Linking DNA damage to medulloblastoma tumorigenesis in ... - Nature

Oncogene (2006) 25, 1165–1173

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Linking DNA damage to medulloblastoma tumorigenesis in patched heterozygous knockout mice S Pazzaglia1, M Tanori1, M Mancuso1, S Rebessi1, S Leonardi1,2, V Di Majo1, V Covelli3, MJ Atkinson4, H Hahn5 and A Saran1 1 Biotechnology Unit, ENEA CR-Casaccia, Rome, Italy; 2Department of Experimental Oncology, Istituto Nazionale Tumori, Milan, Italy; 3Radiation Protection Unit, ENEA CR-Casaccia, Rome, Italy; 4Institute of Pathology, GSF National Research Center, Germany and 5Institute of Human Genetics, University of Goettingen, Germany

Hemizygous Ptc1 mice have many features of Gorlin syndrome, including predisposition to medulloblastoma development. Ionizing radiation synergize with Ptc1 mutation to induce medulloblastoma only in neonatally exposed mice. To explore the mechanisms underlying agedependent susceptibility, we irradiated Ptcneo67/ þ mice at postnatal day 1 (P1) or 10 (P10). We observed a dramatic difference in medulloblastoma incidence, which ranged from 81% in the cerebellum irradiated at P1 to 3% in the cerebellum irradiated at P10. A stricking difference was also detected in the frequency of cerebellar preneoplastic lesions (100 versus 14%). Our data also show significantly lower induction of apoptosis in the cerebellum of medulloblastoma-susceptible (P1) compared to -resistant (P10) mice, strongly suggesting that medulloblastoma formation in Ptc1 mutants may be associated with resistance to radiation-induced cell killing. Furthermore, in marked contrast with P10 mice, cerebellum at P1 displays substantially increased activation of the cell survival-promoting Akt/Pkb protein, and markedly decreased p53 levels in response to radiation-induced genotoxic stress. Overall, these results show that developing cerebellar granule neuron precursors’ (CGNPs) radiosensitivity to radiation-induced cell death increases with progressing development and inversely correlates with their ability to neoplastically transform. Oncogene (2006) 25, 1165–1173. doi:10.1038/sj.onc.1209032; published online 9 January 2006 Keywords: ionizing radiation; apoptosis; LOH; p53; PI3K-Akt/Pkb pathway

Introduction Medulloblastoma is an embryonal tumor arising from cerebellar granule neuron precursors (CGNPs), and it is the most common brain malignancy in children Correspondence: Dr S Pazzaglia, Biotechnology Unit, ENEA, Via Anguillarese 301, CR Casaccia, 00060 Rome, Italy. E-mail: [email protected] Received 7 April 2005; revised 1 July 2005; accepted 25 July 2005; published online 9 January 2006

(reviewed by Raffel, 2004). Despite multimodal therapy that includes surgery, radiation and chemotherapy, its poor prognosis has not improved significantly in the last two decades (MacDonald et al., 2003). Improvements in therapy will benefit greatly by the identification of genes and biological pathways that contribute to medulloblastoma development. The Shh pathway is a major regulator of cerebellar development, since Shh secreted by Purkinje cells promotes proliferation of CGNPs in the external granule layer (EGL) (Dahmane and Ruiz-i-Altaba, 1999; Wallace, 1999; Wechsler-Reya and Scott, 1999). A subset of hereditary and sporadic medulloblastomas show mutation in one copy of Patched (Ptc1) (Hahn et al., 1996; Raffel et al., 1997; Zurawel et al., 2000a), which encodes a cell surface Shh-receptor and negative regulator of the pathway. This suggests that uncontrolled activation of this mitogenic pathway, that has a pivotal role in cerebellar ontogenesis, leads to medulloblastoma formation. Insights for medulloblastoma tumorigenesis arise from recently developed animal models. In the mouse, deletion of one Ptc1 allele results in spontaneous medulloblastoma development, underlying the critic tumorigenic role of Ptc1 as a key modulator of signalling in the Shh pathway (Goodrich et al., 1997; Hahn et al., 1998). Various combinations of targeted deletions in genes controlling cell cycle checkpoints, apoptosis and DNA repair, such as Ptc1, Rb, DNA Ligase IV, PARP-1 and Ink4c in combination with p53 loss also result in murine medulloblastoma (Marino et al., 2000; Wetmore et al., 2001; Lee and McKinnon, 2002; Pazzaglia et al., 2002; Tong et al., 2003; Zindy et al., 2003). While these models are informative, simultaneous mutations in these pathways are rare in human medulloblastoma. An alternative approach is to complement Ptc1 mutation with ionizing radiation. DNA damage induced by radiation strongly promotes medulloblastoma development in newborn Ptcneo67/ þ mice in which the EGL is still proliferating (Pazzaglia et al., 2002). Since the quasi-totality of neurogenesis from the EGL is neonatal, and occurs during the first 2 weeks of postnatal life in mice, here we have further investigated the age-dependence of susceptibility by irradiating Ptcneo67/ þ mice at P1 and P10. Our data show

