New insights in the relative radiobiological effectiveness of proton irradiation

Clinical practice assumes a fixed proton RBE of 1.1, but it has been postulated that higher RBEs occur at the distal edge of proton spread out Bragg peak (SOBP). However, apart from the advantages offered by depth-dose profile of protons, they also show an enhanced biological effectiveness in cell killing [7]. This is related to the increased LET compared to X-rays when protons are close to the Bragg peak. Therefore, the use of ion beams for radiation therapy is currently undergoing investigation at different institutions. In 2008 a meeting on RBE in ion beam therapy dealt primarily with a review of experimental measuring of RBE and approaches to the clinical use of the concept of RBE based on experimental findings, theoretical models and previous clinical experience with protons and heavy ions [14]. Although the physical aspects of proton beam radiobiology are well understood, the biological aspects, particularly the complex biological endpoints need further attention. The current estimates of RBE depend on the cell type and also on the detection methods because it has been shown that DNA damage and apoptotic responses vary greatly between gamma radiation and proton therapy in a tissue- and dose-dependent fashion [15]. Experimental data emerging from recent studies suggest that, for several endpoints of clinical relevance, the biological response is differentially modulated by protons compared to photons. However, up to date only few studies have been performed to understand the differential response on the molecular and cellular levels between proton and photon irradiation. Several studies reported an increased induction of double strand breaks (DSBs) and more complex DNA damage induced by protons in comparison to photon irradiation [16, 17]. DNA DSB induction by different radiation qualities shows that, even though similar patterns of initial induced DSBs are produced by photons and protons, there are differences when looking at the rejoining process [18]. Another study demonstrated that lesions induced by proton irradiation were preferentially repaired by homologous recombination, a much slower repair mechanism than Non-Homologous End Joining, which could be attributed to the increased complexity after proton irradiation [19]. This also affects the number of residual lesions measured late after irradiation. Another study found differences between photon and proton irradiation reactive oxygen species dependent mechanism by which proton radiation induces DNA damage and cell apoptosis [4]. In the study of Di Pietro et al., lower percentage of apoptotic cells was found after photon irradiation and apoptosis was induced in a temporally delayed fashion compared to protons [20]. The study of Manti et al., showed increased amounts of complex chromosomal aberrations as well as increased frequency of sister chromatid exchanges after proton irradiation [21]. The study of Green et al., found that micronuclei formation and apoptosis induction were higher in thyroid follicular cells after proton irradiation compared to photon irradiation [22]. Also different epigenetic changes where reported after proton and photon irradiation. Exposure to X-rays was associated with hypo-methylation, while proton irradiation produced mainly hyper-methylated DNA, both in normal and cancer cells [23]. For the gold standard on the cellular level, the colony formation assay, many in vitro studies were published up to now. Using the colony formation assay an average RBE of 1.1–1.2 can be associated to the middle of the SOBP [6, 7, 24, 25]. A lower level of migration and a reduced invasion potential has been reported after proton irradiation in comparison to X-rays [11]. Interestingly, protons show anti-invasive and anti-migration behavior. The studies of Girdhani et al., showed lower levels of migration and invasion after proton irradiation in comparison to X-rays [26, 27]. Unfortunately, there are still no randomized trials available for second cancer induction in patients treated with proton vs. photon radiation. There are only very few studies which suggest that the rate of second cancer induction is less than 50% after proton irradiation compared to photon radiation [28].

