Akt Promotes Post-Irradiation Survival of Human Tumor Cells Through Initiation, Progression, and Termination of DNA-Pkcs–Dependent DNA Double-Strand Break Repair

Published OnlineFirst May 17, 2012; DOI: 10.1158/1541-7786.MCR-11-0592

Molecular Cancer DNA Damage and Cellular Stress Responses Research

Akt Promotes Post-Irradiation Survival of Human Tumor Cells through Initiation, Progression, and Termination of DNA-PKcs–Dependent DNA Double-Strand Break Repair

Mahmoud Toulany1, Kyung-Jong Lee3, Kazi R. Fattah3, Yu-Fen Lin3, Brigit Fehrenbacher2, Martin Schaller2, Benjamin P. Chen3, David J. Chen3, and H. Peter Rodemann1

Abstract Akt phosphorylation has previously been described to be involved in mediating DNA damage repair through the nonhomologous end-joining (NHEJ) repair pathway. Yet the mechanism how Akt stimulates DNA-protein kinase catalytic subunit (DNA-PKcs)-dependent DNA double-strand break (DNA-DSB) repair has not been described so far. In the present study, we investigated the mechanism by which Akt can interact with DNA-PKcs and promote its function during the NHEJ repair process. The results obtained indicate a prominent role of Akt, especially Akt1 in the regulation of NHEJ mechanism for DNA-DSB repair. As shown by pull-down assay of DNA-PKcs, Akt1 through its C-terminal domain interacts with DNA-PKcs. After exposure of cells to ionizing radiation (IR), Akt1 and DNA-PKcs form a functional complex in a first initiating step of DNA-DSB repair. Thereafter, Akt plays a pivotal role in the recruitment of AKT1/DNA-PKcs complex to DNA duplex ends marked by Ku dimers. Moreover, in the formed complex, Akt1 promotes DNA-PKcs kinase activity, which is the necessary step for progression of DNA-DSB repair. Akt1-dependent DNA-PKcs kinase activity stimulates autophosphorylation of DNA-PKcs at S2056 that is needed for efficient DNA-DSB repair and the release of DNA-PKcs from the damage site. Thus, targeting of Akt results in radiosensitization of DNA-PKcs and Ku80 expressing, but not of cells deficient for, either of these proteins. The data showed indicate for the first time that Akt through an immediate complex formation with DNA-PKcs can stimulate the accumulation of DNA-PKcs at DNA-DSBs and promote DNA-PKcs activity for efficient NHEJ DNA-DSB repair. Mol Cancer Res; 10(7); 945–57. 2012 AACR.

Introduction double-strand break (DNA-DSB) repair and subsequent The serine/threonine kinase Akt/PKB is expressed as 3 reproductive cell death (6). isoforms, Akt1/PKBa, Akt2/PKBb and Akt3/PKBg. Akt DNA-DSBs are the most lethal type of DNA lesions that activation is efficiently induced by ionizing radiation (IR) or lead to cell death following IR exposure. Two processes are by growth factors, such as EGF receptor (EGFR) ligands, primarily involved in DNA-DSB repair, nonhomologous through the activation of phosphoinositide 3-kinase (PI3K; end-joining (NHEJ) and homologous recombination (7), ref. 1). Our results and accumulated reports from other but NHEJ is the predominant process in higher eukaryotes laboratories indicate that radiosensitization by targeting and mammals. The catalytic subunit of the DNA-dependent PI3K or Akt1 (2–5) is a consequence of impaired DNA protein kinase complex (DNA-PKcs) is a key enzyme in the NHEJ process. Activation of DNA-PKcs requires the phosphorylation of specific amino acid residues, among which the fi Authors' Affiliations: 1Division of Radiobiology and Molecular Environ- T2609 cluster and S2056 have been identi ed as essential for mental Research, Department of Radiation Oncology, 2Department of efficient rejoining of DNA-DSBs during NHEJ (8). Like- Dermatology, Eberhard Karls University Tubingen,€ Tubingen,€ Germany; and 3Division of Molecular Radiation Biology, Department of Radiation wise, mutations in these phosphorylation sites result in Oncology, University of Texas Southwestern Medical Center at Dallas, enhanced cellular sensitivity to IR (9, 10). In this context, Dallas, Texas our previous work provided the first new insights into the Note: Supplementary data for this article are available at Molecular Cancer possible function of Akt1 in modulating post-irradiation Research Online (http://mcr.aacrjournals.org/). survival. This occurs most likely via phosphorylation of D.J. Chen and H.P. Rodemann shared corresponding authorship. DNA-PKcs and consequently via the NHEJ repair pathway (2). We showed that an Akt antagonist inhibits radiation- Corresponding Author: H. Peter Rodemann, Eberhard Karls University fi Tubingen,€ Roentgenweg 11, Tubingen€ 72076, Germany. Phone: 49-7071- induced phosphorylation of DNA-PKcs. This nding cor- 298-5962; Fax: 49-7071-29-5900; E-mail: hans-peter.rodemann@uni- relates with cellular radiosensitization after treatment with an tuebingen.de Akt inhibitor due to suppression of DNA-DSB repair, as doi: 10.1158/1541-7786.MCR-11-0592 measured by g-H2AX foci. We also discovered that DNA- 2012 American Association for Cancer Research. PKcs co-immunoprecipitates with either Akt1 or p-Akt (2).

www.aacrjournals.org 945

Toulany et al.