Medulloblastoma in Ptcneo67/ þ mice S Pazzaglia et al

1166

a striking enhancement of medulloblastoma by irradiating mice at P1, compared to no enhancement after irradiation at P10. This indicates that the period of granule cell competence in which radiation-induced stochastic events required for medulloblastoma tumorigenesis may take place in a setting of Shh pathway deregulation is very limited and ends within P10. To elucidate the mechanisms of age-dependent susceptibility of the cerebellum to medulloblastoma induction, we have characterized the apoptotic response to radiation of P1 and P10 cerebellum in Ptc1 mutant mice. We present data to show that medulloblastoma formation is associated with CGNPs resistance to radiation-induced cell killing, regulated by the dynamic interplay of p53 and the cell survival-promoting Akt/ protein kinase B (Pkb).

Results Survival and tumorigenesis after X-ray irradiation of newborn Ptcneo67/ þ mice at P1 and P10 The survival of Ptcneo67/ þ heterozygous mice irradiated at P1 or P10 is reported in Figure 1. A striking increase in mortality was observed in Ptcneo67/ þ heterozygous mice irradiated at P1 (Figure 1a). By 36 weeks of age, none of these mice had survived compared to 36% of mice irradiated at P10 (Po0.0001), confirming that survival is strongly dependent on mouse age at irradiation. For comparison, the survival curve of mice irradiated at P4 is reported from a previous publication (Pazzaglia et al., 2002). The cumulative tumor incidence was 100% in mice irradiated at P1 compared to 85% in mice irradiated at P10. This difference, however, was not statistically

significant. Life shortening of mice irradiated at P1 was mainly due to development of medulloblastomas, which represented the major cause of death with an incidence of 81% (Figure 1b). In this group, medulloblastoma showed a very short latency, developing between 7 and 20 weeks postirradiation. There was a large difference between medulloblastoma incidence in mice irradiated at P1 (81%) and the incidence previously observed in mice irradiated at P4 (51%) (Pazzaglia et al., 2002) (P ¼ 0.0330), and a striking difference with the incidence observed in mice irradiated at P10 (3%) (Po0.0001), indicating that the ability of ionizing radiation to synergize with Ptc1 mutation to promote medulloblastoma formation decreases very sharply as neurogenesis progresses. The incidence of medulloblastoma in mice irradiated at P10 was similar to that of unirradiated controls (6.7%), suggesting the existence of a short timewindow of radiosensitivity. The predominant tumor type in mice irradiated at P10 was soft tissue sarcoma, which developed in 42% of mice, an incidence very similar to that of unirradiated Ptcneo67/ þ heterozygotes (40%). In contrast, mice irradiated at P1 did not develop soft tissue sarcomas – that have a late onset – due to early mortality for medulloblastoma. No further significant difference in tumor incidence and spectrum was observed between mice irradiated at P1 or P10 (Figure 1c). Cerebellar hyperproliferation of granule neurons after irradiation at P1 and P10 Figure 2 shows H&E-stained sagittal sections of cerebella from unirradiated Ptcneo67/ þ mice at P1 and P10 (a and b), and proliferating CGNPs, detected by BrdU incorporation, in P1 and P10 cerebellum (Figure 2c and d). While BrdU incorporation was evident

Figure 1 Mortality and medulloblastoma incidence of Ptcneo67/ þ mice irradiated with 3 Gy of X-rays at P1 or P10. (a) High mortality was observed in P1- compared to P10-irradiated mice. (b) A 81% medulloblastoma incidence was observed in P1- compared to 3% in P10-irradiated mice. Survival and medulloblastoma incidence in Ptcneo67/ þ mice irradiated at P4 are reported for comparison (Pazzaglia et al., 2002). (c) Tumor spectrum in Ptcneo67/ þ mice irradiated at P1 or P10. Oncogene


1167

a

b IGL IGL

ML ML

c

d IGL ML

Mice with lesions (%)

e 120 100 80

Control P1 RX P10 RX

60 40 20 0

Figure 2 Cerebellum size and BrdU incorporation at P1 and P10. (a) and (b) Low-power views of P1- and P10-cerebellum from Ptcneo67/ þ mice, showing a large size difference depending on developmental stage. Immunostaining for BrdU shows incorporation through the entire EGL at P1 (c) and restriction to the proliferative EGLa outer compartment at P10 (d). (a) and (b) Scale bar ¼ 800 mm. (c) and (d) Scale bar ¼ 20 mm.

