Author: Alejandro Bertolet, Isaac Meyer, Harald Paganetti, Jan PO Schuemann, Wook-Geun Shin 👨🔬
Affiliation: Massachusetts General Hospital, Massachusetts General Hospital and Harvard Medical School 🌍
Purpose: To develop a methodology for Monte Carlo modeling of cell-specific DNA geometries based on Hi-C data using TOPAS-nBio, and to evaluate early DNA damage distributions.
Methods: Hi-C data of the IMR90 human lung fibroblast cell was used to model nuclear DNA. The DNA structure was arranged in base-pairs (bps) forming a double helix and cross-linked solenoidal structure to generate chromatin fiber based on geometrical parameters proposed in previous studies. A spherical nucleus of 5-µm radius was filled with 6.4 Gbps DNA in 15,282 topologically associated domains (TADs). Each TAD is filled with voxels representing straight and turned chromatin fibers using an in-house developed algorithm, so that all the chromatin fiber is connected in the whole nucleus. LET-dependent strand breaks (SBs) and double strand breaks (DSBs) were evaluated.
Results: The bp density in the whole nucleus is 0.012 bp/nm3, with the TADs occupying 15% of the nucleus, resulting in bp density inside TADs of 0.077 bp/nm3. The SB yields agree with results in our previous DNA geometry (fractal). The direct SB yields are independent of LET at 70 SB/Gy/Gbp. The indirect SB yields decrease with LET from 120 to 100 SB/Gy/Gbp in the LET range of 1.24 to 18.12 keV/µm. However, the DSB yields in the Hi-C geometry increase with LET (6-10 DSB/Gy/Gbp), and the yields in the Hi-C geometry were 10% higher than those in the fractal geometry.
Conclusion: This work demonstrated a methodology to calculate cell-specific early DNA damages using Hi-C data. This application can be used for investigating cell-specific radiosensitivity by modeling Hi-C data of different cell types. This framework will be used for the tuning of the damage parameters for the Hi-C-based geometry and the study of the influence of geometrical parameters such as local bp density.