Primary and Secondary Bystander Effects of Proton Microbeam Irradiation on Human Lung Cancer Cells under Hypoxic Conditions.

Tumor hypoxia is the most common feature of radioresistance to the radiotherapy (RT) of lung cancer and results in poor clinical outcomes. High-linear energy transfer (LET) radiation is a novel RT technique to overcome this problem. However, a limited number of studies have been elucidated on the underlying mechanism(s) of RIBE and RISBE in cancer cells exposed to high-LET radiation under hypoxia. Here, we developed a new method to investigate the RIBE and RISBE under hypoxia using the SPICE-QST proton microbeams and a layered tissue co-culture system. Normal lung fibroblast (WI-38) and lung cancer (A549) cells were exposed in the range of 06 Gy of proton microbeams, wherein only ~0.04-0.15% of the cells were traversed by protons. Subsequently, primary bystander A549 cells were co-cultured with secondary bystander A549 cells in the presence or absence of a GJIC and NO inhibitor using co-culture systems. Studies show that there are differences in RIBE in A549 and WI-38 primary bystander cells under normoxia and hypoxia. Interestingly, treatment with a GJIC inhibitor showed an increase in the toxicity of primary bystander WI-38 cells but a decrease in A549 cells under hypoxia. Our results also show the induction of RISBE in secondary bystander A549 cells under hypoxia, where GJIC and NO inhibitors reduced the stressful effects on secondary bystander A549 cells. Together, these preliminary results, for the first time, represented the involvement of intercellular communications through GJIC in propagation of RIBE and RISBE in hypoxic cancer cells.

The COX-2/PGE2 Response Pathway Upregulates Radioresistance in A549 Human Lung Cancer Cells through Radiation-Induced Bystander Signaling.

This study aimed to determine the mechanism underlying the modulation of radiosensitivity in cancer cells by the radiation-induced bystander effect (RIBE). We hypothesized that the RIBE mediates cyclooxygenase-2 (COX-2) and its metabolite prostaglandin E2 (PGE2) in elevating radioresistance in unirradiated cells. In this study, we used the SPICE-QST microbeam irradiation system to target 0.07-0.7% cells by 3.4-MeV proton microbeam in the cell culture sample, such that most cells in the dish became bystander cells. Twenty-four hours after irradiation, we observed COX-2 protein upregulation in microbeam-irradiated cells compared to that of controls. Additionally, 0.29% of the microbeam-irradiated cells exhibited increased cell survival and a reduced micronucleus rate against X-ray irradiation compared to that of non-microbeam irradiated cells. The radioresistance response was diminished in both cell groups with the hemichannel inhibitor and in COX-2-knockout cells under cell-to-cell contact and sparsely distributed conditions. The results indicate that the RIBE upregulates the cell radioresistance through COX-2/PGE2 intercellular responses, thereby contributing to issues, such as the risk of cancer recurrence.

Induction of DNA strand breaks and oxidative base damages in plasmid DNA by ultra-high dose rate proton irradiation.

PURPOSE: Radiation cancer therapy with ultra-high dose rate (UHDR) exposure, so-called FLASH radiotherapy, appears to reduce normal tissue damage without compromising tumor response to therapy. The aim of this study was to clarify whether a 59.5 MeV proton beam at an UHDR of 48.6 Gy/s could effectively reduce the DNA damage of pBR322 plasmid DNA in solution compared to the conventional dose rate (CONV) of 0.057 Gy/s. MATERIALS AND METHODS: A simple system, consisting of pBR322 plasmid DNA in 1× Tris-EDTA buffer, was initially employed for proton beam exposure. We then used formamidopyrimidine-DNA glycosylase (Fpg) enzymes. which convert oxidative base damages of oxidized purines to DNA strand breaks, to quantify DNA single strand breaks (SSBs) and double strand breaks (DSBs) by agarose gel electrophoresis. RESULTS: Our findings showed that the SSB induction rate (SSB per plasmid DNA/Gy) at UHDR and the induction of Fpg enzyme sensitive sites (ESS) were significantly reduced in UHDR compared to CONV. However, there was no significant difference in DSB induction and non-DSB cluster damages. CONCLUSIONS: UHDR of a 59.5 MeV proton beam could reduce non-clustered, non-DSB damages, such as SSB and sparsely distributed ESS. However, this effect may not be significant in reducing lethal DNA damage that becomes apparent only in acute radiation effects of mammalian cells and in vivo studies.

