Main

Human hereditary syndromes with genetic defects in the recognition and/or repair of damage to DNA frequently manifest, besides other clinical symptoms, early onset developmental and progressive neurodegeneration,1, 2 indicating that faulty responses to DNA damage are neuropathological. Most prominent among these syndromes is ataxia-telangiectasia (A-T), a recessive disorder presenting cerebellar ataxia, progressive neuronal degeneration, retinal telangiectasia, immunodeficiency and cancer predisposition.3 The occurrence of ataxia, which typically appears in infancy, reflects the progressive degeneration of the cerebellar cortex due to the loss of Purkinje and granule cells.4, 5 Loss of neurons in striatum and substantia nigra are also observed in A-T.6 A-T results from inactivating mutations in A-T mutated (ATM) gene coding for a kinase with an essential role in responding and transducing the signal arising from DNA double-strand breaks (DSBs).7 ATM, along with the MRN (Mre11-Rad50-Nbs1) complex senses DSBs, leading to ATM autophosphorylation on serine 1981, and phosphorylation of a plethora of substrates involved in cell-cycle checkpoint arrest, repair of breaks, chromatin remodeling and cell death.8, 9, 10 Concordant with the function of ATM in DNA-damage responses (DDR), cells from A-T patients have an impaired response to DSBs associated with defective phosphorylation of ATM substrates.3, 8

Although the neurodegenerative phenotype in A-T has been attributed to a defective DDR in pre- and postmitotic neurons,1 oxidative stress and endoplasmic reticulum stress may also play a causative role.11 Data from ATM-null mice support the notion that the cerebellar degeneration arises from a defective DDR,12 but consequent to oxidative stress.13, 14, 15 As developing neurons are rapidly proliferating and potentially generating high levels of oxidants, which favor the accumulation of DNA lesions, genetic defects of the DDR, by exacerbating genomic instability, may eventually compromise neuron survival particularly at postmitotic level. Patients with A-T show attenuated antioxidant capacity and increased oxidative stress. Furthermore, brains or astrocytes from ATM-deficient mice present elevated production of ROS and increased activation of the redox-sensitive kinases, ERK1/2.14, 16 Interestingly, the antioxidant N-acetyl-cysteine strongly suppresses ERK signaling and protects Purkinje cells from oxidative stress-induced degeneration. Moreover, dietary supplementation of N-acetyl cysteine (NAC) suppresses the mutational signature indicative of oxidative DNA damage in ATM-deficient mice.13

Genetically stable human neural stem cell lines (ihNSCs), which maintain self-renewal capacity and differentiation potential in vitro, can provide an important tool for neurobiological investigations.17, 18, 19 In this study, we showed that in ihNSCs, the regulation of the ATM-dependent pathway in response to ionizing radiation (IR) induced DNA damage and the outcome of ATM protein ablation by short hairpin RNA (shRNA) interference on survival, self-renewal and terminal differentiation. We show that the ablation of ATM in ihNSCs does not affect the proliferative activity, suppresses the response to DNA damage, attenuates the apoptosis associated with differentiation and mildly affects the differentiation and survival of GalC+ oligodendrocytes.

Results

DDR protein levels during terminal differentiation of ihNSCs

The protocol to culture the previously established ihNSCs19 is depicted in Figure 1a. The cells are propagated in suspension as neurospheres (N) and expanded following mechanical dissociation and replating. After 24 h, the cells appear under microscopy as doublets (D0), capable of generating new neurospheres with self-renewal potential. When seeded onto laminin-coated flasks without EGF, doublets mature into progenitors at day 3 (D3) and after the withdrawal of FGF, they undergo further time-dependent differentiation (D10, D17 and D24) into neurons, oligodendrocytes and astrocytes.

