TGFβ1 Protects Cells from γ-IR by Enhancing the Activity of the NHEJ Repair Pathway

TGFβ1 regulates growth, differentiation, migration, adhesion, and apoptosis (1). TGFβ1 initiates signal transduction through two distinct serine-threonine kinase receptors, termed type I (TβRI) and type II (TβRII). TβRI is activated by TβRII upon ligand binding. The activated TβRI phosphorylates receptor-regulated Smads (R-Smad; Smad2 and Smad3), which, in turn, interact with a common mediator Smad (Smad4) and translocate to the nucleus. Nuclear Smad complexes interact with various transcription factors and transcriptional coactivators that regulate the transcription of target genes. The transcriptional target genes of TGFβ signaling vary depending on the cellular context. In nonmalignant cells and very early stages of carcinogenesis, TGFβ1 acts as an epithelial tumor suppressor through growth-inhibiting actions (2), which include activation of the cyclin-dependent kinase inhibitors p15^INK4b and p21^WAF1/CIP1 (p21). p21 mainly inhibits the activity of cyclin/cdk2 complexes and negatively regulates cell-cycle progression. In contrast, epithelial cancer cells are responsive to TGFβ1 control of malignant phenotypes that promote invasion and motility via a program similar to epithelial-to-mesenchymal transition. TGFβ1 can also promote progression and metastasis of tumors through induction of angiogenesis, and repression of the immune system during the advanced stages of cancer (3).

A number of researchers have described a protective role of TGFβ against ionizing radiation (IR), reporting that inhibition of TβRI increases radiosensitivity (4, 5) and attenuates γ-irradiation–induced ATM kinase activity (6). ATM is activated in response to DNA double-strand breaks (DSB) caused by IR and, in turn, phosphorylates numerous substrates, which activate a complex program that controls cell-cycle checkpoints, apoptosis, and genomic integrity. In addition, DNA repair processes are activated by DSBs; the two major pathways are homologous recombination (HR) and nonhomologous end joining (NHEJ; ref. 7). HR is a form of repair that requires an undamaged DNA template and functions only during DNA replication and G₂ phase, whereas NHEJ occurs in the absence of a DNA template throughout the cell cycle. NHEJ is considered the major pathway for the repair of IR-induced DSBs in human cells (8). NHEJ is generally assumed to proceed in a stepwise fashion with binding of Ku and DNA-dependent protein kinase, catalytic subunit (DNA-PKcs), followed by recruitment of the X-ray repair complementing defective repair in Chinese hamster cells 4 (XRCC4) and DNA ligase IV (Lig4) complex. The final step of NHEJ is carried out by Lig4, which joins two separate DNA strands (7).

Here, we sought to identify the cellular alterations induced by TGFβ1 that protect cells from the harmful effects of γ-irradiation. We anticipate that such a detailed understanding will provide novel insight into the relationship between TGFβ and IR-induced genotoxic stress control.

A431 (human epidermoid carcinoma), HaCaT (human keratinocyte), and 293T (human embryonic kidney cell) cell lines were routinely maintained in DMEM supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. SiHa (human cervical carcinoma) cell line was maintained in MEM with 10% FBS, 1 mmol/L sodium pyruvate, 0.1 mmol/L nonessential amino acids, 100 U/mL penicillin, and 100 μg/mL streptomycin. Cells were pretreated with the TβRI inhibitor, SB431542 (0.5 μmol/L; Tocris) for 1 hour, incubated with 0.5 ng/mL TGFβ1 (R&D Systems) for 24 hours, and irradiated (10 Gy) using a Gamma-cell 3000 (Atomic Energy of Canada Ltd.) or BIOBEAM 8000 (Gamma-Service Medical GmbH) with a [¹³⁷Cs] source.

The following antibodies were used in this study: p21 (#2946), cleaved caspase-3 (#9662), phospho-H2AX (γ-H2AX, #2577), Smad2 (#3103), and Smad3 (#9513) antibodies were purchased from Cell Signaling Technology; Lig4 (sc-271299), and β-actin (sc-47778) antibodies were purchased from Santa Cruz Biotechnology; and proliferating cell nuclear antigen (PCNA; M0879) antibody was purchased from Dako.

