoptosis (Hsieh et al., 1997; Phillips et al., 1997). Others and we have shown that E2F1 can directly upregulate the transcription of proapoptotic genes such as APAF1 and the p53 homologue p73 (Irwin et al., 2000; Moroni et al., 2001). In the current study, we show that induction of apoptosis by activated ER-E2F1 in germ cells coincided with the upregulation of p73 and APAF1. These results could suggest that these two E2F1 target genes are involved in the apoptotic response of germ cells in the transgenic testes and suggest that they might be key players in an apoptotic checkpoint responding to excessive E2F1 activity in germ cells. It is possible that during the development of GCT, this putative non-p53 apoptotic checkpoint is mutated in combination with deregulation of the pRB pathway, thus pushing the cells toward malignant transformation. Apoptosis was more evident in premeiotic or mitotic/ less-mature germ cells even though they express the transgene to a lower level than the differentiating cells located more centrally in the tubules. As these cells are either undergoing meiotic division or have completed doing so, we reason that the cells engaged in meiotic processes are protected from the direct apoptotic signals elicited by elevated E2F1 activity. Several explanations can account for this apparent resistance of more Oncogene

differentiated germ cells to E2F1-mediated apoptosis. First, mechanisms ensuring DNA repair and chromosomal stability are more operative in these cells than in mitotic germ cells. This allows them to deal with the frequent double-strand DNA breaks physiologically occurring during meiotic recombination and ensures proper spermatogenesis. Indeed, mice deficient for specific regulators of these mechanisms show defective spermatogenesis and infertility due to increased apoptosis (Celeste et al., 2002; Hsia et al., 2003). Similarly, the CDK inhibitors p18ink4c and p19ink4d, whose absence in specifically engineered knockout mice leads to testicular atrophy and deficient progeny, are particularly highly expressed in the more mature stages of spermatogenesis (Zindy et al., 2000; Bartkova et al., 2003). It is therefore tempting to speculate that DNA repair and cell cycle checkpoints essential for correct regulation of meiotic and postmeiotic stages of spermatogenesis might safeguard the more differentiated germ cells also from the genomic instability and the apoptotic signals prompted by excessive E2F activity. In addition, it is possible that mitotic, less-mature germ cells like spermatogonia and primary spermatocytes be more sensitive than the differentiating cells to the apoptotic signals imposed by activated ER-E2F1, because they already have high levels of endogenous E2F activity induced by the pRb phosphorylation carried out by cyclin D/CDK4 complexes during early stages of spermatogenesis (Bartkova et al., 2003). Thus, even lower levels of ER-E2F1 may be sufficient to reach a critical threshold of E2F activity and trigger apoptosis in spermatogonia and primary spermatocytes. Finally, although merely speculative at this moment, one cannot exclude the possibility that the germ cells with higher expression of ER-E2F1 centrally located in the seminiferous tubules might induce apoptosis of the lessmature, peripheral germ cells expressing lower levels of transgene by a mechanism of cell competition similar to that described for Myc in Drosophila (de la Cova et al., 2004; Moreno and Basler, 2004). When the transgenic mice were exposed to tamoxifen for a period of 21 days, we observed severe atrophy of the seminiferous tubules and intratubular dysplastic changes mimicking CIS lesions described in humans. Both tubular atrophy and CIS-like lesions indicate that overexpression of E2F1 in germ cells is potentially oncogenic. Indeed, testicular atrophy, most often a consequence of testicular retention (criptorchidism), represents a major risk factor for testicular carcinogenesis in humans and is invariably observed in human testes affected by in situ and invasive GCT (Cheville et al., 1998). Therefore, by causing tubular atrophy via increased apoptosis, E2F1 overexpression may indirectly exercise an oncogenic effect on the mouse testis, which may, at least in part, be based on selection of cells resistant to apoptosis. Furthermore, the increased and aberrant mitotic activity of remaining clusters of undifferentiated germ cells and the CIS-like changes seen in transgenic tubules imply that E2F1 may also have a direct oncogenic effect on germ cells by virtue of deregulated cell cycle promoting effects.


