Sparing of DNA irradiated with Ultra-High Dose-Rates under Physiological Oxygen and Salt conditions

TL;DR

This work addresses the molecular basis of the FLASH effect by examining DNA damage in plasmid pUC19 irradiated with 18 MeV electrons at ultra-high dose rates (UHDR, $>10^6$ Gy/s) and conventional low dose rates (LDR, $<0.1$ Gy/s) under physiological oxygen (~5% O2) and salt conditions in PBS. Using a combination of controlled experiments and Geant4-TOPAS-nBio simulations, the study observes a DR- and oxygen-dependent sparing of DNA strand-breaks under physoxia for total doses above 10 Gy, with more SSB and DSB produced under LDR than UHDR, a pattern not seen under ambient oxygen. An extended chemical-kinetics model predicts ROS yields and links the sparing to interplay between O2, OH-induced backbone damage, and dose-rate–driven H3O+ yields that influence beta-elimination processes. These findings provide a chemically plausible explanation for FLASH-related tissue sparing and underscore the importance of physiologically relevant solvent conditions in in vitro radiobiology experiments, potentially informing optimization of radiotherapy protocols. The work combines precise beam physics, chemical Monte Carlo simulations, and DNA damage assays to advance understanding of how dose rate and oxygen modulate DNA damage chemistry at the molecular level.

Abstract

Cancer treatment with radiotherapy aims to kill tumor cells and spare healthy tissue.Thus,the experimentally observed sparing of healthy tissue by the FLASH effect during irradiations with ultra-high dose rates (UHDR) enables clinicians to extend the therapeutic window.However, the underlying radiobiological and chemical mechanisms are far from being understood.DNA is one of the main molecular targets for radiotherapy.Ionizing radiation damage to DNA in water depends strongly on salt,pH,buffer and oxygen content of the solvent.Here we present a study of plasmid DNA pUC19,irradiated with 18MeV electrons at low dose rates (LDR) and UHDR under tightly controlled ambient and physiological oxygen conditions in PBS at pH 7.4.For the first time a sparing effect of DNA strand-break induction between UHDR(>10MGy/s) and LDR(<0.1Gy/s) irradiated plasmid DNA under physiological oxygen, salt and pH is observed for total doses above 10Gy.Under physiological oxygen (physoxia,5%O2,40mmHg),more single (SSB) and double strand-breaks (DSB) are observed when exposed to LDR, than to UHDR.This behaviour is absent for ambient oxygen (normoxia,21%O2,150-160mmHg).The experiments are accompanied by TOPAS-nBio based particle-scattering and chemical MCS to obtain information about the yields of reactive oxygen species (ROS).Hereby,an extended set of chemical reactions was considered, which improved upon the discrepancy between experiment and simulations of previous works, and allowed to predict DR dependent g-values of hydrogen peroxide (H2O2).To explain the observed DNA sparing effect under FLASH conditions at physoxia,the following model was proposed:The interplay of O2 with OH induced H-abstraction at the phosphate backbone,and the conversion of DNA base-damage to SSB,under consideration of the dose-rate dependent H3O+ yield via beta elimination processes is accounted for, to explain the observed behavior.

Figures (12)

• Figure 1: Electron bunches and trains: Comparison of LDR (blue) and UHDR (green) bunch (left) and bunch-train (right) structure in time. Note the differences in y axis scale.
• Figure 2: Experimental and simulated doses: Comparison of measured and simulated dose: Top left: Exemplary spatial dose distribution measured on a Gafchromic film behind the sample tube with 0.5 ml volume filled with 50 $\mu$L sample (left). The region of interest for sample dose determination is marked by light blue dots and lines. Top Center: Histogram of the dose distribution within the measured target volume in the sample tube as indicated by the blue lines in the image on the left. Top right: Simulated dose distribution within the same Gafchromic film as obtained by Geant4/Topas particle scattering simulations for the same conditions as in the experiment. Central image, left: Particle scattering simulations showing the simulated setup. Hereby electrons are shown in red, photons in light green. Sample tubes are grey, water blue, Gafchromic films before and behind the samples are shown in yellow. Before them, the scattering plate in blue and the lead shielding with a hole are shown in yellow. The image in the center right, shows a zoom to the sample tube and Gafchromic films. The electron beam propagated from right to left and for illustration purposes the number of electron simulated is strongly reduced with respect to the real case. Bottom: Shown is the dose deposit within the water volume for LDR (orange) and UHDR (blue) conditions in experiment (circles and a linear regression as broken lines) and simulation (solid lines) in dependence of the total charge.
• Figure 3: Oxygen stability: Shown is the oxygen content over time for different tubes and sealing approaches for closed sample tubes stored under ambient oxygen at room temperature, measured directly with an Oxford Oxylite oxygen sensor after opening the respective tube. The inset shows the Thermo and Eppendorf tubes without additional sealing and with Parafilm and Teflon sealing, from left to right respectively.
• Figure 4: Plasmid conformations, damage and their position in an exemplary CGE measurement. Shown is a CGE densiogram of plasmid pUC19 with 2686 bp and three clearly visible isoforms. From left to right as linear (Lin: red, around 400 s), supercoiled/covalently-closed circular (SC: yellow, around 500 s) and open-circular (OC: green, around 750 s) conformation. To visualize the microscopic structure of the different plasmid DNA isoforms, atomic-force microscopy (AFM) images are shown as insets next to the respective band in the densiogram. The images were taken and combined with permission from previous works.cordsmeierdna2022hahnrapid2025
• Figure 5: Properties of the electron beam: Energy dependent stopping power of electrons from the NIST Elstar database (top) and the related LET values (center) for different electron energies and simulation settings, and the respective kinetic energy spectra originating from particle scattering simulations (bottom). For simulation details see Sec. \ref{['sec:methods']} and the discussion.