In recent years, modeling of RBE as a function of LET receives much attention in the proton therapy community [29]. However, these LET-RBE parametrizations are ion type specific and their application is restricted by large uncertainties associated with the biological input parameters from proton experiments [29]. The RBE is defined as the ratio of a dose of sparsely ionizing radiation, mostly photons to a dose of any other radiation quality to produce the same biological effect. High LET radiation has an increased biological effectiveness compared to photons of low LET. Carbon or oxygen ions offer a higher RBE due to the severe radiation damage produced within the beam track. However, data on in-vitro RBE evaluation of high-LET irradiations are still sparse. Recently, our group reported RBE-datasets for carbon and oxygen ion and examined the effect of additional anti-tumorigenic substances [30‐33]. The main reason for an increased biological effectiveness is the clustered damages to the DNA structure within one nucleus, which is more difficult for the cell to repair and which leads to increased cell killing [34]. As a result, the RBE varies spatially within the patient and increases toward the distal end of a SOBP, as LET values increases with the depth of the beam [35]. It is known that the RBE is highly dependent on both cell type and the studied endpoint but also on particle species, due to the different dose deposition profiles on microscopic scale [36]. The study of Rorvik et al., developed linear as well as and non-linear RBE models for protons by applying the LET spectrum as a parameter for the radiation quality [35]. The study demonstrated that non-linear models give a better representation of the RBE-LET relationship for protons compared to linear models. Therefore, the LET is not sufficient as a predicting factor of RBE. In general, the RBE depends on the microdose distribution formed by a single ion track and the areal ion track density determining the total dose. Due to the complex RBE dependency, biophysical models are essential for the estimation of clinically relevant RBE values in treatment planning [37]. There are some approaches to model radiobiological endpoints based directly on the microdose distribution [38‐40] the three-dimensional dose distribution with nanometer resolution deposited by a single particle. An important biophysical prediction model that is currently implemented in the treatment panning systems for the heavy ion radiotherapy in Europe is the local effect model (LEM) [37, 41]. This model is used to predict the RBE for cell killing in order to correct the physical dose required for the tumor irradiation with heavy ions. According to the latest version of the LEM (LEM IV) [42, 43] the spatial DNA DSB distribution and their local density within a cell nucleus are assumed to be the most relevant factors that influence the cell fate following radiation.

It is known that the energy deposition for high LET radiation is much more inhomogeneous in time and space than that of low LET radiation [44]. The energy deposition of a single ion hit into a biological cell runs on the femtosecond to picoseconds time scale, while the spatial dose distribution peaks at the center of the ion track [45]. It was shown already in the 70ies and 80ies of the last century that spatial distributions of energy deposition events and the resulting DSB distributions do affect the outcome as shown using spatially correlated ions which were produced from diatomic ions [46, 47]. Recently, the influence of spatial dose distribution on the RBE with respect to different biological endpoints has been investigated using an experimental approach where low LET 20 MeV protons (LET = 2.65 keV/m) were focused to sub-micrometer spots in cell nuclei [44, 45, 48]. Here, the authors reported on an enhanced RBE with regard to induction of dicentric chromosomes and micronuclei in hybrid human-hamster A_L cells after spot application of a bunch of 20 MeV protons compared to a quasi-homogeneous irradiation [45]. In another manuscript A_L cells have been irradiated with 20 MeV (2.6 keV/m) protons quasi-homogeneously distributed or focused to 0.5 × 1 μm² spots on regular matrix patterns (point distances up to 10.6 × 10.6 μm), with pre-defined particle numbers per spot to provide the same mean dose of 1.7 Gy [44]. The yields of dicentrics and their distribution among cells have been scored. The yields of dicentric chromosomes increased by focusing up to a factor of 2 for protons compared to quasi-homogeneous irradiation (Fig. 1). The local density of DNA DSBs increased at the irradiated spots enhancing also the probability for the interaction of the DSBs and thus increasing the probability of connecting the wrong ends. The reported study improved the understanding of the mechanisms by which radiation induces these lethal chromosome aberrations [44].

Furthermore, variation of the spatial DSB distribution within a cell nucleus by focusing low LET protons resulted in a higher cell killing compared to quasi homogeneous proton application [48]. These results indicate that the sub-micrometer proton focusing, which affects the DSB distribution within the cell nucleus leads to decreased cell survival [44, 48]. Thus significant variations in RBE can be expected if low LET protons are applied in a spatially correlated manner. Moreover, these results strongly support the assumption of the LEM model that the spatial DNA damage distribution is the source of relative biological effectiveness [45].