Figure 1
figure 1figure 1

Culture modality of ihNSC and expression of DDR proteins during differentiation. The scheme (a, top) illustrates the maintenance and differentiation of ihNSCs. Cells are grown as neurospheres (N) and are mechanically disaggregated for expansion, and replated in complete medium (plus EGF and FGF). After 24 h, a time point referred to as day 0 (D0), the cells appear under microscopy as doublets. To induce differentiation, doublets are plated on laminin-coated slides and cultured without EGF for 3 days, giving rise to neural progenitors (D3). The medium without EGF and FGF is then replaced every week to promote generation of neurons, oligodendrocytes and astrocytes, and three different time points are considered (D10, D17 and D24). The different cell types are detected by IF with lineage-specific phenotypic markers (a, bottom). In (b), the expression of DDR proteins was evaluated by western blot on lysates from ihNSCs harvested at the indicated days. Protein loading per lane was verified with antibodies anti-β-actin and vinculin. Band intensities relative to ATM, ATR, DNA-PK and p53 proteins, normalized to β-actin and vinculin, are depicted in the graphs (data from three independent experiments; mean±S.D.). (c) Immunoblot for p53 and p21waf1 performed on extracts from ihNSCs cells harvested at the indicated time points following differentiation. Normalized p53 and p21waf1 band intensities are shown in the graphs. (d) Immunoblot, evidencing induction of the p53-responsive p21waf1 and increase of the apoptosis-associated cleaved PARP following ihNSCs differentiation. β-actin verified protein loading per lane. (e) DNA flow cytometry histograms showing the reduction of S-phase and increase of subdiploid apoptotic cells during differentiation

Full size image

To validate the ihNSCs model, we examined the expression of components of the DDR on immunoblots. ATM, the ATM-related kinase ATR, DNA-PKcs, the members of the MRN complex NBS1-Mre11-RAD50, Chk1, Chk2 and Smc1, showed basal levels of expression at D0 stage, whereas p53 was almost undetectable (Figure 1b). At D3, the levels of these molecules did not change appreciably, whereas at D10, p53 markedly increased, Chk2 began to decline and Chk1 sharply dropped. Likewise, the levels of ATM, and even more those of DNA-PK, decreased, whereas the levels of ATR dropped sharply. Between D17 and D24, ATR was virtually undetectable, ATM, particularly DNA-PKcs, increased significantly, whereas p53 returned to baseline (Figure 1b, right graphs for quantified immunoblot data for ATM, ATR, DNA-PK and p53). A narrower time–course analysis showed a rise in p53 protein beginning at D6 and reaching a plateau at D10, paralleled by a drastic increase in the levels of the p53-regulated p21waf1 (Figure 1c). To determine whether the differentiation involved the endogenous activation of ATM, we analyzed the autophosphorylation of ATM-Ser1981 and phosphorylation of its substrates Smc1-S966, Chk2-T68 and p53-S15. On western blots, the ATM-pS1981 signal was very faint (Supplementary Figure S1), and required a long exposure time to be detected, and when normalized for the total amounts of ATM, it appeared not to change appreciably during differentiation. On the other hand, SMC1-pS966, Chk2-pT68 and p53-pS15 signals became detectable at D10 (Supplementary Figure S1).

Interestingly, the accumulation of p53 and p21waf1 during differentiation was paralleled by the appearance of cleaved PARP (a marker of apoptosis; Figure 1d), and according to flow cytomery analysis, associated with the loss of DNA replication and increased apoptosis, accounting for >25% and <2% at D0-D3, and <1% and >20% from D10 onward (Figure 1e), respectively. Thus, the differentiation of ihNSCs is accompanied by a modulation of proteins of the DDR pathway and the onset of apoptosis in a fraction of cells that correlates with the functional induction of p53.

ATM protein depletion in ihNSCs does not affect the proliferation but attenuates the DDR

To establish an in vitro model system that recreates the A-T phenotype, we used the lentiviral shRNA interference, as described in Materials and Methods. Of note, in ihNSCs gene silencing with lentiviruses was more effective than with retro- or adenoviruses, whereas silencing with synthetic siRNAs was totally ineffective, irrespective of the transfection reagents used. For ATM, we generated ihNSCs separately transduced with two different shRNA sequences targeting nt 268–286 and nt 1267–1285 of the ATM transcript. As in both cases, ATM was equally depleted and the biological properties were similar, the results presented hereafter refer to the cells denoted shATM silenced with the sequence targeting nt 1267–1285. shATM expressed 5–10% of the ATM levels seen in control interfered cells (shCon; Figure 2a). The depletion of ATM did not affect the proliferative rate (Figure 2b), the DNA replication as assessed by the BrdU incorporation (Figure 2c), or the chromosomal stability, as revealed by cytogenetic analysis showing a normal karyotype in both control and ATM silenced clones (Figure 2d).