A similar interaction was reported by Bozulic and colleagues HT1080 cells were maintained in medium containing (11). On the basis of these data, we conclude that Akt might 250 mg/mL of G418. Xrs6 þ Ku80 cells were maintained be partially necessary for the repair of DNA-DSBs in in medium containing 150 mg/mL hygromycin B. irradiated cells via activation of DNA-PKcs. Rad51D1-deficient Chinese hamster ovary (CHO) cells In the present study, using different approaches in cells originally established in the laboratory of Dr. Larry H. lackingorexpressingDNA-PKcsorKu80,weprovide Thompson (Lawrence Livermore National Laboratory, detailed information for the function of Akt, especially Livermore, CA) were received from the Laboratory of Dr. Akt1, in DNA-PKcs–dependent repair of DNA-DSBs. Eckhardt Dikomey (Radiobiology and Experimental Radio- Ourresultsindicateforthefirst time that Akt1 is not only oncology, University Medical Center Hamburg-Eppendorf, involved in DNA-DSB repair but also directly promoting Hamburg, Germany). and regulating repair process through the 2 complemen- tary mechanisms. First, Akt, especially Akt1, facilitates IR- Glutathione S-transferase pull-down assay induced Ku/DNA-PKcs complex formation and accumu- Glutathione S-transferase (GST), GST-Akt1 full-length, lation of DNA-PKcs to DNA damage site; and second, GST-Akt1 N-terminal fragment (1–150 a.a.), and GST- Akt induces DNA-PKcs kinase activity and its autopho- Akt1 C-terminal fragment (151–480 a.a.) were expressed in sphorylation that is needed for release of DNA-PKcs from BL21 Escherichia coli bacteria by induction of cultures at damage site. optical density (OD)600 of 0.6 with 0.2 mmol/L isopropyl- b-D-thiogalactopyranoside (IPTG) at 37 C for 4 hours. Bacterial lysates were applied to Glutathione Sepharose Materials and Methods beads (GE Healthcare Life Sciences), and the GST-tagged Reagents and antibodies proteins were pulled down and washed with PBS. Then, the The Akt pathway inhibitor (API) and antibodies against beads were incubated with 1 mg of purified human DNA- Akt1, p-Akt (S472/3), p-H2AX (S139), DNA-PKcs, and PKcs in HEPES buffer [50 mmol/L HEPES, pH 7.5, 150 phospho-DNA-PKcs (S2056) have been previously mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 10% described (2, 3). The DNA-PKcs inhibitors NU7441 and glycerol, 1% Triton X-100, 1 mmol/L dithiothreitol (DTT), NU7026 were purchased from Tocris Bioscience and Sigma- 0.5 mmol/L phenylmethylsulfonylfluoride (PMSF), 1 mg/ Aldrich, respectively. Hygromycin B was purchased from mL Pepstatin, and 2 mg/mL leupeptin] supplemented with Invitrogen. G418 and puromycin were purchased from 0.01% bovine serum albumin (BSA) and 40 ng/mL ethi- Biochrom. Sepharose bead–conjugated IgG antibody was dium bromide (EtBr) for 2 hours at 4C. Beads were washed purchased from Cell Signaling. Active Akt1 and calf thymus 3 times with HEPES buffer and then subjected to Western DNA cellulose were purchased from Sigma-Aldrich. Control blot analysis. nontargeting siRNA (catalog no: D-001810-10) and AKT1- siRNA (catalog no: NM-0010144219) were purchased from Clonogenic assay, immunoprecipitation, Western Thermo Fisher Scientific. blotting, g-H2AX foci formation assay, siRNA transfection, fluorescence-activated cell-sorting analysis, Cell lines immunostaining, and microscopy The colon carcinoma cell lines HCT116 wild-type These assays have been described previously (2, 13). (HCT116wt), HCT116 DNA-PKcs–deficient (HCT116- DNA-PKcs / ), and HCT116 DNA-PKcs–deficient com- DNA-PKcs in vitro kinase assay plemented with GFP-tagged DNA-PKcs (GFP-DNA-PKcs- A DNA-PKcs kinase assay was conducted using the HCT116) were kindly provided by Dr. Eric Hendrickson recombinant protein substrate GST-X4 as described earlier (Department of Biochemistry, Molecular Biology and Bio- (14). Each phosphorylation reaction contained 25 mmol/L physics, University of Minnesota Medical School, Minnea- Tris-HCl (pH 7.9), 25 mmol/L MgCl2, 1 mmol/L DTT, 25 polis, MN; ref. 12). HT1080 is a human fibrosarcoma cell mmol/L KCl, 10% glycerol, [g-32P]ATP (6,000 Ci/mmol), line that stably expresses yellow fluorescent protein (YFP)- the indicated concentration of DNA-PKcs, 24 nmol/L Ku tagged DNA-PKcs (YFP-DNA-PKcs-HT1080). The ham- dimer, and the indicated concentration of inhibitors or ster Xrs6 cell line lacking Ku80 (Xrs6 Ku80-def) and its dimethyl sulfoxide (DMSO) in a final volume of 10 mL. Ku80 stably transfected counterpart (Xrs6 þ Ku80) was also Reactions were incubated for 30 minutes at 30C and then used. A549 cells are human non–small cell lung cancer cells. terminated by the addition of SDS-PAGE sample buffer. Except for the A549 cells, all cell lines were cultured in Reaction products were analyzed using 8% SDS-PAGE and Minimum Essential Medium routinely supplemented with detected by PhosphorImager analysis (Amersham Pharmacia 10% fetal calf serum and 1% penicillin/streptomycin. A549 Biotech). Thereafter, the gels were stained with 0.1% Coo- cells were cultured in Dulbecco's Modified Eagle's Media massie blue. (DMEM). Cells were incubated in a humidified atmosphere The ability of Akt to activate DNA-PKcs was also inves- fi of 93% air and 7% CO2 at 37 C. Stably transfected cells tigated using kinase assay. Puri ed DNA-PKcs was incu- were maintained in medium containing various antibiotics. bated with active Akt and the reaction mixture described GFP-DNA-PKcs-HCT116 cells were maintained in medi- above in a final volume of 10 mL for 30 minutes on ice. um containing 1 mg/mL puromycin. YFP-DNA-PKcs- Following electrophoresis and blotting, the phosphorylation

946 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

of DNA-PKcs was analyzed using a phospho-specific S2056 tated after radiation exposure or stimulation of cells with antibody. EGF (2). These data suggest a regulatory interaction between Akt1 and DNA-PKcs. To analyze the nature of this potential DNA-binding assay interaction, we further investigated whether an interaction A DNA-binding assay was conducted as described previ- between p-Akt and the DNA-PK regulatory subunits Ku70/ ously (15). Briefly, calf thymus DNA cellulose was sus- 80 can also be observed by co-immunoprecipitation (co-IP). pended in 3 mL of binding buffer (15) and incubated with After radiation exposure or stimulation with EGF or insulin, rotation at 4C overnight. Cellulose suspension was then enhanced co-IP of Ku70/80 can be observed (Fig. 1A and B). washed with fresh binding buffer and centrifuged. The Importantly, in this experiment as calculated by the ratio of supernatant was replaced cell lysate, suspended, incubated Ku80/Akt1, a marked increase by about 50% to 60% of the with rotation for 1 hour at 4C, and centrifuged. The beads proportion of Ku80 in complex with Akt1 after irradiation as were washed with binding buffer, and the amounts of DNA- well as EGF treatment was observed. After insulin treatment PKcs and Akt1 retained on the resin were determined by an approximately 100% increase in the proportion of Ku80 Western blotting. was apparent. In a control immunoprecipitation, a slight binding of DNA-PKcs to the control antibody IgG is Live-cell imaging and laser microirradiation apparent which is not affected by radiation exposure (Fig. The indicated cells were treated with DMSO or 2.5 mmol/ 1C). Because Akt1/DNA-PK complex formation was L API for 48 hours, trypsinized, and cultured on glass slides. induced by ligand stimulation (EGF or insulin) as well as After 24 hours, live-cell imaging combined with laser micro- by IR, we concluded that the presence of activated Akt is a irradiation was conducted (16). The fluorescence of living prerequisite for Akt/DNA-PK complex formation. Next, we cells was monitored with an Axiovert 200M microscope, showed that IR induces immediate Akt1/DNA-PKcs com- MicroImaging Inc. A 365-nm pulsed nitrogen laser (Spec- plex formation as well as Ku70/80 co-IP with Akt1 in the tra-Physics) was directly coupled to the epifluorescence path nucleus (Fig. 1D), but not in the cytoplasm. Co-IP of DNA- of the microscope. DNA-DSBs were generated in a defined PKcs with Akt1 was observed immediately after irradiation area of the nucleus by microirradiation with the 365-nm and was enhanced in a time-dependent manner (Fig. 1D and laser. For quantitative analyses, we used standardized irra- E). Interaction between Akt1 and DNA-PKcs might be diation conditions (80% laser output at 10 Hz for 400 ms) to mediated by direct or indirect binding of both proteins to generate the same amount of DNA-DSBs in each experi- DNA. To rule out this possibility, we conducted immuno- ment. Time lapse images were obtained, and the fluores- precipitation by p-Akt (S472/3) antibody in the nuclear cence intensities of GFP/YFP-tagged DNA-PKcs before and fraction of A549 cells after mock irradiation or irradiation after microirradiation at the site of irradiation within the cell with 4 Gy. Thereafter, immunoprecipitates from irradiated nucleus were determined using Axiovision Software, version samples were not treated or treated for 30 minutes with 50 4.5 (Carl Zeiss). All measurements were corrected for non- mg/mL of EtBr, which is known to disrupt protein–DNA specific bleaching during monitoring. interactions. The results presented in Fig. 1F (top) show that DNA did not mediate Akt1/DNA-PKcs complex formation. Electrophoretic mobility shift assay The IR-induced Akt1/DNA-PKcs complex formation A standard electrophoretic mobility shift assay (EMSA) showed for A549 cells could also be observed in HCT116 was conducted using purified Ku70/80 or nuclear protein wild-type cells (Fig. 1G). fractions. The effects of API on the DNA binding of the Because we could show that Akt1 forms a complex with Ku70/80 heterodimer were investigated using either 600 DNA-PKcs, we analyzed whether binding of Ku70/80 to nmol/L purified protein or 4 mg nuclear protein fraction. Akt1 is DNA-PKcs–dependent. As shown in Fig. 1G, Ku70/ The effect of API on Ku/DNA complex formation in 80 mainly co-IPs with Akt1 in HCT116-DNA-PKcs wild- nuclear fractions was investigated either by directly incu- type, but not in HCT116-DNA-PKcs / cells. These data bating protein samples with the inhibitor or in samples indicate that the Ku70/80 dimer may not directly interact isolated from cells pretreated with the inhibitor. Reactions with Akt1, and its appearance in the complex with Akt1 most were incubated with [g-32P]-labeled 30-bp DNA oligo- likely occurs through binding to DNA-PKcs. Ku-indepen- nucleotides for 30 minutes on ice. DNA–protein com- dent interaction of Akt with DNA-PKcs was further inves- plexes were separated from unbound oligonucleotides by tigated in Ku80-deficient (Xrs6) and Ku80-complemented electrophoresis and detected using a PhosphorImager. For cells (Xrs6 þ Ku80). The results shown in Fig. 1H indicate supershift analysis, the reactions were incubated with the that IR induces Akt1/DNA-PKcs complex formation in Ku80 antibody. both Ku80-deficient and -proficient cells. These data pro- vide evidence that Ku80 is not necessary for Akt1/DNA- PKcs complex formation. Results We further investigated whether a direct physical inter- IR induces Akt1/DNA-PKcs complex formation in the action between Akt1 and DNA-PKcs can be detected. A nucleus GST pull-down experiment in the presence of EtBr using Previously, we have shown that Akt1 or phospho-Akt (p- different Akt1 constructs was carried out. As shown in Akt; S472/3) and DNA-PKcs can be co-immunoprecipi- Fig. 1E, full-length Akt1 and the C-terminal fragment

