Of Men and Mice: Using Terrestrial Radiation Epidemiology Methods to Inform Analysis of Animal Models for Space Radiation Risk Assessment.

Prediction of cancer risk from space radiation exposure is critical to ensure spaceflight crewmembers are adequately informed of the risks they face when accepting assignments to ambitious long-duration exploratory missions. Although epidemiological studies have assessed the effects of exposure to terrestrial radiation, no robust epidemiological studies of humans exposed to space radiation exist to support estimates of the risk from space radiation exposure. Mouse data derived from recent irradiation experiments provides valuable information to successfully develop mouse-based excess risks models for assessing relative biological effectiveness for heavy ions that can provide information to scale unique space radiation exposures so that excess risks estimated for terrestrial radiation can be adjusted for space radiation risk assessment. Bayesian analyses were used to simulate linear slopes for excess risk models with several different effect modifiers for attained age and sex. Relative biological effectiveness values for all-solid cancer mortality were calculated from the ratio of the heavy-ion linear slope to the gamma linear slope using the full posterior distribution and resulted in values that were substantially lower than what is currently applied in risk assessment. These analyses provide an opportunity to improve characterization of parameters used in the current NASA Space Cancer Risk (NSCR) model and generate new hypotheses for future animal experiments using out-bred mouse populations.

Ablating putative Ku70 phosphorylation sites results in defective DNA damage repair and spontaneous induction of hepatocellular carcinoma.

Multiple pathways mediate the repair of DNA double-strand breaks (DSBs), with numerous mechanisms responsible for driving choice between the pathways. Previously, we reported that mutating five putative phosphorylation sites on the non-homologous end joining (NHEJ) factor, Ku70, results in sustained retention of human Ku70/80 at DSB ends and attenuation of DSB repair via homologous recombination (HR). In this study, we generated a knock-in mouse, in which the three conserved putative phosphorylation sites of Ku70 were mutated to alanine to ablate potential phosphorylation (Ku703A/3A), in order to examine if disrupting DSB repair pathway choice by modulating Ku70/80 dynamics at DSB ends results in enhanced genomic instability and tumorigenesis. The Ku703A/3A mice developed spontaneous and have accelerated chemical-induced hepatocellular carcinoma (HCC) compared to wild-type (Ku70+/+) littermates. The HCC tumors from the Ku703A/3A mice have increased γH2AX and 8-oxo-G staining, suggesting decreased DNA repair. Spontaneous transformed cell lines from Ku703A/3A mice are more radiosensitive, have a significant decrease in DNA end resection, and are more sensitive to the DNA cross-linking agent mitomycin C compared to cells from Ku70+/+ littermates. Collectively, these findings demonstrate that mutating the putative Ku70 phosphorylation sites results in defective DNA damage repair and disruption of this process drives genomic instability and accelerated development of HCC.

Assessing Nonlinearity in Harderian Gland Tumor Induction Using Three Combined HZE-irradiated Mouse Datasets.

Recently reported studies considering nonlinearity in the effects of low-dose space radiation have assumed a nontargeted mechanism. To date, few analyses have been performed to assess whether a nontargeted term is supported by the available data. The Harderian gland data from Alpen et al. (published in 1993 and 1994), and Chang et al. (2016) provide the most diversity of ions and energies in a tumor induction model, including multiple high-energy and charge particles. These data can be used to investigate various nonlinearity assumptions against a linear model, including nontargeted effects in the low-dose region or cell sterilization at high doses. In this work, generalized linear models were used with the log complement link function to analyze the binomial data from the studies independently and combined. While there was some evidence of nonlinearity that was best described by a cell-sterilization model, the linear model was adequate to describe the data. The current data do not support the addition of a nontargeted effects term in any model. While adequate data are available in the low-dose region (<0.5 Gy) to support a nontargeted effects term if valid, additional data in the 1-2 Gy region are necessary to achieve power for cell-sterilization analysis validation. The current analysis demonstrates that the Harderian gland tumor data do not support the use of a nontargeted effects term in human cancer risk models.

Radiation Exposure and Mortality from Cardiovascular Disease and Cancer in Early NASA Astronauts.

