Radiotherapy is a part of the standard treatment for more than 50% of cancer patients and has a documented contribution to local tumor control and improved overall survival. Clinical radiotherapy aims to achieve maximal tumor control rates while reducing the risk for dose-limiting adverse late effects in normal tissues [1
]. Depending on the disease stage, radiotherapy is therefore either given alone or as a part of multimodal combinations with surgery, chemotherapy and molecular-targeted drug therapy or immunotherapy [1
]. Overcoming dose-limiting toxicity to normal tissues is still a major challenge in clinical radiotherapy, particularly when tumors grow adjacent to critical structures or within tissues or organs with pronounced radiation sensitivity. Enhancing the accuracy of dose delivery to the tumor volume, e.g., by stereotactic radiotherapy or intensity-modulated radiation therapy, has allowed clinicians to improve the safety profile of radiotherapy for many solid tumors [4
]. Moreover, recent technical developments have led to broader application of particle therapy in clinical practice. It is expected that the favorable depth-dose curves and linear energy transfer (LET) characteristics of charged particles will allow a precise targeting of deep-seated tumors while reducing the ionizing dose and irradiated volume of normal tissue. It is expected that these attempts to achieve a precise and more target-specific irradiation will help to lower the toxicity rates and the risk of developing secondary tumors [5
]. In fact, particles pass through normal tissues on their track without losing much energy but instead releasing the main carried energy at a specific tissue depth shortly before their complete stop. This maximized dose deposition at the end of the particle’s range is represented as a Bragg peak of the depth-dose curve. In contrast, photons lose their energy exponentially, with higher values at the entry point and lower values at the deep-seated tissues [12
]. Additionally, the rising number of particle therapy centers worldwide facilitates the use of protons or heavy ions in the clinical care of cancer patients and allows the validation of the suspected superiority of particle beams compared to photon beams in terms of normal tissue protection [15
To allow for irradiation of a three-dimensional tumor volume, therapeutic proton beam therapy is performed using a so called spread-out Bragg peak (SOBP), which is composed of several single Bragg peaks with slightly varying energies [17
]. Despite divergent physical properties of proton and photon beams, it is assumed that the relative biological effectiveness (RBE) of proton beams resemble those of photon beams, mostly when put in relation to high LET particle beams such as carbon ions [18
]. As a consequence, treatment planning for proton beam radiotherapy has been developed based on data originating from therapy with gamma-ray photons generated by a 60
Co source, using a correction factor for RBE of 1.1 [18
]. Herein, RBE is defined as the ratio of biological effectiveness of two radiation modalities, measured by absorbed dose for a given effect (reference irradiation type/irradiation type under investigation), which is inversely related to a given dose (Equation (3) in Material and Methods) [13
]. Generally, the higher the deposited energy, the higher the density of ionizing events, and the higher the resulting RBE per unit of dose, as defined by more severe DNA damage [17
However, it is increasingly appreciated that not only the physical characteristics of the beam but also the microscopic pattern of energy deposition differs between photons and protons, particularly at the distal edge of the Bragg-peak, with a potential impact on the resulting biological effects [14
]. In fact, depending on the tissue, the measured endpoint, dose and LET of the beam, the RBE values reported for protons vary between ~ 1.1–1.7, with increasing RBE values for protons along the distal edge of the Bragg peak [18
]. Moreover, first in vitro studies implicated that cancer-associated genetic defects in DNA repair—homologous recombination repair (HRR) or the Fanconi Anemia (FA) pathway—are associated with an increase in the RBE values for irradiation with proton beams compared to irradiation with gamma-ray photons or X-ray photons [28