DNA strand break induction of aqueous plasmid DNA exposed to 30 MeV protons at ultra-high dose rate.

Radiation cancer therapy with ultra-high dose rate exposure, so called FLASH radiotherapy, appears to reduce normal tissue damage without compromising tumor response. The aim of this study was to clarify whether FLASH exposure of proton beam would be effective in reducing the DNA strand break induction. We applied a simple model system, pBR322 plasmid DNA in aqueous 1 × TE solution, where DNA single strand breaks (SSBs) and double strand breaks (DSBs) can be precisely quantified by gel electrophoresis. Plasmid DNA were exposed to 27.5 MeV protons in the conventional dose rate of 0.05 Gy/s (CONV) and ultra-high dose rate of 40 Gy/s (FLASH). With both dose rate, the kinetics of the SSB and DSB induction were proportional to absorbed dose. The SSB induction of FLASH was significantly less than CONV, which were 8.79 ± 0.14 (10-3 SSB per Gy per molecule) and 10.8 ± 0.68 (10-3 SSB per Gy per molecule), respectively. The DSB induction of FLASH was also slightly less than CONV, but difference was not significant. Altogether, 27.5 MeV proton beam at 40 Gy/s reduced SSB and not DSB, thus its effect may not be significant in reducing lethal DNA damage that become apparent in acute radiation effect.

Hypoxia and Proton microbeam: Role of Gap Junction Intercellular Communication in Inducing Bystander Responses on Human Lung Cancer Cells and Normal Cells.

Radiation-induced bystander effect (RIBE) has been identified as an important contributing factor to tumor resistance and normal tissue damage. However, the RIBE in cancer and normal cells under hypoxia remain unclear. In this study, confluent A549 cancer and WI-38 normal cells were subjected to condition of hypoxia or normoxia, before exposure to high-LET protons microbeam. After 6 h incubation, cells were harvested and assayed for colony formation, micronucleus formation, chromosome aberration and western blotting. Our results show that there were differences of RIBE in bystander A549 and WI-38 cells under hypoxia and normoxia. The differences were also observed in the roles of HIF-1α expression in bystander A549 and WI-38 cells under both conditions. Furthermore, inhibition of gap junction intercellular communication (GJIC) showed a decrease in toxicity of hypoxia-treated bystander A549 cells, but increased in bystander WI-38 cells. These findings clearly support that GJIC protection of bystander normal cells from toxicity while enhancing in bystander cancer cells. Together, the data show a promising strategy for high-LET radiation in designing an entire new line of drugs, either increase or restore GJIC in bystander cancer cells which in turn leads to enhancement of radiation accuracy for treatment of hypoxic tumors.

Radiation engenders converse migration and invasion in colorectal cancer cells through opposite modulation of ANXA2/AKT/GSK3ß pathway.

Radiation therapy is an effective non-surgical means to achieve local control for various solid tumors including colorectal cancer (CRC), but metastasis and recurrences after conventional radiotherapy remains a major obstacle in clinical practice, and the knowledge concerning the changes of metastatic potential after heavy ion radiation is still limited. This study investigated how radiation, including γ- and carbon ion radiation, would change the metastatic capacity of two CRC cell lines, HCT116 and DLD-1, and examined the underlying molecular mechanisms. We found that the migration and invasion was enhanced in DLD-1 cells but impaired in HCT116 cells in vitro and in vivo after radiation of γ-rays or carbons, and radiation induced epithelial mesenchymal transition (EMT) in DLD-1 cells but mesenchymal epithelial transition (MET) in HCT116 cells. The expression of snail, a key inducer of EMT, was significantly enhanced by inhibition of glycogen synthase kinase-3ß (GSK3ß) in both cell lines, suggesting the modulation of snail was alike in the two CRC cell lines. However, radiation inactivated GSK3ß through stimulating the phosphorylation of AKT and GSK3ß at Ser473 and Ser9 in DLD-1 cells respectively, but activated GSK3ß by decreasing the expression of pAKTSer473 and pGSK3ßSer9 or increasing the phosphorylation of GSK3ß at Tyr216 in HCT116 cells. Therefore, the above inverted motility changes was due to the opposite modulation of AKT/GSK3ß signaling pathway by radiation, which was further verified in other type of cancer cell lines including MCF-7, U251 and A549 cells. Moreover, it was found that annexin A2 (ANAX2) directly bound with GSK3ß and acted as a negative regulator of GSK3ß upon radiation. Knocking-down ANXA2 gene reversed the enhanced migration of the irradiated DLD-1 cells and strengthened radiation-impaired migration of HCT116 cells. Collectively, this study reveals that the change of cellular motility after radiation is independent of radiation type but is correlated with the inherent of cells.

