FLASH Proton Pencil Beam Scanning Irradiation Minimizes Radiation-Induced Leg Contracture and Skin Toxicity in Mice

Abstract

Ultra-high dose rate radiation has been reported to produce a more favorable toxicity and tumor control profile compared to conventional dose rates that are used for patient treatment. So far, the so-called FLASH effect has been validated for electron, photon and scattered proton beam, but not yet for proton pencil beam scanning (PBS). Because PBS is the state-of-the-art delivery modality for proton therapy and constitutes a wide and growing installation base, we determined the benefit of FLASH PBS on skin and soft tissue toxicity. Using a pencil beam scanning nozzle and the plateau region of a 250 MeV proton beam, a uniform physical dose of 35 Gy (toxicity study) or 15 Gy (tumor control study) was delivered to the right hind leg of mice at various dose rates: Sham, Conventional (Conv, 1 Gy/s), Flash60 (57 Gy/s) and Flash115 (115 Gy/s). Acute radiation effects were quantified by measurements of plasma and skin levels of TGF-β1 and skin toxicity scoring. Delayed irradiation response was defined by hind leg contracture as a surrogate of irradiation-induced skin and soft tissue toxicity and by plasma levels of 13 different cytokines (CXCL1, CXCL10, Eotaxin, IL1-beta, IL-6, MCP-1, Mip1alpha, TNF-alpha, TNF-beta, VEGF, G-CSF, GM-CSF and TGF- β1). Plasma and skin levels of TGF-β1, skin toxicity and leg contracture were all significantly decreased in FLASH compared to Conv groups of mice. FLASH and Conv PBS had similar efficacy with regards to growth control of MOC1 and MOC2 head and neck cancer cells transplanted into syngeneic, immunocompetent mice. These results demonstrate consistent delivery of FLASH PBS radiation from 1 to 115 Gy/s in a clinical gantry. Radiation response following delivery of 35 Gy indicates potential benefits of FLASH versus conventional PBS that are related to skin and soft tissue toxicity.

Keywords: FLASH; normal tissue toxicity; proton beam scanning; proton therapy; skin and soft tissue; ultra-high dose rate.

Figure 1

Irradiation approach for PBS radiation of the murine hind leg. (A) Representative picture of the animal holder, ion chamber and beam blocker set up in the proton gantry used for all irradiations. (B) Representative pictures of the alignment strategy, showing the mouse being first positioned on gafchromic film in the removable holder bed and validation of alignment by the gantry X-ray system. (C) Description of beam delivery parameters for the different mouse groups and experimentations. Values are expressed as the average ± SD. PBS: pencil beam scanning.

Figure 2

Decreased early TGF-Β1 production following FLASH PBS versus Conv irradiation. Right hind legs of the mice were PBS-irradiated with 35 Gy or left untreated (Sham) with dose rates of 1 Gy/s (Conv), 57 Gy/s (Flash60) and 115 Gy/s (Flash115). (A) Plasma measurement of total TGF-Β1 by ELISA at 24 and 96 h post irradiation of the mice at the different indicated dose rates (n = 8/group). (B) Detection of latent (45 kD) and activated (12 kD) TGF-β1 by Western blot analysis in individual sham or irradiated mouse leg skin at 96 h post treatment is shown. Vinculin was used as a loading control. Differences between conditions were determined by a one-way ANOVA test with a multiple-comparison Tukey correction. * p ≤ 0.05 vs. Sham and # p ≤ 0.05 vs. Conv. Bars represent the mean +/− SD. The whole western blot figures can been see in the Supplementary Materials.

Figure 3

Benefits of FLASH PBS with regards to leg contracture and skin toxicity. Right hind legs of the mice were either PBS-irradiated with 35 Gy or left untreated (Sham) with dose rates of 1 Gy/s (Conv), 57 Gy/s (Flash60) and 115 Gy/s (Flash115). (A) Irradiated leg contracture measurements at 3, 7 and 12 weeks post irradiation. Leg contracture was calculated as described in Materials and Methods using the non-irradiated left leg of each animal as a reference (n = 8/group). (B) Representative pictures of leg extension measurements with an in-house 3D-printed jig at 84 days post IR. (C) Irradiated skin toxicity score of Conv (red square) and Flash60 (blue triangle) mouse groups (n = 8/group) as a function of time after radiation. Skin toxicity was scored as: Score 1 = normal, Score 2 = alopecia, Score 3 = erythema, Score 4 = dry desquamation, Score 5 = <30% moist desquamation and Score 6 = >30% moist desquamation. (D,E) Percentage of mice with (D) moist desquamation (scores 5 + 6) or (E) irradiated leg moist desquamation >30% of irradiated area (score 6) after Conv (red square) and Flash60 (blue triangle) irradiation. Differences between conditions were determined by a two-way ANOVA test with multiple-comparison Tukey correction. Bars represent mean +/− SD. * p ≤ 0.05 vs. Flash60 and # p ≤ 0.05 vs. Flash115.

Figure 4

Cytokine response to FLASH PBS- versus Conv-treated mice. Right hind legs of the mice were PBS-irradiated with 35 Gy or left untreated (Sham control mice) with dose rates of 1 Gy/s (Conv), 57 Gy/s (Flash60) and 115 Gy/s (Flash115). Blood was collected at 12 weeks post radiation and the levels of 12 cytokines were quantified by multiplex array (n = 8/group). Differences between conditions were determined by a one-way ANOVA test with multiple-comparison Tukey correction. * p ≤ 0.05 vs. Sham and # p ≤ 0.05 vs. Conv. Bars represent mean +/− SD.

Figure 5

FLASH PBS and Conv irradiation induce equivalent control of indolent and aggressive squamous cell carcinomas. MOC1 (2 × 10⁶ cells) or aggressive MOC2 (2 × 10⁵ cells) murine squamous cell carcinoma cells were injected subcutaneously into the right flanks of C57Bl/6 immunocompetent mice (n = 8/group). Twenty-one days post injection, the mice were randomized in groups of equivalent average tumor sizes. Tumors were irradiated with 15 Gy of Conv or FLASH PBS. (A) MOC1 and (B) MOC2 tumor volume measurements as a function of post irradiation time. Differences between conditions were determined by a two-way ANOVA test with multiple-comparison Tukey correction. Bars represent mean +/− SD. * p ≤ 0.05 vs. Conv and # p ≤ 0.05 vs. Flash60.