The four categories of risks from radiation in space are defined by the NASA Bioastronautics Roadmap (BR). They are: 1) Carcinogenesis, 2) Acute and late effects to the Central Nervous System (CNS), 3) Degenerative Tissue Effects such as heart disease and cataracts, and 4) Acute Radiation risks. The number of safe days currently predicted for an astronaut’s career is less than required by mission planning, due to the large uncertainties in risk prediction. In particular, a projection uncertainty below + or - 50% is the goal for the 1000-day Mars mission because the high level of risk will require high precision risk evaluations. The current approach used to project risk is based on epidemiology data and on phenomenological models used to derive risk prediction from them. This approach cannot lead to improvements in the accuracy of risk prediction beyond a factor of approximately 2. New approaches using molecular biology and genetics are the only viable ones for achieving the level of accuracy required by space exploration and a robust program to obtain the required data is supported by the Space Radiation Program. However, how to incorporate these data into risk prediction and assessment models is not well understood. This Project Plan describes the approaches that will be used to develop models of risk assessment based on mechanistic space radiobiology research funded by the Space Radiation Program, leading to incremental uncertainty reduction based on new experimental data, and to the development of application software to be used in the NASA operational radiation protection program. To accomplish these goals, we will establish new molecular based models of risk. The molecular pathways that are the hallmarks of genomic instability and cancer, and the perturbation of these pathways by radiation will be described using systems biology approaches and Monte-Carlo simulation. We will develop descriptive models of such pathways utilizing track structure models of biomolecular damage, and deterministic and stochastic kinetic models of dominant molecular pathways causative of BR radiation risks. These simulations will make maximum use of results from mechanistic space radiobiology, and will replace traditional hazard functions and their inherent uncertainties due to reliance on epidemiological or phenomenological approaches.

Models of cancer, CNS, heart disease, acute and other risks developed by the Space Radiation Risk Assessment Project provide NASA with the ability to project risks and develop cost-effective mitigation approaches for future exploration missions.

Our recent focus is the confounding role of tobacco on cancer and circulatory disease risk estimates. Understanding the effects of tobacco usage on radiation risk estimates benefits ground based use of diagnostic procedures that utilize radiation.

Project B: Cancer Risk Projection Model and Uncertainties: New findings and knowledge from NSRL and other sources were used to revise the NASA’s risk model for space radiation cancer risks:

1) Revised values for low LET risk coefficients for tissue specific cancer incidence.

2) An analysis of lung cancer and other smoking attributable cancer risks for never-smokers show significantly reduced lung cancer risks as well as overall cancer risks for astronauts as compared to the risk estimated for the average U.S. population.

3) Derivation of track structure based radiation quality functions that depend on charge number, Z, and kinetic energy, E, in place of a dependence on LET alone. The assignment of a smaller maximum in the quality function for leukemia than for solid cancers.

4) Revised uncertainty assessments for all model coefficients in the risk model (physics, low LET risk coefficients, dose and dose-rate effectiveness factor (DDREF), and quality factors), and an alternative uncertainty assessment that considers deviation from linear responses due to non-targeted effects (NTE).

Results of calculations for the average U.S. population show more restrictive dose limits for astronauts above age 40 y as compared to National Council on Radiation Protection and Measurements (NCRP) Report 132, and a modest narrowing of uncertainties if NTEs are not included and much broader uncertainties with NTEs. Risks for never-smokers compared to the average U.S. population are estimated in a mixture model to be reduced by more than 20% and 30% for males and females, respectively. A larger reduction is possible if purely multiplicative risk transfer is assumed.