Cytoplasmic Radiation Induced Radio-Adaptive Response in Human Lung Fibroblast WI-38 Cells.

It has been reported that in cells exposed to low-dose radiation, radio-adaptive responses can be induced which make irradiated cells refractory to subsequent high-dose irradiation. However, whether adaptive responses are possible when only the cytoplasm, not the nucleus, of the cell is exposed to radiation is still unclear. In this study, using the proton microbeam facility at the National Institute of Radiological Sciences (Japan), we found that cytoplasmic irradiation activates radio-adaptive responses in normal human lung fibroblast WI-38 cells. Our results showed that when cells received cytoplasmic irradiation with 500 protons prior to 2 Gy or 6 Gy X-ray broad-beam irradiation, the DNA double-strand break levels were significantly reduced. In contrast, at cytoplasmic irradiation with less than 100 protons, the radio-adaptive response was not detected. Moreover, the time interval between cytoplasmic irradiation and whole-cell X-ray irradiation should be longer than 6 h for the induction of adaptive responses. In addition, cytoplasmic irradiation elevated the level of cellular mitochondrial superoxide, which enhanced the phosphorylation of extracellular signal-regulated kinases 1/2 (ERK 1/2) and its mediated nuclear accumulation of nuclear factor (erythroid-derived 2)-like 2 (NRF2). This signaling pathway contributed to cytoplasmic irradiation-induced adaptive response as supported by the observations that treatment with the mitochondrial superoxide scavenger mito-tempol, ERK 1/2 inhibitor U0126 or NRF2 inhibitor ML385 could repress the adaptive response. Overall, we showed that cytoplasmic irradiation induces radio-adaptive responses and that mitochondrial superoxide/ERK 1/2/NRF2 signaling is a mechanism. Our results provide new information on the biological effects induced by cytoplasm-targeted irradiation.

Role of Endoplasmic Reticulum and Mitochondrion in Proton Microbeam Radiation-Induced Bystander Effect.

It is well known that mitochondria and the endoplasmic reticulum (ER) play important roles in radiation response, but their functions in radiation-induced bystander effect (RIBE) are largely unclear. In this study, we found that when a small portion of cells in a population of human lung fibroblast MRC-5 cells were precisely irradiated through either the nuclei or cytoplasm with counted microbeam protons, the yield of micronuclei (MN) and the levels of intracellular reactive oxygen species (ROS) in nonirradiated cells neighboring irradiated cells were significantly increased. Mito/ER-tracker staining demonstrated that the mitochondria were clearly activated after nuclear irradiation and ER mass approached a higher level after cytoplasmic irradiation. Moreover, the radiation-induced ROS was diminished by rotenone, an inhibitor of mitochondria activation, but it was not influenced by siRNA interference of BiP, an ER regulation protein. While for nuclear irradiation, rotenone-enhanced radiation-induced ER expression, and BiP siRNA eliminated radiation-induced activation of mitochondria, these phenomena were not observed for cytoplasmic irradiation. Bystander MN was reduced by rotenone but enhanced by BiP siRNA. When the cells were treated with both rotenone and BiP siRNA, the MN yield was reduced for nuclear irradiation but was enhanced for cytoplasmic irradiation. Our results suggest that the organelles of mitochondria and ER have different roles in RIBE with respect to nuclear and cytoplasmic irradiation, and the function of ER is a prerequisite for mitochondrial activation.

Emerging Role of Secondary Bystander Effects Induced by Fractionated Proton Microbeam Radiation.

Increased understanding of radiation-induced secondary bystander effect (RISBE) is relevant to radiation therapy since it likely contributes to normal tissue injury and tumor recurrence, subsequently resulting in treatment failure. In this work, we developed a simple method based on proton microbeam radiation and a transwell insert co-culture system to elucidate the RISBE between irradiated human lung cancer cells and nonirradiated human normal cells. A549 lung cancer cells received a single dose or fractionated doses of proton microbeam radiation to generate the primary bystander cells. These cells were then seeded on the top of the insert with secondary bystander WI-38 normal cells growing underneath in the presence or absence of gap junction intercellular communication (GJIC) inhibitor, 18-α-glycyrrhetnic acid (AGA). Cells were co-cultured before harvesting and assayed for micronuclei formation. The results of this work showed that fractionated doses of protons caused less DNA damage in the secondary bystander WI-38 cells compared to a single radiation dose, where the means differ by 20%. However, the damaging effect in the secondary bystander normal cells could be eliminated when treated with AGA. This novel work reflects our effort to demonstrate that GJIC plays a major role in the RISBE generated from the primary bystander cancer cells.