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 947

Toulany et al.

948 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

containing the kinase domain of Akt1 were able to pull- The function of Akt in the accumulation of DNA-PKcs at down DNA-PKcs, whereas the N-terminal fragment of Akt1 DNA-DSB sites was further investigated by live-cell imag- was unable to pull-down DNA-PKcs (Fig. 1I). ing. GFP-DNA-PKcs–expressing HCT116 cells were treated with API (2.5 mmol/L) and irradiated with a microbeam Akt regulates the accumulation of DNA-PKcs at damage laser. Immediately after irradiation, DNA-PKcs accumulat- sites ed at sites of DNA damage; however, this accumulation was NHEJ repair pathway is initiated by detection of DSBs markedly reduced by API treatment (Fig. 2C). Likewise, through binding of the Ku70/80 heterodimer to the ends of downregulating Akt1 by siRNA (Western blotting data) in the DNA-DSBs, which results in the recruitment and HT1080 cells markedly reduced the accumulation of YFP- binding of DNA-PKcs to the Ku70/80 dimer (17–19). DNA-PKcs at DNA-DSBs (Fig. 2D). In this context and to IR-induced Ku/DNA-PKcs complex formation occurred in further prove the appearance of p-Akt at DNA damage sites, HCT116 wild-type in a time-dependent manner after irra- co-localization of p-Akt foci with g-H2AX foci was inves- diation (Fig. 2A). Because Akt1 forms a direct complex with tigated by confocal microscopy after laser and is shown DNA-PKcs through its kinase domain and is also present in a in Fig. 2E. complex with Ku70/80, we examined whether targeting of The data from the pull-down assays (Fig. 2B, top) indicate Akt interferes with Ku/DNA-PKcs complex formation. To that DNA binding of Ku80 was not affected by Akt inhi- this aim the Akt inhibitor, API-59CJ-OH (API) was used bition. This aspect of Ku/DNA complex formation was (20). In A549 cells, API (2.5 mmol/L) markedly inhibited further investigated more specifically by EMSA using both IR-induced phosphorylation of Akt at S473 and T308 as purified proteins and nuclear fractions of whole-cell lysates. well as the phosphorylation of Akt substrate PRAS40 at As a specificity control, we used the Ku80 antibody at various T246 (Supplementary Fig. S1). Most importantly, however, concentrations to show the presence of a supershift of the API disturbed radiation-induced Ku/DNA-PKcs complex Ku/DNA complex using purified Ku or nuclear lysates of formation (Fig. 2A, bottom). In line with our previous A549 cells irradiated with a dose of 4 Gy (Supplementary reports (2), radiation-inducible phosphorylation of DNA- Fig. S2A). Thereafter, the effect of 2.5 mmol/L API on Ku/ PKcs at T2609 was reduced to about 50% by API. As DNA complex formation was analyzed using purified Ku70/ radiation-induced ataxia telangiectasia mutated (ATM) 80 dimers. In parallel, we examined the effect of API at phosphorylation was not affected, we can most likely exclude concentrations of 0, 2.5, 5, and 12.5 mmol/L on the Ku/ unspecific targeting of PI3K-like kinases by API. On the DNA complex in the nuclear fraction of irradiated A549 basis of these results, we propose that Akt inhibition also cells. As shown in Supplementary Fig. S2B, Ku binding to interferes with DNA binding of DNA-PKcs. We conducted DNA was not affected by API at concentrations up to 12.5 a DNA-PKcs pull-down assay with cells pretreated with or mmol/L. As a specificity control, the supershift induced by a without the Akt inhibitor and showed that Akt inhibition Ku80 antibody is shown in Supplementary Fig. S2B, lane 3. also blocks DNA-PKcs binding to Ku and accumulation of In an additional experiment, cells were pretreated with DNA-PKcs at damage sites (Fig. 2B, top). Importantly, in API and mock-irradiated or irradiated with a dose of 4 Gy. this experiment, Akt inhibition did not affect the binding of Five minutes after irradiation, nuclear extracts were prepared Ku80 to DNA. Yet although the EGFR molecule does not and used for an EMSA. As shown in Supplementary Fig. contain a DNA-binding domain (21), the observed binding S2C, API (2.5 mmol/L) did not affect Ku/DNA complex of EGFR to DNA cellulose, which, however, is not altered by formation. As a specificity control, the Ku80 antibody the experimental conditions tested, is most likely due to the induced a supershift (Supplementary Fig. S2C, lane 1). high amount of EGFR loaded (see input) and contamination of the pull-down assay. In a similar experiment, IR-induced Radiation-induced DNA-PKcs activity is partially Akt1- DNA binding of Akt1 was blocked by the Akt inhibitor. The dependent levels of total Akt1 in whole-cell lysates remained constant On the basis of the direct interaction between Akt1 and (Fig. 2B, bottom). DNA-PKcs, the potential of Akt1 to activate DNA-PKcs was

Figure 1. IR induces Akt1/DNA-PKcs complex formation in the nucleus. A, A549 cells were irradiated with 4 Gy or treated with 100 ng/mL of EGF or insulin, and immunoprecipitation (IP) of phospho-Akt (S473) was conducted. Whole-cell lysates from control and EGF-stimulated cells were used as inputs. After SDS-PAGE, membranes were stained with Ponceau S and then Western blotted for p-Akt (S473), Akt1, Ku70, and Ku80. B, the intensities of p-Akt (S473), Akt1, Ku70, and Ku80 were analyzed by densitometry and normalized to the nonstimulated control (), set to 1. Ins, insulin. C and D, A549 cells were mock-irradiated/irradiated. At the indicated time points post-IR, cytoplasmic and nuclear fractions were isolated, and IP was conducted by using IgG antibody (C) or antibody against Akt1 (D). Co-IP of DNA-PKcs, Ku70, and Ku80 was veriﬁed. Total lysates from cytoplasmic and nuclear fractions, isolated 5 minutes post-IR were used as inputs. Lamin A/C and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were detected as purity controls. E, the densitometric values represent the ratio of DNA-PKcs to Akt1 in nuclear fraction, normalized to nonirradiated control (), set to 1. F, A549 cells were mock-irradiated/irradiated, and 10 minutes post-IR, cytoplasmic and nuclear fractions were isolated. IP was conducted using Akt1 antibody. Immunoprecipitations were mock-treated or treated with 50 mg/mL of EtBr for 30 minutes and followed by washing with washing buffer (2). As purity control for nuclear (Nuc.) and cytoplasmic (Cyt.) preparations, the levels of lamin A/C and GAPDH were detected from whole lysates of both fractions (bottom). G and H, cells were mock-irradiated/irradiated. At indicated time (G) and 10 minutes after IR (H), nuclear fractions were isolated and IP of Akt1 was conducted. Co-IP of DNA-PKcs and Ku70/80 was analyzed. Akt1 was detected as loading. In 100-mg protein samples from the same samples used for IP, the levels of DNA-PKcs and Ku80 were analyzed. I, GST pull-down experiment was conducted using different Akt1 constructs. Pull-down of DNA-PKcs was analyzed by Western blotting. Fl, full length.

