Contribution of the Atm Protein to Maintaining Cellular Homeostasis Evidenced by Continuous Activation of the AP-1 Pathway in Atm-deficient Brains*

Maintenance of genome stability is essential for keeping cellular homeostasis. The DNA damage response is a central component in maintaining genome integrity. Among of the most cytotoxic DNA lesions are double strand breaks (DSBs) caused by ionizing radiation or radiomimetic chemicals. ATM is missing or inactivated in patients with ataxia-telangiectasia. Ataxia-telangiectasia patients display a pleiotropic phenotype and suffer primarily from progressive ataxia caused by degeneration of cerebellar Purkinje and granule neurons. Additional features are immunodeficiency, genomic instability, radiation sensitivity, and cancer predisposition. Disruption of the mouse Atm locus creates a murine model of ataxia-telangiectasia that exhibits most of the clinical features of the human disease but very mild neuronal abnormality. The ATM protein is a multifunctional protein kinase, which serves as a master regulator of cellular responses to DSBs. There is growing evidence that ATM may be involved in addition to the DSB response in other processes that maintain processes in cellular homeostasis. For example, mounting evidence points to increased oxidative stress in the absence of ATM. Here we report that the AP-1 pathway is constantly active in the brains of Atm-deficient mice not treated with DNA damaging agents. A canonical activation (increased phosphorylation of mitogen-activated protein kinase kinase-4, c-Jun N-terminal kinase, and c-Jun) of the AP-1 pathway was found in Atm-deficient cerebra, whereas induction of the AP-1 pathway in Atm-deficient cerebella is likely to mediate elevated expression of c-Fos and c-Jun. Although Atm+/+ mice are capable of responding to ionizing radiation by activating stress responses such as the AP-1 pathway, Atm-deficient mice display higher basal AP-1 activity but gradually lose their ability to activate AP-1 DNA-binding activity in response to ionizing radiation. Our results further demonstrate that inactivation of the ATM gene results in a state of constant stress.

deficiency, premature aging, gonadal dysgenesis, extreme radiosensitivity, and a high incidence of lymphoreticular malignancies (for review, see Ref. 1). Functional inactivation of the ATM gene product accounts for this complex phenotype (2). The ATM protein, the product of the ATM gene, is a member of a family of large protein kinases found in various organisms, which share phosphatidylinositol 3 kinase-like domain. ATM is a serine/threonine protein kinase that mediates the activation of multiple signaling transduction pathways following the induction of DNA double strand breaks (DSBs) (reviewed in Refs. 3

and 4).
Atm-null mice that were created by disrupting the Atm locus (5-8) recapitulate most of the features of the human A-T phenotype and display growth retardation, male and female infertility, extreme predisposition to thymic lymphomas, and acute sensitivity to ionizing radiation (IR). Of note, the neurological symptoms of the disease are very mild in the animal model and can be observed only after challenge (5,8,9).
Cells derived from A-T patients and Atm-deficient mice exhibit genomic instability, extreme sensitivity to DNA breaking agents, and a broad defect in the activation of the cell cycle checkpoints and other pathways induced by DSBs (3,10). However, A-T cells are not sensitive to UV light or to replication inhibitors. To activate the DNA damage response, ATM phosphorylates a number of downstream targets, including p53, Mdm2, Brca1, Chk2, and Nbs1 (reviewed in Refs. 3 and 4). ATM is activated by DSBs induced by internal or external DNA damaging agents, such as IR, radiomimetic chemicals, and reactive oxygen species and is probably activated also by DSBs that occur normally during meotic and V(D)J recombination. ATM can bind to DNA in vitro (11,12), and a fraction of ATM binds tightly to the chromatin following the induction of DSBs, in accordance with the suggested role of ATM in the early steps of the DNA damage response.
Evidence is accumulating that ATM is also involved in maintaining genomic stability under normal conditions of apparently no stress. Because it appears to mediate IR-induced death in the murine-developing nervous system (13), it was hypothesized that it functions during nervous system development to eliminate neuronal cells afflicted with genomic damage (13). ATM was also shown to be required for the execution of apoptosis in all areas of the developing nervous system in a model of a DNA ligase IV-deficient mouse (14).