Radiation quality effects alteration in COX-2 pathway to trigger radiation-induced bystander response in A549 lung carcinoma cells.

This study aimed to determine whether the radiation-induced bystander effect (RIBE) is affected by radiation quality. To mimic the different radiation qualities of the direct action (D)/indirect action (ID) ratio, A549 cells were exposed to X-rays, with either 100 mM of the radical scavenger, thio-urea (TU+), or null (TU-). Biological responses in irradiated and bystander cells were compared at equal lethal effects of a 6% survival dose, which was estimated from the survival curves to be 8 Gy and 5 Gy for TU+ and TU-, respectively. Cyclooxygenase-2 (COX-2) expression in TU- irradiated cells increased up to 8 h post-irradiation, before decreasing towards 24 h. The concentration of prostaglandin E2 (PGE2), a primary product of COX-2 and known as a secreted inducible factor in RIBE, increased over 3-fold compared with that in the control at 8 h post-irradiation. Conversely, COX-2 expression and PGE2 production of TU+ irradiated cells were drastically suppressed. These results show that the larger D/ID suppressed COX-2 expression and PGE2 production in irradiated cells. However, in contrast to the case in the irradiated cells, COX-2 expression was equally observed in the TU- and TU+ co-cultured bystander cells, which showed the highest expression levels at 24 h post-irradiation. Taken together, these findings demonstrate that radiation quality, such as the D/ID ratio, may be an important factor in the alteration of signalling pathways involved in RIBE.

Equivalency of the quality of sublethal lesions after photons and high-linear energy transfer ion beams.

The quality of the sublethal damage (SLD) after irradiation with high-linear energy transfer (LET) ion beams was investigated with low-LET photons. Chinese hamster V79 cells and human squamous carcinoma SAS cells were first exposed to a priming dose of different ion beams at different LETs at the Heavy Ion Medical Accelerator in the Chiba facility. The cells were kept at room temperature and then exposed to a secondary test dose of X-rays. Based on the repair kinetics study, the surviving fraction of cells quickly increased with the repair time, and reached a plateau in 2-3 h, even when cells had received priming monoenergetic high-LET beams or spread-out Bragg peak beams as well as X-ray irradiation. The shapes of the cell survival curves from the secondary test X-rays, after repair of the damage caused by the high-LET irradiation, were similar to those obtained from cells exposed to primary X-rays only. Complete SLD repairs were observed, even when the LET of the primary ion beams was very high. These results suggest that the SLD caused by high-LET irradiation was repaired well, and likewise, the damage caused by the X-rays. In cells where the ion beam had made a direct hit in the core region in an ion track, lethal damage to the domain was produced, resulting in cell death. On the other hand, in domains that had received a glancing hit in the low-LET penumbra region, the SLD produced was completely repaired.

Enhanced DNA double-strand break repair of microbeam targeted A549 lung carcinoma cells by adjacent WI38 normal lung fibroblast cells via bi-directional signaling.

Understanding the mechanisms underlying the radiation-induced bystander effect (RIBE) and bi-directional signaling between irradiated carcinoma cells and their surrounding non-irradiated normal cells is relevant to cancer radiotherapy. The present study investigated propagation of RIBE signals between human lung carcinoma A549 cells and normal lung fibroblast WI38 cells in bystander cells, either directly or indirectly contacting irradiated A549 cells. We prepared A549-GFP/WI38 co-cultures and A549-GFP/A549 co-cultures, in which A549-GFP cells stably expressing H2BGFP were co-cultured with either A549 cells or WI38 cells, respectively. Using the SPICE-NIRS microbeam, only the A549-GFP cells were irradiated with 500 protons per cell. The level of γ-H2AX, a marker for DNA double-strand breaks (DSB), was subsequently measured for up to 24h post-irradiation in three categories of cells: (1) "targeted"/irradiated A549-GFP cells; (2) "neighboring"/non-irradiated cells directly contacting the "targeted" cells; and (3) "distant"/non-irradiated cells, which were not in direct contact with the "targeted" cells. We found that DSB repair in targeted A549-GFP cells was enhanced by co-cultured WI38 cells. The bystander response in A549-GFP/A549 cell co-cultures, as marked by γ-H2AX levels at 8h post-irradiation, showed a decrease to non-irradiated control level when approaching 24h, while the neighboring/distant bystander WI38 cells in A549-GFP/WI38 co-cultures was maintained at a similar level until 24h post-irradiation. Surprisingly, distant A549-GFP cells in A549-GFP/WI38 co-cultures showed time dependency similar to bystander WI38 cells, but not to distant cells in A549-GFP/A549 co-cultures. These observations indicate that γ-H2AX was induced in WI38 cells as a result of RIBE. WI38 cells were not only involved in rescue of targeted A549, but also in the modification of RIBE against distant A549-GFP cells. The present results demonstrate that radiation-induced bi-directional signaling had extended a profound influence on cellular sensitivity to radiation as well as the sensitivity to RIBE.

