The purpose of this paper is to characterize the skin deterministic damage due to the effect of proton beam irradiation in mice occurred during a long-term observational experiment. This study was initially defined to evaluate the insurgence of myelopathy irradiating spinal cords with the distal part of a Spread-out Bragg peak (SOBP). To the best of our knowledge, no study has been conducted highlighting high grades of skin injury at the dose used in this paper. Nevertheless these effects occurred. In this regard, the experimental evidence of significant insurgence of skin injury induced by protons using a SOBP configuration will be shown. Skin damages were classified into six scores (from 0 to 5) according to the severity of the injuries and correlated to ED50 (i.e. the radiation dose at which 50% of animals show a specific score) at 40 days post-irradiation (d.p.i.). The effects of radiation on the overall animal wellbeing have been also monitored and the severity of radiation-induced skin injuries was observed and quantified up to 40 d.p.i.
All relevant data are within the manuscript.
In recent years, the use of proton therapy is emerging as an alternative to conventional radiotherapy with photons. This is mainly due to the presence of the Bragg peak (BP), which allows higher dose conformation to tumor regions while sparing the surrounding organs reducing side effects [
Several
The research shown in this paper was originally designed to perform a long observational study investigating the dose at which 50% of small animals develop myelopathy when the spinal cords were irradiated in the distal part of a SOBP. For this purpose, the animals were irradiated with a range of proton doses at spinal cord level to evaluate a possible onset of proton-induced myelopathy. The proton irradiation doses have been chosen taking into account a work published by Yeh Chi Lo
In agreement with these assumptions, in this paper a skin deterministic damage has been adopted as endpoint in order to assess the biological impact at the last few mm of a proton SOBP. Healthy mice were irradiated using a range of proton doses to evaluate radiation-induced skin deterministic damage, as well as to verify their possible involvement to animal wellbeing.
The experiments were performed in accordance with the European Communities Council directive and Italian regulations (EEC Council 2010/63/EU and Italian D.Lgs. 26/2014). The project was approved by Italian Ministry of Health (n. 248/2018-PR of 30/03/2018).
Efforts were employed to replace, reduce and refine the use of laboratory animals. To avoid irrelevant suffering to treated mice, euthanasia was performed as soon as the final score was reached. All reasonable efforts were made to reduce suffering, avoiding the most painful procedures. The endpoint used to determine if animals should undergo to euthanasia was reached when lesions showed dimension higher than 0.5 cm.
All experiments were performed on 6 weeks old C57BL/6 male mice (Charles River Laboratory), weighing 27 ± 3 gr. Animals were housed in IVC-cages for 9 weeks using a stocking density of maximum 4 mice per cage. In order to identify mice stored within the same cage, the animals were randomly marked on the tail as follows: mouse 1 = NT (No Tag); mouse 2 = 1 tag; mouse 3 = 2 tags; mouse 4 = 3 tags. Animals were fed ad libitum and maintained in the same room under a 12:12-hour light/dark photoperiod at 24°C. To minimize suffering and distress of mice, standard environmental enrichment of nest paper, a cardboard fun tunnel and one wooden chew block were provided.
A total number of 40 animals was used in this study. Mice were randomly assigned to sham-control group (n = 8) and treated group (n = 32). The treated groups were randomly divided into 4 subgroups as follows: Dose 12 Gy (n = 8); Dose 15 Gy (n = 8); Dose 17 Gy (n = 8) and Dose 19 Gy (n = 8). Before each irradiation, the mice were anesthetized with Zoletil (tiletamine) 40 mg/kg and Sedastart (medetomidine) 50 μg/kg and shaved in the treatment region. Animal health and behavior were monitored twice a week. All animals were scanned with an Albira Si microPET/CT to define the irradiation setup as it will be described later on this paper.
All proton treatments were performed at the INFN-LNS CATANA proton therapy facility in Catania (Italy) using a passive proton beam line. Since 2002, CATANA is a proton therapy facility where radiotherapeutic treatments of eye tumours are performed. In this experimental hall, a fixed horizontal beam line is installed and clinical proton beams can be delivered with a maximum energy of 62 MeV. A beam shaping system is used to obtain a uniform dose distribution at the isocenter. Indeed, when the proton beam reaches the experimental hall, it goes out in air and flies for three meters before hitting the target. Along its path, the beam is intercepted by various elements in order to obtain flat transversal dose distribution at the isocenter, as described elsewhere [
The A indicates the skin position along SOBP.
The dosimetry was performed using a Markus ionization chamber (PTW Freiburg GmbH, Germany) and Gafchromic EBT3 films (ISP Corp., New York, USA). During each irradiation, a Gafchromic EBT3 film was placed just before the target to check the beam flatness and the released dose. The films were scanned 24 hours after irradiation with an Epson Expression 10000 XL Scanner (Epson, Germany) and analysed by homemade MatlabTM script. The dose delivery was monitored by a transmission ionization chamber placed along the beam line that automatically switched off the beam when the requested number of monitor units (MU) were reached. The calibration of MU as a function of the absolute dose to water was determined by measurements with the Advanced Markus IC at middle SOBP position. A constant dose rate of 5 Gy/min was set for all the irradiations. All animal irradiations were performed at the same time interval of the day to avoid differences related to circadian rhythm of mice.