One of the major stress responses is mediated by the AP-1 transcription factor. AP-1 is a superfamily of proteins consisting of several subfamilies, including Jun, Fos, and ATF-2, all of which possess a leucine zipper domain that assists their dimer-ization (15). The AP-1 proteins form homo-and heterodimers before binding to their DNA target sites. The activation of AP-1 is regulated by c-Jun N-terminal kinase (JNK), ERK, and p38 MAPKs (mitogen-activated protein kinases) (16,17). JNKs are a family of serine-threonine kinases related to the MAPKs with at least 10 isoforms encoded by three genes (JNK1, JNK2, and JNK3), with additional diversification resulting from alternative splicing (18 -28). The JNK1 and JNK2 subfamilies each comprise four isoforms (46 or 55 kDa), whereas the JNK3 contains two different isoforms (45)(46)(47)(48). Although JNK1 and JNK2 are expressed in most cell types (23), expression of JNK3 is limited to neuronal cells (18). All JNKs possess a conserved TPY tripeptide motif in their kinase domain (Thr-183 and Tyr-185) (22), and their activation is mediated by phosphorylaion of the threonine and tyrosine residues within this motif by the upstream JNK kinases MKK4 (29 -31) and MKK7 (32). Among the known subsrates of the JNKs are the c-Jun protein, a component of the AP-1 transcription factor (33), and the transcription factors ATF-2 (34), Elk1 (35), and p53 (36). Although all JNKs are capable of phosphorylating c-Jun on serine residues 63 and 73, isoforms exhibit different substrate-binding specificities (37,38).
In vitro studies have revealed a defect in IR-induced activation of the JNK-signaling pathway in lymphoblastoid cells from individuals with A-T. Although IR induced c-Jun phosphorylation in normal cells, no such phenomenon was observed in A-T fibroblasts. The 46-kDa isoform of JNKs was dually phosphorylated in the nuclei of both normal and A-T fibroblast following exposure to IR, whereas c-Jun activity was detected in normal cells but not in A-T fibroblasts. Furthermore, an exogenous-purified active JNK protein was able to phosphorylate endogenous c-Jun in nuclear extracts only of normal but not A-T cells and only after the normal cells were irradiated, (39).
Here we report that the AP-1 pathway is constantly active in untreated Atm-deficient brain. Atm ϩ/ϩ mice are capable of responding to IR by activating the AP-1 pathway, whereas Atm-deficient mice display higher basal AP-1 activity but gradually lose their ability to activate AP-1 DNA-binding activity in response to IR.

MATERIALS AND METHODS
Mice-Atm ϩ/Ϫ mice (5) were a generous gift from Dr. Wynshaw-Boris. Offspring of these mice were genotyped by PCR-based assays using mouse-tail DNA, prepared with the GeneReleaser TM reagent (Bio-Ventures Co., Murfreesboro, TN).
Nuclear Extracts-Tissues (cerebellum, cerebrum, and liver) were washed with ice-cold phosphate-buffer saline and maintained in 0.32 M sucrose. The organs were homogenized in ice-cold hypotonic buffer B (0.5 M sucrose, 10 mM HEPES pH 7.9, 1.5 mM MgCl 2 , 10% glycerol, 1 mM EDTA, 1 mM dithiothreitol, 1 mM PMSF and protease inhibitor mixture) and 0.2 mM Na 3 VO 4 . The lysates were allowed to incubate for 15 min on ice and then centrifuged for 10 min at 7,500 ϫ g. The supernatants were removed (cytoplasmic extracts), and the nuclear pellets were rinsed once with ice-cold buffer B and resuspended in buffer C (0.42 M NaCl, 20 mM Tris-HCl, pH 7.9, 25% glycerol, 0.2 mM EDTA, 1 mM dithiothreitol, and protease inhibitor mixture (Roche Molecular Biochemicals), 0.2 mM NaVO 3 (100 l for cerebellum, 400 l for cerebrum, and 0.5 ml for liver)). The nuclear extracts were incubated for 45 min on ice and centrifuged for 20 min at 15,500 ϫ g. The supernatants were collected and used for electrophoretic mobility shift assay (EMSA) and Western blot analyses.