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 949

Toulany et al.

Figure 2. Akt1 regulates the accumulation of DNA-PKcs to damage sites. A, cells were irradiated, and following IP of Ku80, co-IP of DNA-PKcs was veriﬁed. The densitometric values represent the ratio of DNA-PKcs to Ku80 normalized to the nonirradiated control (), set to 1. Total lysates from DNA-PKcs / cells were used as inputs. Bottom, A549 cells were pretreated with or without API (2.5 mmol/L for 72 hours) before IR. Thereafter, IP of DNA-PKcs was conducted, and co- IP of Ku80 with DNA-PKcs was analyzed. Hundred micrograms of whole-cell lysates from DMSO-treated nonirradiated cells was used as input. B, A549 cells were treated with or without API (2.5 mmol/L) for 72 hours and were mock-irradiated/irradiated with 10 Gy. At indicated times post-IR (top) or 5 minutes post-IR (bottom), protein samples were prepared, and a DNA-binding assay was conducted. DNA-bound proteins were subjected to SDS-PAGE. After Ponceau staining, levels of Ku80, DNA-PKcs, and Akt1 were analyzed. EGFR with a lack of DNA-binding domain was detected as negative control. Lysates of nonirradiated, DMSO-treated cells were used as input. Levels of Akt1 in whole-cell lysates were used to indicate constant amounts of Akt under different

950 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

Figure 3. Radiation-induced DNA-PKcs activity is partially Akt-dependent. A, purified human DNA-PKcs was incubated with or without active recombinant human Akt1 in the presence of ATP, and a kinase reaction was conducted. Levels of p-DNA-PKcs (S2056), DNA-PKcs, and Akt1 were detected by Western blotting. B, cells were treated with/without API (2.5 mmol/L) for 72 hours and mock-irradiated/irradiated with 4 Gy. Five minutes post-IR, nuclear fractions were isolated, and 4 mg of protein was used in the assay. Following the detection of phospho-GST-X4 (p-GST-X4), gels were stained with 0.1% Coomassie blue. The DNA-PKcs inhibitor NU7026 (10 mmol/L, 1-hour pretreatment) was used as positive control. analyzed. In an in vitro kinase assay, active Akt1 was protein, GST-X4 was highly phosphorylated. API did not incubated with DNA-PKcs in the presence of ATP and markedly affect the phosphorylation of the DNA-PKcs kinase buffer. Active Akt1 clearly induced autophosphoryla- substrate, but the DNA-PKcs inhibitor NU7441 completely tion of DNA-PKcs at S2056 (Fig. 3A). blocked GST-X4 phosphorylation (Supplementary Fig. Furthermore, the effect of API on DNA-PKcs kinase S3A). activity was analyzed in a [g-32P]ATP-based in vitro kinase To investigate whether Ku-independent and DNA- assay. A GST-tagged recombinant XRCC4 protein fragment PKcs–dependent phosphorylation of GST-X4 were affect- (251–334 a.a., GST-X4) was used as a DNA-PKcs–specific ed by API, DNA-PKcs kinase assays were conducted in the substrate (14). A549 cells were either pretreated or untreated absence of Ku heterodimer. Reactions were incubated with API (2.5 mmol/L) and mock-irradiated or irradiated with different concentrations of DNA-PKcs (0, 12.5, with a dose of 4 Gy. Nuclear lysates were obtained from these 30, or 60 nmol/L). A comparison of the phosphorylation cells and were used in a kinase assay. As shown in Fig. 3B, of GST-X4 in either the presence or absence of Ku70/80 API treatment reduced basal and radiation-induced DNA- indicated that approximately 80% of DNA-PKcs activity PKcs activity in both the cell lines analyzed. As a positive is Ku70/80-dependent (Supplementary Fig. S3B, lane 1). control, the DNA-PKcs inhibitor NU7026 completely However, even in the absence of Ku70/80, a minor blocked phosphorylation of GST-X4 (Fig. 3B). Specific proportion of phosphorylated GST-X4 was observed as DNA-PKcs–dependent phosphorylation of GST-X4 was a function of the DNA-PKcs concentration (Supplemen- confirmed by a lack of substrate phosphorylation under tary Fig. S3B). On the basis of these results, we investi- basal conditions or post-irradiation in HCT116-DNA- gated the effect of API on GST-X4 phosphorylation by PKcs / cells (Fig. 3B). DNA-PKcs in the absence of Ku70/80. As shown in Because Akt inhibition markedly affected DNA-PKcs Supplementary Fig. S3C, incubation of the reaction mix- kinase activity in cell culture (Fig. 3B), we investigated tures with API at concentrations of 2.5 to 25 mmol/L did whether API had any off-target effects on Ku-dependent not affect GST-X4 phosphorylation. Again, in the pres- DNA-PKcs kinase activity in the absence of Akt. As shown ence of the Ku dimer, strong phosphorylation of GST-X4 in Supplementary Fig. S3A, in the presence of Ku70/80 was observed (Supplementary Fig. S3C, lane 1).

treatment conditions (bottom). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control. C, GFP-DNA-PKcs-HCT116 cells were treated with API (2.5 mmol/L) for 48 hours and plated onto glass slides. After 24 hours, cells were irradiated with a microbeam laser. The fluorescence intensity of GFP-tagged DNA-PKcs at the irradiation site before and at 0 to 320 seconds after irradiation was determined and tested for statistical significant differences (P < 0.001; Student t test). Arrows point to the microirradiated site. D, YFP-DNA-PKcs-HT1080 cells were transfected with control (ctrl) or AKT1-siRNA. Four days after transfection, levels of Akt1 and actin were determined. In parallel, 3 days after transfection with AKT1-siRNA, cells were transferred to glass slides and irradiated with microbeam laser 24 hours later. Intensity of YFP-DNA-PKcs accumulation at the irradiation site was quantified. Each data point in C and D represents the average of 15 independent measurements. Data points represent the mean fluorescence intensity values SEM. Statistical significance (P < 0.01) was analyzed by Student t test. E, GFP-DNA-PKcs–expressing HCT116 cells were laser-irradiated, incubated for 30 minutes, fixed, and immunostained with p-Akt (S473) and g-H2AX. Ctrl, control.

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 951

Toulany et al.