Cells were plated in 60 mm dishes at a density of 2 × 10⁵ cells per dish, irradiated with γ-radiation, incubated for 24 hours, and then harvested by trypsinization. Cells were resuspended in 1× binding buffer, and labeled with Apoptosis Detection Kit (BD Pharmingen). Apoptosis were quantified using a FACScan flow cytometer (BD Biosciences) fitted with CellQuest PRO software (BD Biosciences).

Cells were extracted using lysis buffer (Cell Signaling Technology). Protein concentrations were measured with the BCA Protein Assay Kit (Pierce). Equal amounts of proteins were used. Immunoreactive proteins were detected using ECL reagents (Amersham Pharmacia Biotechnology) and X-ray films (AGFA).

Transfections with siRNA were done using LipofectAMINE 2000 (Invitrogen) according to the instructions of the manufacturer. The cells were transfected with control siRNA, p21 siRNA, Smad2/3 siRNA (Santa Cruz Biotechnology), or Lig4 siRNA (Bioneer Inc.) at a final concentration of 10 nmol/L.

Cells were transfected with p21 siRNAs (Santa Cruz Biotechnology) using LipofectAMINE 2000 (Invitrogen) according to the instructions of the manufacturer. The next day, cells were incubated with TGFβ1 (0.5 ng/mL) for 24 hours and irradiated (10 Gy). After 8 hours, cells were harvested by trypsinization, and an aliquot (1 × 10⁶ cells) was washed once with ice-cold PBS, followed by fixation with 70% ethanol at 4°C overnight. The fixed cells were then washed with PBS and DNA was stained with a solution containing 20 μg/mL propidium iodide (PI) and 10 μg/mL RNase A in PBS. The cell-cycle distribution was analyzed with a FACS flow cytometer (BD Biosciences) at an excitation wavelength of 488 nm.

Three-dimensional culture was performed as described previously (9). A dermal equivalent was prepared by mixing human dermal fibroblasts (1 × 10⁵ cells/mL) from foreskin with a type I collagen gel matrix, which was reconstituted according to the manufacturer's specification (Nitta Gelatin). Thirty minutes or 7 days after γ-irradiation, three-dimensional cultures were fixed in Carnoy solution (ethanol:chloroform:acetic acid; 6:3:1) for 30 minutes at 4°C. Fixed samples were embedded in paraffin and sectioned (3 μm).

Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assays were performed using ApopTag kit (Millipore) in accordance with the manufacturer's instructions. Apoptotic bodies were visualized using the DAB reagent set (KPL).

Immunohistochemistry was performed as previously described (10). Tissue sections were deparaffinized in a xylene-to-ethanol gradient and then rinsed with PBS. The slides were placed in a domestic pressure cooker containing 1 mmol/L EDTA buffer (pH 8.0) and boiled for 15 minutes. Endogenous peroxidase was blocked by incubating the sections for 10 minutes with 3% (v/v) H₂O₂ in methanol. After rinsing with PBS, the tissues were blocked with PBS containing 20% (v/v) horse serum followed by an overnight incubation at 4°C with PCNA or Lig4 antibody. After incubating for 30 minutes with diluted biotinylated secondary antibody, immunoreactivity was detected using the Vectastain ABC Kit (Vector Laboratories). After washing with PBS, antibody binding was visualized using the DAB reagent set. The sections were counterstained with hematoxylin and examined using a light microscope.

Comet Assay Kit (Trivigen) was used according to the manufacturer's instructions. Comet images were obtained with a Nikon fluorescence microscope with an attached CCD camera. Images were analyzed using Andor Komet software (Andor). The Olive tail moment was determined with 100 cells in each sample.

Cells were plated on 12 mm, noncoated glass coverslips, and fixed with 3.7% formaldehyde in PBS at room temperature for 20 minutes. Fixed cells were permeabilized with 0.2% Triton X-100 in PBS for 10 minutes and blocked with 20% normal horse serum in PBS for 1 hour. After blocking, cells were incubated overnight with γ-H2AX antibody and then incubated for 1 hour with FITC-conjugated secondary antibodies (Vector Laboratories). Coverslips were mounted in DAPI-included Vectashield mounting medium (Vector Laboratories) and observed by laser-scanning confocal microscopy (LSM710; Carl Zeiss).

