Mechanisms governing inducible resistance to ionizing radiation in untransformed epithelial cells pre-exposed to low-dose ionizing radiation (LDIR; ≤10 cGy) are not well understood. The present study provides evidence that pre-exposure to 10 cGy X-rays increases clonogenic survival of mouse skin JB6P+ epithelial cells subsequently exposed to 2 Gy doses of γ-rays. To elucidate the molecular pathways of LDIR-induced adaptive radioresistance, the transcription factor nuclear factor-κB (NF-κB) and a group of NF-κB–related proteins [i.e., p65, manganese superoxide dismutase (MnSOD), phosphorylated extracellular signal-regulated kinase, cyclin B1, and 14-3-3ζ] were identified to be activated as early as 15 min after LDIR. Further analysis revealed that a substantial amount of both 14-3-3ζ and cyclin B1 accumulated in the cytoplasm at 4 to 8 h when cell survival was enhanced. The nuclear 14-3-3ζ and cyclin B1 were reduced and increased at 4 and 24 h, respectively, after LDIR. Using YFP-fusion gene expression vectors, interaction between 14-3-3ζ and cyclin B1 was visualized in living cells, and LDIR enhanced the nuclear translocation of the 14-3-3ζ/cyclin B1 complex. Treatment of JB6P+ cells with the NF-κB inhibitor IMD-0354 suppressed LDIR-induced expression of MnSOD, 14-3-3ζ, and cyclin B1 and diminished the adaptive radioresistance. In addition, treatment with small interfering RNA against mouse MnSOD was shown to inhibit the development of LDIR-induced radioresistance. Together, these results show that NF-κB, MnSOD, 14-3-3ζ, and cyclin B1 contribute to LDIR-induced adaptive radioresistance in mouse skin epithelial cells. [Cancer Res 2007;67(7):3220–8]
Mammals are continuously exposed to low doses of ionizing radiation (LDIR; ≤10 cGy) that may contribute to biological responses that can affect upon disease processes. Recent data strongly imply that pre-exposure to LDIR is able to activate specific proteins that may increase cellular tolerance to subsequent radiation injuries (1). Adaptive resistance to radiation and/or chemotherapeutic agents is also induced in several tumor or transformed cell lines (2, 3). An enhanced cell survival rate in cells treated with fractionated ionizing radiation (FIR) has also been observed (4). We have reported an adaptive radiation resistance in human breast cancer MCF-7 cells after a long-term exposure to FIR (2 Gy per fraction, for total 30 fractions; ref. 5). All of the information suggests that specific molecular pathways are required for the development of an adaptive response. Mammalian cells have shown the ability to induce adaptive responses to ionizing radiation via activation of different cellular and molecular mechanisms leading to genomic instability (6) as well as cell cycle arrest and radioresistance (7, 8). DNA microarray profiles have revealed that irradiated mammalian cells can activate many stress-responsive genes, including many known genes that encode proteins for the regulation of cell cycle, cell growth, and DNA repair (9, 10). The specific signaling network that influence overall cell survival in adaptive radioresistance have not been clearly delineated, but nuclear factor-κB (NF-κB) is believed to be an active participant in this process. Elucidating the molecular mechanisms involved in the NF-κB–induced signaling networks that contribute to LDIR-induced radioresistance may provide insights for understanding LDIR-mediated biological responses.
NF-κB can be quickly activated by any stress that induces the phosphorylation and proteolysis of IκB or by IκB-independent pathway (11, 12). Stress-induced NF-κB is able to antagonize apoptotic pathways (13) and increases cell survival causing cellular resistance to radiation and chemotherapy (14). Ionizing radiation–induced NF-κB activity is associated with the enhanced radiation resistance in human breast carcinoma MCF-7 cells (5) and the papillomavirus-transformed human keratinocytes (15). Irradiated HeLa cells show a transient NF-κB activation following IκB phosphorylation (16). Interestingly, NF-κB is also linked with the pathways of mitochondrial stress resulting the activation of calcineurin and the mitochondrial antioxidant enzyme manganese superoxide dismutase (MnSOD; refs. 17, 18). Gene cluster analysis further shows that NF-κB is responsible for at least a fraction of the gene expression profile induced by ionizing radiation, including MnSOD, indicating that mitochondrial redox imbalance may be an important factor contributing to ionizing radiation–induced adaptive responses (5). Blocking ionizing radiation–induced NF-κB activation inhibits the expression of MnSOD (5) and other NF-κB-associated proteins and enhances ionizing radiation–mediated cell death (5, 15). These observations support the hypothesis that NF-κB and MnSOD co-operate in ionizing radiation–induced adaptive responses.