The radiosensitizing effect of API depends on the siRNA also resulted in a marked reduction in radiation- expression of DNA-PK induced DNA-PKcs autophosphorylation at S2056 in A549 Because Akt1 directly interacts with DNA-PKcs and cells (Fig. 5B). Moreover, in A549 and HCT116 cells, the regulates its activation following radiation exposure, pattern frequency of residual DNA-DSBs was significantly of radiosensitization was investigated in DNA-PKcs and enhanced by API treatment (Fig. 5C). Ku70/80-proficient and -deficient cells after Akt targeting. Autophosphorylation of DNA-PKcs at 2056 is an API increased radiation sensitivity in HCT116wt cells and essential step in the release of DNA-PKcs from damage Xrs6 þ Ku80 cells, but not in NHEJ-deficient HCT116- sites after DSB repair (16, 22) and is reviewed by Dobbs DNA-PKcs / and Xrs6 þ Ku80 / (Ku80-def.). In CHO and colleagues (23). Therefore, we investigated whether Rad51D1 / homologous recombination cells (Rad51D1- Akt1 expression is required for dissociation of DNA-PKcs def.), a similar radiosensitization by API as seen for NHEJ from damage sites. Cells were transfected with AKT1- repair–proficient cells (HCT116, Xrs6 þ KU80, A549, and siRNA and irradiated with a dose of 4 Gy. The cells were HT1080) was observed (Fig. 4B and D). Moreover, as subsequently stained with antibodies specific for p-S2056 expected, the high intrinsic radiosensitivity of DNA-PKcs- and g-H2AX foci 24 hours after irradiation. As shown and Ku80-deficient cells was not affected after high-dose (1– in Fig. 5D, 24 hours after radiation exposure, phosphor- 4 Gy) or low-dose (0.2–1 Gy) irradiation in combination ylation of DNA-PKcs at S2056 appears as foci and co- with API pretreatment (Fig. 4B and C). These control localizes with g-H2AX foci. Interestingly, similar to the experiments support the assumption that targeting Akt increase in g-H2AX foci after Akt reduction, the frequen- results in an impairment of NHEJ repair and consequently cy of p-S2056 foci after downregulation of Akt1 was radiosensitization. significantly enhanced (Fig. 5D and E). The magnified Moreover, we investigated whether API treatment picture of p-DNA-PKcs (S2056) resulting from ctrl- affected cell proliferation and consequently a redistribu- siRNA–transfected cells indicates that phosphorylated tion of cells through the cell cycle. As shown by the slope DNA-PKcs appears in the focus as well as throughout of the growth curve in Fig. 4E, API did not affect the the nucleus, whereas in AKT1-siRNA–transfected cells, doubling time but prolonged the lag phase of cells before the S2056 signal is mainly co-localized with g-H2AX foci entering logarithmic growth. This led to a significant but not throughout the nucleus. These data indicate difference in the number of cells at the indicated time that targeting of Akt impairs DNA-DSB repair and the points (Fig. 4E). Moreover, cell-cycle analysis indicated subsequent release of DNA-PKcs from the damage sites that API treatment alone did not affect the cell-cycle (Fig. 5D). distribution significantly but resulted in combination with irradiation in a significant enhancement of the IR-medi- Discussion ated G1 arrest 24 hours after irradiation. The percentage of sub-G1cells indicating the fraction of apoptotic cells was Previously, it was described that phosphorylation of not increased after treatment with API alone or in com- DNA-PKcs mediated by IR is markedly antagonized by bination with irradiation (control, 0.21% 0.01%; API, specific inhibitors of EGFR, PI3K, and Akt such as 0.17% 0.03%; IR, 0.24% 0.02%; API þ IR, 0.21% erlotinib, LY294002, PI-103, and Akt inhibitor VIII 0.03%; data not shown). These data suggest that (3, 6). Consequently, these inhibitors have been shown inhibition of Akt activity by API (Supplementary Fig. to impair DNA-DSB repair and to induce radiosensitiza- S1) does not result in a potential IR-induced apoptosis. tion of a variety of tumor cells in vitro and in vivo (3, 6, 24). Increased cellular radiosensitivity following Targeting of Akt1 impairs the repair of radiation- PI3K/Akt targeting coincides with enhanced residual induced DNA-DSBs and prevents dissociation of DNA- DNA-DSBs(2,5,6,25,26)mostlikelythroughinter- PKcs from damage sites ference with DNA-PKcs phosphorylation during the Because Akt inhibition markedly reduced DNA-PKcs NHEJ repair process (2). However, until now, the specific kinase activity in living cells (see Fig. 3B), we investigated regulatory function of Akt in NHEJ repair mechanism has whether Akt inhibition reduces autophosphorylation of not been investigated. Yet this topic was specifically DNA-PKcs at S2056. This was accomplished using A549 addressed in the present study. The results presented and HCT116wt cells. As shown in Fig. 5A, 30 minutes after herein indicate that Akt in general and specifically Akt1 irradiation, both cells lines showed similar levels of total through 3 functions can regulate DNA-DSB repair. First, DNA-PKcs; however, radiation-induced phosphorylation of complex formation of Akt and DNA-PKcs stimulates DNA-PKcs was much stronger in A549 cells than in binding of DNA-PKcs to DNA duplex ends marked by HCT116wt cells. Because the reduction in S2056 phos- Ku dimers. Second, Akt in the complex with DNA-PKcs phorylation mediated by API was much stronger in irradi- promotes kinase activity of DNA-PKcs, which is necessary ated A549 cells than in HCT116wt cells, this result could for executing an efficient DNA-DSB repair. Third, Akt- explain the stronger API-induced radiosensitization of A549 stimulated autophosphorylation of DNA-PKcs facilitates ¼ þ cells than HCT116wt cells (A549: SF3 API 0.23, SF3 thereleaseofDNA-PKcsfromthedamagesitethatis ¼ ¼ þ ¼ API 0.09; HCT116wt: SF3 API 0.06, SF3 API known to be a necessary step for ligation and termination 0.03). In addition, we showed that Akt1 knockdown by of DNA-DSB repair (16, 23).

952 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

ABHCT116 DNA-PKcs 1.00 HCT116 wt Xrs6 + Ku80 Rad51D1-def Wild-type –/– wt + API + Ku80 + API Ctrl API DNA-PKcs DNA-PKcs–def Ku80–def DNA-PKcs–def Ku80–def + API +API

Actin SF ± SEM 0.1 Xrs6 - Ku80 + Ku80

Ku80

Actin 0.01 0 12345012345012345

1.00 HCT116 Ku80 DNA-PKcs–def. Xrs6-def. 10 ) 6 Ctrl API SF ± SEM 0.1 1

-API -API 0.1 +API +API Cell numbers ± SEM (10 Cell numbers 0.01 D 0 D 2 D 4 D 6 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 DF

1.00 A549 HT1080 70 G1 S 60 G2–M

50 SF ± SEM

0.1 40

-API -API 20 +API +API of cells ± SEM Percentage 0.01 10 012345 60123456 Ctrl API IR API/IR X-ray (Gy) X-ray (Gy)

Figure 4. Radiosensitizing effect of API depends on DNA-PKcs and Ku80 expression. A, whole-cell lysates were applied for analysis of the expression levels of DNA-PKcs and Ku80. Actin was used as a loading control. B–D, cells were treated with/without API (2.5 mmol/L) for 72 hours and plated to assess colony formation. After 24 hours, cells were irradiated with the indicated doses of IR. Colonies that formed after 10 days were counted, and the clonogenic fraction of irradiated cells was normalized to the plating efficiency of nonirradiated controls. The data bars in B to D represent the mean surviving fraction (SF) SEM from 6 parallel experiments in HCT116, Xrs6, and A549 cells and 3 parallel experiments in HT1080 cells. , statistically significant radiosensitization in response to API (, P < 0.05; , P < 0.01; , P < 0.001; Student t test). E, A549 cells were treated with API (2.5 mmol/L) for 72 hours and plated in 60-mm culture dishes. After the time periods, indicated cells were counted and graphed. Data points shown represent the mean SEM of 12 data points from 2 independent experiments. , statistically significant antiproliferative effect of API (, P < 0.01; , P < 0.001; Student t test). F, API-pretreated A549 cells were seeded in 6-cm dishes and after 24 hours were irradiated. After irradiation for 24 hours, cells were collected, and fluorescence-activated cell-sorting (FACS) analysis was conducted. Percentage of cells in different cell cycles were calculated and graphed. A significant IR-induced G1 arrest ( , P < 0.001) and enhancement of IR-induced G1 arrest by API ( , P < 0.001) was observed. Combination of API with radiation reduced the population of cells in S and G2/M significantly ( , P < 0.01). Ctrl, control.