Plasmid-based NHEJ assay was performed as described previously (11). In brief, plasmid pGL2-control (Promega) was linearized using either HindIII or EcoRI. The linearized plasmid was mixed with an internal control plasmid, pRL-SV40 (Promega), and the mixture was transfected into cells. Twenty-four hours after transfection, firefly and Renilla luciferase activities were assayed with “Dual Luciferase Assay System” (Promega), and the firefly luciferase activity was normalized to the Renilla luciferase activity. Modified plasmid-based assay was also performed using the GFP vector, pEGFP-N1 (Clontech), linearized by HindIII. The linearized plasmid was mixed with an internal control plasmid, pRFP-N1 (Clontech), which yields red fluorescence. The mixture was then transfected into cells, and the transfectants were harvested 24 hours after transfection. GFP and RFP intensities were measured by fluorescence microscopy using PhotoShop CS software (Adobe).

Total RNA was isolated using TRI reagent (Molecular Research Center). Reverse transcription-PCR (RT-PCR) was performed using superscript II (Invitrogen). Primers targeted the coding region of Lig4 cDNA (upstream primer: 5′-GGC AAC TGC ATG ATC CTT CT-3′; downstream primer: 5′-GGG CTT CTC TGC TAC TGC AC-3′); GAPDH was used as a control (upstream primer: 5′-ACC AAC TAT TGC TTC AGC TC-3′, downstream primer: 5′-TTA TGC TGG TTG TAC AGG-3′). PCR cycling conditions were 25 cycles of 95°C for 2 minutes, 55°C for 1 minute, and 72°C for 1 minute. PCR products were analyzed by electrophoresis in 1.5% ethidium bromide-containing agarose gels and visualized by UV detection.

293T cells were transfected with pLig4-Luc (SwitchGear Genomics), containing Lig4 promoter region, or empty vector along with Renilla control plasmids. After transfection, cells were incubated for 24 hours in the absence or presence of TGFβ1 (0.5 ng/mL). Firefly and Renilla luciferase activities were assayed with “Dual Luciferase Assay System” (Promega), and the firefly luciferase activity was normalized to the Renilla luciferase activity.

In situ detergent extraction and immunofluorescence was performed as described previously (12). Thirty minutes after γ-irradiation (10 Gy), cells were extracted in situ by extraction buffer (50 mmol/L HEPES pH 7.5, 150 mmol/L NaCl, 1 mmol/L EDTA) containing 0.1% Triton X-100 for 2 minutes. Cells were fixed in 3.7% paraformaldehyde for 20 minutes followed by a 5-minute incubation with 0.2% Triton X-100 in PBS. Cells were blocked by incubating with PBS containing 10% FBS, 2% bovine serum albumin, 0.2% Triton X-100, and 100 μg/mL RNase A for 1 hour. Cells were then incubated overnight at 4°C with Lig4 antibody. 1:200 dilution of FITC-conjugated horse anti-mouse IgG antibody (Vector Laboratories) was used for secondary antibodies. Coverslips were mounted in DAPI-included Vectashield mounting medium (Vector Laboratories) and observed by LSM710(Carl Zeiss). FITC intensity was measured using PhotoShop CS (Adobe).

All data were analyzed using Microsoft Office Excel (Microsoft Corp.) and are presented as means ± SDs. P values (Student t test) less than 0.05 were considered significant.

The effects of TGFβ1 have been the subject of controversy, with some reports arguing that TGFβ1 promotes apoptosis and others reaching different conclusions (13). To clarify the role of TGFβ1 in the cellular response to γ-irradiation, we tested whether TGFβ1 regulates IR-induced apoptosis in cells using Annexin V-FITC (fluorescein isothiocyanate) staining. As expected, γ-irradiation increased apoptotic cell death (Fig. 1A). Pretreatment with 0.5 ng/mL TGFβ1 alone did not change the number of apoptotic cells, suggesting that TGFβ1 pretreatment did not induce apoptosis in these cells under these experimental conditions. However, TGFβ1 pretreatment greatly reduced the number of apoptotic cells induced by γ-irradiation (Fig. 1A). We also assessed the levels of cleaved caspase-3, which is an indicator of apoptotic cell death. As expected, caspase-3 cleavage was induced by γ-irradiation. This effect was markedly attenuated in TGFβ1-pretreated cells (Fig. 1B), supporting the idea that TGFβ1 pretreatment protects cells from IR-induced apoptosis. To further confirm this antiapoptotic action of TGFβ1, we examined IR-induced apoptosis in the presence of the TβRI inhibitor, SB431542. IR-induced apoptotic response was restored in the presence of SB431542 and TGFβ1 pretreatment (Fig. 1A and B), suggesting that TβRI activity is important for the suppression of IR-induced apoptosis.

