Preliminary Investigation of Using TOPAS-nBio to Simulate DNA Double-Strand Breaks from Volumetric Radioisotope Sources

Introduction: Radionuclide therapy encompasses a broad range of medical therapies that utilize radioactive isotopes to deliver dose in close proximity to tumors and other tissues. Determining the effectiveness of different radionuclides in these therapies can be achieved by analyzing their emissions via radiation transport simulations. One computer simulation package, TOol for PArticle Simulations (TOPAS), is a standard radiation transport simulator used to study radiation therapies and is based on the Geant4 simulation toolkit.[1] Recent developments of TOPAS, especially its biophysics-modeling extension TOPAS-nBio, allow for radiation simulations on the scale of cells, DNA, and other subcellular structures.[2] One method of determining the effectiveness of a certain radioisotope can be achieved by counting the number of DNA double-strand breaks (DSBs) it causes relative to its activity. Because cells are less likely to successfully repair DSBs compared to other damages, DSBs are the main method by which ionizing radiation induces cell death. Here, we explore using TOPAS-nBio to count DSBs in cell nuclei from two theranostic radionuclides.

Methods: TOPAS-nBio was used to simulate DSBs resulting from radiation emitted by 64Cu and 47Sc. We modeled a tumor cell as a spherical shell with a 5 μm radius and a nucleus of radius 4 μm.[3] Radioisotopes were uniformly distributed in the cell, whose decay products would induce DSBs in the nucleus. The DNA is modeled as a length of stacked cylinders and coiled into subdomains inside the nucleus. In addition to physical processes, chemical processes were also enabled to allow for breaks caused by chain ionization events from chemical hydroxyl and hydroxide reactions in the DNA. Two parameters (HilbertCurve3DRepeat and HilbertCurveLayer) control DNA fractal recursion and should be set to 30 and 4, respectively, as per the recommendation from TOPAS-nBio. In our investigation, however, simulations were unstable unless HilbertCurve3DRepeat was set to 27 (this issue is under current investigation). The resulting DNA model has approximately 6 billion base pairs. The number of DSBs in the nucleus is counted with the native scoring algorithm DNAScorer. It estimates hits to the DNA via a linear threshold where the probability of a break is 0 for energies less than or equal to 5 eV, and 1 for energies greater than or equal to 37.5 eV. The probability increases linearly between these two values. If two or more hits are seen within 10 base pairs, a DSB is recorded.

Results: Our simulations employed PCs with easy access but limited computing power. Even with a limited number of simulated particles, we found that DSBs were at the same scale as those published using the DBSCAN algorithm to count DSBs from 64Cu,[3] a positron-emitting radionuclide for PET imaging. Both cytoplasm-distributed (N <- Cy) and nucleus-distributed (N <- N) isotopes were simulated inducing nucleus DSBs. We found that DSBs per total activity for 64Cu distributed inside the nucleus was 0.075 DSBs/Bq·࣪sec. For 64Cu distributed inside the cytoplasm, our results were 0.013 DSBs/Bq·࣪sec. For 47Sc, a beta-emitter with gamma-ray emission, our results were 0.024 DSBs/Bq·࣪sec when distributed in the nucleus.

Conclusions: Our preliminary limited computing power studies showed that the TOPAS-nBio simulation of DSBs was reasonably comparable to published data for 64Cu. We have used this approach for a preliminary evaluation of the DSBs resulting from 47Sc. These initial investigations are being expanded to utilize more powerful computing clusters to perform the simulations, and to evaluate additional imaging, therapeutic, and theranostic radioisotopes.

• Download figure
• Open in new tab
• Download powerpoint