Figure 2
figure 2

ATM depletion does not affect the proliferation or chromosomal stability of ihNSCs. Characterization of shATM and shCon cells stably expressing short hairpin RNAi targeting ATM and a control sequence, respectively. (a) Western blot analysis for ATM protein. (b) Proliferative rate over a period of 30 days. (c) DNA replication assessed by dual fluorescence flow cytometry analysis of cells labeled with BrdU-FITC and propidium iodide to stain the S-phase and total DNA, respectively. (d) karyotypic analysis showing no evidence for chromosomal differences or abnormalities following ATM depletion

Full size image

IR-induced DSBs activate the ATM-dependent DDR pathway, causing the autophosphorylation of ATM on Ser1981 and phosphorylation of its substrates.20, 21, 22 In ATM-deficient cells, the phosphorylation of the ATM targets is markedly attenuated.23 To determine the effect of ATM depletion in our cells, we analyzed the genotoxic response on western blots using antibodies specific for ATM phosphoresidues. At D0, a dose of IR as low as 0.5 Gy induced in shCon the phosphorylation of ATM-S1981 and of its substrates Smc1-S966, Chk2-T68 and p53-S15 (Figure 3), though this response was more vigorous with higher doses of IR. A more detailed time- and dose-dependent IR response at D0 (Supplementary Figure S2) showed that the activity of ATM is elicited by even 0.25 Gy and abrogated by 5–10 μM of KU-55933 (Supplementary Figure S2A), a selective inhibitor of ATM. The response to IR in shATM cells was considerably attenuated, as revealed by the reduced phosphorylation of ATM-S1981, Smc1-S966, Chk2-T68 and p53-S15, particularly at low doses of IR (Figure 3).

Figure 3
figure 3

The response to IR-induced DNA damage by ihNSC-shCon and ihNSC-shATM. The autophosphorylation of ATM-S1981 and phosphorylation of the ATM targets Smc1-S966, Chk2-T68 and p53-S15 were analyzed by western blot on cells at D0 (a) irradiated with 0.5 to 5.0 Gy and harvested 45 min later. Normalized band intensities for ATM, Smc1, Chk2 and p53 phosphoresidues are shown in the histograms (b). For each IR dose, the difference between shCon and shATM was statistically significant (P<0.001) except for those indicated with # (analysis performed by the Student's t-test)

Full size image

IR-induced formation and loss of γ-H2AX nuclear foci and genotoxic response

Several proteins of the DDR are rapidly recruited to sites of breaks, leading to the formation of nuclear foci22, 23 that are cleared upon DNA repair. We enumerated γ-H2AX foci in ihNSCs at D0 treated with 0.25 or 1 Gy IR and collected 5 and 60 min and 24 h later (Figure 4a). Unirradiated shCon and shATM cells showed 2÷3 basal foci/cell, but 5 min after 0.25 and 1.0 Gy IR the number of foci in the former increased to 9.2±2.4 and 21±2.9, respectively, to decline at 60 min and even further at 24 h, whereas in the latter, it increased more modestly, accounting for 14±4 at 60 min after 1 Gy, and persisting at 24 h (Figure 4b). These findings thus indicate that ihNSCs are highly proficient in signaling and repairing DNA lesions, and that ATM deficiency impairs these responses, as seen in other cell types.24

Figure 4
figure 4

Formation and loss of IR-induced γ-H2AX nuclear foci. Cells at D0 were irradiated with 0.25 and 1.0 Gy, collected at the indicated times and IF labeled for γ-H2AX. (a) Representative γ-H2AX foci present in untreated and treated cells. (b) For each treatment, the number of foci were scored from >100 cell nuclei per duplicate preparations and from three independent experiments (mean±S.D.). For each time point and IR dose, the difference between shCon and shATM was statistically significant (P<0.001) except for those indicated with # (analysis performed by the Student's t-test)

Full size image

The time- and dose-dependent genotoxic response assessed in D0 cells by DNA flow cytometry showed the presence of subdiploid apoptotic cells at 24 h even in response to 0.25 Gy (Figure 5a), and this effect was more pronounced in shCon than shATM cells (Figure 5b), indicating that ATM deficiency attenuates the short-term apoptosis, concordant with the findings in lymphoblastoid cells from A-T patients.25