Many researchers have argued that cell-cycle arrest facilitates cellular recovery from external stimuli that induce DNA damage (14). Consistent with this, slowly propagating cells are more resistant to cytotoxic stimuli than are actively propagating cells. In our experimental condition, TGFβ1 pretreatment increased G₁ population (Supplementary Fig. S1). Therefore, we hypothesized that TGFβ1 pretreatment might confer resistance to γ-irradiation because of its ability to cause cell-cycle arrest. To test this hypothesis, we evaluated the effects of siRNA-mediated p21 knockdown on IR-induced apoptosis. As expected, the knockdown of p21 eliminated the increase in the G₁ population (15), indicating that TGFβ1-induced G₁ arrest was dependent on p21 (Fig. 2A). Surprisingly, p21 knockdown had no effect on IR-induced apoptosis in the presence of TGFβ1 (Fig. 2B). We also performed the similar experiment with therapeutic dose of γ-irradiation (2 Gy, Supplementary Fig. S2), and the results were consistent. These data suggest that G₁ arrest and the protection against γ-irradiation are separate and independent processes, although TGFβ1 induced both events.

Our data indicated that TGFβ1 pretreatment suppresses IR-induced apoptosis in a conventional cell culture system (Fig. 1A and B). Next, we investigated the effect of TGFβ1 in a three-dimensional culture system, which is more analogous to the in vivo condition (9). In this system, human cervical cancer SiHa cells were cultured at the air–liquid interface on a three-dimensional matrix and then, after the indicated treatments, were examined histologically for the presence of apoptosis. Consistent with the response in conventional cell culture, TGFβ1 pretreatment suppressed IR-induced apoptosis (Fig. 3A). The effect was eliminated by concomitant SB431542 treatment.

In addition, we also investigated cell survival in a three-dimensional system. Cells were examined immunohistochemically using an antibody against PCNA, which has essential roles in DNA synthesis and thus can be used as a marker for proliferating cells (16). Consistent with the refractory nature of cancer cells to TGFβ1, unirradiated samples showed no difference in PCNA expression, regardless of TGFβ1 pretreatment. However, seven days after γ-irradiation, TGFβ1-pretreated samples showed an increase in PCNA expression relative to samples exposed to γ-irradiation alone, suggesting that TGFβ1 treatment protected cells from γ-irradiation and ultimately facilitated long-term cell proliferation (Fig. 3B). The increase in the PCNA-positive cell population is consistent with the observed suppression of apoptosis (Fig. 3A). Overall, these results indicate that TGFβ1 pretreatment protects cells from apoptosis and enhances cell proliferation.

Because TGFβ1 pretreatment decreased γ-irradiation–induced apoptosis, the effect of TGFβ1 might be mediated through the well-known canonical pathway involving R-Smad proteins. TGFβ1 specifically activates both Smad2 and Smad3. Therefore, we tested whether Smad2/3 knockdown interfered with the effect of TGFβ1. To do so, we transfected cells with Smad2/3 siRNA, then pretreated them with TGFβ1 followed by exposure to γ-irradiation. Western blot analyses showed that TGFβ1 pretreatment suppressed IR-induced apoptosis and Smad2/3 siRNA blocked the suppression (Fig. 4). We also performed the similar experiment with therapeutic dose of γ-irradiation (2 Gy, Supplementary Fig. S3), and the results were consistent. These data indicate that TGFβ1 and subsequent R-Smad activation are important for the suppression of IR-induced apoptosis.