Two key stress signaling elements (14-3-3ζ and cyclin B1) are also closely related to NF-κB activation and MnSOD expression (5). The 14-3-3 chaperon protein family with seven members is involved in a variety of cellular functions, including the phosphoserine-binding ability that inhibits proapoptotic partner proteins, and the ATP-dependent targeting of precursor proteins to mitochondria (19). Ras can also function as a radiation resistance molecule by activating 14-3-3 (20), and the activated Ras further inhibits the activity of protein kinase Raf-1 (21) and KSR-1 (22), both of which seem to be essential for signaling apoptotic cell death. Further evidence indicates that transfection of cultured fibroblasts with the R56A and R60A double mutant forms of 14-3-3ζ increases the activation of pro-apoptotic c-Jun NH2-terminal kinase 1 (JNK1) and p38 mitogen-activated protein kinase (MAPK; ref. 23). Until recently, 14-3-3 proteins were thought to be present only in the cytoplasm and on the plasma membrane. It has now been shown that 14-3-3 proteins are not only found in the nucleus (24), where they serve as transcriptional regulators, but also function in protein import into mitochondria to regulate ATP synthesis (25). Furthermore, it has recently been shown that 14-3-3 proteins may play a role in modulating the activity of uncoupling factors at the mitochondrial inner membrane (26). Most interestingly, both 14-3-3 proteins and cyclins (cyclin B1) are inhibited by genistein (4,5,7-trihydroxyisoflavone), an antitumor agent that induces cell cycle arrest and apoptosis in different cell lines (27). These results highlight the tight relationship between 14-3-3 proteins and cyclin B1 in the protection of cells against radiation-induced cytotoxicity.
Cyclin B1 is a rate-limiting component for the transition of cells going from G2 to M phase (28, 29). When cyclin B1 is induced, G2 delay is decreased (28), which is related to the phenotypic alternations in radioresistant MCF-7 cells (30). If cyclin B1 is inhibited by antisense oligonucleotides directed against cyclin B1, cell radiosensitivity is increased (30, 31). Most interestingly, clinical data suggest that head and neck tumors treated by radiotherapy show enhanced expression of cyclin B1 with an decreased radiosensitivity and increased local recurrence (32). Despite the evidence that cyclin B1 is important in signaling tumor-adaptive radioresistance, the role of cyclin B1 in LDIR-induced adaptive radioprotection in normal cells remains to be defined.
Epithelial cell lines obtained from mouse and human skin origin have been extensively used as surrogate in vitro models for studies relevant to oxidative stress. However, LDIR-induced adaptive responses have not been well described in epithelial cells. The mouse JB6P+ cells, showing a tumor promotion sensitive phenotype, have been well characterized in studying the molecular mechanisms of skin carcinogenesis (33, 34). Here, we report a novel finding of LDIR-induced adaptive radioprotection in JB6P+ cells. We also find that the NF-κB–associated signaling network is linked with the activation of the mitochondrial antioxidant enzyme MnSOD as well as 14-3-3ζ and cyclin B1 in the LDIR-induced adaptive radioprotection in these epithelial cells. These results identify molecular pathways of LDIR-induced radioresistance that may be relevant to studies of normal epithelial tissue responses in biologically relevant environmental and clinical exposures to LDIR.
Materials and Methods
Cell culture and radiation. JB6P+ mouse epidermal cell line [Cl 41, promotable by 12-O-tetradecanoylphorbol-13-acetate (TPA) treatment] were maintained in Eagle's MEM (EMEM; Cellgro, Herndon, VA) supplemented with 2 mmol/L l-glutamine and 5% fetal bovine serum under standard culture conditions. Cells were subcultured by trypsinization, and all experiments were done within 20 to 60 passages. Exponentially growing cells (∼80% confluence) in six-well plates were irradiated at ∼0.028 Gy/min with total doses of 5 or 10 cGy X-ray using a Cabinet X-ray System Faxitron Series (Hewlett-Packard, McMinnville, OR) at room temperature. For study on radiation-induced adaptive response, exponentially growing JB6P+ cells were exposed to 60Co-γ irradiation with GR-12 irradiator (dose rate: 2.3 Gy/min; U.S. Nuclear Corp., Burbank, CA) 6 h after pre-irradiation with 10 cGy X-ray. Cells shielded from the ionizing radiation source were used as the sham-IR controls.