Monte Carlo (MC) simulation is usually applied for routine clinical applications as it allows to perform accurate and efficient prediction of dose distribution inside the target. Moreover, when small animals are irradiated, the power of this tool becomes more important because the target dimensions are less than a few centimetres. The application developed by our group allows the simulation of
The GEANT4-based application was used to define the irradiation setup and in detail to determine which combination of modulator wheel and range shifter was suitable to correctly position the target along the SOBP. In addition, it was used to calculate the dose distributions and the LETd value occurring by SOBP in the target. Indeed, our GEANT4 application permits to simulate entirely the CATANA proton beamline and defines voxel-by-voxel the real composition of the target using the real DICOM micro-CT [
Proton-induced skin injuries were evaluated paying attention to erythema, microlesions and ulcers. To monitor the skin alterations, mice were weighed and photographed once a week. The mice pictures acquired once a week were grouped per dose level and ordered in a photographic time-lapse. The assignment of the skin injury score was carried out during the check and independently confirmed using recorded images.
According to a modified scoring system [
SCORE 0 = healthy skin;
SCORE 1 = light redness;
SCORE 2 = redness, erythema principle, alopecia;
SCORE 3 = extended microlesions, early-stage ulcers);
SCORE 4 = confluent moist desquamation, ulcers;
SCORE 5 = open wound, necrosis.
Skin injury was estimated by a semi-quantitative scale and mean score was calculated at different time-points (7, 14, 21, 30 and 40 d.p.i.) for each treatment group.
Body weight changes were quantified as a function of dose considering the mean weight of each treatment group compared to the control group.
The impact of skin injury progression on mice survival was assessed calculating survival curves using GraphPad Software version 5 (GraphPad Software, Inc.). This software uses the product limit method of Kaplan and Meier and compares survival curves using both the logrank test and the Gehan-Wilcoxon test. Survival curves were plotted as a function of time (7, 14, 21, 30 and 40 d.p.i.) classifying the animals as follows: “1” mouse deceased at a certain time-point and “0” if it survived until the end of the observation time (40 d.p.i.).
All statistical results were plotted using GraphPad Prism 5. The median skin injury score was calculated for each experimental group and the statistically significant differences between irradiated and control groups were analysed using one-way analysis of variance (ANOVA).
Bonferroni and Holm multiple comparison method was used to evaluate the differences between experimental groups as a
Specific dosimetric evaluations were computed for each animal used in this work thanks to the use of our GEANT4-based application. In detail, it allowed to define the best irradiation configuration and to calculate the dose distributions and the mean total LET value occurring by SOBP on skin. The mean LET value calculated at skin surface was equal to 6.68 keV/μm. In
Four groups of mice were irradiated with a range of proton absorbed dose (12, 15, 17 and 19 GyRBE) and one group was used as a negative control (sham-CTRL). As expected, the monitoring of mice has highlighted a dose-dependent relationship between the onset of skin damage and the amount of radiation dose delivered.
A qualitative heat map per dose is shown in
Mice exposed to 12 Gy developed first symptoms of skin damage at 7 d.p.i. as shown in
This condition evolved in extended micro-lesions (SCORE 3) at 30 d.p.i only in 33.3% of animals, but without reaching the SCORE 5. Additionally, a complete recovery was registered in 16.7% of the animals.
Skin injuries induced by 15 Gy arose with SCORE 2 at 7 d.p.i in 50.0% of the animals, as shown in
Qualitative heat map of mice exposed to 17 Gy (see
Moreover, the mice irradiated with the highest proton dose (19 Gy) showed the first sign of skin injury at 7 d.p.i.. The corresponding qualitative heat map (see
The box plot in
SCORE | 12 Gy | 15 Gy | 17 Gy | 19 Gy |
---|---|---|---|---|
0 | 1 | 2 | 4 | |
3 | 3 | 5 | 5 | |
1.5 | 3 | 3 | 5 |
Statistical analysis of mean scores was calculated with ANOVA test using Bonferroni and Holm multiple comparison and confirms that first symptoms of skin injury arise at 7 d.p.i for all treatment groups without statistically significant differences among the experimental groups (
In addition, no statistically significant differences were found between the doses of 12 and 15 Gy (*p-value < 0.05,
Dose (Gy) | p-value | T-value | SD |
---|---|---|---|
p>0.05 p = 0.3478593 | 2.0112 | 0.51639 | |
**p<0.01 p = 0.0003867 | 5.0280 | 0.51639 | |
**p<0.01 p = 9.5629e-06 | 6.7040 | 0.5164 | |
*p<0.05 p = 0.040872 | 3.0168 | 0.8164 | |
**p<0.01 p = 0.0008384 | 4.6928 | 0.8164 | |
p>0.05 p = 0.6558068 | 1.6760 | 1.3291 |
The dose–response curves for skin reaction induced by protons is shown in
SCORE | 1 | 2 | 3 | 4 | 5 |
---|---|---|---|---|---|
NC | NC | 13.6 Gy | 17 Gy | 19 Gy |
The data shown suggest that the onset of the skin injury started just at 7 d.p.i in all treated mice regardless of the single dose received. In detail, the doses of 12 Gy and 15 Gy induced slightly reddened skin (SCORE 1) accompanied, in some cases, by diffuse erythema (SCORE 2). Severer effects were found in the animals irradiated with single doses of 17 Gy and 19 Gy which caused diffuse microlesions (SCORE 3). The photographic time-lapse, shown in
Panels with lettering “no significant difference” indicated that no significant variations between the previous photo are observed.