]. These observations point to potential differences in the biology of DNA damage induced by irradiation with proton beams compared to irradiation with gamma-ray photons or X-ray photons. Moreover, these findings suggest that genetic alterations affecting DNA damage response (DDR) and DNA repair pathways may not only contribute to heterogeneity in cancer cell radiosensitivity per se, but might also cause variations in proton RBE [22
Independent of the radiation quality, the cytotoxic effects of ionizing radiation are based on its ability to cause damage to cellular macromolecules, particularly the DNA. Herein, DNA double strand breaks (DSB) are considered to be the most toxic lesions induced by ionizing radiation so that the capacity of cells to repair radiation-induced DNA DSB constitutes a major determinant of cellular radiosensitivity [32
]. Radiation-induced DNA DSB are mainly repaired by non-homologous end joining (NHEJ) and HRR [34
]. However, if both pathways are disrupted, alternative end joining (alt-EJ) can be activated [35
]. NHEJ is a cell cycle independent and rapid, but error-prone DNA repair pathway that is mediated by the DNA-PK (DNA-dependent protein kinase) complex consisting of the KU70/80 heterodimer and a catalytic subunit (DNA-PKcs) [34
]. DNA-PKcs is essential for the recruitment of repair proteins of the NHEJ complex, including XRCC4 (X-ray repair cross-complementing protein 4), Lig IV (DNA ligase IV), the nuclease Artemis as well as the stabilizing factors XLF (XRCC4-like factor) and PAXX (paralog of XRCC4 and XLF) [37
]. In contrast, HRR is a precise, but slow and cell cycle-dependent DNA repair mechanism requiring homology of a sister chromatid [34
]. HRR efficiency relies on numerous proteins and protein complexes, e.g., BRCA2 (breast cancer 2) and Rad54. BRCA2 is a tumor suppressor and one of the most essential proteins regulating HRR; it promotes binding of Rad51 to single-stranded DNA during DNA damage processing [43
]. Rad54 induces DNA synthesis and promotes dissociation of Rad51 from the DNA strand, ensuring annealing of the synthesized DNA to the ends of the damaged DNA strand [44
]. In this context, 53BP1 (p53 binding protein 1) functions as a sensor of DNA damage and has been proposed to mediate DNA repair pathway choice by promoting NHEJ [44
Though an increasing number of reports point to different patterns of DNA damage induced by photon and proton irradiation with potential relevance for DNA repair [14
] there is a gap of knowledge about potential microscopic differences in the DNA damage pattern induced by protons from the entrance plateau (EP) versus protons from SOBP. Moreover, only limited data are available on variations in RBE caused by genetic deficiencies in components of the two major DNA DSB repair pathways, NHEJ and HRR, which are also observed in human cancer. Such investigations are particularly important in view of the increasing interest in combining DNA repair pathway inhibitors with radiotherapy for inducing synthetic lethality in tumors with intrinsic or acquired DNA repair-deficiencies [47
In the present study, we used STED (stimulated emission depletion) microscopy to compare the microscopic pattern of DNA damage induced by irradiation with equal physical doses of SOBP protons, EP protons and X-ray photons. Moreover, we used fibroblasts and cancer cells without and with deficiencies in specific proteins of NHEJ, HRR or both, to explore potential differences in RBE upon irradiation with a similar physical dose of protons (SOBP, EP) and X-ray photons for the endpoints clonogenic cell survival and DNA DSB repair kinetics. We observed the induction of more clustered γH2A.X foci in cells exposed to irradiation with SOBP protons. Moreover, we noticed a significant increase in cell death in fibroblasts and cancer cells with deficiency in a single core protein of NHEJ upon irradiation with both protons and photons. Instead, MEFs with deficiency in a single core protein of HRR displayed enhanced radiosensitivity to proton irradiation but not to X-ray photons when compared to the respective HRR-proficient MEF cells.
Further comprehensive work in additional repair-deficient cell lines and patient-derived cancer cell lines are required to define signaling molecules specifically enhancing RBE values upon irradiation with proton beams. Such investigations are needed to provide a scientific basis for the identification of patients that might particularly benefit from proton beam radiotherapy according to the molecular characteristics of their tumors for the definition of combinatorial treatments suited to harness the full potential of radiotherapy with SOBP protons whilst avoiding increased toxicity.