EMSA-The binding reaction mixture containing 10 mM Tris-HCl, pH 7.9, 60 mM KCl, 0.4 mM dithiothreitol, 10% glycerol, 2 g of bovine serum albumin, 1 g of poly(dI-dC), 15,000 cpm of 32 P-labeled AP-1 oligonucleotides (5Ј-CGCTTGATGAGTCAGCCGGAA-3Ј, respectively) (Promega, Madison, WI) was incubated for 30 min with 6 g of nuclear extract. For AP-1-binding activity the reaction was done on ice. For specificity control a 50-fold excess of unlabeled probe was applied. Products were analyzed on a 5% acrylamide gel made up in 1 ϫ TGE (50 mM Tris, 400 mM glycine, 2 mM EDTA). Dried gels were exposed to x-ray film or to phosphor screen (Molecular Dynamics, Sunnyvale, CA). Quantitative data were obtained using phosphoimaging analysis (Molecular Dynamics). Western Blotting-Western blotting was performed as described by Harlow and Lane (61), using 12.5% polyacrylamide gels (PAGE). Each lane was loaded with an identical amount of protein extract (50 or 100 g), which, following electrophoresis, were transferred to an Immobilon polyvinyldene disulfide membrane for 1.5 h. Blots were stained with Ponceau to verify equal loading and transfer of proteins. Membranes were then probed with anti-phospho-JNK (1:1000) antibodies (Cell Signaling Technology, Beverly, MA), mouse anti-c-Jun (1:1000) antibodies (BD Biosciences), mouse anti-c-Fos (1:400) antibodies (Santa Cruz, CA), mouse anti-phospho-c-Jun (Ser-63) (1:1000) antibodies (Santa Cruz Biotechnology), and rabbit anti-phospho-MKK4 (1:1000) antibodies (Cell Signaling Technology, Beverly, MA). For the detection of phospho-JNK and phospho-c-Jun, EnVision TM ϩ/horseradish peroxidase (rabbit, 1:100) (DAKO, Glostrup, Denmark) was used to enhance sensitivity. The rest of the antibodies were detected using horseradish peroxidaseconjugated antibodies (1:25000 for mouse and 1:15000 for rabbit) purchased from Jackson Laboratories. Intensity of the signal was determined by the ECL-Plus detection system (Amersham Biosciences). Protein concentration was determined according to the method of Bradford (62) using BSA as standard.
Statistical Analysis-Results were analyzed using two-tailed Student's t test. Data were expressed as mean ϮS.E., and p values Յ 0.05 were considered significant.

Atm Deficiency Leads to Gradual Induction of AP-1 DNAbinding Activity-
The AP-1-signaling pathway is one of the most prominent pathways known to respond to various stress conditions. Alterations in the activity of this pathway in a particular tissue imply that it is under stress. To test whether tissues of Atm-deficient mice are under constant stress, we analyzed the basal level of AP-1 DNA-binding activity in the brains of Atm Ϫ/Ϫ -and Atm ϩ/ϩ -untreated mice. In 1-month-old animals, the differences in AP-1 activity between Atm Ϫ/Ϫ and Atm ϩ/ϩ cerebella and cerebra were significant but rather small (45% induction in Atm Ϫ/Ϫ animals, p Ͻ 0.025 for the cerebellum and p Ͻ 0.0023 for the cerebrum). In 4-month-old animals, there was a 3-fold constant induction of AP-1 activity in Atmdeficient tissues (p Ͻ 0.0079 for the cerebellum and p Ͻ 0.0264 for the cerebrum) (Fig. 1A).
