A better understanding of the relationship between ionising radiation (IR) dose, dose rate and radiation quality, and the risk of stochastic effects would improve risk extrapolation from atomic bomb survivors’ data. Owing to insufficient statistical power of epidemiological studies to detect excess incidence of cancer following low doses of IR delivered at low dose rates (LDLDR), as typically encountered in most common human exposure scenarios, radiobiological experiments are fundamental to describe the biological effectiveness of LDLDR and to define the underlying molecular mechanisms. DNA damage and downstream effects are major contributors to radiation carcinogenesis, and as such, these processes have been investigated in the context of plausible mechanisms of radiation-induced health effects in the studies compiled in this thesis, using different cell models and appropriate radiation sources. 

In Paper I, we characterized the energy, activity and dose rate of new low activity gamma and alpha sources of IR, expected to promote small-scale radiation protection research, and used to demonstrate that LDLDR led to an increased micronucleus frequency, a marker of DNA damage, in human osteosarcoma cells as compared to control cells. 

In Paper II, we used blood from patients undergoing radiological imaging procedures, i.e. PET-CT and scintigraphy, to investigate whether candidate IR biomarkers, i.e. ROS, γH2AX, and expression of a panel of radiation-responsive genes, are altered following in vivo low dose exposure as compared to control samples obtained before the diagnostic procedure. We showed that radiological imaging generally induced weak γH2AX, ROS, and gene expression fold changes at the selected timepoints, although few donors presented stronger responses. The observed mild increase in DNA damage was, nevertheless, coherent with a subsequent DNA damage response. This study also indicated that owing to the heterogeneity of the response across individuals, the discrimination of exposed samples might be complicated in the absence of a control for low dose exposures. 

The current risk assessment approach for mixed beam exposures, as encountered in space and other exposure scenarios, assumes additivity of effects of each radiation quality component, but some reports, which show synergistic effects instead, are in conflict with this assumption and indicate a potential underestimation of the corresponding cancer risk. In Paper III, we investigated the consistency of the interaction between low and high LET IR in two healthy donors who presented the largest inter- and intra-donor variability following mixed beam exposure in a previous study.  Based on nine biological replicates, this study confirmed that combined alpha particles and photon radiation led to a higher cytogenetic damage and gene expression responses than those expected based on simple additivity of effects, but that the interaction was prone to seasonal intra-donor and inter-donor variation for both endpoints. This study additionally showed that IR exposure modified alternative transcription of FDXR and MDM2 in a radiation quality-dependent manner, albeit alternative transcription did not coincide with the mode of interaction between the different radiation qualities. In light of these results, we suggest that the possible interaction between low and high LET IR should be considered in calculating uncertainty of risk for mixed exposures. 

In Paper IV, we investigated the early- and long-term biological effects of LDLDR gamma radiation as compared to the same doses delivered acutely in human AHH-1 lymphoblasts, using relevant endpoints related to carcinogenesis, i.e. cell viability, clonogenic survival, chromosomal aberrations, cell growth and global gene expression. The results presented in this study are coherent with a potential detrimental effect of 100 mGy, delivered either chronically or acutely, with a clear dose rate effect for chromosomal aberrations and gene expression, which may modulate cancer risk by dose rate-dependent mechanisms.