While the physical characteristics of photon irradiation have been investigated in much detail during recent decades, potential specificities in biological effects of proton irradiation are less well understood. Dose-deposition profiles of therapeutic proton beams are characterized by a low-dose plateau at small depths EP and the SOBP. In the SOBP, the major part of energy is deposited shortly before the sharp distal energy fall-off. First, by using STED we reveal here that irradiation with X-ray photons resulted in homogenously distributed small DNA damage foci, whereas SOBP proton irradiation induced clusters of several smaller γH2A.X foci in closer proximity that we termed “foci clusters”. Interestingly, EP proton irradiation caused both clustered and homogenously distributed small γH2A.X foci within the nucleus of MEFs. Second, we demonstrated that loss of important components of HRR has a more severe impact on DNA repair kinetics and survival of cancer cells and MEFs exposed to a similar physical radiation dose of mid SOBP or EP protons than upon irradiation with X-ray photons. We speculate that the suggested increasing importance of repair by HRR in cancer cells irradiated with SOBP protons compared to X-ray photon irradiation might be due to the induction of more clustered DNA lesions composed of multiple DNA damage sites in close proximity, as demonstrated in MEFs by visualizing γH2A.X foci using STED microscopy. Yet further mechanistic investigations and investigations in patient-derived cells are needed to verify that cancer cells exposed to proton irradiation rely more on the integrity of the HRR pathway.
In more detail, we first hypothesized that irradiation with protons (SOBP, EP) and X-ray photons may induce distinct microscopic patterns of DNA damage because the physical characteristics of proton beams include a higher LET and a deposition of more dose per path length than gamma-ray or X-ray photons [63
]. Indeed, we observed that irradiation with SOBP protons induced γH2A.X DNA foci-clusters, whereas DNA damage foci induced by X-ray photons were smaller and more randomly distributed over the whole nucleus. These observations are consistent with the described action of particles, including protons, that induce a direct DNA damage within their defined track and therefore tend to induce accumulated DNA damage in closer proximity [13
]. It has been previously suggested that the high energy of particles, e.g., protons, strongly correlates with clustered DNA damage in the form of coalesced DNA DSB due to denser ionizing events [6
]. Moreover, a prediction analysis done by a new computational track structure model which simulates complexity of DNA damage after proton irradiation, endorses our assumption of a more complex DNA damage induced by proton irradiation [67
]. Interestingly, we observed that EP proton irradiation caused both clustered and randomly distributed DNA damage. Therefore, we speculate that high energy protons of the EP region deposit their energy more randomly causing less ionizing events than the ‘precise’ SOBP protons and result therefore in lower probability to hit the cell nucleus and to induce only clustered DNA lesions [68
]. Our findings support assumptions proposed by others that differences in dose deposition between photons and protons induce different DNA lesions. Thereby, more complex DNA damage induced by protons may result in less efficient repair and more effective eradication of clonogenic cells [14
]. Furthermore, the time-dependent increase in distance between γH2A.X foci observed in our study suggests that the initial clusters are removed over time leaving smaller but longer-lasting foci [65
Since X-ray photons, SOBP and EP protons induced different patterns of DNA damage, we further determined radiosensitivity of wildtype MEFs to all three radiation modalities and observed higher RBE values for 10% of clonogenic survival upon SOBP proton than upon EP proton irradiation. We therefore conclude that the effects of EP on survival are more similar to X-ray photons, at least as long as the MEFs have an intact DNA repair machinery (NHEJ, HRR). These observations corroborate findings obtained by others describing higher RBE values in a zebrafish model at mid SOBP compared to EP proton irradiation [71
However, when analyzing MEFs harboring deficiencies in specific DNA repair proteins associated with NHEJ (Lig IV−/−
), HRR (Rad54−/−
) or both (Rad54−/−
), we observed divergent results. In fact, MEFs deficient in Lig IV turned out to be highly radiosensitive, independent of the radiation quality used. The high radiosensitivity of Lig IV-deficient MEFs was associated with high levels of residual DNA damage foci at 24 h after irradiation upon both photon and proton irradiation. These observations underline the importance of NHEJ for the DNA DSBs repair induced by both photon and proton irradiation. In contrast, Rad54−/−
MEFs responded only to proton irradiation with increased rates of clonogenic cell death, whereas they were similarly sensitive to photon irradiation as wild type MEFs. This is consistent with observations from others reporting that in cells with intact NHEJ a Rad54-deficiency is not associated with a detectable defect in DNA DSBs repair induced by high dose irradiation of X-ray photons [72
]. The authors concluded that cells harboring NHEJ defects repair the majority of X-ray photon irradiation-induced DNA DSBs by using the slower alt-EJ, which is suppressed by NHEJ and mostly not impaired by mutations in HRR [72
]. Nevertheless, for both cell lines, Lig IV−/−
, the RBE values determined for SOBP proton irradiation were higher than for EP, as observed in wildtype MEFs.