787

Our observations are interesting because spontaneous testicular tumourigenesis is extremely rare in mice and usually represented by Leydig cell tumours or by teratomas, which are complex GCTs composed of ecto-, endo- and mesodermal elements (Mostofi et al., 1994). Since we detected neither teratomatous structures nor alterations of interstitial Leydig cells, it is more likely that nonteratoma-GCT may arise after long-term overexpression of E2F1 in mice. Alterations similar to those reported here have been observed in transgenic mice overexpressing glial cell line-derived neurotrophic factor (GDNF) in undifferentiated spermatogonia (Meng et al., 2000). In these mice, clusters of spermatogonia appear in seminiferous tubules at 2–4 weeks of age, testicular atrophy occurs at 8 weeks and invasive seminomas after 1 year reportedly preceded by CIS-like changes (Meng et al., 2000), suggesting a long latency and additional oncogenic events before the appearance of invasive tumours. Our results may have important implications for a better understanding of the early molecular events involved in human testicular carcinogenesis. Indeed the occurrence of CIS-like lesions caused by deregulated E2F1 activity in mouse testis is consistent with the defects of the pRB pathway recently described in human CIS (Bartkova et al., 2003). In these reports, human CIS lesions were characterized by transcriptional repression of pRB expression, lack of p19ink4d expression and cyclin D2 overexpression. As pRB expression appears to be physiologically absent in fetal gonocytes, while it is present throughout adult spermatogenesis, CIS is thought to arise from transformed gonocytes that maintain the developmentally repressed pRB status, which would imply that high levels of E2F activity in fetal gonocytes might predispose to the transition into CIS cells (Bartkova et al., 2003). Taken together, these data and our findings in the testis of ER-E2F1 transgenic mice indicate that deregulated E2F activity generated by defects of the pRB pathway is a very early event in testicular carcinogenesis. In conclusion, we have developed an in vivo model system in which deregulated E2F1 activity in testicles is readily obtained by simple intraperitoneal administration of the exogenous inducer tamoxifen. Activation for 3 weeks indeed resulted in excessive apoptosis and aberrant mitotic activity as well as appearance of premalignant changes. Targeting ERE2F1 to other organs will allow for the analysis of oncogenic effects of deregulated E2F1 in the tissue of choice.

fertilized eggs were collected in M2 medium (Sigma). Cumulus cells were removed enzymatically by brief incubation in M2 medium containing hyaluronidase, washed three times in fresh medium and subsequently transferred to M16 medium (Sigma) and cultured at 371C in a humid 5% CO2 atmosphere until injection. Eggs were microinjected with DNA in the male pronucleus under the microscope with a Zeiss micromanipulator. Injected eggs were transferred back to M16 at 371C for later implantation. Pseudopregnant female mice were anesthetized by injection of 2.5% avertin in PBS IP approximately 0.7 ml for 24 g mice. A total of 20–25 eggs were implanted into the infundibulum under dissection microscope in each pseudopregnant mice. Implantation was carried out the same day as the injections or the following day when the embryos were at the two-cell stage. Generation of ER-E2F1;p53/ compound mice ER-E2F1 transgenic mice were crossed into a p53 null background. The genotyping of the mice was carried out by PCR on tail DNA. p537 mice from an inbred 129 genetic background were generous gifts from Alan R Clarke, Cancer Research Campaign Laboratories, Edinburgh, UK. The mutant p53 gene in these mice carries a deletion from exon 2 to exon 6 made by replacement with a neomycin expression cassette (Clarke et al., 1993). After one or two backcrosses to C57Bl/6, p537 mice were crossed with our transgenic mice. Southern blotting Genomic DNA extracted from tail biopsies using standard methods was digested with BamHI. The DNA was run on agarose gels and blotted onto nylon membranes. To detect the transgene, the membranes were hybridized to an E2F1-specific 32P-labelled probe generated from the XhoI–SalI restriction fragment of the human E2F-1 cDNA, using a random prime DNA labelling kit from Boehringer Mannheim according to instructions from the manufacturer. Real-time RT–PCR Frozen tissue samples were pulverized in a mortar and total RNA was prepared using the RNA easy midi kit from Quiagen. RNA was reverse transcribed using random hexamers, and real-time PCR was performed on an Applied biosystems 7000 Sequence Detection System, using the SYBR green PCR master mix from the same manufacturer. The following primers were used in this study: Beta actin-F: Beta actin-R: Ccne1-F: Ccne1-R: p73-F: p73-R: APAF1-F: APAF1-B:

CCGGGACCTGACAGACTACC GGTGGTGAAGCTGTAGCCACG TGTTACAGATGGCGCTTGCTC TTCAGCCAGGACACAATGGTC TGACCGACATTGTTAAGCGCTG TGTCCTTCATTGAAGTCCCTTCC TGGGTCCTAAGCATGTTGTCC CCGCTGGCAGATCTTCTTTC

Materials and methods Western blotting Generation of transgenic mice C57Bl/6 CBA F1 hybrid mice, 5 weeks old, were superovulated by i.p. injections of 5 U PMS (pregnant mare serum) around 1500 hours followed 46–48 h later by 5 U of HCG (human chorionic gonadotropin). The superovulated mice were mated with C57Bl/6 CBA F1 males. Successfully mated superovulated females were identified by visual inspection for the presence of plugs. Plugged females were killed and

Protein lysates were made from tissue samples snap-frozen in liquid nitrogen. The frozen samples were pulverized and incubated 5 min on ice with three volumes of swelling buffer (20 mM HEPES pH 7.9, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM DTT, 25% glycerol). After incubation, NaCl was added to a final concentration of 0.42 M. The samples were homogenized using a polytron and incubated for 20 min on ice. Subsequently the samples were centrifugated for Oncogene


788 30 min at 100 000 g, and the supernatants were used for Western blotting. Protein concentration was determined by the Bradford method according to instructions from the manufacturer (Biorad). After SDS–polyacrylamide gel electrophoresis (SDS–PAGE) of the extracts, semidry transfer to a nitrocellulose membrane was performed. The blotted membranes was blocked in PBS supplemented with 0.05% Tween20 and 5% nonfat dry milk (PBS-Tween) for 1 h prior to addition of the primary antibody. The blot was incubated 1 h with monoclonal anti-E2F-1 antibody KH95 diluted 1 : 1000 in PBS-Tween. After incubation with the primary antibody, the blot was washed and incubated with secondary antibodies coupled to horseradish peroxidase (HRP). Final visualization was obtained using ECL reagents and film exposure according to the manufacturer (Amersham).

paraformaldehyde/PBS. The coverslips were stained by standard methods using the HA-specific antibody 12CA5 and counterstained with DAPI. TUNEL labelling For the TUNEL procedure, the tissue sections were pretreated as for immunohistochemistry. The TUNEL assay was performed as described in Gavrieli et al. (1992) and Holmberg et al. (1998). Haematoxylin eosin staining Haematoxylin eosin staining was carried out by standard methods.

Immunohistochemistry

Induction with tamoxifen

Testes were fixed using 4% paraformaldehyde in PBS, dehydrated and embedded in paraffin. The fixed samples were sectioned on a microtome and moved onto glass slides. The tissue sections were rehydrated and the endogenous peroxidase activity was blocked in 3% H2O2 in methanol for 15 min. For immunohistochemical detection of E2F-1, the sections were incubated with KH95 diluted 1 : 200 in 5% BSA and 0.5% Tween-20 in PBS. The primary antibody was detected using secondary HRP-coupled anti-mouse antibody and the Vectorstain ABC kit (Vector Laboratories) as described by the manufacturer. The sections were briefly counterstained in haematoxylin and mounted with moviol.

Tamoxifen was dissolved in peanut oil (Sigma) by sonication to a final concentration of 10 mg/ml. Mice were injected intraperitoneally with 2 mg tamoxifen per animal.

Staining of coverslips NIH3T3 cells stably expressing HA-tagged ER-E2F1 were grown on coverslips, induced for 1 h and fixed in 4%

Acknowledgements We thank Claire Attwooll and Eros Lazzerini Denchi for critical reading of the manuscript, and members of the Helin laboratory for discussions. KA was supported by Dagmar Marchalls Fond, Harboefonden, Fonden til fremme af medicinsk behandling af cancer, Direktr Jacob Madsen and Hustru Olga Madsens Fond, Fabrikant Einar Willumsens Mindelegat and Beckett-Fonden during the course of this work. This work was supported by grants from the Associazione Italiana per la Ricerca sul Cancro (AIRC) and Fondazione Italiana per la Ricerca sul Cancro (FIRC), the Danish Cancer Society and the Danish Ministry of Research.