Expression vectors. Human cyclin B1 open reading frames (ORF) were amplified by reverse transcription-PCR (RT-PCR) using the forward primer cyclin B1-FXbaI (5′-GCTCTAGAATGGCGCTCCGAGTCACCAGGA-3′) and the reverse primer cyclin B1-RXbaI (5′-GCTCTAGACACCTTTGCCACAGCCTTGGCT-3′). The amplified DNA products were digested with the restriction enzyme XbaI and ligated individually into the pFlag-CMV-YN155 vector to produce the pFlag-CMV-cyclin B1-YN155 expression plasmids. Human 14-3-3ζ ORF was also amplified by RT-PCR using the forward primer 14-3-3ζ-FEcoRI (5′-GCGAATTCGGATGGATAAAAATGAGCTGGTT-3′) and the reverse primer 14-3-3ζ-REcoRI (5′-GCGAATTCGATTTTCCCCTCCTTCTCCTGC-3′). The amplified DNA products were digested with the restriction enzyme EcoRI and ligated into the pHA-YC155 vector to produce the pHA-14-3-3ζ-YC155 expression plasmid.
Anchorage-independent growth assay. To determine TPA-induced cell transformation, exponentially growing JB6P+ cells were washed thrice with PBS, trypsinized, diluted, and suspended in 1.5 mL of 0.4% agar in EMEM over 7 mL of 0.5% agar EMEM containing with or without TPA (3 ng/mL). The cultures were maintained in 37°C incubator for 14 days, and a computerized image analyzer scored anchorage-independent colonies >25 cells.
Preparation of cytoplasmic and nuclear extracts. JB6P+ cells were rinsed with PBS containing 1 mmol/L EDTA, collected by centrifugation, and resuspended in ice-cold hypotonic lysis buffer supplemented with protease inhibitors. Following 15 min of incubation on ice, the cell lysates were vortexed and centrifuged at 15,000 × g for 1 min to obtain the nuclear pellet and the cytosolic protein-containing supernatant. The nuclear pellets were washed using the washing buffer and then resuspended in the ice-cold extraction buffer. After 30 min of incubation on ice, the samples were centrifuged at 15,000 × g for 5 min, and the resulting supernatant was saved as nuclear extracts and stored in the −80°C freezer until use for gel shift or Western analysis.
Electrophoresis mobility shift analysis. Gel shift analysis was done as previously described (35).
Reporter transfection and luciferase assay. To measure the basal and radiation-induced (10 cGy X-rays) transactivation activity of transcription factors NF-κB and activator protein-1 (AP-1), reporter luciferase assays were done, as described previously (5).
3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide and clonogenic survival assays. JB6P+ cells were seeded into 60-mm diameter cell culture plates and 96-well plate for clonogenic and 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays, respectively, as described previously (35).
Immunoblotting and antibodies. Immunoblotting was done as described previously (35). The antibodies for NF-κB (p65) and MnSOD were obtained from Upstate Biotechnology (Lake Placid, NY). The antibodies 14-3-3ζ (sc-1019), extracellular signal-regulated kinase 2 (ERK2; sc-154), phosphorylated ERK1/2 (pERK1/2; sc-7383), and cyclin B1 (sc-752) were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The antibodies β-actin (A5441), histone H3 (ab1791), and Pan-Ras (OP38) were received from Sigma (St. Louis, MO), Abcam (Cambridge, MA), and Calbiochem (San Diego, CA), respectively.
MnSOD enzymatic activity assay. JB6P+ cells (∼80% confluent) in 10-cm dishes were irradiated with 0, 5, or 10 cGy of X-rays. Cells were scrape harvested at 4°C after 4, 8, and 24 h radiation and lysed in 50 mmol/L potassium phosphate buffer containing 1.34 mmol/L diethylenetriaminepentaacetic acid, and protein concentration was determined. MnSOD activity was determined as previously described (36).
Imaging of 14-3-3ζ/cyclin B1 in living JB6P+ cells. Plasmids with full-length sequence encoding human cyclin B1 and 14-3-3ζ were fused to NH2- (1-155) and COOH-terminal (156-238) fragments of EYFP, respectively, with linker sequences described as before (37). Cotransfection of pHA-CMV-14-3-3ζ-YC156 and pFlag-CMV-cyclin B1-YN155 were done as previously detailed (5). Briefly, cells in 24-well plates were cotransfected with 0.4 μg of each expression vector using LipofectAMINE 2000 reagent (Invitrogen, Carlsbad, CA). The fluorescence emissions were clearly observed in living cells 16 h after transfection using a Nikon TE300 inverted fluorescence microscope with a cooled CCD camera.