The nervous system, especially the cerebellum, is highly affected in A-T; other organs such as the liver are not known to be involved. This coincides with our earlier findings that the cerebellum but not the liver is under constant oxidative stress (40). To correlate the degree of stress to AP-1 DNAbinding activity, we measured that activity in livers isolated from 4-month-old Atm Ϫ/Ϫ -and Atm ϩ/ϩ -untreated animals. Atm-deficient mice and their control littermates displayed similar levels of AP-1 DNA-binding activity (Fig. 1B). Taken together, the results presented in Fig. 1 show that organs most affected in A-T display higher levels of AP-1 DNAbinding activity, whereas non-affected organs behave like controls.
Atm Loss Leads to Constant Induction of JNK Activity-The earlier step in the signaling pathway that leads to AP-1 induction is JNK-mediated phosphoryaltion of the N terminus of c-Jun. This phosphorylation activates its dimerization (either with Jun or Fos isoforms) and binding to its target consensus sequence. To test whether Atm loss leads to increased JNK activity, we measured the levels of phospho-JNK, the activated form of this enzyme (Fig. 2). There was a gradual increase in phospho-JNK levels in Atm-deficient brains, from no change compared with normal tissue at the age of 1 month to close to 3-fold induction at 4 months (p Ͻ 3.05 ϫ 10 Ϫ5 for the cerebellum and p Ͻ 0.0037 for the cerebrum). The level of JNK induction correlates well with the induction of AP-1 DNAbinding activity (compare Figs. 1 and 2).
Elevated Phospho-MKK4 Levels in Atm-deficient Cerebra-MKK4 is among the kinases capable of activating JNK. To explore whether Atm loss leads to constant activation of MKK4, we examined the levels of activated MKK4 in 4-monthold cerebra and cerebella. Interestingly, we found elevation (67% increase, p Ͻ 0.03) only in Atm-deficient cerebra and not in the cerebella (Fig. 3).
Atm Ϫ/Ϫ Mice Gradually Lose their Ability to Activate AP-1 DNA-binding Activity in Response to IR-Atm Ϫ/Ϫ mice and their control littermates were exposed to 20 Gy of x-rays for different time periods, after which the levels of AP-1 DNAbinding activity were assayed. At 4 months of age, Atm ϩ/ϩ brains responded to the irradiation treatment with 2.5-fold induction in AP-1 DNA-binding activity (Fig 4). A different pattern of AP-1 induction was observed between Atm ϩ/ϩ cerebellum and cerebrum. The Atm ϩ/ϩ cerebellum displayed a clear peak of AP-1 induction around 60 min post-irradiation, whereas the cerebrum showed a rapid and steady induction in AP-1 levels. In contrast, 4-month-old Atm Ϫ/Ϫ mice failed to activate AP-1 DNA-binding activity in response to IR, and the levels of AP-1 remained unchanged (Fig. 5, right panel). Interestingly, young Atm Ϫ/Ϫ mice retained their ability to induce AP-1 DNA-binding activity in response to IR (Fig. 4, left panel). Similar to Atm ϩ/ϩ , 2-fold induction in AP-1 DNA-binding activity was measured in these Atm Ϫ/Ϫ cerebella and cerebra. These results suggest that damage accumulation is gradually accompanied by loss of the ability of Atm-deficient mice to activate stress responses as a result of noxious stimuli such as IR.
Atm Ϫ/Ϫ Mice Are Capable of Activating JNK in Response to IR-Next we assayed the ability of Atm Ϫ/Ϫ and Atm ϩ/ϩ mice to activate JNK in response to 20 Gy x-ray. Similar results were obtained in younger (1-month-old) and older (4-month-old) Atm Ϫ/Ϫ and Atm ϩ/ϩ mice. In both cases, the peak of phospho-JNK induction was around 60 min post-irradiation (Fig. 5). Four-month-old Atm-deficient mice are incapable of activating AP-1 DNA-binding activity, but notably they were capable of elevating phospho-JNK levels in response to IR to a level similar to their control littermates. These results suggest that accumulating damage may not affect JNK activity in response to IR but rather events downstream to it.
