Effect of fluorescence in situ hybridization detection threshold on chromosome aberration counting: a simulation...

Radiation-induced carcinogenesis remains one of the main hurdles for long duration missions in deep space. The space radiation environment is diverse and includes high linear energy transfer (LET) ions that are particularly effective at inducing adverse health outcomes including cancer. Quantifying the health effects of these high-LET ions is difficult, and large uncertainties remain in cancer risk projections. Chromosome aberrations are a biomarker of radiation-induced cancer used to assess radiation quality effects. Fluorescence in situ hybridization (FISH) measurements of simple and complex exchanges have inherent detection limitations that might underestimate the overall number of chromosomal rearrangements, possibly affecting estimates of the relative biological effectiveness of high-LET ions.

In this work, we introduced a new chromosome aberration classification approach in the simulation code RITCARD (Radiation induced tracks, chromosome aberrations, repair, and damage), that accounts for FISH detection threshold and the use of different chromosome painting probes. We also modified our 3D nuclear architecture model using Hi-C data to generate the DNA distribution within cell nuclei with the tool G-NOME. This new approach allowed the discrimination of true simple and complex exchanges from apparently simple exchanges (complex exchanges detected as simple), as well as undetected exchanges.

We compared the results of this new classification method in the RITCARD tool with experimental FISH data obtained for the staining of 3 pairs of chromosomes (referred to as 3-FISH), and found an overall good agreement of the total exchanges for fibroblasts (hTERT 82-6) and lymphocytes (whole blood) for high LET ions, a slight underestimation in the low LET range (< ∼ 20 keV/µm), and a slight imbalance between simple and complex exchanges for lymphocytes. The model reproduced well the higher yield of aberrations for lymphocytes, compared to fibroblasts. Remarkably, in our model, this higher yield was solely due to differences in nuclear geometries and repair time between the two cell types, both derived from experimental data. For both cell types, we observed an increased number of complex exchanges detected as simple, and an increased number of undetected simple exchanges for high LET ions when we increased the detection threshold. For lymphocytes, this resulted in an overall increased number of simple exchanges, while, for fibroblasts, simple exchanges remained largely unchanged. Overall, the number of total exchanges decreased with increased detection threshold for both cell types. We also found that, for high LET ions, the majority of detected simple exchanges were true complex exchanges, due to many intra-chromosomal rearrangements that are undetected with traditional FISH technique.

Our new chromosome aberration classification approach allows us to go beyond FISH detection limitations and quantify how they impact aberration yields. Our simulation results suggest that, for high LET exposure, 3-FISH underestimates the total number of exchanges as well as their complexity, due to the inability to detect small fragments and intra-chromosomal rearrangements. Future work will focus on optimizing the model parameters to better reproduce low LET measurements. Once validated, RITCARD predictions may be used in the NASA cancer model to inform radiation quality factors as part of an ensemble framework. We also intend to investigate how predictions obtained with partial chromosome staining (3-FISH) compares with predictions obtained with whole genome staining (mFISH), and how both compare with predictions of true exchanges, where all exchanges are accounted for, included those undetectable by traditional FISH such as inversion or small deletions.