IMD-0354 treatment. IMD-0354 (1 μmol/L; an efficient IKKβ inhibitor; Sigma; ref. 38) that inhibits activation of NF-κB via IκB-α phosphorylation was used to specifically block NF-κB activation. JB6P+ cells grown in complete medium with or without IMD-0354 were exposed to 2 Gy ionizing radiation with or without pre-irradiation (10 cGy X-ray). NF-κB activity and radiation sensitivity were measured by luciferase assay and MTT and clonogenic survival assays, respectively. Western blotting was done to measure the inhibited expression of NF-κB effector genes (e.g., MnSOD, 14-3-3ζ, and cyclin B1).
Small interfering RNA fragment construction, transient transfection, and immunostaining. MnSOD small interfering RNA (siRNA; sense, 5- AAGAATAAGGCCTGTTGTTCCCCTGTCTC -3; antisense, 5-AAGGAACAACAGGCCTTATTCCCTGTCTC-3) targeted to the mRNA for mouse MnSOD was designed with the program on Ambion web site, and then duplex oligonucleotides were synthesized using Silencer siRNA Construction kit (Abcam). A scrambled siRNA was used as an siRNA transfection control. JB6P+ cells (30% confluent) were transfected with siRNAs (100 nmol/L) using LipofectAMINE 2000 reagent following the manufacturer's instructions, which is specific for siRNA. The cells were exposed to 10 cGy X-ray at 6 h before a high dose of 2 Gy ionizing radiation. For immunostaining, JB6P+ cells transfected with MnSOD siRNA or scrambled siRNA were fixed in ice-cold 3.7% formaldehyde for 20 min followed by permeabilization in ice-cold 0.2% Triton X-100 for 5 min. Cells then were incubated with anti-MnSOD antibody for 1 h followed by three washes, and the incubation with the secondary antibody conjugated with Texas Red (Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 h.
Test of JB6P+ cells as a suitable model for assessing adaptive responses. The AP-1 and NF-κB transcription factors have been found to be activated by UV and ionizing radiation in human and mouse cells (35, 39). We sought to determine whether NF-κB and/or AP-1 are differently induced in JB6P+ cells in response to LDIR. Supplementary Fig. S1A shows that the tumor promoter TPA was capable of inducing a significant increase in anchorage-independent growth in JB6P+ cells. The basal (sham-LDIR) and 10 cGy X-ray–induced AP-1 reporter activity measured in JB6P+/AP-1-Luci stable transfectants (40) shows that no significant AP-1 activation was induced (Supplementary Fig. S1B). In contrast with this result, gel shift analysis of JB6P+ cells treated with known inducers of anchorage-independent transformation (i.e., TPA or IL-6) shows an enhanced DNA-binding activity of both NF-κB and AP-1 (Supplementary Fig. S1C). These results suggest that JB6P+ cells provide an appropriate normal cell model system for the test of LDIR-induced adaptive responses.
LDIR activates NF-κB and induces an adaptive radioprotection in JB6P+ cells. In Fig. 1A, the basal (sham-LDIR) and LDIR-induced NF-κB reporter activity measured in JB6P+/NF-κB-Luci stable transfectants (40) shows that NF-κB is activated by exposure to a single dose of 10 cGy X-ray. This result indicates that NF-κB could play a role in signaling adaptive responses induced by exposure to LDIR. To determine if adaptive radioresistance to the challenge high dose of ionizing radiation (2 Gy) could be induced by prior exposure to LDIR, cell proliferation, clonogenic cell survival, and apoptosis assays were done in JB6P+ cells exposed to 10 cGy X-rays followed 6 h later by exposure to 2 Gy γ-rays. Pre-exposure to 10 cGy X-ray clearly induced a significant cellular resistance to the cytotoxicity of 2 Gy (Fig. 1B and C; Supplementary Fig. S2). These results support the conclusion that prior exposure of immortalized normal mouse skin epithelial cells to LDIR activates NF-κB and renders cells resistance to subsequent cytotoxic effect of ionizing radiation.