In recent years, the fixed RBE value of 1.1 is being questioned with respect to its safety, because if the dose to the tumor is too low, the risk of tumor recurrence increases. On the other hand, if the dose is too high, the chances for acute and last side effects will increase. Disregarding this RBE and LET variations could have negative clinical implications, especially when an organ at risk is located near the distal end of a tumor [35]. A fixed RBE during fractionated exposures disregards any effects due to the variation of dose per fraction and the total number of fractions delivered in relation to the LET. However, a number of recent in vitro studies have reported that the RBE within the SOBP is not constant and the RBE increases at the distal end of the SOBP. Table 1 summarizes these in vitro studies. The study of Britten et al., demonstrated that the RBE of the proton beam at certain depths is greater than 1.1 and therefore there is an increased potential for cell killing and normal tissue damage in the distal regions of the Bragg peak [10]. Proton beam therapy has a higher LET rate, particularly toward the distal edge of the SOBP, compared with conventional X-ray radiation. An enhanced efficiency in the induction of cell inactivation can be measured at different positions along the SOBP [49, 50]. Differences in the RBE which are depending on the position along the SOBP were reported in several studies. The study of Petrovic et al., found an increased killing ability at the SOBP distal edge, which was the consequence of increasing proton LET [51]. Another study reported on the variation of the RBE with depth in the SOBP of the 76 MeV proton beams, where they found that, despite a homogeneous physical dose, the tumor cells at the distal end receives a higher biologically equivalent dose than at the proximal end [16]. More recent, the study of Hojo et al., demonstrated that the RBE using an high-energy proton beam, differed according to the position on the SOBP in two human esophageal cancer cell lines with differing radiosensitivities [52]. Also the number of unrepaired double-stranded DNA breaks, as assessed by the number of γ-H2AX foci assay 24 h after irradiation was higher for irradiation at the distal end of the SOBP. In a theoretical study of Carante and Ballarini, a biophysical model of radiation-induced cell death and chromosome aberrations called Biophysical Analysis of Cell death and chromosome Aberrations (BIANCA) was used in order to predict the cell death and the yield of dicentric chromosomes at different depth positions along a SOBP dose profile of therapeutic protons [53]. These simulation data are consistent with the experimental cell survival data as reported in Chaudhary et al. [11] and for both investigating endpoints an increased beam effectiveness was shown along the plateau, implying that the assumption of a constant RBE along a proton SOBP may be sub-optimal [53]. The results of an ex vivo study, where the intestine of mice was irradiated with 200 MeV clinical proton beam are consistent with in vitro data showing an increased proton RBE with depth in an SOBP for both investigated biological endpoints, the intestinal crypt regeneration and lethal dose 50% (LD₅₀) [54]. The study of Marshall et al. have analyzed clinical implications of a variable RBE on proton dose fractionation in human skin fibroblast (AG01522) cells using pencil scanned proton clinical beam of maximum energy 219.65 MeV. Their findings have shown significant variations in the cell killing RBE for both acute and fractionated exposures along the proton dose profile, with a sharp increase in RBE toward the distal position [55]. The study of Chaudhary et al. used the same cell line and investigated the DNA damage response after irradiation with a modulated SOBP and a pristine proton beam, as this new delivery technique was applied in form of intensity-modulated particle therapy (IMPT) in more and more proton therapy centers worldwide [56]. A significantly higher frequency of persistent DNA damage foci was observed at the distal end of the SOBP, whereas the irradiation with a monoenergetic proton beam resulted in significantly increased number of foci at Bragg peak position 24 h after irradiation [56]. In the study of Guan et al. clonogenic cell survival has been mapped as a function of LET along pristine scanned proton beam and the findings indicated that the measured biological effects are greater than reported in previous studies [57]. Furthermore a non-linear RBE for cell survival as a function of LET near and beyond the Bragg peak was observed in this study.

It is important to note, that the RBE predicted by the LEM is in better agreement with the experimental data within the SOBP region than with the constant RBE of 1.1 that is currently applied in the clinics [58]. However, the LEM predictions and experimental data show only a weak dependence of RBE on the tissue type, which is considered insignificant with regard to the general uncertainties of RBE [58].

Recently, clinical evidence for variations in proton RBE was demonstrated by the study of Peeler et al., where the authors analyzed correlation of the tissue damage with increased biological dose effectiveness in pediatric ependymoma patients after proton therapy [59]. Their findings have shown that voxel-based changes on post-treatment MR images are associated with increased LET and dose.