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 953

Toulany et al.

Figure 5. Targeting Akt1 impairs radiation-induced DNA-DSB repair and prevents dissociation of DNA-PKcs from damage sites. A, HCT116wt and A549 cells were pretreated with/without API (2.5 mmol/L) for 72 hours and irradiated with 4 Gy. At 10 and 30 minutes post-IR, protein samples were prepared, and levels of p-DNA-PKcs (S2056) and DNA-PKcs were analyzed by Western blot analysis. B, A549 cells were transfected with control (ctrl)- or AKT1-siRNA and irradiated with 4 Gy, 4 days later. Expression levels of Akt1, p-DNA-PKcs (S2056), and DNA-PKcs were analyzed by Western blotting. Actin was used as a loading control. C, A549 and HCT116wt cells were treated with/without API (2.5 mmol/L) for 72 hours and irradiated with 2 or 4 Gy. Twenty-four hours post-IR, the cells were stained with an antibody-specific for p-H2AX (S139). Using a fluorescence microscope, the number of g-H2AX foci was counted in 160 to 260 cells per treatment condition. , statistically significant enhancement of residual g-H2AX foci following API treatment (, P < 0.001; Student t test). D1–4, A549 cells were transfected with control (ctrl)-siRNA or AKT1-specific siRNA. Three days after transfection, the cells were irradiated with 4 Gy and were stained with antibodies specific for p-H2AX (S139) and p-DNA-PKcs (S2056) 24 hours post-IR. DNA was stained with YO-PRO (green). E, the number of S2056 foci in 100 to 135 nuclei was counted and graphed. , statistically significant enhancement of residual S2056 foci following AKT1-siRNA transfection (, P < 0.001; Student t test).

In line with the report by Park and colleagues (27), in our showing that the kinetics of DNA-DSB repair through the study, we showed that complex formation of Akt1 with NHEJ mechanism is dominated by "fast" and "slow" com- DNA-PKcs requires the C-terminal domain of Akt1. ponents. Approximately 85% of DNA-DSBs induced by IR Because of this direct interaction, exposure of cells to IR are repaired within the ﬁrst 2 to 3 hours of post-irradiation leads to an immediate Akt1/DNA-PKcs complex formation. via the fast component, which has been shown to be Bozulic and colleagues (11) also reported a complex forma- independent of ATM function (28). The remaining 15% tion of Akt1 and DNA-PKcs 30 minutes after radiation of DNA-DSBs, which are mainly composed of complex exposure. However, in that study, no further investigation of lesions, are repaired in an ATM-dependent manner via the the regulatory function of this complex in DNA repair was slow component (29, 30). Thus, in this context, the imme- conducted. Convincing data have been published by Jeggo diate post-irradiation formation of the nuclear Akt1/DNA- and colleagues (28) and Goodarzi and colleagues (29) PKcs complex can be assumed to be primarily an important

954 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

Akt

DNA-PKcs

Akt Akt Ku70/80 (1) (2) DNA-PKcs DNA-PKcs

Ligases and processing enzymes

Activated Activated Activated Activated Akt Akt Akt Akt p (3) Activated Activated (4) Activated Activated DNA-PKcs DNA-PKcs DNA-PKcs DNA-PKcs

S2056 S2056

p Autophosphorylation

Figure 6. Proposed interaction of Akt and DNA-PKcs during NHEJ repair process. On the basis of the data shown, the scheme illustrates the potential interaction of Akt with DNA-PKcs with respect to the described classical NHEJ (18, 23). Step 1, after induction of DNA-DSBs by IR, DNA break-sites are detected and marked by Ku70/80 heterodimer. Step 2, for recruitment of DNA-PKcs and binding to Ku70/80, direct interaction and binding of Akt to DNA-PKcs is needed. Step 3, DNA-PKcs undergoes trans/autophosphorylation, resulting in the release of DNA-PKcs from the repaired damage site. Step 4, various additional factors, such as XLF, XRCC4, DNA-ligase IV, and polynucleotide kinase/phosphatase process the final steps of NHEJ repair. step in the execution of DNA-DSB repair via the fast can be assumed that Akt promotes the kinase activity of component. This assumption is further substantiated by DNA-PKcs, which seems to be prerequisite for binding of our immunostaining results showing that activated Akt, that DNA-PKcs to Ku70/80 (34). On the basis of the observa- is, p-Akt (S473), co-localizes with g-H2AX foci in micro- tion that inhibition of Akt interferes with the binding of beam laser–irradiated cells. This observation is in agreement DNA-PKcs to Ku70/80, we addressed the question whether with our previous report (31) and the report by Fraser and Akt targeting interferes with the recruitment of DNA-PKcs colleagues (32) reporting co-localization of p-Akt (S473) to the damage site. Results from live-cell imaging indicate with g-H2AX foci after exposure to IR. However, it is that Akt targeting prevents accumulation of DNA-PKcs at unclear whether the IR-induced S473 signal is solely orig- the site of DNA damage. Yet the repression of DNA-PKcs inated from p-Akt1 or from the phosphorylated isoforms accumulation is much more prominent in cells treated with Akt2 and Akt3. Because the commercially available p-Akt API than in cells transfected with AKT1-siRNA. This (S473) antibodies are used detect the phosphorylation of differential effect can most likely be explained by the fact Akt2 at S474 as well as the phosphorylation of Akt3 at S472, that API inhibits all 3 isoforms of Akt (20), whereas the the involvement of Akt2 and Akt3 in the regulation of DNA- siRNA approach depletes specifically Akt1. Thus, these data DSB repair cannot be ruled out at present and needs to be cannot exclude the involvement of Akt2 and/or Akt3 in the further investigated. regulation of NHEJ. Nevertheless, because AKT1-siRNA The first step in the initiation of the DNA-DSB repair via results in an approximately 50% reduction of DNA-PKcs NHEJ pathway requires the binding of the Ku70/80 hetero- accumulation, it can also be argued that the Akt1 is the major dimer to the 30 and 50 ends of the DNA break-site which isoform of Akt, which is directly involved in the efficient marks the DNA-DSB for binding of DNA-PKcs and its activation step of DNA-PK–dependent DSB repair. further activity (33). As shown, binding of DNA-PKcs to As reported by Viniegra and colleagues (35), full activation Ku70/80 can be significantly inhibited by pretreating cells of Akt in response to IR is mediated through ATM. Regard- with the Akt inhibitor API (see Fig. 2A, bottom). Thus, it ing the role of ATM in Akt phosphorylation, Fraser and

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 955

Toulany et al.