Figure 3 Histological features of early cerebellar lesions (arrows) in unirradiated Ptcneo67/ þ (a)–(c), and in asymptomatic P1irradiated Ptcneo67/ þ mice (d) of 3 weeks of age. Morphologic features ranged from hyperproliferation of granule neurons to overt nodules with variable size and location. (e) Percent incidence of microscopic lesions: such lesions were evident in 6/6 of P1irradiated mice, in 1/7 P10-irradiated mice and in 6/10 of unirradiated mice. IGL, internal granular layer; ML, molecular layer. (a), (b), (d) Scale bar ¼ 100 mm. (c) Scale bar ¼ 200 mm.

through the entire EGL at P1 (Figure 2c), it was clearly restricted to the proliferative EGLa outer compartment at P10 (Figure 2d). A further important feature of the cerebellum of Ptcneo67/ þ mice at 3 weeks was the presence of abnormal cerebellar EGL regions ranging from small areas of hyperproliferation of granule neurons to overt nodules (Figure 3a–d); these abnormalities were present in both control and irradiated mouse cerebella, although with different frequencies (Figure 3e). In unirradiated Ptcneo67/ þ mice these lesions were evident in 3/5 (60%) mice at P21 and 3/5 (60%) at P31 (Figure 3e). However, only 7% of unirradiated mice developed medulloblastomas, indicating that only a subset of hyperproliferating areas progress to medulloblastoma. Most of these spontaneous lesions will regress by adulthood, since they are not detected in mice autopsied at later times (data not shown). The frequency of these areas increased to 100% (6/6) in cerebellum irradiated at P1, but was only 14% (1/7) after irradiation at P10 (Figure 3e). This suggests that irradiation of the cerebellum at P10 accelerated regression of these lesions, probably through a mechanism of cell killing.

Apoptotic response to ionizing radiation of Ptcneo67/ þ cerebellum at P1 and P10 To investigate the mechanisms of different susceptibility to radiation-induced medulloblastoma tumorigenesis of Ptcneo67/ þ mouse cerebellum of different developmental stages, we examined the apoptotic response to radiationinduced DNA damage in P1 and P10 mice (Figure 4a and b). Quantitation of pyknotic nuclei in the EGL of irradiated cerebellum revealed widespread cell death in both Ptcneo67/ þ heterozygous and wt mice, with no significant difference (wt is not shown). Cell death induced by 3 Gy of X-rays was 2.5-fold lower in the cerebellum of P1 compared to P10 Ptcneo67/ þ heterozygotes 3 h after irradiation (8.9% apoptotic cells versus 20.0%) (Figure 4c). This difference was reduced to 1.3fold 6 h after irradiation (36.7% apoptotic cells versus 49.5%). Differences in cell death between P1 and P10 mice were highly statistically significant (Po0.001) at both postirradiation times analysed. Results of cell death assay led us to hypothesize that radiation might activate caspase-dependent apoptotic pathways. Thus, we investigated the expression of active caspase-3 in the cerebellum of mice irradiated at P1 or Oncogene


1168

c

b

50

Number of pycnotic nuclei (%)

a

40

P1 P10

30 20 10 0

f

3h RX 3h P1 P10

6h RX 6h P1 P10

Parp

d

Actin

e

Parp/Actin protein (%)

g 250 200

P1 P10

150 100 50 0

3h

6h

neo67/ þ

Figure 4 Radiation-induced apoptotic response in the cerebellum of Ptc mice irradiated at P1 and P10. H&E-stained sections of P1- (a) and P10-cerebellum (b) 6 h after irradiation, showing a lower frequency of pyknotic nuclei in the EGL in P1- compared to P10cerebellum. (c) Quantitative representation of radiation-induced pyknotic nuclei out of the total number of cells in the EGL, showing that P1-cerebellum is significantly less sensitive to radiation compared to P10-cerebellum at both postirradiation times analysed (Po0.001). Immunohistochemical analysis of caspase-3 activation in irradiated P1- (d) and P10-cerebellum (e), showing a lower number of positively stained cells in P1- compared to P10-cerebellum, 3 h post-irradiation. (f) Western blot of cerebellar homogenates from P1- and P10-irradiated mice 3 and 6 h postirradiation, showing increased levels of the 85 kDa PARP fragment in P10- compared to P1-cerebellum at both postirradiation times. (g) Graphic representation of densitometric immunoblot analysis. (a), (b), (d), (e) Scale bar ¼ 20 mm.

P10. A strong expression of active caspase-3, as revealed by immunohistochemistry 3 h postirradiation, was observed in the EGL of the cerebellum irradiated at P10 (Figure 4e), compared to a slight expression in the cerebellum irradiated at P1 (Figure 4d). One of the essential substrates cleaved by caspase-3 is poly (ADP-ribose) polymerase (PARP). Cleavage of the entire 116-kDa PARP protein into two fragments of 85 and 26 kDa has been considered indicative of functional caspase-3 activation. We used Western blot analysis to determine generation of caspase-3-cleaved PARP in the cerebellum of P1 and P10 mice at 3 and 6 h after irradiation with 3 Gy of X-rays (Figure 4f). Our results show 2.2 (3 h) and 1.7-fold (6 h) increased intensity of the 85 kDa PARP antibody band corresponding to the major PARP fragment in P10- compared to P1-irradiated cerebellum (Figure 4g). P53 expression in P1 or P10 cerebellum following irradiation Since the p53 gene product is required for radiationinduced apoptosis in cerebellar granule neurons (Wood and Youle, 1995), we characterized immunohistochemically the p53 pattern of expression after radiation exposure. We show a dramatic difference in p53 expression in the EGL between cerebellum of mice irradiated at P1 or P10 (Figure 5a and b). While a strong p53 expression was detected in the cerebellum irradiated at P10 (Figure 5b), lack of detectable p53 expression was shown in the cerebellum irradiated at P1 (Figure 5a). Oncogene