Panels with lettering “no significant difference” indicated that no significant variations between the previous photo are observed.
Mice body weight changes were monitored once a week in order to evaluate whether the health conditions and the observed increase in lethality were influenced by proton-induced skin injury.
The mean weight of each treated group is shown in
7 | **p<0.01 p = 0.024757 | 4.0800 | 0.9710 | |
14 | **p<0.01 p = 3.7860e-06 | 6.5916 | 1.1463 | |
21 | **p<0.01 p = 0.0007880 | 4.7110 | 1.1176 | |
30 | **p<0.01 p = 0.0001608 | 5.3258 | 1,9568 | |
40 | **p<0.01 p = 0.0005106 | 4.8784 | 1.3889 | |
7 | **p<0.01 p = 0.002019 | 4.1528 | 1.3680 | |
14 | *p<0.05 p = 0.0110938 | 3.6344 | 1.4142 | |
21 | **p<0.01 p = 0.0031802 | 4.1722 | 1.4142 | |
30 | **p<0.01 p = 0.0009967 | 4.6204 | 0.7778 | |
40 | *p<0.05 p = 0.036357 | 3.2089 | 1.6263 |
On the other hand, statistically significant changes in body weight were observed for mice irradiated with single dose of 12 Gy and 15 Gy when compared to the sham-CTRL group only at 14, 21, 30 and at 40 d.p.i. as confirmed by ANOVA test results shown in
7 | p>0.05 p = 0.3151779 | 2.2403 | 1.7054 | |
14 | **p<0.01 p = 0.0005691 | 4.7370 | 1.2873 | |
21 | **p<0.01 p = 0.0017150 | 4.4110 | 1.4961 | |
30 | **p<0.01 p = 0.006810 | 3.8764 | 1.3918 | |
40 | **p<0.01 p = 0.0004174 | 4.9562 | 1.2325 | |
7 | p>0.05 p = 0.2072 | 2.4225 | 1.4918 | |
14 | **p<0.01 p = 1.8353e-05 | 5.9989 | 1.4257 | |
21 | **p<0.01 p = 0017150 | 4.4110 | 1.4961 | |
**p<0.01 p = 1.3639e-05 | 6.2991 | 1.4159 | ||
**p<0.01 p = 1.2712e-05 | 6.3274 | 1.3659 |
The differences in body weight between the two highest-dose groups (17 and 19 Gy) and sham-CTRL group were quantified and plotted in
Finally, skin injury-related survival proportions of the two highest dose (17 and 19 Gy) and the two lowest dose (12 and 15 Gy) groups were quantified in percentage and shown in
Proton therapy is quickly diffusing worldwide due to many advantages compared to conventional radiotherapy. The balance between efficiency and efficacy of proton therapy and the appearance of side effects is related to several issues including a correct assessment of the variation of the RBE along the BP. Currently during proton therapy treatment planning the proton RBE is assumed equal to constant and equal to 1.1 [
Recently,
This work is part of a long observational study which pursued the study of the dose at which 50% of animals develop myelopathy effects by irradiating the spinal cord in order to assess the biological impact of distal SOBP. For this purpose, animals were irradiated with a range of proton doses at spinal cord level to evaluate a possible onset of myelopathy proton beam-induced. Proton doses were calculated taking into account a previous work of Yeh Chi Lo et al. [
Significant insurgence of skin injury was observed in this work from 7 days post-irradiation. Specifically, we observed that healthy mice treated using a modulated proton beam at different doses of radiation show different severity grade of skin damage, even though previous works have shown that this side effect occurs at higher doses [
In conclusion, although this study would need further investigation to specify in detail the variation of the RBE along a Bragg peak, also by using data from x-ray (i.e. 250 keV) radiation as reference, it suggests to analyse the side effects thoroughly, especially the early ones, such as skin damage laying the bases for more dedicated preclinical research.
The results showed in this paper have been achieved within the framework of the Catania Animal Facility thanks to the collaboration among the University of Catania, the IBFM-CNR, the Cannizzaro Emergency Hospital and the LNS-INFN.
We are grateful to Vincenzo Zimmitti, Marco Abbate, Elisa Giuffrida, Marco Durante, Marco Schwarz and Christian P. Karger for useful intellectual contributions.