Understanding space radiation health effects is critical due to potential increased morbidity and mortality following spaceflight. We evaluated whether there is evidence for excess cardiovascular disease or cancer mortality in early NASA astronauts and if a correlation exists between space radiation exposure and mortality. Astronauts selected from 1959-1969 were included and followed until death or February 2017, with 39 of 73 individuals still alive at that time. Calculated standardized mortality rates for tested outcomes were significantly below U.S. white male population rates, including all-cardiovascular disease (n = 7, SMR = 33; 95% CI, 14-65) and all-cancer (n = 7, SMR = 43; 95% CI, 18-83), as anticipated in a healthy worker population. Space radiation doses for cohort members ranged from 0-78 mGy. No significant associations between space radiation dose and mortality were found using logistic regression with an internal reference group, adjusting for medical radiation. Statistical power of the logistic regression was <6%, remaining <12% even when expected risk level or observed deaths were assumed to be 10 times higher than currently reported. While no excess radiation-associated cardiovascular or cancer mortality risk was observed, findings must be tempered by the statistical limitations of this cohort; notwithstanding, this small unique cohort provides a foundation for assessment of astronaut health.

Defining the Biological Effectiveness of Components of High-LET Track Structure.

During space travel, astronauts are exposed to a wide array of high-linear energy transfer (LET) particles, with differing energies and resulting biological effects. Risk assessment of these exposures carries a large uncertainty predominantly due to the unique track structure of the particle's energy deposition. The complex damage elicited by high charge and energy (HZE) particles results from both lesions along the track core and from energetic electrons, δ rays, generated as a consequence of particle traversal. To better define how cells respond to this complex radiation exposure, a normal hTERT immortalized skin fibroblast cell line was exposed to a defined panel of particles carefully chosen to tease out track structure effects. Phosphorylation kinetics for several key double-strand break (DSB) response proteins (γ-H2AX, pATF2 and pSMC1) were defined after exposure to ten different high-LET radiation qualities and one low-LET radiation (X ray), at two doses (0.5-2 Gy) and time points (2 and 24 h). The results reveal that the lower energy particles (Fe 300, Si 93 and Ti 300 MeV/u), with a narrower track width and higher number and intensity of δ rays, cause the highest degree of persistent damage response. The persistent γ-H2AX signal at lower energies suggests that damage from these exposures are more difficult to resolve, likely due to the greater complexity of the associated DNA lesions. However, different kinetics were observed for the solely ATM-mediated phosphorylations (pATF2 and pSMC1), revealing a shallow induction at early times and a higher level of residual phosphorylation compared to γ-H2AX. The differing phospho-protein profiles exhibited, compared to γ-H2AX, suggests additional functions for these proteins within the cell. The strong correspondence between the predicted curves for energy deposition per nucleosome for each ion/energy combination and the persistent levels of γ-H2AX indicates that the nature of energy distribution defines residual levels of γ-H2AX, an indicator of unrepaired DSBs. Our results suggest that decreasing the energy of a particle results in more complex damage that may increase genomic instability and increase the risk of carcinogenesis.

Induction of chromosomal aberrations at fluences of less than one HZE particle per cell nucleus.

The assumption of a linear dose response used to describe the biological effects of high-LET radiation is fundamental in radiation protection methodologies. We investigated the dose response for chromosomal aberrations for exposures corresponding to less than one particle traversal per cell nucleus by high-energy charged (HZE) nuclei. Human fibroblast and lymphocyte cells were irradiated with several low doses of <0.1 Gy, and several higher doses of up to 1 Gy with oxygen (77 keV/µm), silicon (99 keV/µm) or Fe (175 keV/µm), Fe (195 keV/µm) or Fe (240 keV/µm) particles. Chromosomal aberrations at first mitosis were scored using fluorescence in situ hybridization (FISH) with chromosome specific paints for chromosomes 1, 2 and 4 and DAPI staining of background chromosomes. Nonlinear regression models were used to evaluate possible linear and nonlinear dose-response models based on these data. Dose responses for simple exchanges for human fibroblasts irradiated under confluent culture conditions were best fit by nonlinear models motivated by a nontargeted effect (NTE). The best fits for dose response data for human lymphocytes irradiated in blood tubes were a linear response model for all particles. Our results suggest that simple exchanges in normal human fibroblasts have an important NTE contribution at low-particle fluence. The current and prior experimental studies provide important evidence against the linear dose response assumption used in radiation protection for HZE particles and other high-LET radiation at the relevant range of low doses.