References Bartkova J, Rajpert-De Meyts E, Skakkebaek NE, Lukas J and Bartek J. (2003). Apmis, 111, 252–265, discussion 265–266. Cam H and Dynlacht BD. (2003). Cancer Cell, 3, 311–316. Cartwright P, Muller H, Wagener C, Holm K and Helin K. (1998). Oncogene, 17, 611–623. Celeste A, Petersen S, Romanienko PJ, Fernandez-Capetillo O, Chen HT, Sedelnikova OA, Reina-San-Martin B, Coppola V, Meffre E, Difilippantonio MJ, Redon C, Pilch DR, Olaru A, Eckhaus M, Camerini-Otero RD, Tessarollo L, Livak F, Manova K, Bonner WM, Nussenzweig MC and Nussenzweig A. (2002). Science, 296, 922–927. Chaganti RS and Houldsworth J. (2000). Cancer Res., 60, 1475–1482. Cheville JC, Sebo TJ, Lager DJ, Bostwick DG and Farrow GM. (1998). Am. J. Surg. Pathol., 22, 1361–1367. Clarke AR, Maandag ER, van Roon M, van der Lugt NM, van der Valk M, Hooper ML, Berns A and te Riele H. (1992). Nature, 359, 328–330. Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Hooper ML and Wyllie AH. (1993). Nature, 362, 849–852. Conner EA, Lemmer ER, Omori M, Wirth PJ, Factor VM and Thorgeirsson SS. (2000). Oncogene, 19, 5054–5062. de Bruin A, Maiti B, Jakoi L, Timmers C, Buerki R and Leone G. (2003). J. Biol. Chem., 278, 42041–42049. de la Cova C, Abril M, Bellosta P, Gallant P and Johnston LA. (2004). Cell, 117, 107–116. DeGregori J, Kowalik T and Nevins JR. (1995). Mol. Cell. Biol., 15, 4215–4224.

Oncogene

Di Stefano L, Jensen MR and Helin K. (2003). EMBO J., 22, 6289–6298. Gaubatz S, Wood JG and Livingston DM. (1998). Proc. Natl. Acad. Sci. USA, 95, 9190–9195. Gautier C, Mehtali M and Lathe R. (1989). Nucleic Acids Res., 17, 8389. Gavrieli Y, Sherman Y and Ben-Sasson SA. (1992). J. Cell Biol., 119, 493–501. Hickman ES, Moroni MC and Helin K. (2002). Curr. Opin. Genet. Dev., 12, 60–66. Holmberg C, Helin K, Sehested M and Karlstrom O. (1998). Oncogene, 17, 143–155. Hsia KT, Millar MR, King S, Selfridge J, Redhead NJ, Melton DW and Saunders PT. (2003). Development, 130, 369–378. Hsieh JK, Fredersdorf S, Kouzarides T, Martin K and Lu X. (1997). Genes Dev., 11, 1840–1852. Humbert PO, Rogers C, Ganiatsas S, Landsberg RL, Trimarchi JM, Dandapani S, Brugnara C, Erdman S, Schrenzel M, Bronson RT and Lees JA. (2000). Mol. Cell, 6, 281–291. Irwin M, Marin MC, Phillips AC, Seelan RS, Smith DI, Liu W, Flores ER, Tsai KY, Jacks T, Vousden KH and Kaelin Jr WG. (2000). Nature, 407, 645–648. Jacks T, Fazeli A, Schmitt EM, Bronson RT, Goodell MA and Weinberg RA. (1992). Nature, 359, 295–300. Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993). Nature, 365, 349–352.