Western blot analysis was used to determine p53 protein levels in the cerebellum after irradiation with 3 Gy of X-rays (Figure 5c). As shown in Figure 5d, a sevenfold lower protein level was observed in the cerebellum of mice irradiated at P1 compared to mice irradiated at P10. This difference was reduced to 1.6fold 6 h after irradiation, as a result of consistent increase in p53 protein level in the cerebellum irradiated at P1, and of concurrent sharp decrease after irradiation at P10. These findings indicate a significant delay and a consistent decrease in the activation of p53 in response to cellular stress induced by radiation in cerebellum of mice at P1. As expected, p53 protein was undetectable in unirradiated P1 and P10 cerebellum (not shown). Activation of the phosphatidylinositide 30 -OH kinase (PI3K)-Akt/Pkb pathway in P1 and P10 cerebellum We asked whether activation of PI3K-Akt/Pkb pathway is involved in lack of p53 accumulation in the mouse cerebellum irradiated at P1. Thus, we performed Western blot analysis to determine total and pS437 Akt/Pkb protein levels in the cerebellum of unirradiated and irradiated mice (Figure 5e). In unirradiated cerebella, we observed a very large difference in Akt/ Pkb phosphorylation status between P1 and P10 mice (Figure 5f). In fact, 100% of phosphorylated Akt/Pkb was observed at P1 compared to about 50% at P10. Interestingly, irradiation decreased the expression of pS473 Akt, which was reduced to 47% in P1- and 13% in P10-irradiated cerebellum 3 h after irradiation. The


1169

a

b

d

RX 6h P1 P10

p53/Actin protein (%)

RX 3h P1 P10 p53 Actin

e pAKT

Control RX 3h RX 6h P1 P10 P1 P10 P1 P10 P10

AKT

750 600 450 300 150 0

P1 P10

3h

6h

f pAKT/AKT (%)

c

150

P1 P10

100 50 0

0h

3h

6h

Figure 5 Dramatic differences in p53 levels and in Akt/Pkb phosphorylation status were observed in the cerebellum of Ptcneo67/ þ mice irradiated at P1 and P10 with 3 Gy of X-rays. Immunohistochemical analysis showing lack of detectable p53 expression in P1irradiated mice (a), compared to strong p53 levels in the EGL of P10-irradiated mice (b) at 3 h postirradiation. (c) Immunoblot analysis of protein extracted from P1- and P10-cerebellum revealed a very low level of P53 expression in P1- compared to P10-cerebellum, 3 and 6 h after irradiation. (d) Graphic representation of densitometric immunoblot analysis in (c). (e) Akt/Pkb phosphorylation status was evaluated by immunoblotting in cerebellum from control P1 and P10 mice and in cerebellum irradiated at P1 or P10, 3 or 6 h postirradiation. Phosphorylation was immunodetected with phospho-Akt (Ser473) antibody, and membranes were re-blotted with an antibody that detected total Akt (lower panel). (f) Graphic representation of densitometric immunoblot analysis in (e). (a) and (b) Scale bar ¼ 20 mm.

Akt/Pkb phosphorylation status analysed 6 h after irradiation remained substantially unchanged in P1-, while it returned to the level of unirradiated controls in P10-irradiated cerebellum.

contrast, amplification of genomic DNA from normal tissue of Ptcneo67/ þ heterozygotes showed the presence of both Ptc1 alleles (Figure 6b).

Allelic imbalance analysis at the Ptc1 locus in medulloblastomas Loss of the wt Ptc1 allele was examined in medulloblastomas. This was performed by discriminating the nucleotide sequence of the targeted Ptc1 allele – derived from 129Sv ES cells – from the sequence of the wt allele – derived from CD1 mice – based on the presence of a polymorphism (T/C) in position 4016 of the Ptc1 gene (Calzada-Wack et al., 2002). Sequence analysis of cDNA (two samples) showed expression of the 129Svderived mutant allele and lack of expression of the CD1-derived wt mRNA in radiation-induced medulloblastomas from Ptcneo67/ þ mice. In agreement, sequence analysis of genomic DNA (seven samples) from radiation-induced medulloblastomas showed lack of the CD1 wt Ptc1 sequence (Figure 6a). This indicates that mutation rather than epigenetic silencing was responsible of Ptc1 loss of function in medulloblastomas. In

Discussion Ionizing radiation synergize with Ptc1 mutation to induce medulloblastoma at high incidence. The underlying molecular events, however, remain unclear. Induction of medulloblastoma by radiation in Ptcneo67/ þ mice offers the advantage of controlled timing of DNA damage induction, and allows in vivo analysis of early alterations in cerebellar signalling that can help to clarify the steps of the tumorigenic process from initial gene deregulation to medulloblastoma development. In this study, we have identified a narrow timewindow of radiosensitivity in Ptcneo67/ þ cerebellum, since even small age-differences at the time of irradiation produced drastic changes in medulloblastoma tumorigenesis. Medulloblastoma incidence, in fact, ranged from 81 to 3% after irradiation at P1 or P10. A striking difference was also observed in the frequency of Oncogene