How safe is safe enough? Radiation risk for a human mission to Mars.

Astronauts on a mission to Mars would be exposed for up to 3 years to galactic cosmic rays (GCR)--made up of high-energy protons and high charge (Z) and energy (E) (HZE) nuclei. GCR exposure rate increases about three times as spacecraft venture out of Earth orbit into deep space where protection of the Earth's magnetosphere and solid body are lost. NASA's radiation standard limits astronaut exposures to a 3% risk of exposure induced death (REID) at the upper 95% confidence interval (CI) of the risk estimate. Fatal cancer risk has been considered the dominant risk for GCR, however recent epidemiological analysis of radiation risks for circulatory diseases allow for predictions of REID for circulatory diseases to be included with cancer risk predictions for space missions. Using NASA's models of risks and uncertainties, we predicted that central estimates for radiation induced mortality and morbidity could exceed 5% and 10% with upper 95% CI near 10% and 20%, respectively for a Mars mission. Additional risks to the central nervous system (CNS) and qualitative differences in the biological effects of GCR compared to terrestrial radiation may significantly increase these estimates, and will require new knowledge to evaluate.

Radiation carcinogenesis risk assessments for never-smokers.

Cigarette smoking, which is presently associated with more than 20% of adult deaths in the United States, is a large confounder to radiation risk estimates derived from epidemiology data. Astronauts and other exposed groups are classified as never-smokers (NS), defined as lifetime use of less than 100 cigarettes. In the past, radiation risk estimates have been made using average U.S. population rates for cancer and all causes of death, which may lead to overestimation of radiation risks for NS. In this report, age- and gender-specific radiation carcinogenesis risk calculations for NS and the average U.S. population are compared. Lung is the major tissue site for smoking and radiation-related cancer. However, other radiogenic cancers where tobacco has been shown to increase population cancer rates are esophagus, oral cavity, salivary gland, bladder, stomach, liver, colorectal, and leukemia. After adjusting U.S. cancer rates to remove smoking effects, radiation risks for lung and other cancers were estimated using the multiplicative risk model and a mixture model, with weighted contributions for additive and multiplicative risk transfer. Radiation mortality risks for NS were reduced compared to the average U.S. population by more than 20% and 50% in the mixture model and multiplicative transfer models, respectively. The authors discuss possible mechanisms of cancer risks from radiation and tobacco that suggest multiplicative effects could occur. These results suggest that improved understanding of possible synergisms between cancer initiators and promoters, such as radiation and tobacco, would greatly improve risk estimates and reduce uncertainties for differentially exposed groups, including NS.

NASCA report 2: Longitudinal study of relationship of exposure to space radiation and risk of lens opacity.

The NASA Study of Cataract in Astronauts (NASCA) was designed to measure the impact of exposure to space radiation on progression rates of cortical, nuclear, and posterior subcapsular cataract in U.S. astronauts who have flown in space and comparison groups of astronauts who had not flown in space, and subjects with a history of military aviation. We present our analyses of 5 years of data with an average of 3.8 exams per subject. All subjects had digital lens images with the Nidek EAS 1000 Lens Imaging System. Because of high variability and skewness of opacity measures, nonparametric methods were used to test for association between rates of opacification and space radiation exposure. First, median regression was used to collapse longitudinal data into robust estimates of progression rates (opacity severity compare to time for each eye of each subject). To quantify and test for a radiation effect, median regression with the dependent variable being the maximum of the two slopes (OD and OS) per subject was then used, adjusting for the confounding variables of age, nutritional, and sun-exposure histories. Median regression showed evidence of an association between the rate of cortical progression in the worse eye with radiation dose and age. The estimated median progression rate from space radiation being 0.25 ± 0.13% lens area/Sv/year (P = 0.062). We found no relationship between radiation exposure and progression of aggregate area of posterior subcapsular cataract or nuclear progression rates. However, longer follow-up may be needed to further understand any impact of space radiation on progression rates for posterior subcapsular cataracts and nuclear cataracts, and to characterize changes to visual acuity.

Updates to astronaut radiation limits: radiation risks for never-smokers.