Atm-deficient Cerebra Display High Basal Levels and Irradiation-induced Elevation of Ser(P)-63-c-Jun-Atm-deficient mice were capable of activating JNK in response to IR but failed to induce AP-1 DNA-binding activity. To test whether Atm loss disrupts the ability of JNK to phosphorylate its target proteins, we analyzed the levels of phospho-c-Jun in the cerebra and cerebella of untreated and irradiated animals. Using antibody directed against c-Jun phosphorylated on Ser-63, we observed a 50% increase in phospho-c-Jun levels (p Ͻ 0.017) in cerebra of 4-month-old untreated Atm-deficient mice. In contrast, no change in Ser(P)-63-c-Jun levels was observed in Atm-deficient cerebella (Fig. 6, upper panel). Exposing the an- imals to 20 Gy x-ray resulted in a gradual increase (2-3-fold) in Ser(P)-63-c-Jun levels only in the cerebra of 1-month-and 4-month-old Atm ϩ/ϩ mice, but not in the same tissues of Atm Ϫ/Ϫ mice.
Increased c-Fos and c-Jun Expression in Untreated Atmdeficient Cerebella-Elevated levels of either c-Jun or c-Fos proteins can lead to constant AP-1 activation. To test whether Atm deficiency leads to alterations in c-Jun and c-Fos levels, we analyzed the basal levels of these proteins in untreated Atmdeficient cerebella and cerebra. Using antibody directed against c-Fos, we observed a more than 3-fold increase in c-Fos levels (p Ͻ 0.019) in cerebella of 4-month-old untreated Atmdeficient mice compared with Atm ϩ/ϩ animals. Smaller induction in c-Fos levels (1.8-fold, p Ͻ 0.032) was observed in the cerebra. Elevated levels of c-Jun were observed only in 4-month-old untreated Atm-deficient cerebella (2.5-fold, p Ͻ 0.0242). No change in c-Jun levels was observed in Atm-deficient cerebra (Fig. 7). DISCUSSION Alterations in the activity of the AP-1-signaling pathway in a given tissue imply that the tissue may be under stress. The data reported here, together with other reports (40 -43), further substantiate the facts that organs most affected in A-T patients gradually develop stress conditions. Several of our findings support this notion: (a) Atm-deficient brains, but not Atm-deficient livers, display higher basal AP-1 DNA-binding activity. (b) Untreated Atm-deficient cerebra and cerebella display elevated levels of phospho-JNK. (c) Atm-deficient mice gradually lose their ability to activate AP-1 DNA-binding activity in response to IR, suggesting that the system is either saturated and can no longer respond to stress stimuli, or accumulated damage disrupts the signaling pathways mediating radiation-induced AP-1 activation. (d) There are increased levels of phospho-c-Jun in untreated Atm-deficient cerebra. (e) There are elevated levels of c-Fos in untreated Atm-deficient cerebella. Our concept that Atm loss leads to constant stress in organs most affected in A-T is further supported by our recent findings using microarray DNA chips. Initial analysis revealed that unirradiated Atm-deficient cerebella displayed constitutive activation of repression

FIG. 2. Up-regulation of phospho-JNK levels in Atm-deficient mice.
Nuclear extracts were isolated from the cerebrum and cerebellum of 1-month-old (upper panel) and 4-month-old (lower panel) Atm Ϫ/Ϫ and Atm ϩ/ϩ , separated on 12.5% polyacrylamide gel, and blotted onto polyvinyldene disulfide membrane. The membrane was stained with Ponceau, after which the blot was reacted with anti-phospho-JNK polyclonal antibody (1:1000). The blots were developed using the ECL system, and a representative Western blot is shown. To verify that equal amounts of protein were transferred, the blots were stripped and reacted with mouse monoclonal anti-tubulin antibody (1:25,000). Quantitative analysis of bands was performed using TINA software and normalized to the corresponding tubulin. Quantitative analysis of bands is represented by arbitrary units Atm ϩ/ϩ mice (n ϭ 10).

FIG. 3. Up-regulation of phospho-MKK4 levels in Atm-deficient cerebra.