1170

Figure 6 Sequencing for detection of LOH at Ptc1 locus in medulloblastomas. Genomic and tumor DNAs were analysed by sequencing of the CD1-derived (wt) and 129Sv-derived (mu) polymorphism at position 4016 of murine Ptc1 (Calzada-Wack et al., 2002). (a) Representative electropherogram of DNA extracted from medulloblastomas, showing only the 129Sv allele, and demonstrating loss of the wt Ptc1 allele. (b) Genomic DNAs revealed both CD1- and 129Sv-derived polymorphisms, demonstrating retention of both Ptc1 alleles in normal tissue.

cerebellar preneoplastic lesions 3 weeks postirradiation. Marked enhancement of hyperproliferation areas was obtained after irradiation at P1, as opposed to drastic suppression at P10. Difference in frequency of hyperproliferative lesions between mice irradiated at P1 or P10 (100 versus 14%) was highly statistically significant (P ¼ 0.0047). We reported previously a strong enhancement of medulloblastoma (51%) after irradiation of neonatal Ptcneo67/ þ mice at P4, as opposed to lack of enhancement by irradiating adult Ptcneo67/ þ mice (Pazzaglia et al., 2002). In view of the well-known higher radiosensitivity of younger actively dividing cells in very early postnatal cerebellum compared to postmitotic cells in older rodents (Anderson and Stromberg, 1977; Cerda, 1983), we correlated the ability of radiation to induce medulloblastoma in neonatal age with the rapid expansion of CGNPs from which medulloblastoma originates. However, here, we present evidence that irradiation of Ptcneo67/ þ mice at P10 fails to induce medulloblastoma despite evident ongoing proliferation in P10-cerebellum, shown by BrdU incorporation in CGNPs of the EGL (Figure 2d). These observations suggest that additional factors influence CGNPs susceptibility to neoplastic transformation. Interestingly, lower incidences of hyperproliferation areas (14 versus 60%) were detected in mice irradiated at P10 compared to untreated mice. This observation implies that irradiation of P10-cerebellum is not only unable to initiate new lesions, but it also exerts a strong cell killing effect on the existing spontaneous lesions. Consistent with this observation, cellular and biochemical investigation of radiation-induced apoptotic Oncogene

response of cerebellar neurons at P1 and P10 showed that the number of pyknotic nuclei, and the expression of active caspase-3 and of cleaved PARP are substantially lower in CGNPs of P1- compared to P10-irradiated mice, thus revealing development-related differences between the overall levels of cell death in response to ionising radiation. Moreover, these results suggest that medulloblastoma formation in Ptc1 mutants may be associated with resistance to radiationinduced cell killing. A number of studies have shown that radiationinduced cell death in the developing nervous system is p53-dependent (Wood and Youle, 1995; Enokido et al., 1996; Herzog et al., 1998; Johnson et al., 1998). To clarify the molecular mechanisms responsible of the different response to radiation of CGNPs in P1- and P10-cerebellum, we investigated the expression of p53. In normal unstressed cells, p53 is an unstable protein with a half-life of less than 20 min, and it is kept at very low levels owing to continuous degradation. Immunohistochemical analysis showed that p53 induction after irradiation is mostly localized to the EGL (Figure 5a and b), the same region that undergoes apoptosis in response to genotoxic stress, suggesting that the apoptotic process occurring in irradiated cerebellum is largely driven by activation of p53. Moreover, immunoblot and immunohistochemical analysis showed a drastically decreased level of p53 in P1- compared to P10-irradiated cerebellum (Figure 5a–d). Our data show an association between the delayed and decreased p53 expression and the resistance to radiation-induced cell killing observed in CGNPs at P1. Having shown that there is increasing radiationinduced apoptosis as cerebellar development progresses, we focused on the PI3K-Akt/Pkb pathway, which is a critical mediator of neuronal survival (Dudek et al., 1997). We found that the level of phosphorylation of Akt at residue S437, essential for its full activation, is considerably higher in both unirradiated and irradiated cerebellum at P1 than in corresponding cerebellum at P10, thus suggesting that Akt/Pkb activation may be critical for regulation of the balance between cell death and survival. The inverse correlation between the levels of phosphorylated Akt/Pkb and p53 in irradiated cerebellum supports a regulatory interaction between the PI3K-Akt/Pkb pathway and p53, as also suggested by prior reports (Buckbinder et al., 1995; Singh et al., 2002). Different levels of activated Akt/Pkb in P1 and P10 cerebellum may reflect variations in insulin-like growth factors (IGFs). These widely expressed factors are temporally and spatially distributed in the CNS, as they play a significant role in the stimulation of brain development (D’Ercole et al., 1996). IGF-I predominantly expressed in neuronal cells, has been reported to protect CGNPs from apoptosis via PI3K and Akt/Pkb (Dudek et al., 1997) and transgenic mice overexpressing IGF-I, exhibit a doubling of CGNPs number and show decreased expression of caspase-3 in cerebellum (Ye et al., 1996; Chrysis et al., 2001). IGF-II, mainly restricted in non-neuronal cells (meningies and choroids