Figure 5
figure 5

Cell cycle phase distribution and apoptosis after IR treatment. Cells at D0 stage were treated with escalating doses of IR, harvested 24 or 48 h later, and analyzed for the DNA content by flow cytometry (a). The percentage of cells with a subdiploid DNA content (apoptotic cells; b) has been calculated from the analysis of the DNA histograms from three independent experiments. For each IR dose, the difference between shCon and shATM was statistically significant (P<0.01 in all cases except those indicated with ***P<0.05). Those indicated with # showed no statistical significance (analysis performed by the Student's t-test)

Full size image

ATM depletion and linage differentiation

As neurodegeneration is the hallmark of the A-T, we determined whether ATM depletion somehow affected the apoptosis and differentiation of ihNSCs, which, as shown above (Figure 1e), are intrinsically linked. We first assessed the presence of cleaved caspases on western blots, which are markers of apoptosis, and found consistently higher amounts of cleaved caspases 3 and 9 in shCon than in shATM, particularly at D17 (Figure 6a), and likewise, the accumulation of p53 and p21waf1 was less pronounced in the latter (densitometry analysis of the bands are shown in Figure 6b). It should be noticed that for each time point during differentiation, ATR was expressed at similar levels in both shATM and shCon cells, and furthermore, the residual amount of ATM in shATM did not change with differentiation (Supplementary Figure S3). The potential impact of ATM deficiency on differentiation was determined by western blot and immunofluorescence (IF) analysis using antibodies specific for neurons (β-tubulin III and MAP2), astrocytes (glial fibrillary acidic protein (GFAP)) and oligodendrocytes (GalC and CNPase). Upon differentiation, the fraction of neurons positive for MAP2 (Figure 7a) and β-tubulin III (not shown) accounted for about 20% at D10 and D17, both in shCon and shATM. GalC-positive oligodendrocytes, which accounted for 7% at D10, both in shCon and shATM, increased to 11% at D14 in the former, but not in the latter. The presence of GFAP-reactive astrocytes, determined by western analyses, showed no major differences between shCon and shATM at any differentiation stage (Figure7b and Supplementary Figure S4 for representative IF analysis). In addition, CNPase levels augumented between D10 and D17 in shCon, whereas in shATM did not increase (Figure 7b), in agreement with the above data on GalC-positive oligodendrocytes. These results indicate that ATM deficiency appears not to impair neurogenesis, though it partially affects glial cells.

Figure 6
figure 6

Expression of apoptotic markers during differentiation. shCon and shATM cells were collected at various times after differentiation and examined by western blot (a) with antibodies recognizing cleaved caspases (markers of apoptosis), p21waf1 and p53. (b) Relative levels of the indicated markers, determined by densitometric analysis of the western blots from two independent experiments. For each time point, the difference between shCon and shATM was statistically significant (*P<0.01; **P<0.001; ***P<0.05). Those indicated with # showed no statistical significance (analysis performed by the Student's t-test)

Full size image
Figure 7
figure 7

ATM depletion and effects on differentiation. Cells were induced to differentiate on glass coverslips, and then IF-labeled with antibodies specific for neurons (β-tubulin III and MAP2), oligodendrocytes (GalC) and astrocytes (GFAP). Nuclei were counterstained with DAPI and IF-labeled cells scored by fluorescence microscopy. Results in (a) were obtained from the analysis of three independent experiments (mean±S.D.). For each time point and treatment, the difference between shCon and shATM was statistically significant (**P<0.001; *P<0.05). All other differences were not statistically significant (analysis according to χ2-test). (b) Expression of neuron, oligodendrocyte (CNPase) and astrocyte markers were evaluated by western blot. The densitometic analysis of CNPase bands after normalization is reported in the graph. For each time point, the difference between shCon and shATM was statistically significant (**P<0.001). All other differences were not statistically significant (analysis performed by the Student's t-test)

Full size image

As increased oxidative stress is believed to contribute to neural cell death in ATM deficiency, we analyzed the sensitivity of nonreplicating differentiated cells to oxidative stress induced by buthionine sulphoximine (BSO) as well as to the antioxidant NAC. Although neurons cells were not significantly affected by BSO or NAC, irrespective of the expression of ATM (Figure 7a), oligodendrocytes were rather sensitive to BSO when depleted of ATM, according to GalC positivity accounting for 9 and 4% at D17 in shCon and shATM, respectively. Interestingly, this difference could be rescued by treatment with NAC (Figure 7a).