colleagues (32) showed that Akt phosphorylation depends detected and bound by Ku70/80 heterodimers, which on MRE11-ATM-RNF168 signaling in direct response to represents the initial step of NHEJ repair (Fig. 6, step 1). DNA-DSB–induced IR (32). Transphosphorylation of For recruitment of DNA-PKcs and binding to Ku70/80, DNA-PKcs at T2609 seems not only to be dependent on the direct interaction and complex formation of Akt to ATM as reported by Chen and colleagues (36) but also on DNA-PKcs seem to warrant full activity of DNA-PKcs Akt as previously shown by Choi and colleagues (6) and our (Fig.6,step2).AfterprocessingofDNAbreakendsby laboratory (2). Although the involvement of Akt in ATM- various proteins, such as XLF, XRCC4, DNA-ligase IV dependent slow component of DNA-DSB repair needs to be and polynucleotide kinase/phosphatase (Fig. 6, step 3), further investigated, it can be assumed that Akt acts as a link efficient autophosphorylation of DNA-PKcs at S2056 between ATM and DNA-PKcs. This conclusion is further occurs under co-control of Akt. As a consequence of this supported by the results shown in Supplementary Fig. S1, Akt-stimulated autophosphorylation, DNA-PKcs will be which indicates that Akt inhibition without affecting ATM released from the repaired damage site (Fig. 6, step 4). phosphorylation inhibits DNA-PKcs phosphorylation. Thus, in this model, Akt is stimulating the efficient DNA-PKcs kinase activity and its autophosphorylation, accumulation of DNA-PKcs at the damage site, initiation partially activated by Akt, are necessary steps for the pro- of the repair process, as well as autophosphorylation of gression and the termination of DNA-DSB repair. Espe- DNA-PKcs, which is prerequisite for the release of DNA- cially, autophosphorylation at DNA-PKcs at S2056 (36) is PKcs after repair has been executed. necessary for efficient DNA-DSB repair through NHEJ In conclusion, the present study provides the first mech- mechanism (10, 36–39) and for subsequent release of anistic evidence for a stimulatory role of Akt in DNA-PKcs– DNA-PKcs from the damage site (16, 22, 23). Akt-depen- dependent NHEJ repair pathway in human tumor cells after dent DNA-PKcs activation/autophosphorylation implies radiation exposure. On the basis of the results shown, that Akt1 functions as a mediator of the dissociation of specific experiments can be designed to answer open ques- DNA-PKcs from the DNA-DSB which is a prerequisite step tions of specific details of Akt/DNA-PKcs interaction during for ligation and subsequent termination of DNA-DSB DNA-DSB repair in irradiated tumor cells. repair. This conclusion is supported by a significantly Disclosure of Potential Conflicts of Interest enhanced amount of DNA-PKcs foci co-localizing with fl unrepaired DNA-DSBs after Akt1 knockdown. Yet the No potential con icts of interests were disclosed. – AKT1-siRNA mediated increase in the number of residual Authors' Contributions S2056 foci 24 hours after irradiation seems to be in conflict Conception and design: M. Toulany, K.R. Fattah, D.J. Chen, H.P. Rodemann with Akt targeting–mediated inhibition of DNA-PKcs phos- Development of methodology: M. Toulany, K.-J. Lee, B. Fehrenbacher, B.P. Chen, D.J. Chen phorylation at S2056 for up to 30 minutes after IR observed Acquisition of data (provided animals, acquired and managed patients, provided by Western blotting (see Fig. 5A and B). This potential facilities, etc.): M. Toulany, K.-J. Lee, K.R. Fattah, Y.-F. Lin, M. Schaller, B.P. Chen conflict can be explained by a decreased repair efficiency in Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- tational analysis): M. Toulany, K.-J. Lee, K.R. Fattah, B. Fehrenbacher, M. Schaller, AKT-targeted cells, which results 24 hours after radiation in B.P. Chen, D.J. Chen, H.P. Rodemann an increased number of S2056 foci because of nonreleased Writing, review, and/or revision of the manuscript: M. Toulany, K.R. Fattah, M. Schaller, B.P. Chen, D.J. Chen, H.P. Rodemann DNA-PKcs. In contrast, the noncompromised DNA repair Administrative, technical, or material support (i.e., reporting or organizing data, efficacy in control cells results in a decreased number of constructing databases): M. Toulany, B. Fehrenbacher, H.P. Rodemann DNA-PKcs foci, reflecting the termination of DSB repair Study supervision: M. Toulany, H.P. Rodemann and released DNA-PKcs from the damage site. With respect Acknowledgments to the stimulation of DNA-PKcs by Akt, it needs to be The authors thank Dr. Rainer Kehlbach and Dr. Stephan M. Huber for further investigated whether phosphorylation of DNA-PKcs fluorescence-activated cell-sorting (FACS) analysis and Tim-Andre Schickfluß and is taking place before or after recruitment of Akt/DNA-PKcs Shih-Ya Wang for technical assistance. complex to DNA damage site. Furthermore, it needs to be fi fi Grant Support clari ed whether Akt is modi ed upon recruitment of the The study was supported by grants from the Deutsche Forschungsgemeinschaft Akt/DNA-PKcs complex to DNA. On the basis of the data [Ro527/5-1 (DFG-PAK190); SFB-773-TP B02] awarded to H.P. Rodemann; 03- presented herein and in reflection of the current model of MTO-80000903 awarded to M. Toulany/H.P. Rodemann; and SFB 773-TP Z2 awarded to M. Schaller. how DNA-PKcs is acting at the DNA damage site described The costs of publication of this article were defrayed in part by the payment of page by Weterings and Chen (18) and Dobbs and colleagues (23), charges. This article must therefore be hereby marked advertisement in accordance with we suggest an expanded model by implementing the func- 18 U.S.C. Section 1734 solely to indicate this fact. tion of Akt in the regulation of NHEJ repair (Fig. 6). After Received December 12, 2011; revised April 19, 2012; accepted May 4, 2012; induction of DNA-DSBs by IR, DNA-DSB sites are published OnlineFirst May 17, 2012.

References 1. Mukohara T, Kudoh S, Matsuura K, Yamauchi S, Kimura T, Yoshimura 2. Toulany M, Kehlbach R, Florczak U, Sak A, Wang S, Chen J, et al. N, et al. Activated Akt expression has signiﬁcant correlation with EGFR Targeting of AKT1 enhances radiation toxicity of human tumor cells by and TGF-alpha expressions in stage I NSCLC. Anticancer Res inhibiting DNA-PKcs-dependent DNA double-strand break repair. Mol 2004;24:11–7. Cancer Ther 2008;7:1772–81.