Highly cytotoxic DSBs could be the most significant consequence of γ-irradiation (7). DSBs elicit a prototypic DNA damage response, which activates cell-cycle arrest and DNA repair processes. Our data indicated that TGFβ1 pretreatment alleviated the adverse effects of γ-irradiation in cells, suggesting the possibility that TGFβ1-treated cells had less IR-induced DNA damage or improved DNA damage recovery. To clarify the effects of TGFβ1 on DNA damage response, γ-H2AX expression level was analyzed by Western blotting after γ-irradiation in the presence or absence of TGFβ1 (Supplementary Fig. S4). The TGFβ1-induced suppression of γ-H2AX was not obvious in 10 minutes after γ-irradiation, but noticeable after 1 hour. These results suggest that TGFβ1 pretreatment does not change the IR-induced DNA damage itself, but rather improves DNA damage recognition. To further test this possibility, we measured the extent of DNA damage caused by γ-irradiation in the presence of TGFβ1 using three different methods.

We first used a single-cell gel electrophoresis technique known as the Comet assay to detect DNA strand breaks (Fig. 5A and Supplementary Fig. S5). TGFβ1 pretreatment alone did not alter the basal level of damaged DNA compared with controls. However, TGFβ1 pretreatment enhanced the restoration of DNA integrity in irradiated cells after 6 hours. Next, exploiting the fact that DSBs elicit the formation of nuclear γ-H2AX foci, we assessed the number of γ-H2AX–positive cells after γ-irradiation (Fig. 5B and C). As expected, TGFβ1-pretreated samples showed a reduced number of γ-H2AX–positive cells compared with untreated samples. Finally, the amount of γ-H2AX was measured by Western blot analyses. γ-H2AX levels were increased upon γ-irradiation, but this increase was blunted by TGFβ1 pretreatment (Fig. 5D lane 1–4 and Supplementary Fig. S4). It is clear from the Olive tail moment that the number of DSBs in both cell lines appears to increase over 24 hours, while the γ-H2AX detection indicates the opposite. It is possible that the DSB detected by the comet assay are partially due to radiation and partially due to apoptosis, while the γ-H2AX detection reflects the repair of radiation-induced DSB.

Because R-Smad activation is involved in the suppression of IR-induced apoptosis, we next tested whether the R-Smad activation is also important for the TGFβ1-induced reduction of residual DSB. Cells were transfected with Smad2/3 siRNA, treated as previously described, and analyzed by Western blotting. siRNA-mediated knockdown of Smad2/3 abolished the reduction of IR-induced γ-H2AX formation by TGFβ1 pretreatment (Fig. 5D), suggesting that the response to IR-induced DNA damage is regulated, at least in part, by the TGFβ/R-Smad canonical signaling pathway. Taken together, these results indicate that TGFβ1 pretreatment improves recognition and repair of IR-induced DSBs in a Smad-dependent way.

The TGFβ1-induced effect on IR-induced DNA damage response is consistent with the idea that TGFβ1 may directly regulate DNA repair process. Of the two major DNA repair process, HR and NHEJ, NHEJ is considered the major repair pathway for IR-induced DSBs in human cells (8). Accordingly, we focused on this pathway, adopting a previously published in vivo NHEJ assay (11) to measure NHEJ activity. The NHEJ activity detected using HindIII-digested pGL2 represents overall NHEJ activity, because the restriction site for this enzyme is located between the promoter and the coding sequence (Supplementary Fig. S6A). Thus, NHEJ activity that resulted in small deletions or insertions would still restore firefly luciferase expression. On the other hand, precise rejoining is required for exact restoration of the firefly luciferase gene in pGL2 linearized with EcoRI, which cleaves at a site located within the firefly luciferase coding sequence. As shown in Fig. 6A and B, both overall and precise NHEJ activities were increased by TGFβ1 pretreatment.

To confirm these findings, we used a modified version of the same strategy employing a GFP vector (pEGFP-N1) linearized with HindIII restriction enzyme as the repair substrate (Supplementary Fig. S6B). The results were consistent with those obtained using the luciferase reporter genes: pretreatment with TGFβ1 greatly enhanced GFP expression, suggesting that TGFβ1 pretreatment upregulates the NHEJ process (Fig. 6C). Overall, we conclude that TGFβ1 pretreatment enhances DNA repair activity in cells.