Induction of NF-κB–related stress-responsive proteins. A group of stress responsive proteins as well as adaptive radioresistance have been found to be associated with the coactivation of NF-κB and MnSOD in human breast cancer cells (5, 30). Figure 2A shows a significant induction of 14-3-3ζ immunoreactive protein in JB6P+ cells 2 to 8 h following a single exposure to 5 or 10 cGy X-rays. Cyclin B1 immunoreactive protein was also found to be increased as early as 15 min following 5 or 10 cGy irradiation (Fig. 2B) and persisted 8 h (5 cGy) to 24 h (10 cGy) after radiation. These results indicate an important correlation between the accumulation of 14-3-3ζ and cyclin B1 in the same timeframe as adaptive radioresistance is being induced by LDIR in JB6P+ cells. To determine if NF-κB transactivation following LDIR was accompanied by the increase of major subunit of NF-κB (p65), immunoblotting for p65 was accomplished following exposure to 5 and 10 cGy of X-rays (Fig. 2C). Compared with the control cells with sham (0 cGy) LDIR, a clear increase of p65 was observed at 15 min, and this level was slightly decreased 24 h after LDIR. Because the MAP/ERK kinase (MEK)/ERK pathway has been shown to be activated by ionizing radiation, and because ERK activity is tightly linked with NF-κB regulation (35), ERK phosphorylation (pERK) was measured. Figure 2D shows a two-wave phosphorylation of ERK (i.e., increases at 15 min and 8 h after either dose of LDIR). Therefore, the results in Fig. 2C and D indicate that expression of NF-κB subunit p65 and ERK phosphorylation are sensitive to LDIR. The mechanism causing the two-wave–like activation of ERK is not clear but may be indicative of metabolic changes associated with shifts in oxidative metabolism following exposure to LDIR.
Induction of MnSOD by LDIR. The mitochondrial superoxide scavenging enzyme MnSOD is believed to play a key role in adaptive responses to oxidative stress, including fractionated exposure to ionizing radiation in breast cancer cells (5, 8). Figure 2E shows that as early as 15 min following exposure to 5 or 10 cGy X-ray, MnSOD immunoreactive protein levels were increased, and this increase persisted at least 24 h after LDIR. Figure 2F shows that MnSOD enzymatic activity was increased 2- to 2.5-fold 4 to 24 h after LDIR compared with control (sham-LDIR) cells. These results clearly show that exposure to LDIR induced the accumulation of MnSOD immunoreactive protein and activity during the same timeframe as NF-κB activation and adaptive radioresistance was being induced. These results support that NF-κB activation, and increases in mitochondrial antioxidant MnSOD activity may contribute to the adaptive radioresistance induced by exposure to LDIR.
Induction of p65, pERK, 14-3-3ζ, cyclin B1, and MnSOD by chronic exposure of JB6P+ cells to 0.5 cGy X-ray. To determine if chronic exposure to LDIR was capable of altering steady-state levels of p65, pERK, 14-3-3ζ, cyclin B1, and MnSOD proteins, exponentially growing JB6P+ cells were irradiated (or sham-irradiated) daily with 0.5 cGy X-ray for 0, 10, 20, and 30 fractions (total doses of 5, 10, and 15 cGy, respectively). The sham-LDIR control JB6P+ cells were maintained with the same passage numbers as their irradiated counterparts. The Western blot analysis in Fig. 3 shows the enhanced expression of p65, pERK, 14-3-3ζ, cyclin B1, and MnSOD proteins in cells exposed to 30 fractions of 0.5 cGy (total dose, 15 cGy). MnSOD, p65, pERK, and 14-3-3ζ (but not cyclin B1) expressions were also increased by 20 fractions of 0.5 cGy (total dose, 10 cGy). In contrast, cells irradiated with 10 fractions of 0.5 cGy (total dose, 5 cGy) showed increases in steady-state levels of p65, pERK, and 14-3-3ζ but not MnSOD and cyclin B1. Total ERK levels were not changed in any LDIR group, but pERK seemed to be increased in a manner dependent on the number of fractions and total dose. These results indicate that chronic LDIR exposure activates the same set of signaling pathways that are activated by single LDIR exposures, supporting the hypothesis that p65, MnSOD, pERK, 14-3-3ζ, and cyclin B1 may be components of a coordinated adaptive response.
Intracellular translocation of 14-3-3ζ and cyclin B1 after LDIR. It has been hypothesized that in response to radiation, the intracellular localization of cell cycle proteins may be an important factor determining cellular responses. To examine the dynamic translocation of 14-3-3ζ and cyclin B1 in response to LDIR, cytoplasmic and nuclear proteins were isolated from JB6P+ cells irradiated with 10 cGy X-ray. Figure 4A shows that a substantial amount of immunoreactive protein corresponding to 14-3-3ζ and cyclin B1 increased in the cytoplasm 4 h after LDIR, with an obvious reduction of 14-3-3ζ in the nucleus (no significant reduction of cyclin B1 was observed in the nucleus). Cytoplasmic accumulation of both proteins started to decrease at 8 h and reached to the levels of control cells 24 h after LDIR. However, an increase of both proteins was detected in the nucleus 24 h after LDIR. Based on the well-defined function of 14-3-3 chaperon proteins in subcellular localization of their partner proteins, Fig. 4A indicates an active role of 14-3-3ζ in translocation of cyclin B1 between the nucleus and cytoplasm following LDIR.