789 Lee EY, Chang CY, Hu N, Wang YC, Lai CC, Herrup K, Lee WH and Bradley A. (1992). Nature, 359, 288–294. Lindeman GJ, Dagnino L, Gaubatz S, Xu Y, Bronson RT, Warren HB and Livingston DM. (1998). Genes Dev., 12, 1092–1098. Littlewood TD, Hancock DC, Danielian PS, Parker MG and Evan GI. (1995). Nucleic Acids Res., 23, 1686–1690. Lomazzi M, Moroni MC, Jensen MR, Frittoli E and Helin K. (2002). Nat Genet, 31, 190–194. Lukas J, Petersen BO, Holm K, Bartek J and Helin K. (1996). Mol. Cell. Biol., 16, 1047–1057. Mehtali M, LeMeur M and Lathe R. (1990). Gene, 91, 179–184. Meng X, Lindahl M, Hyvonen ME, Parvinen M, de Rooij DG, Hess MW, Raatikainen-Ahokas A, Sainio K, Rauvala H, Lakso M, Pichel JG, Westphal H, Saarma M and Sariola H. (2000). Science, 287, 1489–1493. Moreno E and Basler K. (2004). Cell, 117, 117–129. Morkel M, Wenkel J, Bannister AJ, Kouzarides T and Hagemeier C. (1997). Nature, 390, 567–568. Moroni MC, Hickman ES, Denchi EL, Caprara G, Colli E, Cecconi F, Muller H and Helin K. (2001). Nat. Cell Biol., 3, 552–558. Mostofi FK, Davis Jr J and Rehm S. (1994). IARC Sci. Publ., 111, 407–429. Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, Grassilli E, Prosperini E, Vigo E, Oliner JD and Helin K. (2001). Genes Dev., 15, 267–285. Muller H and Helin K. (2000). Biochim. Biophys. Acta, 1470, M1–M12. Nikitin A and Lee WH. (1996). Genes Dev., 10, 1870–1879. Pan H, Yin C, Dyson NJ, Harlow E, Yamasaki L and Van Dyke T. (1998). Mol. Cell, 2, 283–292.

Phillips AC, Bates S, Ryan KM, Helin K and Vousden KH. (1997). Genes Dev., 11, 1853–1863. Pierce AM, Fisher SM, Conti CJ and Johnson DG. (1998a). Oncogene, 16, 1267–1276. Pierce AM, Gimenez-Conti IB, Schneider-Broussard R, Martinez LA, Conti CJ and Johnson DG. (1998b). Proc. Natl. Acad. Sci. USA, 95, 8858–8863. Poon GK, Walter B, Lonning PE, Horton MN and McCague R. (1995). Drug Metab. Dispos., 23, 377–382. Rempel RE, Saenz-Robles MT, Storms R, Morham S, Ishida S, Engel A, Jakoi L, Melhem MF, Pipas JM, Smith C and Nevins JR. (2000). Mol. Cell, 6, 293–306. Sage J, Miller AL, Perez-Mancera PA, Wysocki JM and Jacks T. (2003). Nature, 424, 223–228. Sherr CJ and McCormick F. (2002). Cancer Cell, 2, 103–112. Symonds H, Krall L, Remington L, Saenz-Robles M, Lowe S, Jacks T and Van Dyke T. (1994). Cell, 78, 703–711. Trimarchi JM, Fairchild B, Verona R, Moberg K, Andon N and Lees JA. (1998). Proc. Natl. Acad. Sci. USA, 95, 2850–2855. Trimarchi JM and Lees JA. (2002). Nat. Rev. Mol. Cell. Biol., 3, 11–20. Vigo E, Muller H, Prosperini E, Hateboer G, Cartwright P, Moroni MC and Helin K. (1999). Mol. Cell. Biol., 19, 6379–6395. Wu L, Timmers C, Maiti B, Saavedra HI, Sang L, Chong GT, Nuckolls F, Giangrande P, Wright FA, Field SJ, Greenberg ME, Orkin S, Nevins JR, Robinson ML and Leone G. (2001). Nature, 414, 457–462. Yamasaki L, Bronson R, Williams BO, Dyson NJ, Harlow E and Jacks T. (1998). Nat. Genet., 18, 360–364. Zindy F, van Deursen J, Grosveld G, Sherr CJ and Roussel MF. (2000). Mol. Cell. Biol., 20, 372–378.

Oncogene