G(2)-M phase-correlative bystander effects are co-mediated by DNA-PKcs and ATM after carbon ion irradiation.

Accumulated evidence has shown that radiation-induced bystander effect (RIBE) may have significant implications to the efficiency of radiotherapy. Although cellular radiosensitivity relies on cell cycle status, it is largely unknown how about the relationship between RIBE and cell cycle distribution, much less the underlying mechanism. In the present study, the lung cancer A549 cells were synchronized into different cell cycle phases of G1, S and G2/M and irradiated with high linear energy transfer (LET) carbon ions. By treating nonirradiated cells with the conditioned medium from these irradiated cells, it was found that the G2-M phase cells had the largest contribution to RIBE. Meanwhile, the activity of DNA-PKcs but not ATM was increased in the synchronized G2-M phase cells in spite of both of them were activated in the asynchronous cells after carbon ion irradiation. When the G2-M phased cells were transferred with DNA-PKcs siRNA and ATM siRNA individually or treated with an inhibitor of either DNA-PKcs or ATM before carbon ion irradiation, the RIBE was effectively diminished. These results provide new evidence linking cell cycle to bystander responses and demonstrate that DNA-PKcs and ATM are two associated factors in co-regulating G2-M phase-related bystander effects.

Rescue of Targeted Nonstem-Like Cells from Bystander Stem-Like Cells in Human Fibrosarcoma HT1080.

Cancer stem-like cells (CSCs) have been suggested to be the principal cause of tumor radioresistance, dormancy and recurrence after radiotherapy. However, little is known about CSC behavior in response to clinical radiotherapy, particularly with regard to CSC communication with bulk cancer cells. In this study, CSCs and nonstem-like cancer cells (NSCCs) were co-cultured, and defined cell types were chosen and irradiated, respectively, with proton microbeam. The bidirectional rescue effect in the combinations of the two cell types was then investigated. The results showed that out of all four combinations, only the targeted, proton irradiated NSCCs were protected by bystander CSCs and showed less accumulation of 53BP1, which is a widely used indicator for DNA double-strand breaks. In addition, supplementation with c-PTIO, a specific nitric oxide scavenger, can show a similar effect on targeted NSCCs. These results, showed that the rescue effect of CSCs on targeted NSCCs involves nitric oxide in the process, suggesting that the cellular communication between CSCs and NSCCs may be important in determining the survival of tumor cells after radiation therapy. To our knowledge, this is the first report demonstrating a rescue effect of CSCs to irradiated NSCCs that may help us better understand CSC behavior in response to cancer radiotherapy.

Target irradiation induced bystander effects between stem-like and non stem-like cancer cells.

Tumors are heterogeneous in nature and consist of multiple cell types. Among them, cancer stem-like cells (CSCs) are suggested to be the principal cause of tumor metastasis, resistance and recurrence. Therefore, understanding the behavior of CSCs in direct and indirect irradiations is crucial for clinical radiotherapy. Here, the CSCs and their counterpart non stem-like cancer cells (NSCCs) in human HT1080 fibrosarcoma cell line were sorted and labeled, then the two cell subtypes were mixed together and chosen separately to be irradiated via a proton microbeam. The radiation-induced bystander effect (RIBE) between the CSCs and NSCCs was measured by imaging 53BP1 foci, a widely used indicator for DNA double strand break (DSB). CSCs were found to be less active than NSCCs in both the generation and the response of bystander signals. Moreover, the nitric oxide (NO) scavenger c-PTIO can effectively alleviate the bystander effect in bystander NSCCs but not in bystander CSCs, indicating a difference of the two cell subtypes in NO signal response. To our knowledge, this is the first report shedding light on the RIBE between CSCs and NSCCs, which might contribute to a further understanding of the out-of-field effect in cancer radiotherapy.