1171

plexus), has been strictly involved in Shh-dependent tumor formation. Deletion of the Igf2 gene, in fact, completely eliminates medulloblastoma and rhabdomyosarcoma development in Ptcneo67/ þ /Igf2 double knockout mice (Hahn et al., 2000), and coexpression with IGF2 or Akt/Pkb strongly enhances the effect of Shh-induced medulloblastoma formation in mice (Rao et al., 2004). This suggests a synergy between the Shh and PI3K pathways in Shh/Ptc1-associated tumorigenesis. While it was initially proposed that loss of one Ptc1 allele was sufficient for medulloblastoma formation in mice (Wetmore et al., 2000; Zurawel et al., 2000b), recent reports suggest that Ptc1 expression is drastically reduced or lacking in medulloblastomas (Berman et al., 2002; Tong et al., 2003). To analyse this issue, allelic imbalance at the Ptc1 locus was examined in cDNAs and DNAs from radiation-induced medulloblastomas. In agreement with our previous results (Pazzaglia et al., 2002), we show that all tumors are characterized by loss of the remaining wt Ptc1 allele, confirming loss of Ptc1 function as a key event in radiation-induced medulloblastoma tumorigenesis. Double-strand break (DSB) repair involving homologous chromosomes in mitotically dividing cells is a potential mechanism leading to loss of heterozygosity (LOH) (Stark and Jasin, 2003). In view of the recently proposed additional role for the p53 protein as a suppressor of homologous recombination (HR) (Sturzbecher et al., 1996; Linke et al., 2003; Sengupta et al., 2003), and based on evidences showing that functional inactivation of wt p53 results in elevated HR rates (Saintigny and Lopez, 2002; Blackburn et al., 2004), we speculate a role of mitotic recombination (MR) as a pathway to LOH in our tumors. Decreased p53 expression in P1 cerebellum may, in fact, promote DSB repair by MR. In line with this hypothesis, p53 loss has been shown to dramatically accelerate the age of onset and incidence of medulloblastoma in Ptc1 heterozygotes; 95% of the Ptc1 þ //p53/ mice developed medulloblastomas by 12 weeks of age (Wetmore et al., 2001). Experiments aimed to test the contribution of DNA damage signalling and repair in medulloblastoma formation are in progress. In summary, we have shown that the response of the cerebellum to ionizing radiation exposure depends on developmental stage, as demonstrated by the very large difference in medulloblastoma incidence, frequency of cerebellar preneoplastic lesions and apoptotic response between P1 and P10 cerebellum. Notably, this agedependence cannot be ascribed to obvious differences in the proliferation status of CGNPs at the two postnatal ages. Moreover, we have shown that susceptibility to medulloblastoma induction and marked resistance to radiation-induced cell death observed at P1 are associated with p53 inhibition and Akt/Pkb activation in CGNPs. In contrast, resistance of P10-cerebellum to medulloblastoma induction is associated with susceptibility to radiation-induced cell death, probably mediated by strong p53 expression and lower Akt/Pkb activity.

This study helped to elucidate the in vivo molecular mechanisms involved in Ptc1-associated susceptibility to radiation induced medulloblastoma tumorigenesis from early radiation response of CGNPs to acquisition of overt malignancy, providing insights into cerebellar signalling implicated in the pathogenesis of one of the most deadly childhood tumors.

Materials and methods Mice Mice lacking one Ptc1 allele, generated through disruption of exons 6 and 7 in 129/SV ES cells, and maintained on a CD1 background were used for this study. Ptcneo67/ þ mice were genotyped using PCR primers specific to the neo insert and wt regions as described (Hahn et al., 1998). Animals were housed under conventional conditions with food and water available ad libitum and a 12-h light cycle. Experimental protocols were reviewed by the Institutional Animal Care and Use Committee. Irradiation X-ray irradiation was performed using a Gilardoni CHF 320 G X-ray generator (Gilardoni S.p.A., Mandello del Lario, Lecco, Italy) operated at 250 kVp, 15 mA, with filters of 2.0 mm Al and 0.5 mm Cu (HVL ¼ 1.6 mm Cu). For tumorigenesis experiments, 21 Ptcneo67/ þ heterozygous mice and 20 wt littermates were irradiated at P1; 33 Ptcneo67/ þ and 24 wt littermates were irradiated at P10. The X-ray dose was 3 Gy for all groups. Histological analysis and tumor quantification Mice were observed daily for their whole life span. Upon decline of health (i.e., severe weight loss, paralysis, ruffling of fur or inactivity) or when tumors were visible, they were killed and autopsied. Whole brain and any other visible mass were partly fixed in 4% buffered formalin and partly snap frozen. Samples were then embedded in paraffin wax, sectioned and stained according to standard techniques. Tumor incidence was expressed as the percentage of mice with one or more tumors. Incidence of microscopic cerebellar lesions was determined on asymptomatic Ptcneo67/ þ heterozygous mice irradiated with 3 Gy of X-rays at P1 or P10 and sacrificed 3 weeks post irradiation, and on age-matching unirradiated mice. Five to seven mouse brains were collected at each time point. In all, 24 sections were examined for each cerebellum: to ensure a representative sampling, each set of sections (n ¼ 8) in an individual slide was recovered with an interval of 70 mm. Radiosensitivity analysis Brains (two per time point) of Ptc1 heterozygous pups irradiated with 3 Gy of X-rays at P1 or P10 were collected at 3 and 6 h postirradiation and fixed in 4% buffered formalin. Serial sections of cerebellar tissues were cut at 5-mm thickness and stained with H&E. In all, 13–15 digital images/mouse from midsagittal cerebellar sections of P1 pups, covering the entire cerebellar surface, and 25 digital images/mouse from P10 pups, covering most of the cerebellar surface, were collected by IAS image-processing software (Delta Sistemi, Rome, Italy). Cells showing signs of nuclear chromatin condensation and morphologically normal cells in the EGL were counted using a double-blind method. Apoptotic values were calculated as the percentage pyknotic nuclei relative to Oncogene