Nuclear extracts were isolated from the cerebrum and cerebellum of 4-month-old Atm Ϫ/Ϫ and Atm ϩ/ϩ and treated as described in the legend to Fig. 2. The blot was reacted with anti-phospho-MEK4 polyclonal antibody (1:1000). The blots were developed using the ECLϩ system. Verification of equal loading and quantitative analysis were performed as described in the legend to Fig. 3 (n ϭ 8).
of numerous genes that the corresponding wild type cerebella showed only after irradiation. 2 AP-1 DNA-binding activity can be induced in two ways: by increased transcription of c-Jun or Fos, or by phosphorylation of the Jun proteins, mainly by different JNK isoforms (for review see Ref. 44). Canonical AP-1 activation was observed in the cerebra of untreated Atm-deficient mice, but there was no c-Jun phoshorylation either on Ser-63 or Ser-73 in the cerebella (see Fig. 6). In untreated Atm-deficient cerebra, elevated levels of phospho-MKK4, JNK, and Ser(P)-63-c-Jun were observed, whereas higher levels of c-Fos and c-Jun were likely to mediate Atm-induced AP-1 activation in the cerebella. Interestingly, DNA microarray analysis revealed a significant elevation of another member of the Fos family, FosB, in untreated Atm-deficient cerebella, suggesting that Fos proteins play an important role in cellular processes induced by Atm deficiency. Increased and prolonged expression (mRNA and protein) of c-Jun and c-Fos with no apparent elevation in c-Jun (Ser-63) phosphorylation was observed in the hippocampal CA1 pyramidal cell layer in regions undergoing delayed neuronal death (45). Similar to what we observed in Atm-deficient cerebella, these results demonstrate that increased AP-1 DNA-binding activity can stem solely from an increase in c-Fos and c-Jun. The biological role of JNK activation in Atm-deficient cerebella remains unclear because none of its downstream substrates were found to be elevated. It is possible that JNK activation is merely a con-sequence of Atm loss with no actual biological role. Taken together, our results indicate that the cerebella and the cerebra are different entities with regard to AP-1 activation.
The AP-1 pathway can be activated in response to diverse stimuli and a wide range of stresses. However, its role in stress remains largely unknown because both pro-apoptotic and prosurvival effects of AP-1 have been observed (for reviews see Refs. 46 and 47). DNA damage, especially DSBs, poses a critical threat to the cell. The finding of an accumulation of DNA breaks in Atm-deficient cerebella with no signs of cell death (48) argues in favor of a protective role for AP-1. Our data, on the other hand, does not rule out a pro-apoptotic role for AP-1. A pro-survival role for AP-1 activity is also consistent with several reports showing that AP-1 activity is involved in defense against UV radiation and DNA-alkylating agents (49). AP-1 was shown to be involved in DNA repair by its ability to regulate the expression of O 6 -methylguanine-DNA methyltransferase; the promoter of this gene was found to contain two putative sites of AP-1 (50). This enzyme is up-regulated in response to a variety of genotoxic insults (for review see Ref. 51). Stable expression of a non-phosphorylated dominant negative protein c-Jun (S63A, S73A) markedly inhibited AP-1 activity and greatly increased the cytotoxic effects of a variety of DNA-damaging agents (52). In contrast to the previous reports, DNA fragmentation and activation of c-Jun in cerebellar neuronal cells can lead to their demise. Gillardon et al. (53) found increased DNA fragmentation and c-Jun up-regulation in Purkinje cells of weaver and pcd mutant mice. Weaver and Purkinje cell degeneration (pcd) are autosomal mutations that lead to complete loss of granule and Purkinje cells, respectively. Co-localization of DNA fragmentation and c-Jun in dying granule cells (weaver) and Purkinje cells (pcd), and inhibition of the death process by function-blocking c-Jun antibodies, led Gillardon et al. (53) to suggest a pro-apoptotic role for c-Jun. One way to explain the dual role (death or survival) of AP-1 is by their ability to form different dimer combinations. A 35-kDa Fosrelated antigen/JunD dimer was detected in neurons that survived injury, whereas a Jun⅐JunD complex was detected in neurons prior to undergoing apoptosis (54). Because the present study was carried out for the whole brain, it is not clear which of the AP-1 complexes actually functions in specific cell populations.