To further characterize the effect of ATM deficiency on glial cell differentiation, we analyzed the temporal expression of oligodendrocyte markers. NG2, PDGFR-α and Olig2, relatively specific markers for oligodendrocyte precursors,26 were found by quantitative flow cytometry expressed in a similar proportion of cells (>60%) and with similar intensity both in shCon and shATM neurospheres and at D0 (Supplementary Figure S5). The expression of these oligodendrocyte lineage markers in a large proportion of our neural stem cells is in agreement with the findings in murine neurospheres.26 The flow cytometry analysis showed also a faint expression of CNPase, a marker of immature oligodendrocytes,27 and virtual absence of GalC staining (Supplementary Figure S5). The regulation during differentiation of Olig2 and PDGFR-α assessed on western blots showed a modest reduction in Olig2 levels between D0 and D3 in shATM relative to shATM, and a marked drop in PDGFR-α levels at D10, more effective in shATM than shCon, particularly at D17 (Supplementary Figure S6). O4, a marker of differentiated oligodendrocytes, was rarely detected by IF between D10 and D17 (<1 positive cells per glass slide as average; Supplementary Figure S4, bottom), thus precluding any statistical correlations with ATM expression. Altogether, ATM deficiency appears not to impair the differentiation of ihNSCs toward neuronal and glial lineages, though it makes oligodendrocytes more vulnerable to oxidative stress.

Discussion

The link between mutations in genes of the DNA repair machinery and neurodegeneration in various human syndromes, such as A-T illustrates, the importance of a proper DNA repair for homeostasis of the nervous system.1 In A-T, the neuropathological condition results from Purkinje cell death in the cerebellum and malfunction of other cells of the central nervous system.3, 28 ATM, the protein defective in A-T, plays a key role in initiating and propagating the intracellular response to DNA DSBs by phosphorylating several substrates involved in cell cycle control, DNA repair processes, apoptosis, senescence and transcription.7, 10 Besides, ATM is involved in sensing and maintaining the redox status, and indeed cells from patients with A-T exhibit reduced antioxidant capacity and persistent oxidative stress,29 whereas brains from ATM-deficient mice show enhanced production of ROS and increased markers of oxidative stress.16 Whether neurodegeneration in A-T primarily arises from the inappropriate response to DNA damage or increased oxidative stress remains to be clearly established.

A renewable source of human neural stem cells would certainly significantly facilitate studies on neurogenesis and neurodegeneration, particularly in combination with RNA interference strategies to suppress the expression of specific genes. Such cell lines, recently established from fetal brains using v-myc immortalization,17, 19, 30, 31 retain self-renewal potential and capacity to differentiate into the three major neural lineages, astrocytes, oligodendrocytes and neurons. Interestingly, transplantation of a conditionally immortal human neural stem cell line into the brain has allowed functional recovery in a rat model of stroke disability.17

To get an insight into the mechanisms of neuronal degeneration in A-T, we used the human neural stem cell line ihNSC,19 to evaluate the ATM-dependent response to DNA damage and the differentiation potential following ATM depletion. We have shown that these cells constitutively express at D0 stage ATM, ATR, Nbs1, Mre11, RAD50, Chk1, Chk2 and Smc1, but the levels of these proteins change upon differentiation. In particular, ATM and DNA-PK levels rise by D24, whereas ATR as well as Chk1 irreversibly drop at D10. The p53 protein shows a transient increase paralleled with the induction of the p53-responsive gene p21waf1 and onset of apoptosis in a fraction of cells. The accumulation of p53 and induction of p21waf1 actually starts at D6, and persists up to D10 and D17 in the case of p53 and p21waf1 respectively. The DDR of ihNSCs appeared normal, since even low doses of IR (e.g., 0.25 Gy, predicted to induce 9 DSBs/cell) rapidly activated the ATM pathway, revealed by the formation of γ-H2AX nuclear foci, autophosphorylation of ATM-S1981, phosphorylation of the ATM substrates Smc1-S966, Chk2-T68 and p53-S15, and by the abrogation of these events using the ATM inhibitor KU-55933. Likewise, ihNSCs were proficient in repairing the IR-induced DBSs, according to the time-dependent loss of γ-H2AX foci.