Interestingly, co-depletion of proteins involved in both NHEJ and HRR (Rad54−/−
) rendered MEFs more sensitive to X-ray photon irradiation than a single depletion of Rad54−/−
. Interestingly, at a higher dose of 8 Gy, no significant differences in radiation sensitivity were observed between LigIV−/−
single knockout and Rad54−/−
double knockout MEFs. Our results are consistent with previously published data [72
], though higher doses of X-ray photon irradiation (10 Gy, 20 Gy) has been used in this study. Moreover, the authors concluded that Rad54-dependent HRR does not facilitate repair of radiation-induced DNA DSBs in NHEJ-deficient cells in the G2 cell cycle phase [72
]. Instead, we observed a similarly reduced survival of Rad54−/−
MEFs and Rad54−/−
MEFs after proton irradiation and a better survival than LigIV−/−
MEFs. At present we cannot explain this phenomenon. We believe that the discrepancy observed in the response of the MEFs between photon and proton irradiation in the knockout strains is mainly due to the more pronounced sensitivity of Rad54−/−
cells to the cytotoxic effects to proton irradiation, which is also reflected by the higher levels of residual γH2A.X foci observed in Rad54−/−
cells only after proton irradiation.
Interestingly, MRN-complex-dependent end resection of DNA DSBs is necessary for both HRR and alt-EJ [74
]. One might thus speculate that in cells with a severe NHEJ defect a competition between HRR and alt-EJ repair might delay DNA DSBs repair, and thereby further contribute to the higher amounts of residual DNA damage foci, loss of genomic stability and higher radisensitivity of Lig IV−/−
cells. HRR is a cell cycle-dependent pathway, whereas alt-EJ operates regardless of the cell cycle. In cells with NHEJ deficiency, alt-EJ may become more prominent even if an end resection has occurred, so that additional HRR defects will not further compromise radiosensitivity at least upon X-ray photon irradiation [72
]. Instead, the altered biology of the DNA damage observed upon proton irradiation may cause a higher dependency on HRR [28
]. Herein, others revealed that clustered DNA damage induces chromatin destabilization resulting in exclusion of HRR as a possible DNA repair pathway and strongly increasing the contribution and importance of alt-EJ to DSB repair [76
]. Following this, cells that failed in HRR used alt-EJ as a backup for DNA DSBs repair [77
]. Yet, further mechanistic work is required to reveal if activation of alt-EJ as an adaptation process to inactivation of NHEJ or both major DSBs repair pathways [35
] or other mechanisms contribute to the above findings.
Instead, neither loss of XLF nor loss of PAXX had a significant impact on survival of fibroblasts after any type of irradiation, pointing to a redundant function of those NHEJ factors. We speculate that only the loss of key factors of the NHEJ complex, such as Lig IV, XRCC4 or DNA-PKcs, has a significant impact on clonogenic survival. However, XLF-deficiency affected the proper DNA repair kinetics after SOBP proton irradiation, pointing to its role in DNA repair and presumably in stabilizing the NHEJ-complex, but not in long-term survival. We speculate that loss of XLF can be compensated for and is therefore not essential for cells to survive.