In tests of the candidates that are involved in the NHEJ process (7), we found that the amount of DNA ligase IV (Lig4) increased with TGFβ1 pretreatment. TGFβ1 pretreatment upregulated Lig4 mRNA (Fig. 7A) and Lig4 protein (Fig. 7B) level within 24 hours, although this effect disappeared after 48 hours (Fig. 7G). We also measured Lig4 promoter activity in the presence or absence of TGFβ1 with dual luciferase assay system. Treatment of TGFβ1 greatly enhanced the promoter activity of Lig4 (Fig. 7C), implying that TGFβ1 responsive element(s) is present in Lig4 promoter region. Treatment of Smad2/3 siRNA confirmed that Lig4 protein levels were dependent on the canonical TGFβ1 signaling pathway (Fig. 7D). In addition, TGFβ1 pretreatment increased Lig4 protein levels in a three-dimensional system, and the effects were blocked by cotreatment with SB431542 (Fig. 7E). Taken together, we concluded that TGFβ1 treatment enhanced Lig4 transcription and subsequent expression.

It has been reported that DNA repair proteins are mobilized to nuclear aggregates and become resistant to extraction following γ-irradiation (12). In addition, the amount of nuclear aggregates is correlated with its activation. To evaluate the formation of nuclear aggregates, we employed an in situ detergent extraction approach (12), treating cells with 0.1% Triton X-100, which removes nucleoplasmic proteins. We found that the amount of Lig4 aggregates increased with TGFβ1 pretreatment (Fig. 7F). In addition, we adopted a biochemical fractionation method (12). After fractionation with Triton X-100 at 30 minutes after exposure to 10 Gy, Western blot analyses were performed using antibodies against the major repair proteins. In the presence of TGFβ1, Lig4 was increased in the P2 fraction, which represents nuclear aggregates (Supplementary Fig. S7). XRCC4 and Ku 70/80 also tended to increase in the P2 fraction following TGFβ1 treatment. An increase in DNA repair activity can be inferred from the decreased amount of γ-H2AX in the P2 fraction (Supplementary Fig. S7). Taken together, TGFβ1 treatment enhanced nuclear retention of NHEJ proteins that could facilitate the nuclear NHEJ proteins at damaged sites, consistent with functional assays. These data provide a novel mechanism by which TGFβ1 enhances the DNA repair process.

In the previous experiments, we showed that TGFβ1 pretreatment enhanced the expression and nuclear retention of Lig4. Because Lig4 is the only DNA ligase used in NHEJ pathway (8), the knockdown of Lig4 would affect the activity of NHEJ. To assess the importance of Lig4 on the TGFβ1-induced protection, we silenced Lig4 with siRNA. Although Lig4 knockdown was not complete (Fig. 7G), the TGFβ1-induced suppression of cleaved caspase-3 was inhibited by Lig4 knockdown. These data strengthen our argument that Lig4 and NHEJ pathway plays an important role during TGFβ1-induced protection against IR.

TGFβ1 is an extracellular protein that plays a key role in initiating and integrating multiple cellular responses to various stimuli, including γ-radiation and other types of DNA-damaging agents (17). In this article, we investigated whether the IR-induced genotoxic stress program was regulated by TGFβ signaling. We concluded that NHEJ repair mechanism was facilitated by TGFβ1 and, thereby the long-term survival of irradiated cells was enhanced. These conclusions are consistent with previously published data. For example, Buckley and colleagues reported that TGFβ1 promoted survival and repair during recovery after hyperoxic injury (18), suggesting that TGFβ1 protected cells against various forms of DNA damage. Others have also reported similar results. TGFβ1 inhibition before γ-irradiation attenuated the DNA damage response, increased cell death and delayed tumor growth (4, 5), and TGFβ1 pretreatment protected Mv1Lu epithelial cells from γ-irradiation, an effect that was dependent on de novo protein synthesis (19). Moreover, high levels of TGFβ production in glioma-initiating cells in oncosphere culture provide a clonogenic benefit to this population following radiation exposure (20). In addition, as HaCaT and A431 are known to be defective in p53 pathway (21, 22), the effect of TGFβ1 is independent of p53.