A prolonged induction of phospho-JNK and AP-1 was observed in Atm-deficient cerebella and cerebra ( Figs. 1 and 3). Prolonged expression of some AP-1 components, such as Fosrelated antigen 2, was associated with neuronal regeneration and repair (55). Following this line of thinking, it is possible that prolonged AP-1 induction in Atm-deficient brains is important for neuronal survival. Our results are consistent with several studies (56,63) showing prolonged AP-1 activation in Atm-deficient cells. Whereas wild type fibroblast showed transient induction of c-Jun in response to IR, ATM-deficient fibroblasts responded to IR with prolonged c-Jun expression. Lee et al. (56) showed delayed and prolonged JNK activation in response to IR and oxidative stress.
Increased levels of phospho-JNK can stem from two opposing processes: elevated activity of upstream kinases such as MKK4/7, or reduced activity of phosphatases such as MAPK phosphatase-5 (MKP-5). Bar-Shira et al. (57) showed that treatment with the radiomimetic agent neocarzinostatin resulted in elevated levels of MKP-5 in wild type cells, which was attenuated in ATM-deficient human cells. Moreover, the basal level of this phosphatase was lower in ATM-deficient cells. These findings support our hypothesis that Atm-deficiency leads to accumulation of DNA breaks especially in the cerebellum (48), where Atm loss retards the induction of MKP-5 and thereby increases the level of phospho-JNK in untreated Atm-deficient cerebella. In the presence of Atm, DNA damage will lead to a timely controlled cycle of phosphorylation and dephosphorylation of MAPKs; the activation of this cycle is defective in Atm-deficient cells, accounting for the defective ability of these cells to generate correct stress responses. Chen et al. (58) showed that increased oxidative stress induced JNK activity, not by activating its upstream kinases stress-activated protein kinase/ERK kinase (SEK1)/MKK4 or MKK7, but rather by inactivating its phosphatase M3/6. These results are also consistent with our notion that prolonged AP-1 activation stems from DNA damage-induced oxidative stress (40,43,48). DNA damage can activate the AP-1 pathway regardless of oxidative stress.
Using whole cerebral and cerebellar extracts precluded our determining the exact cellular population in which AP-1 is up-regulated in untreated Atm-deficient mice. Besides neurons, a variety of distinct cell populations can activate AP-1 pathways. Masood et al. (59) showed that AP-1 is an important transcription regulator of the astrocyte specific gene, glial fibrillary acidic protein (GFAP), which is up-regulated in proliferating and activated astrocytes. Our observation of elevated levels of GFAP in the cerebella of Atm-deficient mice 3 further supports the notion that AP-1 is up-regulated in astrocytes. Proliferation and activation of astrocytes and microglia cells may contribute to an inflammatory reaction as a result of tissue damage. Our failure to detect increased microglia proliferation or activation argue against cerebral or cerebellar inflammation. It is possible that astrocytes were activated in response to increased oxidative stress, which has been shown to develop in untreated Atm-deficient mice especially in the cerebella. Astrocytes were indeed found capable of altering their phenotype with up-regulation of large number of molecules, including those controlling the protective system (for review see Ref. 60).
In summary, we have demonstrated that ATM loss leads to constant activation of the AP-1-signaling pathway in untreated Atm-deficient brains in an age-dependent manner. Moreover, Atm-deficient mice gradually lose their ability to activate the AP-1-signaling pathway in response to IR. Based on these findings, we suggest that accumulation of unrepaired DNA breaks exerts constant stress on the DNA damage response mechanisms, whose overactivation leads to constant activation of critical signaling pathways, ultimately disturbing cellular homeostasis. Better understanding of the ongoing stress responses in Atm-deficient tissues may point to novel therapeutic approaches to help alleviate some of the symptoms associated with A-T, in particular the progressive neurodegeneration.