1172 the total number of cells. Approximately 5 103 or 104 cells were examined for each time point in the EGL of P1 and P10 cerebellum, respectively. Antibodies Antibodies included rabbit polyclonal antibody against activecaspase-3 (Santa Cruz Biotechnology, Santa Cruz, CA, USA); rabbit polyclonal antibody against p53 (Novocastra Laboratories, Newcastle, UK); monoclonal antibody against PARP (Trevigen, Gaithersburg, MD, USA); rabbit polyclonal antibodies against phospho-Akt (Ser473) and total Akt (Cell Signaling, Beverly, MA, USA); monoclonal antibody against b-actin (Sigma-Aldrich Inc., St Louis, MO, USA). Immunohistochemistry and immunoblotting We carried out immunohistochemical analysis on cerebellum sections as previously described (Mancuso et al., 2004). For immunoblotting, we normalized total cell extracts for protein concentration (Bradford assay, Bio-Rad Laboratories Inc., Hercules, CA, USA), separated equal amounts of protein (30 mg) by SDS–PAGE and transferred them to polyvinylidene difluoride (PVDF) membranes Hybond P þ (Amersham Biosciences, Bucks, UK). Western blot analysis using specific antibodies was carried out as described (Baldi et al., 1995). Blots were developed using horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology). Proteins were visualized by chemiluminescence detection (SuperSignal West Pico Chemiluminescent Substrate, Pierce, Rockford, IL, USA). Protein levels were quantified by densitometric analysis using the Scion Image Beta 4.02 software package (Scion Corporation, Frederick, MD, USA). The level of actin expression was used as a loading control. BrdU incorporation assay To examine proliferation of CGNPs, mice were injected intraperitoneally with BrdU in PBS (50 mg/g of body weight) (Sigma-Aldrich). After 1 h, they were euthanized and their

tissues processed as described for histological analysis. Immunostaining of incorporated BrdU was performed on paraffin sections using BrdU Staining Kit (Zymed Laboratories Inc., San Francisco, CA, USA) according to the manufacturer’s instructions. Isolation of DNA and total RNA Genomic and tumor DNAs were extracted from tail and medulloblastomas by Wizard SV Genomic DNA Purification System (Promega Corporation, Madison, WI, USA). Total RNA was isolated from medulloblastomas using NucleoSpin Extract (Macherey-Nagel, Duren, Germany). Total RNA (2 mg) was reverse-transcribed using RETROscriptt (Ambion Inc., Austin, TX, USA) according to the manufacturer’s instruction. LOH at the Ptc1 locus in medulloblastomas LOH analysis at the Ptc1 locus was performed by amplification and sequencing of exon 23 – from genomic DNA and cDNA samples – of the murine Ptc1 gene, which holds a polymorphic site at position 4016. The primers used for the study were primer pair 8F/9R (Calzada-Wack et al., 2002). Sequencing reactions were performed by means of dye terminator chemistry using a Big Dye Terminator Vers. 3.1 Sequencing Cycle kit and a 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA). Loss of polymorphic sites in neoplastic compared to nonneoplastic tissue was considered LOH at the Ptc1 locus. Acknowledgements We thank Orsio Allegrucci for assistance in handling and treating the mice, and Emanuela Pasquali for excellent technical help. We also thank Vincenzo Cesi for assistance with the experiments shown in Figures 4 and 5. This work was funded in part by a grant from the Commission of European Communities (FI6R-CT-2003-508842 RISC-RAD).