Bimolecular fluorescence complementation analysis provides for the imaging of direct protein interactions in living cells (37). Next, we determined whether direct interactions between 14-3-3ζ and cyclin B1 occur following LDIR. Fusion protein vectors encoding 14-3-3ζ-YC156 and cyclin B1-YN155 were cotransfected in JB6P+ cells. Cells were then exposed to LDIR (10 cGy) or sham-LDIR 6 h after transfection, and protein interactions were monitored in living cells using fluorescence microscopy. Direct interactions of 14-3-3ζ and cyclin B1 were detected in living JB6P+ cells without LDIR (Fig. 4B). Interestingly, nuclear cotranslocation of 14-3-3ζ/cyclin B1 complex was not increased 1 h after LDIR but, compared with the sham-LDIR control, was significantly increased 24 h following LDIR (Fig. 4C). These results reveal a novel interaction between 14-3-3ζ and cyclin B1 in living JB6P+ cells. The enhancement of nuclear cotranslocation of the 14-3-3ζ/cyclin B1 complex by LDIR again suggests that the interaction may contribute to the signal communication required for LDIR-induced adaptive responses. The detailed mechanism underlying 14-3-3ζ/cyclin B1 complex formation, especially their reentering to the nucleus after LDIR, will be investigated in future studies.
Inhibition of NF-κB reduced the expression of MnSOD, 14-3-3ζ, and cyclin B1 and diminished LDIR-induced adaptive resistance. To determine if a causal relationship existed between the induction of NF-κB–associated proteins and the adaptive radioresistance in response to LDIR, an efficient inhibitor of NF-κB activation (IMD-0354; ref. 38) was used (Fig. 5). Figure 5A shows that IMD-0354 inhibited NF-κB–driven reporter activity in LDIR-treated JB6P+ cells, and Fig. 5B and C shows that using the MTT and clonogenic cell survival assays, respectively, IMD-0354 was capable of inhibiting adaptive radioresistance in cells exposed to LDIR. Figure 5D shows that IMD-0354 was capable of inhibiting the accumulation of immunoreactive MnSOD, 14-3-3ζ, and cyclin B1 following exposure to LDIR. These results support the conclusion that NF-κB–mediated expression of MnSOD, 14-3-3ζ, and cyclin B1 are causally related to LDIR-induced adaptive radioresistance in JB6P+ cells.
siRNA-mediated down regulation of MnSOD diminished LDIR-induced adaptive resistance. To determine the potential key role of MnSOD in adaptive radioresistance in response to LDIR, JB6P+ cells were transfected with either MnSOD siRNA or a scrambled negative control siRNA and then exposed to 10 cGy X-ray 6 h before a dose of 2 Gy 48 h after siRNA transfection. A significant siRNA-mediated suppression of MnSOD protein level was observed in JB6P+ cells by Western blotting using anti-MnSOD antibody (Fig. 6A) and immunocytochemistry (Fig. 6B,, top). Figure 6B, (bottom) shows the inhibited cell growth in MnSOD siRNA but not the control or scrambled siRNA-treated cells. As shown in Fig. 6C, cell radiosensitivity measured by clonogenic survival assay, transfection with MnSOD siRNA but not scrambled siRNA totally abolished LDIR-induced increases in cell survival following 2 Gy. These data provide strong evidence indicating that MnSOD function is required for LDIR-induced adaptive radioprotection in epithelial cells, as was found previously for resistance of cells to oxygen toxicity (41).
There is limited information on the molecular response for adaptive radiation protection after exposure of epithelial cells to LDIR (≤10 cGy). Determining the LDIR-induced signaling elements capable of defending cells against subsequent cytotoxic effects of radiation will provide mechanistic insights into the cellular response that cannot necessarily be extrapolated from high-dose radiation studies. Using mouse skin epithelial JB6P+ cell model irradiated with single or long-term multiple low doses of X-rays, this study suggests that LDIR is able to induce an adaptive radioprotection in normal cells via the activation of NF-κB and MnSOD that could involve the chaperon 14-3-3ζ and cyclin B1. The current study also reveals a novel interaction and nuclear cotranslocation of 14-3-3ζ and cyclin B1 in LDIR-exposed cells.