New epidemiology assessments of the life span study (LSS) of the atomic bomb survivors in Japan and of other exposed cohorts have been made by the U.S. National Academy of Sciences, the United Nations Committee on the Effects of Atomic Radiation, and the Radiation Research Effects Foundation in Japan. The National Aeronautics and Space Administration (NASA) uses a 3% risk of exposure-induced death (REID) as a basis for setting age- and gender-specific dose limits for astronauts. NASA's dose limits originate from the report of the National Council on Radiation Protection and Measurements (NCRP) in the year 2000 based on analysis of older epidemiology data. We compared the results of the recent analysis of the LSS to the earlier risk projections from the NCRP. Using tissue-specific, incidence-based risk transfer from the LSS data to a U.S. population to project REID values leads to higher risk and reduced dose limits for older astronauts (>40 years) compared to earlier models that were based on mortality risk transfer. Because astronauts and many other individuals should be considered as healthy workers, including never-smokers free of lifetime use of tobacco, we considered possible variations in risks and dose limits that would occur due to the reference population used for estimates. After adjusting cancer rates to remove smoking effects, radiation risks for lung and total cancer were estimated using a mixture model, with equal weights for additive and multiplicative transfer, to be 20% and 30% lower for males and females, respectively, for never-smokers compared to the average U.S. population. We recommend age- and gender-specific dose limits based on incidence-based risk transfer for never-smokers that could be used by NASA. Our analysis illustrates that gaining knowledge to improve transfer models, which entail knowledge of cancer initiation and promotion effects, could significantly reduce uncertainties in risk projections.

Analysis of flow cytometry DNA damage response protein activation kinetics after exposure to x rays and high-energy iron nuclei.

We developed a mathematical method to analyze flow cytometry data to describe the kinetics of γ-H2AX and pATF2 phosphorylation in normal human fibroblast cells after exposure to various qualities of low-dose radiation. Previously reported flow cytometry kinetics for these DSB repair phospho-proteins revealed that distributions of intensity were highly skewed, severely limiting the detection of differences in the very low-dose range. Distributional analysis revealed significant differences between control and low-dose samples when distributions were compared using the Kolmogorov-Smirnov test. Differences in radiation quality were found in the distribution shapes and when a nonlinear model was used to relate dose and time to the decay of the mean ratio of phospho-protein intensities of irradiated samples to controls. We analyzed cell cycle phase- and radiation quality-dependent characteristic repair times and residual phospho-protein levels with these methods. Characteristic repair times for γ-H2AX were higher after exposure to iron nuclei compared to X rays in G(1) cells and in S/G(2) cells. The RBE in G(1) cells for iron nuclei relative to X rays for γ-H2AX was 2.1 ± 0.6 and 5.0 ± 3.5 at 2 and 24 h after irradiation, respectively. For pATF2, a saturation effect was observed with reduced expression at high doses, especially for iron nuclei, with much slower characteristic repair times (>7 h) compared to X rays. RBEs for pATF2 were 0.7 ± 0.1 and 1.7 ± 0.5 at 2 and 24 h, respectively. Significant differences in γ-H2AX and pATF2 levels when irradiated samples were compared to controls were noted even at the lowest dose analyzed (0.05 Gy). These results show that mathematical models can be applied to flow cytometry data to identify important and subtle differences after exposure to various qualities of low-dose radiation.

Non-targeted effects and the dose response for heavy ion tumor induction.

Non-targeted effects (NTE), including bystander effects in neighbor cells of cells directly hit by radiation tracks and genomic instability in the progeny of irradiated cells, challenge traditional radiation protection paradigms on Earth. It is thus of interest to understand how NTE could impact our understanding of cancer risks from galactic cosmic rays (GCR), which are comprised of high-energy protons and heavy ions. The most comprehensive data set for tumor induction by heavy ions is the induction of Harderian gland tumors in mice by high-energy protons, helium, neon, iron and niobium ions after doses of 0.05 to several Gy. We report on an analysis of these data that compares a dose response model motivated by the conventional targeted effects (TE) model to one which includes a dose response term descriptive of non-targeted effects (NTE) in cell culture. Results show that a NTE model provides an improved fit to the Harderian gland data over the TE model. Relative biological effectiveness (RBE) factors are shown to have much larger values at low doses based on a NTE model than the maximum RBE estimates based on estimates of the ratio of initial linear slopes of heavy ions compared to gamma-rays in the TE model. Our analysis provides important in vivo support for the deviation from linear dose responses at low doses for high LET radiation, which are best explained by a NTE model.