To determine the effect of ATM deficiency on the biological properties of ihNSCs, we have generated cells stably expressing shRNA sequences targeting ATM or a control sequence. Although depletion of ATM in shATM cells did not affect the growth rate, DNA replication or chromosomal stability, it severely attenuated the response to IR-induced DNA damage, shown by the reduced phosphorylation of Smc1-S966, Chk2-T68 and p53-S15, particularly in response to low doses of IR. Moreover, ATM deficiency markedly impaired the kinetics of repair of DSBs, evidenced by the reduced number of γ-H2AX foci formed immediately after irradiation and persistence up to 24 h, in contrast with control silenced ihNSCs showing rapid formation and the loss of most foci by 24 h. ATM deficiency, in addition, attenuated the apoptotic response to IR, like what observed in lymphoblastoid cells from A-T patients,25 and which reflects the attenuated ATM-dependent activation of p53, required for the DNA damage-induced apoptosis.7

We then analyzed the requirement of ATM during differentiation, and found similar numbers (∼19%) of MAP2+ neurons at D10 and D17 in shCon and shATM cells, suggesting a dispensable role for ATM in neurogenesis. No major numerical differences in GFAP-positive astrocytes (∼70–75%) were seen, either. Conversely, the proportion of GalC-positive oligodendrocytes, accounting for ∼7% at D10 in both shCon and shATM cultures, significantly increased at D17 in the former (∼11%) but not in the latter. These data indicate that ATM deficiency does not impair the differentiation of ihNSCs toward neuron or astrocyte lineages, although it mildly compromises the oligodendrocyte lineage. Whether this mild in vitro phenotype by shATM cells reflects the expression of residual ATM kinase (around 10% of the normal levels) remains as a possibility. In this respect, it should be noticed that genotype–phenotype analysis in A-T patients show a correlation between slow neurological progression and residual ATM kinase activity.28

As oxidative stress can contribute to neural cell death upon ATM deficiency, we analyzed the effects of BSO, which induces oxidative stress by depletion of glutathione, and of the antioxidant NAC during differentiation. Although these treatments spared the neuronal lineage, they clearly affected the oligodendrocytes in shATM, where BSO caused a significant reduction of GalC-positive cells, particularly at D17, whereas NAC had a protective effect. The increased vulnerability of ATM-deficient differentiating oligodendrocytes to oxidative stress could reflect a more critical role of the antioxidant defence in these cells, which contain high amounts of iron than can evoke radical formation.27

Although a substantial cell death involving the activation of caspases and cleavage of PARP accompanied the differentiation of ihNSCs, this apoptotic activity was less pronounced in shATM. Although this could reflect a unique behavior of our cell model, it cannot be excluded that it might reflect a more physiological phenomenon linked to neurogenesis. Indeed, during the early phases of cortical neurogenesis, neural stem cells are regulated in vivo by both lineage differentiation and apoptosis, the latter reducing their contribution to the developing cytoarchitecture.32

An earlier work with ATM−/− animals has shown an essential role for ATM in adult neurogenesis and neuronal survival in vivo and in vitro.33 Accordingly, ATM−/− mice show a reduced number of neural progenitors that proliferate and survive in response to running stress compared with wild type mice, and furthermore, ATM−/− neural progenitors in vitro are unable to differentiate into oligodendrocytes or neurons, though capable of differentiating into astrocytes. These findings obtained with murine cells concord with our results as regards the effects of ATM deficiency on oligodendrocytes, but not neurons.

A factor accounting for neurodegeneration is the re-entry of terminally differentiated postmitotic neurons into cell cycle because of chronic or acute insults associated with DNA damage and oxidative stress, which culminates in apoptosis.34, 35 In this context, ATM deficiency while on one hand suppresses the unscheduled re-entry of neurons into S-phase and protects against apoptosis,35 on the other increases the yield of unrepaired DNA that may eventually be lethal. An additional contributing factor may be the downregulation of the ATR-Chk1 pathway in differentiated neural cells (e.g., stages D10–D24 of ihNSCs; see Figure 1a). Chk1, along with its upstream kinase ATR, constitutes an S-phase checkpoint that in unperturbed cells regulates the replication origin activation and contributes to S-phase progression.36, 37 Unprogrammed initiation of DNA synthesis in terminally differentiated neurons concomitant with the constitutive downregulation of Chk1 could thus contribute to aberrant replicative structures, fork collapse and ultimately cell death.

