Gene Expression Patterns in Pancreatic Neuroendocrine Tumors Suggest Resistance to Radiation Therapy

Purpose/Objective(s): The advent of detailed proximity mapping of the human genome provides a rigorous framework to view the interaction of ionizing radiation with its principle target. The goal of this study was to reanalyze signature radiation-induced rearrangements in the context of this framework. In particular, the conformation of DNA as a fractal globule packaged into physically discrete transcriptionally active and inactive compartments was used as the basis for analysis. Materials/Methods: Published data on the generation of chromosome aberrations by irradiation was collated and reassessed using chromosome capture-based proximity mapping data sets. In particular, established models of radiation-induced rearrangements are validated against available proximity mapping data. Results: Rings are formedmore frequently than dicentrics in irradiated cells Contact proximity is required for strand interchange. Intra-chromosomal contact is highest for locations that are close together and which are in the active compartment. However, contact for large chromosomes 1-6, 8 & 10 does not extend across the centromere, limiting ring formation for these. Partially confirmed. Deletions and inversions are generated by irradiation. Proximity mapping shows transcriptionally active, adjacent locations on the same chromosomewill physically interact more often. Such an interaction is observed in papillary thyroid tumors linked to radiation exposure where an inversion between the RET gene and NCOA4 or CCDC6 partners is seen on chromosome 10. The link to radiation is confirmed from epidemiological studies and the spatial co-localization of RETand CCDC6 genes inferred by the proximity of individual breakpoints in both genes. Confirmed. Chromosome rearrangements have an equal probability to occur between any pair of autosomes. Proximity mapping of all genomic locations reveals two compartments transcending individual chromosome identities, likely correlating with transcribing or quiescent DNA. Fragmentation of adjacent chromosomes (a type damage in the LQ formalism) and subsequent rearrangement will therefore preferentially occur within, rather than across, compartments. Further, the random rearrangement of autosomes subsequent to irradiation, described in FISH studies, is likely incorrect in that population-based data modelling of interaction frequencies is consistent with multiple different genomic states, such as cell cycle stage or transcriptional status, all of which will have a unique interaction profile questioned. Conclusions: Radiation-induced rearrangements are key elements in the killing of carcinoma. Proximity mapping data shows that the DNA target is subject to specific and mathematically predictable levels of organization that indicate lethal rearrangementsmaynot begenerated as simple randomevents. Author Disclosure: A. Vaughan: None.