Interestingly, our data indicated that p21, which is important for TGFβ1-induced G₁ arrest, was not involved in the cellular protection by TGFβ1 (Fig. 2B). This is surprising because G₁ arrest is known to protect cells against DNA-damaging agents (14). However, consistent with our data, Glick and colleagues reported that exogenous TGFβ1 could suppress genomic instability, regardless of G₁ growth arrest (23) and Kirshner and colleagues showed that the DNA damage response was compromised in both growth arrested and exponentially growing epithelial cells (6). Therefore, we could argue that the TGFβ1-induced protection was not dependent on G₁ arrest, at least in these cells under our experimental conditions. These data indicate the presence of an alternate mechanism by which TGFβ signaling protects cells from IR-induced DNA damage. Various lines of research have provided clues about this missing link between IR-induced cellular responses and TGFβ1. Kanamoto and colleagues reported that TGFβ1 was involved in DNA repair mechanisms by modulating Rad51 expression in Mv1Lu epithelial cells (24). Chiba and colleagues recently reported that TGFβ accelerated the DNA damage response and subsequently conferred radioresistance (25). Marx also suggested a DNA repair defect associated with mutations in a TGFβ receptor (26).

Here, we proposed mechanisms to account for the relationship between IR-induced cellular responses and TGFβ1 treatment, suggesting that TGFβ signaling could control the NHEJ process (at least) via two separate mechanisms. First, we demonstrated that Lig4 was upregulated by TGFβ1 treatment (Fig. 7), which is dependent on Smad2/3 (Fig. 7D). We speculate that the resulting increase in Lig4 level could ultimately facilitate the NHEJ process. Second, TGFβ1 treatment activated the NHEJ process itself. Drouet and colleagues reported that the mobilization to nucleus and resistance to extraction of NHEJ proteins are correlated with NHEJ activation (12). Our data indicated that TGFβ1 pretreatment enhanced formation of nuclear aggregates of Lig4. In addition, aggregates of Ku70/80 and XRCC4, which are mediators of NHEJ pathway, were enhanced by TGFβ1 pretreatment. These observations were most likely a reflection of their assembly onto sites of DSBs. Consistent with this speculation, TGFβ1-treated cells showed reduced γ-H2AX only noticeable in 30 minutes after γ-radiation (Figs. 5 and Supplementary Fig. S4). This delayed effect of TGFβ1 raises a possibility that the decrease in foci represent the accelerated NHEJ process in the presence of TGFβ1, as there is a close correlation between the rate of foci loss and DSB repair (27). Similarly, high levels of TGFβ1 and reduced frequency of γ-H2AX foci in glioma-initiating cells correlates with better clonogenic survival (20).

Glick and colleagues reported that the lack of TGFβ signaling causes genomic instability (23). A Tgfb1-null cell line exhibited a strong tendency toward chromosomal abnormalities and tumor progression. Maxwell and colleagues reported that inhibiting TGFβ signaling in nonmalignant human epithelial cells increased genomic instability (28). Interestingly, defect of Lig4 also causes cellular radiosensitivity, neuronal apoptosis, and genomic instability (29), bearing a strong resemblance to the defect of TGFβ signaling. On the basis of our data, we could, at least partially, explain the relationship between TGFβ signaling and genomic stability. We speculate that the lack of TGFβ signaling might interfere with NHEJ process and end up causing various chromosomal defects.

In conclusion, our study emphasized a functional link between TGFβ signaling and the IR-induced genotoxic stress program. TGFβ1 treatment facilitated the DNA repair process and subsequently enhanced long-term survival upon γ-irradiation in cells. In clinical applications using genotoxic damage, TGFβ1 at cancer site can sustain cancer cell survival after cancer therapy (4, 20); Blockade of TGFβ signaling by the TGFβR-I kinase inhibitor LY2109761 enhances radiation response and prolongs survival in glioblastoma (5). This idea is currently being tested in clinical trials for glioblastoma. Our data support that the inhibition of TGFβ has multiple targets that may improve radiotherapy. Therefore, the detailed understanding of the interaction between TGFβ1 and the genotoxic stress program could provide valuable insight into the carcinogenic process and, ultimately, improve cancer treatment options.

No potential conflicts of interest were disclosed.

Conception and design: J.Y. Yi, J. Lee

Development of methodology: J.Y. Yi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.Y. Yi, M.-R. Kim, J. Lee, E. Chung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.Y. Yi, M.-R. Kim, J. Lee, I.-C. Park, M.H. Barcellos-Hoff

Writing, review, and/or revision of the manuscript: J.Y. Yi, J. Lee, I. Shin, M.H. Barcellos-Hoff

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y.B. Jin, Y.S. An

Study supervision: J.Y. Yi

This work was supported by Basic Science Research Program (2012R1A1A2050560) through the National Research Foundation of Korea (NRF) funded by Ministry of Education.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.