In conclusion, we have shown that ATM plays a role not only in DDR but also in terminal differentiation of ihNSCs. Moreover, this study supports the use of ihNSCs as an in vitro model system for assessing neurodegenerative events through systematic silencing of genes of the DDR pathway.

Materials and Methods

Cell cultures conditions and stable shRNA interference

The human neural stem cell line, ihNSC, was established by immortalization with v-myc of fetal brain-derived cultures.19 These cells are grown as neurospheres in tissue culture flasks in the presence of 20 and 10 ng/ml of human recombinant EGF and FGF2, respectively, in NS-A basal serum-free medium (DMEM-F12 medium with nutrients optimized for neural stem cell growth; Euroclone, Irvine, Scotland) containing 2 mM L-glutamine, 0.6% glucose, 9.6 g/ml putrescine, 6.3 ng/ml progesterone, 5.2 ng/ml sodium selenite, 0.025 mg/ml insulin and 0.1 mg/ml transferrin (sodium salt, grade II, Sigma), as described by Vescovi et al.38 For stable gene ablation, the vector-based shRNA interference from Invitrogen (BLOCK-iT Lentiviral RNAi) was used. Sequences targeting ATM at nt 268–286 (5′-CACCGGTTGAGAAATTTAAGCGCCCGAAGGCGCTTAAATTTCTCAAC) and nt 1267–1285 (5′-CACCGGATACCAGATCCTTGGAGATCGAATCTCCAAGGATCTGGTATC), and a negative control (5′-CACCGCTACACAAATCAGCGATTTCGAAAAATCGCTGATTTGTGTAG) were cloned in pLenti6/BLOCK-iT-DEST vector. The generated constructs were transfected in the 293-FT cells, and 48 h later, the culture supernatant containing viral particles was recovered and stored at −80°C. ihNSC neurospheres were incubated with lentiviral supernatant for 12 h, returned to fresh GF-containing medium and expanded for three passages before selection with 1 μg/ml blasticidin. To induce differentiation, neurospheres were mechanically dissociated and transferred onto laminin-coated glass coverslips at a density of 104cells/cm2 and grown with 20 ng/ml FGF2. Cultures were shifted after 72 h to NS-A control medium (CM), and further incubated for up to 2 weeks to obtain a mixture of neurons, astrocytes and oligodendrocytes.19, 38 The ATM kinase inhibitor KU-55933 (KuDOS Pharmaceuticals, UK) was added to cells 1 h before irradiation. Cells were irradiated with an IBL437CO instrument (Oris Industries, France) equipped with a 137Cs source providing 675 cGy/min. Treatments with 1 mM NAC and1 μM BSO, a glutathione-depleting and oxidative stress-inducing agent, were performed as described39 by replenishing cultures with the redox modulators in every 48 h.

Immunofluorescence labelling

Cells grown on coverslips were fixed in 4% buffered paraformaldehyde, blocked with 10% normal goat serum and incubated overnight at 4°C with monoclonal antibodies against class III β-tubulin (MMS-435P, Babco, 1 : 400), microtubule-associated protein 2 (MAP2, Chemicon 1 : 400), galactocerebroside C (Gal-C, MAB345 Chemicon 1 : 100), O4 (MAB342, Chemicon 1 : 100) and with the polyclonal antiGFAP (Dako 1 : 400) as reported by De Filippins et al.19 After rinsing in PBS and incubation for 45 min with Alexa 546-conjugated antimouse or antirabbit antibodies (A11030 Molecular Probe, 1 : 1000), coverslips were extensively washed, counterstained with DAPI, mounted on glass slides with Vectashield antifading (Vector Laboratories, Burlingame, CA, USA), and analyzed with a Zeiss Axiovert 200 epifluorescence microscope. The percentage of positive cells was calculated in relation to the number of DAPI-stained nuclei present. At least 1000 cells/slide were scored, and experiments were independently replicated at least three times. For the nuclear foci, cells were cytocentrifuged onto glass slides, air dried, fixed and labeled with an antiphospho-S139-H2AX antibody (clone JBW301, Upstate Biotechnology) as reported by Buscemi et al.23 Images were collected on a Nikon fluorescence microscope equipped with a CCD camera, and foci enumerated by two operators, on three independent experiments and duplicate slides. For flow cytometry analysis, single cell suspensions from neurospheres and D0 cells were made by gently trypsin treatment and trituration. The cells were then fixed for 20 min in 1% formaldehyde, washed twice in Tris-buffered saline, permeabilized with 0.05% Triton and labeled by indirect IF as above, using as negative controls isotype-matched mouse immunoglobulins.