956 Mol Cancer Res; 10(7) July 2012 Molecular Cancer Research

Function of Akt in DNA-DSB Repair

3. Toulany M, Kasten-Pisula U, Brammer I, Wang S, Chen J, Dittmann K, 21. Lo HW, Hsu SC, Ali-Seyed M, Gunduz M, Xia W, Wei Y, et al. Nuclear et al. Blockage of epidermal growth factor receptor-phosphatidylino- interaction of EGFR and STAT3 in the activation of the iNOS/NO sitol 3-kinase-AKT signaling increases radiosensitivity of K-RAS pathway. Cancer Cell 2005;7:575–89. mutated human tumor cells in vitro by affecting DNA repair. Clin Cancer 22. Hammel M, Yu Y, Mahaney BL, Cai B, Ye R, Phipps BM, et al. Ku and Res 2006;12:4119–26. DNA-dependent protein kinase dynamic conformations and assembly 4. Golding SE, Morgan RN, Adams BR, Hawkins AJ, Povirk LF, Valerie K. regulate DNA binding and the initial non-homologous end joining Pro-survival AKT and ERK signaling from EGFR and mutant EGFRvIII complex. J Biol Chem 2010;285:1414–23. enhances DNA double-strand break repair in human glioma cells. 23. Dobbs TA, Tainer JA, Lees-Miller SP. A structural model for regulation Cancer Biol Ther 2009;8:730–8. of NHEJ by DNA-PKcs autophosphorylation. DNA Repair 2010;9: 5. Kao GD, Jiang Z, Fernandes AM, Gupta AK, Maity A. Inhibition of 1307–14. phosphatidylinositol-3-OH kinase/Akt signaling impairs DNA repair in 24. Konstantinidou G, Bey EA, Rabellino A, Schuster K, Maira MS, Gazdar glioblastoma cells following ionizing radiation. J Biol Chem AF, et al. Dual phosphoinositide 3-kinase/mammalian target of rapa- 2007;282:21206–12. mycin blockade is an effective radiosensitizing strategy for the treat- 6. Choi EJ, Ryu YK, Kim SY, Wu HG, Kim JS, Kim IH, et al. Targeting ment of non-small cell lung cancer harboring K-RAS mutations. Can- epidermal growth factor receptor-associated signaling pathways in cer Res 2009;69:7644–52. non-small cell lung cancer cells: implication in radiation response. Mol 25. Prevo R, Deutsch E, Sampson O, Diplexcito J, Cengel K, Harper J, et al. Cancer Res 2010;8:1027–36. Class I PI3 kinase inhibition by the pyridinylfuranopyrimidine inhibitor 7. IliakisG,WangH,PerraultAR,Boecker W, Rosidi B, Windhofer F, PI-103 enhances tumor radiosensitivity. Cancer Res 2008;68: et al. Mechanisms of DNA double strand break repair and chromo- 5915–23. some aberration formation. Cytogenet Genome Res 2004;104: 26. Westhoff MA, Kandenwein JA, Karl S, Vellanki SH, Braun V, Eramo A, 14–20. et al. The pyridinylfuranopyrimidine inhibitor, PI-103, chemosensitizes 8. Povirk LF, Zhou RZ, Ramsden DA, Lees-Miller SP, Valerie K. Phos- glioblastoma cells for apoptosis by inhibiting DNA repair. Oncogene phorylation in the serine/threonine 2609–2647 cluster promotes but is 2009;28:3586–96. not essential for DNA-dependent protein kinase-mediated nonhomol- 27. Park J, Feng J, Li Y, Hammarsten O, Brazil DP, Hemmings BA. DNA- ogous end joining in human whole-cell extracts. Nucleic Acids Res dependent protein kinase-mediated phosphorylation of protein kinase 2007;35:3869–78. B requires a specific recognition sequence in the C-terminal hydro- 9. Douglas P, Sapkota GP, Morrice N, Yu Y, Goodarzi AA, Merkle D, et al. phobic motif. J Biol Chem 2009;284:6169–74. Identification of in vitro and in vivo phosphorylation sites in the catalytic 28. Jeggo PA, Geuting V, Lobrich M. The role of homologous recombi- subunit of the DNA-dependent protein kinase. Biochem J nation in radiation-induced double-strand break repair. Radiother 2002;368:243–51. Oncol 2011;101:7–12. 10. Chen BP, Chan DW, Kobayashi J, Burma S, Asaithamby A, Morotomi- 29. Goodarzi AA, Jeggo P, Lobrich M. The influence of heterochromatin on Yano K, et al. Cell cycle dependence of DNA-dependent protein kinase DNA double strand break repair: getting the strong, silent type to relax. phosphorylation in response to DNA double strand breaks. J Biol DNA Repair 2010;9:1273–82. Chem 2005;280:14709–15. 30. Beucher A, Birraux J, Tchouandong L, Barton O, Shibata A, Conrad S, 11. Bozulic L, Surucu B, Hynx D, Hemmings BA. PKBalpha/Akt1 acts et al. ATM and Artemis promote homologous recombination of radi- downstream of DNA-PK in the DNA double-strand break response and ation-induced DNA double-strand breaks in G2. EMBO J 2009;28: promotes survival. Mol Cell 2008;30:203–13. 3413–27. 12. Ruis BL, Fattah KR, Hendrickson EA. The catalytic subunit of DNA- 31. Toulany M, Schickfluss TA, Fattah KR, Lee KJ, Chen BP, Fehrenbacher dependent protein kinase regulates proliferation, telomere length, and B, et al. Function of erbB receptors and DNA-PKcs on phosphorylation genomic stability in human somatic cells. Mol Cell Biol 2008;28: of cytoplasmic and nuclear Akt at S473 induced by erbB1 ligand and 6182–95. ionizing radiation. Radiother Oncol 2011;101:140–6. 13. Toulany M, Dittmann K, Fehrenbacher B, Schaller M, Baumann M, 32. Fraser M, Harding SM, Zhao H, Coackley C, Durocher D, Bristow RG. Rodemann HP. PI3K-Akt signaling regulates basal, but MAP-kinase MRE11 promotes AKT phosphorylation in direct response to DNA signaling regulates radiation-induced XRCC1 expression in human double-strand breaks. Cell Cycle 2011;10:2218–32. tumor cells in vitro. DNA Repair 2008;7:1746–56. 33. Lieber MR. The mechanism of human nonhomologous DNA end 14. Lee KJ, Jovanovic M, Udayakumar D, Bladen CL, Dynan WS. Iden- joining. J Biol Chem 2008;283:1–5. tification of DNA-PKcs phosphorylation sites in XRCC4 and effects of 34. Weterings E, van Gent DC. The mechanism of non-homologous end- mutations at these sites on DNA end joining in a cell-free system. DNA joining: a synopsis of synapsis. DNA Repair 2004;3:1425–35. Repair (Amst) 2004;3:267–76. 35. Viniegra JG, Martinez N, Modirassari P, Losa JH, Parada Cobo C, Lobo 15. Johnson LR, Scott MG, Pitcher JA. G protein-coupled receptor kinase VJ, et al. Full activation of PKB/Akt in response to insulin or ionizing 5 contains a DNA-binding nuclear localization sequence. Mol Cell Biol radiation is mediated through ATM. J Biol Chem 2005;280:4029–36. 2004;24:10169–79. 36. Chen BP, Uematsu N, Kobayashi J, Lerenthal Y, Krempler A, Yajima H, 16. Uematsu N, Weterings E, Yano K, Morotomi-Yano K, Jakob B, et al. Ataxia telangiectasia mutated (ATM) is essential for DNA-PKcs Taucher-Scholz G, et al. Autophosphorylation of DNA-PKCS reg- phosphorylations at the Thr-2609 cluster upon DNA double strand ulates its dynamics at DNA double-strand breaks. J Cell Biol 2007; break. J Biol Chem 2007;282:6582–7. 177:219–29. 37. Chan DW, Chen BP, Prithivirajsingh S, Kurimasa A, Story MD, Qin J, 17. Mahaney BL, Meek K, Lees-Miller SP. Repair of ionizing radiation- et al. Autophosphorylation of the DNA-dependent protein kinase induced DNA double-strand breaks by non-homologous end-joining. catalytic subunit is required for rejoining of DNA double-strand breaks. Biochem J 2009;417:639–50. Genes Dev 2002;16:2333–8. 18. Weterings E, Chen DJ. DNA-dependent protein kinase in nonhomol- 38. Douglas P, Cui X, Block WD, Yu Y, Gupta S, Ding Q, et al. The DNA- ogous end joining: a lock with multiple keys? J Cell Biol 2007;179: dependent protein kinase catalytic subunit is phosphorylated in vivo on 183–6. threonine 3950, a highly conserved amino acid in the protein kinase 19. Weterings E, Chen DJ. The endless tale of non-homologous end- domain. Mol Cell Biol 2007;27:1581–91. joining. Cell Res 2008;18:114–24. 39. Ding Q, Reddy YV, Wang W, Woods T, Douglas P, Ramsden DA, et al. 20. Tang HJ, Jin X, Wang S, Yang D, Cao Y, Chen J, et al. A small molecule Autophosphorylation of the catalytic subunit of the DNA-dependent compound inhibits AKT pathway in ovarian cancer cell lines. Gynecol protein kinase is required for efficient end processing during DNA Oncol 2006;100:308–17. double-strand break repair. Mol Cell Biol 2003;23:5836–48.

www.aacrjournals.org Mol Cancer Res; 10(7) July 2012 957

Akt Promotes Post-Irradiation Survival of Human Tumor Cells through Initiation, Progression, and Termination of DNA-PKcs− Dependent DNA Double-Strand Break Repair

Mahmoud Toulany, Kyung-Jong Lee, Kazi R. Fattah, et al.

Mol Cancer Res 2012;10:945-957. Published OnlineFirst May 17, 2012.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-11-0592

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2012/05/16/1541-7786.MCR-11-0592.DC1

Cited articles This article cites 39 articles, 21 of which you can access for free at: http://mcr.aacrjournals.org/content/10/7/945.full#ref-list-1

Citing articles This article has been cited by 11 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/10/7/945.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/10/7/945. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.