References Anderson WJ, Stromberg MW. (1977). J Comp Neurol 171: 17–37. Baldi A, De Luca A, Claudio PP, Baldi F, Giordano GG, Tommasino M et al. (1995). J Cell Biochem 59: 402–408. Berman DM, Karhadkar SS, Hallahan AR, Pritchard JI, Eberhart CG, Watkins DN et al. (2002). Science 297: 1559–1561. Blackburn AC, McLary SC, Naeem R, Luszcz J, Stockton DW, Donehower LA et al. (2004). Cancer Res 64: 5140–5147. Buckbinder L, Talbott R, Velasco-Miguel S, Takenaka I, Faha B, Seizinger BR et al. (1995). Nature 377: 646–649. Calzada-Wack J, Kappler R, Schnitzbauer U, Richter T, Nathrath M, Rosemann M et al. (2002). Carcinogenesis 23: 727–733. Cerda H. (1983). Int J Radiat Biol Relat Stud Phys Chem Med 44: 113–121. Chrysis D, Calikoglu AS, Ye P, D’Ercole AJ. (2001). J Neurosci 21: 1481–1489. D’Ercole AJ, Ye P, Calikoglu AS, Gutierrez-Ospina G. (1996). Mol Neurobiol 13: 227–255. Dahmane N, Ruiz-i-Altaba A. (1999). Development 126: 3089–3100. Dudek H, Datta SR, Franke TF, Birnbaum MJ, Yao R, Cooper GM et al. (1997). Science 275: 661–665. Oncogene

Enokido Y, Araki T, Tanaka K, Aizawa S, Hatanaka H. (1996). Eur J Neurosci 8: 1812–1821. Goodrich LV, Milenkovic L, Higgins KM, Scott MP. (1997). Science 277: 1109–1113. Hahn H, Wojnowski L, Specht K, Kappler R, Calzada-Wack J, Potter D et al. (2000). J Biol Chem 275: 28341–28344. Hahn H, Wojnowski L, Zimmer AM, Hall J, Miller G, Zimmer A. (1998). Nat Med 4: 619–622. Hahn H, Wicking C, Zaphiropoulous PG, Gailani MR, Shanley S, Chidambaram A et al. (1996). Cell 85: 841–851. Herzog KH, Chong MJ, Kapsetaki M, Morgan JI, McKinnon PJ. (1998). Science 280: 1089–1091. Johnson MD, Xiang H, London S, Kinoshita Y, Knudson M, Mayberg M et al. (1998). J Neurosci Res 54: 721–733. Lee Y, McKinnon PJ. (2002). Cancer Res 62: 6395–6399. Linke SP, Sengupta S, Khabie N, Jeffries BA, Buchhop S, Miska S et al. (2003). Cancer Res 63: 2596–2605. MacDonald TJ, Rood BR, Santi MR, Vezina G, Bingaman K, Cogen PH et al. (2003). Oncologist 8: 174–186. Mancuso M, Pazzaglia S, Tanori M, Hahn H, Merola P, Rebessi S et al. (2004). Cancer Res 64: 934–941. Marino S, Vooijs M, van Der Golden H, Jonkers J, Berns A. (2000). Genes Dev 14: 994–1004. Pazzaglia S, Mancuso M, Atkinson MJ, Tanori M, Rebessi S, Di Majo V et al. (2002). Oncogene 21: 7580–7584.


1173 Raffel C, Jenkins RB, Frederick L, Hebrink D, Alderete B, Fults DW et al. (1997). Cancer Res 57: 842–845. Raffel C. (2004). Neoplasia 6: 310–322. Rao G, Pedone CA, Valle LD, Reiss K, Holland EC, Fults DW. (2004). Oncogene 23: 6156–6162. Saintigny Y, Lopez BS. (2002). Oncogene 21: 488–492. Sengupta S, Linke SP, Pedeux R, Yang Q, Farnsworth J, Garfield SH et al. (2003). EMBO J 22: 1210–1222. Singh B, Reddy PG, Goberdhan A, Walsh C, Dao S, Ngai I et al. (2002). Genes Dev 16: 984–993. Stark JM, Jasin M. (2003). Mol Cell Biol 23: 733–743. Sturzbecher HW, Donzelmann B, Henning W, Knippschild U, Buchhop S. (1996). EMBO J 15: 1992–2002. Tong WM, Ohgaki H, Huang H, Granire C, Kleihues P, Wang ZQ. (2003). Am J Pathol 162: 343–352.

Wallace VA. (1999). Curr Biol 9: 445–448. Wechsler-Reya RJ, Scott MP. (1999). Neuron 22: 103–114. Wetmore C, Eberhart DE, Curran T. (2000). Cancer Res 60: 2239–2246. Wetmore C, Eberhart DE, Curran T. (2001). Cancer Res 61: 513–516. Wood KA, Youle RJ. (1995). J Neurosci 15: 5851–5857. Ye P, Xing Y, Dai Z, D’Ercole AJ. (1996). Brain Res Dev Brain Res 95: 44–54. Zindy F, Nilsson LM, Nguyen L, Meunier C, Smeyne RJ, Rehg JE et al. (2003). Cancer Res 63: 5420–5427. Zurawel RH, Allen C, Chiappa S, Cato W, Biegel J, Cogen P et al. (2000a). Genes Chromosomes Cancer 27: 44–51. Zurawel RH, Allen C, Wechsler-Reya R, Scott MP, Raffel C. (2000b). Genes Chromosomes Cancer 28: 77–81.

Oncogene