Western blot analysis

Immunoblots were performed as described.23 Briefly, cells were washed with PBS plus 0.1 mM Na3VO4 (Sigma), pelleted and lysed in Laemli buffer (0.125 M Tris-HCl, pH 6.8, 5% SDS) containing as inhibitors 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/ml pepstatin, 100 KIU/ml aprotinin, 10 μg/ml leupeptin (all from Calbiochem, San Diego, CA, USA) and 1 mM Na3VO4. Aliquots containing 50 μg of protein plus 5% β-mercaptoethanol were size-fractionated on SDS-PAGE and electroblotted onto PVDF membranes (Millipore, Bedford, MA, USA). For the analysis of ATM, ATR and DNA-PK, two gradient SDS-PAGE were used, as described by Delia et al.40 After blocking with 4% non-fat dried milk, membranes were incubated with monoclonal antibodies for Chk2 (clone 44D4/2123), ATM (clone 4D223) (both made in-house), p53 (clone DO-7), p21waf1 (Biomarkers), β-actin (Sigma, Italy), vinculin (Sigma), RAD50 and Mre11 (GeneTex), class III β-tubulin, MAP2, CNP and GFAP (Chemicon), and with rabbit antibodies to Olig2 (Chemicon), PDGFR-α (Santa Cruz), the phosphorylated residues of Chk2-T68 and p53-S15 (all from Cell Signaling Technology, Beverly, MA, USA), ATM-S1981 (Rockland Inc., Gilbertsville, PA, USA), Smc1-S966 and total Smc1 (Bethyl Laboratories, Montgomery, TX), ATR (ABR-Affinity Bioreagents Inc., Golden, CO, USA), DNA-PKcs (Neomarkers, LabVision, CA, USA), NBS1 (Novus Biologicals), Chk1, cleaved PARP and cleaved Caspases 3 and 9 (all from Cell Signaling). PVDF membranes were incubated with antibodies in sealed bags using the XBLOT P100 hybridization instrument (Isenet, Milan, Italy). Antibody binding was detected with ECL Super Signal (Pierce, Rockford, IL, USA), and bands quantitated with ImageQuant.

Karyotype analyses and flow cytofluorimetry

Neurospheres were metaphase arrested by incubation for 12 h with 40 ng/ml colcemid, hence dissociated and subjected to hypotonic swelling with 0.075 M KCl, EDTA and Hepes, pH 7.4 for 60 min at 37°C. After buffer removal, the cells were fixed with 3 : 1 methanol/acetic acid solution. Cells were dropped onto humidified glass slides, dried overnight and chromosomes were stained with Giemsa or DAPI and mounted. Cell cycle analysis were performed on cells disaggregated by trypsin treatment, fixed in 70% ethanol, treated with RNAse and stained with propidium iodide and analyzed by FACSvantage flow cytometry (Becton Dickinson, CA). For BrdU incorporation, cells were incubated with 10 mM bromodeoxyuridine for 10 min, detached by trypsin treatment, fixed and labeled according to the manufacturer's protocol with an FITC-labeled anti BrdU (Becton Dickinson) and analyzed by dual fluorescence flow cytometry.

Statistical analysis

Statistical significance was assessed by χ2- or Student's t-tests where appropriate, and P-values were determined by data analysis at www2.unipr.it/~bottarel/epi/assoc/t_stu.htm and www2.unipr.it/~bottarel/epi/assoc/chi_qua.htm.