Ionizing radiation is well documented to induce DNA double-strand breaks, protein oxidation and lipid peroxidation with an enhanced genomic instability (42). The fate of irradiated cells is believed to be dependent on activation of the signaling network leading to different types of cell death (e.g., apoptosis and necrosis) or survival via pro-survival pathways (43). Clones isolated from high-dose irradiated GM10115 human-hamster hybrid cells showed a high chromosome instability and a broader distribution of radiosensitivity detected by the ability of DNA double-strand break rejoining repair efficiency (44). On the other hand, exposure to the low-dose radiation from X-ray has been shown to offer cell resistance to radiation-induced genomic instability (45), indicating a divergence in cellular response to high and low doses of ionizing radiation. Although the exact signaling network of LDIR-induced radioprotection is unclear, several transcription factors and radiation responsive proteins have been found to be activated in cells after LDIR (1). It has long been observed that neoplastic transformation can be inhibited by exposure to a small priming dose of radiation. For instance, chronic exposure of plateau-phase C3H 10T1/2 cells to γ-radiation at doses as low as 10 cGy protected the cells against neoplastic transformation by a subsequent large acute radiation exposure (46). These results suggest that a single low dose of LDIR, at background or occupational exposure levels, may reduce rather than increase cancer risk. This phenomenon, termed an adaptive response to ionizing radiation, is believed to depend on protein synthesis (47). Significant adaptive responses to radiation are also observed in several mammalian cells exposed to fractionated ionizing radiation (3). We have previously shown that radioresistance was induced in breast carcinoma MCF-7 cells (30) and in virus-transformed human skin keratinocytes (15) following chronic exposure to the therapeutic doses of ionizing radiation. Most importantly, the basal NF-κB activity was elevated in the high-dose radiation–induced radioresistant tumor cells (5). In transformed human lymphoblast cell line, activation of NF-κB has been reported after exposure to 10 cGy along with induction of other early responsive genes (i.e., Fos, Jun, Myc, and H-ras) in the range of 25 cGy to 2 Gy (48). The present study reveals that NF-κB but not AP-1 is a sensitive factor to as low as 5 cGy X-ray in normal mouse skin epithelial JB6P+ cells. The skin is the largest organ of the body and, because of its location (i.e., a protective covering of the whole body), highly susceptible to occupational doses of radiation. There is very little data on LDIR-induced adaptive responses in skin tissue. Here, we showed for the first time that a group of stress-sensitive genes (i.e., NF-κB, MnSOD, pERK, 14-3-3ζ, and cyclin B1) are activated by LDIR with 5 or 10 cGy X-rays. The induction of these NF-κB–regulated stress-responsive proteins is correlated with the LDIR-induced adaptive radioresistance.
Among the three types of MAPKs (JNK, p38, and ERK), JNK and p38 kinase function to promote apoptosis, whereas the activation of ERK tends to be antiapoptotic. The pro-survival function of ERK1/2 is shown by the fact that inhibition of ERK signaling leads to increased sensitivity of ovarian cancer cell lines to cisplatin-induced apoptosis (49). However, we have observed an inhibited ERK activity in radioresistant breast cancer cells derived from long-term therapeutic ionizing radiation (50), which contrasts to the ERK activation by a single high dose (35), or single or multiple doses of LDIR observed in the present study. These diverse cellular responses may be the result of differences between tumor and normal cells. The results reported by Suzuki et al. (51) have shown that ionizing radiation in the low dose range (i.e., between 2 and 5 cGy) is able to stimulate the proliferation of both normal human diploid cells and tumor cells. The present study shows a two-wave–like activation of ERK1/2 (Fig. 2D) in mouse skin cells following LDIR, suggesting that the activation of the MEK/ERK may contribute to the NF-κB–mediated pro-survival pathway. Further studies looking at the role of ERK activation induced by LDIR and its role in LDIR-induced adaptive responses are currently under way.