To corroborate our findings on the impact of DNA repair-deficiencies on radiation response in cancer cell models, we used the DNA-PKcs-deficient M059J and DNA-PKcs-proficient M059K glioblastoma cell lines as well as the pancreatic cell lines BxPC3 with intact BRCA2 and Capan-1 with reported deficiency in BRCA2 expression. Similar to fibroblasts with impaired NHEJ due to loss of Lig IV, DNA-PKcs-deficient glioblastoma cells were more sensitive than DNA-PKcs-proficient M059K cells to irradiation with both X-ray photons and protons (EP and SOBP). However, calculated RBE values for 10% of clonogenic survival for EP and SOBP were approx. 1, which has already been observed by others [28
]. These findings reveal that deficiency in NHEJ caused by the loss of DNA-PKcs already causes a dramatic radiosensitization to X-ray photon irradiation that overrides potential small differences in survival caused by the distinct biology of the DNA damage induced by X-ray photon and proton irradiation, respectively.
Though BRCA2-proficient BxPC3 and BRCA2-deficient Capan-1 pancreatic cancer cell lines were both more sensitive to irradiation with SOBP protons compared to irradiation with X-ray photons, the increase in the cytotoxic effects of proton irradiation was more pronounced in Capan-1 pancreatic cancer cells. However, BxPC3 and Capan-1 cells are not isogenic but constitute distinct cell lines and thus have a different genetic background [52
]. Therefore, it is highly likely that the differences in the radiation response to photon and proton irradiation observed in BxPC3 and Capan-1 cells may be due to further, yet unknown, variations in other proteins or pathways involved in the regulation or execution of DNA DSB repair by NHEJ, HRR, or both [52
]. Thus, mechanistic investigations with matched control cell lines, e.g., Capan-1 cells and Capan-1 cells with reconstituted BRCA2, or down-regulation of BRCA2 in HRR-proficient cancer cells, will be necessary to investigate functional relevance of BRCA2-deficiency for the observed radiosensitizing effect towards proton irradiation.
Despite the severe impact of DNA-PKcs-deficiency on DNA repair kinetics in M059J cells upon proton and photon irradiation, X-ray photons had the highest impact on the number of the induction and removal of γH2A.X foci, as revealed by increased number of initial γH2A.X foci at 4 h and 8 h after irradiation. Moreover, deficiency in DNA-PKcs resulted in more pronounced damage persistence, as revealed by enhanced residual DNA damage foci in M059J cells 24 h upon irradiation with X-ray photons compared to SOBP protons. Our observations corroborate the major role of NHEJ for survival of irradiated cancer cells irrespective of the radiation quality, thereby confirming previous observations by others [46
]. The importance of NHEJ in ensuring cell survival upon both proton and photon irradiation also corresponds to the reported involvement of NHEJ in the repair of approx. 80% of all DNA DSB occurring upon X-ray photon irradiation [46
Taken together, while fibroblasts and cancer cells deficient in NHEJ (Lig IV−/−
) responded to proton and X-ray photon irradiation with a comparable reduction of clonogenic cell survival, the increase in residual DNA damage was higher after X-ray photon irradiation. We speculate that the repair of smaller DNA lesions induced by X-ray photons will mainly rely on NHEJ, so that cells with a deficiency in proper NHEJ will accumulate residual DNA damage [28
]. In contrast, a delay in DNA repair kinetics of HRR-deficient cell lines was only observed after SOBP proton irradiation underlining the increased importance of HRR for the repair of clustered DNA damage induced by irradiation with SOBP protons [14
]. However, the delay was less pronounced than the delay observed in NHEJ-deficient cells after X-ray photon irradiation. This suggests that NHEJ is also involved in repair of clustered DNA lesions induced by SOBP proton irradiation and that HRR plays an additional and supportive role in processing the induced DNA DSB, as suggested by others [46
Generally, the cellular genetic make-up impacts radiosensitivity resulting in uncertainties in RBE, which are estimated to be 10–20% [19
]. Furthermore, the tissue-dependent uncertainty in RBE is another limiting factor in tissue-related RBE estimation [80
A better understanding of the characteristics and consequences of DNA damage induced by proton irradiation in normal tissues and tumors, and potential different molecular requirements for repair of DNA damage induced by proton irradiation is needed to harness the full potential of proton irradiation for clinical radiotherapy by combining proton radiotherapy with chemotherapy, or any other therapy in the future. Further joint efforts of the research community and properly consolidated basic research data are indispensable in order to define predictive markers allowing a patient stratification for proton radiotherapy based on molecular markers in the future.