Co-visualization of DNA damage and ion traversals in live mammalian cells using a fluorescent nuclear track detector.

The geometric locations of ion traversals in mammalian cells constitute important information in the study of heavy ion-induced biological effect. Single ion traversal through a cellular nucleus produces complex and massive DNA damage at a nanometer level, leading to cell inactivation, mutations and transformation. We present a novel approach that uses a fluorescent nuclear track detector (FNTD) for the simultaneous detection of the geometrical images of ion traversals and DNA damage in single cells using confocal microscopy. HT1080 or HT1080-53BP1-GFP cells were cultured on the surface of a FNTD and exposed to 5.1-MeV/n neon ions. The positions of the ion traversals were obtained as fluorescent images of a FNTD. Localized DNA damage in cells was identified as fluorescent spots of γ-H2AX or 53BP1-GFP. These track images and images of damaged DNA were obtained in a short time using a confocal laser scanning microscope. The geometrical distribution of DNA damage indicated by fluorescent γ-H2AX spots in fixed cells or fluorescent 53BP1-GFP spots in living cells was found to correlate well with the distribution of the ion traversals. This method will be useful for evaluating the number of ion hits on individual cells, not only for micro-beam but also for random-beam experiments.

Damaging and protective bystander cross-talk between human lung cancer and normal cells after proton microbeam irradiation.

Most of the studies of radiation-induced bystander effects (RIBE) have been focused on understanding the radiobiological changes observed in bystander cells in response to the signals from irradiated cells in a normal cell population with implications to radiation risk assessment. However, reports on RIBE with relevance to cancer radiotherapy especially investigating the bidirectional and criss-cross bystander communications between cancer and normal cells are limited. Hence, in present study employing co-culture approach, we have investigated the bystander cross-talk between lung cancer (A549) and normal (WI38) cells after proton-microbeam irradiation using γ-H2AX foci fluorescence as a measure of DNA double-strand breaks (DSBs). We observed that in A549-A549 co-cultures, irradiated A549 cells exert damaging effects in bystander A549 cells, which were found to be mediated through gap junctional intercellular communication (GJIC). However, in A549-WI38 co-cultures, irradiated A549 did not affect bystander WI38 cells. Rather, bystander WI38 cells induced inverse protective signalling (rescue effect) in irradiated A549 cells, which was independent of GJIC. On the other hand, in response to irradiated WI38 cells neither of the bystander cells (A549 or WI38) showed significant increase in γ-H2AX foci. The observed bystander signalling between tumour and normal cells may have potential implications in therapeutic outcome of cancer radiotherapy.

Differential effects of p53 on bystander phenotypes induced by gamma ray and high LET heavy ion radiation.

High LET particle irradiation has several potential advantages over γ-rays such as p53-independent response. The purpose of this work is to disclose the effect of p53 on the bystander effect induced by different LET irradiations and underlying mechanism. Lymphocyte cells of TK6 (wild type p53) and HMy2.CIR (mutated p53) were exposed to either low or high LET irradiation, then their mitochondrial dysfunction and ROS generation were detected. The micronuclei (MN) induction in HL-7702 hepatocytes co-cultured with irradiated lymphocytes was also measured. It was found that the mitochondrial dysfunction, p66(Shc) activation, and intracellular ROS were enhanced in TK6 but not in HMy2.CIR cells after γ-ray irradiation, but all of them were increased in both cell lines after carbon and iron irradiation. Consistently, the bystander effect of MN formation in HL-7702 cells was only triggered by γ-irradiated TK6 cells but not by γ-irradiated HMy2.CIR cells. But this bystander effect was induced by both lymphocyte cell lines after heavy ion irradiation. PFT-µ, an inhibitor of p53, only partly inhibited ROS generation and bystander effect induced by 30 keV/µm carbon-irradiated TK6 cells but failed to suppress the bystander effect induced by the TK6 cells irradiated with either 70 keV/µm carbon or 180 keV/µm iron. The mitochondrial inhibitors of rotenone and oligomycin eliminated heavy ion induced ROS generation in TK6 and HMy2.CIR cells and hence diminished the bystander effect on HL-7702 cells. These results clearly demonstrate that the bystander effect is p53-dependent for low LET irradiation, but it is p53-independent for high LET irradiation which may be because of p53-independent ROS generation due to mitochondrial dysfunction.