An important feature in cellular responses to ionizing radiation is the generation of redox imbalances, and many signaling proteins are sensitive to redox alterations (52). The mitochondrial antioxidant enzyme MnSOD is actively involved in redox regulation (53). Several lines of investigation have shown that NF-κB is also activated by radiation via redox imbalances (54). Therefore, it is highly possible that LDIR-mediated NF-κB and MnSOD activation is due to redox imbalance. Our previous data suggest that a group of radiation-responsive elements, induced by high dose chronic ionizing radiation, including Ku86, 14-3-3ζ, and cyclin B1, are up-regulated following radiation due to activation of MnSOD (5). In addition, the presence of free radical scavengers during radiation is shown to reduce the incidence of radiation-induced genomic instability (55). Using MnSOD-knockout mice (Sod2−/−), MnSOD has been shown to protect against reactive oxygen species (ROS)–induced injury during O2 metabolism (56). A link between NF-κB activation and MnSOD expression is further evidenced by the fact that MnSOD is controlled by NF-κB in the antiapoptotic response to tumor necrosis factor-α (TNF-α; ref. 57). Overexpression of MnSOD can also restore cell resistance to TNF-α–induced cell death. NF-κB binding sites have been identified in the regulatory regions of the SOD2 gene that encodes MnSOD, and NF-κB is actively involved in regulation of MnSOD expression. Notably, both proapoptotic and antiapoptotic pathways can be activated by NF-κB via expression of MnSOD (57). In addition, we have previously shown that MnSOD is induced by high-dose ionizing radiation via NF-κB activation, and blocking NF-κB or MnSOD expression down-regulated a group of ionizing radiation–induced genes, resulting increased cell radiosensitivity (5). In the present study, MnSOD enzymatic activity was enhanced 2- to 2.5-fold by exposures of 5 or 10 cGy LDIR. Surprisingly, MnSOD protein level was greatly increased within 15 min of LDIR treatment. This indicates that mouse skin cells are very sensitive to induction of MnSOD by LDIR and suggests that MnSOD may play an important role in LDIR-induced adaptive radioresistance. This was confirmed in the experiments shown in Fig. 6 using siRNA to MnSOD. Overall, the results of the current study lead us to hypothesize that NF-κB–mediated induction of MnSOD contributes to the scavenging of radiation-induced increases in the metabolic production of ROS and significantly contribute to the signaling network leading to LDIR-induced adaptive radioresistance in mouse skin epithelial cells.
Activation of 14-3-3ζ and cyclin B1 have been suggested to be regulated by MnSOD expression in radioresistant cells (5). The current results with mouse skin epithelial cells suggests a coordinately regulated molecular response involving 14-3-3ζ and cyclin B1 may contribute to LDIR-induced adaptive radioprotection. The 14-3-3 family proteins are actively involved in a wide range of cellular processes, including apoptosis, checkpoint activation, and cell cycle progression. The ζ isoform of the 14-3-3 protein is believed to play a major role in regulating apoptosis. 14-3-3ζ inhibits the proapoptosis effect of apoptosis signal-regulating kinase 1 (ASK1) through a mechanism that may involve disruption of the interaction between ASK1 and its death effectors. In addition, interaction between 14-3-3ζ and the pro-survival phosphatidylinositol 3-kinase stimulates its kinase activity (58). Hence, 14-3-3ζ is believed to act to reduce sensitivity to apoptosis by interaction with several different important regulators of apoptosis. In this study, we observed that induction of 14-3-3ζ and cyclin B1 is tightly correlated to MnSOD expression in low-dose irradiated JB6P+ cells. Using molecular imaging technology, we found the physical interaction of 14-3-3ζ and cyclin B1 visualized in living cells. The 14-3-3ζ/cyclin B1 complex showed an enhanced nuclear translocation after LDIR. These results suggest that both 14-3-3ζ and cyclin B1 are required to communicate the essential signals between cytoplasm and nucleus after LDIR.
In summary, we report that exposure to LDIR (5 and 10 cGy X-rays) induced an adaptive radiation protection in mouse skin epithelial JB6P+ cells. The transcription factor NF-κB, the mitochondrial antioxidant MnSOD, ERK, cyclin B1, and chaperon protein 14-3-3ζ were also activated by LDIR. 14-3-3ζ and cyclin B1 formed a complex, and LDIR enhanced the nuclear translocation of this complex. Inhibition of NF-κB diminished LDIR-induced adaptive radioresistance with the reduction of LDIR-induced MnSOD, cyclin B1, and 14-3-3ζ. In addition, inhibition of MnSOD by siRNA reversed LDIR-induced adaptive protection. Overall, the present results reveal a NF-κB/MnSOD–mediated pro-survival network in mouse skin epithelial cells induced by exposure to LDIR. These results may have significant implications to the biological responses seen in mammalian cells following exposure to LDIR in the context of environmental and diagnostic/therapeutic radiation.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
M. Fan and K.M. Ahmed contributed equally to this work.
Grant support: NIH grant RO1 101990 (J.J. Li) and Department of Energy grants DE-FG02-03ER63634 (J.J. Li) and DE-FG02-05ER64050 (D.R. Spitz).
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.
We thank Dr. Nancy H. Colburn (National Cancer Institute, NIH) for providing JB6P+ cells and data discussion and Dr. C-D. Hu (Purdue University School of Pharmacy) for the gifts of pHA-CMV-YC156 and pFlag-CMV-YN155 plasmids.