Project C: Biochemical Kinetics Models of Molecular Pathways: A system biology model (Cucinotta et al., 2008) of the non-homologous end joining (NHEJ) pathway was developed and used to make quantitative descriptions of the gamma H2AX foci and double strand break (DSB) rejoining experiments. The model is extended to consider the radiation quality dependence of the relative fraction of simple and complex DSB, rejoining and associated repair defects, and the kinetics of various radiation induced foci (RIF). In further work, the addition of ataxia telangiectasia mutated (ATM) and the MRN complex to the model was achieved and the role of processing damaged ends by the Artemis proteins is being modeled (Li and Cucinotta, 2011). The interaction of several growth factors with NHEJ components was studied, including the interaction of the growth factors EGFR, IGF1, and TGFbeta-Smad with ATM and DNA-PK. New approaches to Green’s functions for stochastic treatment of molecular diffusion processes were developed (Plante et al. 2011). Flow cytometry or immune-staining considers signals in individual cells and thus provides several unique capabilities to support computation modeling using stochastic approaches. In contrast, methods that average the values of many cells such as Western blots, gene arrays, etc. are limited in elucidating events at low dose where fluctuations are expected to be important. To improve our understanding of DNA repair complexes numerical approaches to simulate immunohistochemistry (Ponomarev et al., 2008, 2009) and flow cytometry experiments (Cucinotta , in preparation, Chappell et al., 2010) were developed. These models embed a basic understanding of track structure with statistical approaches of flow cytometry data sorted by cell cycle phase, and fluorescence intensity taking into account background levels. Following flow cytometry analysis, we were able to distinguish the kinetics of these phospho-proteins in relationship to the cell cycle and to each other in an individual cell. Results revealed a unique pattern of kinetics for high vs low LET radiation, with a failure to initiate full activation of the ATM pathway being evident following High LET exposure. In the process of these studies we have noted that different populations of cells making up normal human tissues can be sorted based on intrinsic qualities of the cells and differences in their radiation sensitivity have been noted. In addition, an increase in proliferation of a specific mammary cell population was observed with low doses of radiation.

Project D: DNA Damage in Cancer Initiation and Genomic Instability: Foci size and clustering including strings of foci along high-Z high-energy (HZE) tracks, were modeled for the first time and provide a useful analysis tool of NSRL experiments on DNA damage foci. These models have been extended to predict chromosomal aberration formation including the distribution of small rings normally below detection levels with fluorescence in-situ hybridization (FISH) (<5 Mbp), and to describe complex aberrations. Stochastic track structure models were combined with a human genome model that considers random walk polymer models of each chromosomes pair built from 2 kbp monomers and constrained to nuclear territories in interphase (Ponomarev et al., 2007, 2009).

Project E: Acute Radiation Risk Models: We extended the granulopoietic model for rodents for the prediction of solar particle effects in canines, non-human primates, and humans (Hu and Cucinotta, 2011). An important feature of this approach is that the dynamics of cell populations are described by non-linear differential equations. Validation of the model was achieved by comparison to comprehensive data sets for animals exposed to acute and chronic radiation and to the Chernobyl and other radiation accident victims.

The four categories of risks from radiation in space are defined by the NASA Bioastronautics Roadmap (BR). They are: 1) Carcinogenesis, 2) Acute and late effects to the Central Nervous System (CNS), 3) Degenerative Tissue Effects such as heart disease and cataracts, and 4) Acute Radiation risks. The number of safe days currently predicted for an astronaut’s career is less than required by mission planning, due to the large uncertainties in risk prediction. In particular, a projection uncertainty below + or - 50% is the goal for the 1000-day Mars mission because the high level of risk will require high precision risk evaluations. The current approach used to project risk is based on epidemiology data and on phenomenological models used to derive risk prediction from them. This approach cannot lead to improvements in the accuracy of risk prediction beyond a factor of approximately 2. New approaches using molecular biology and genetics are the only viable ones for achieving the level of accuracy required by space exploration and a robust program to obtain the required data is supported by the Space Radiation Program. However, how to incorporate these data into risk prediction and assessment models is not well understood. This Project Plan describes the approaches that will be used to develop models of risk assessment based on mechanistic space radiobiology research funded by the Space Radiation Program, leading to incremental uncertainty reduction based on new experimental data, and to the development of application software to be used in the NASA operational radiation protection program. To accomplish these goals, we will establish new molecular based models of risk. The molecular pathways that are the hallmarks of genomic instability and cancer, and the perturbation of these pathways by radiation will be described using systems biology approaches and Monte-Carlo simulation. We will develop descriptive models of such pathways utilizing track structure models of biomolecular damage, and deterministic and stochastic kinetic models of dominant molecular pathways causative of BR radiation risks. These simulations will make maximum use of results from mechanistic space radiobiology, and will replace traditional hazard functions and their inherent uncertainties due to reliance on epidemiological or phenomenological approaches.

Models of cancer, CNS, heart disease, acute and other risks developed by the Space Radiation Risk Assessment Project provide NASA with the ability to project risks and develop cost-effective mitigation approaches for future exploration missions.

Cancer risk model updates were completed including: comparison of the BEIR VII, UNSCEAR and NCRP models, development of new tissue weighting factors appropriate for astronauts, preliminary recommendations of new solid cancer and leukemia specific quality factors, and new developments of dose response models for non-targeted effects for heavy ion carcinogenesis. The non-targeted effects model suggest much higher cancer risks at low doses of heavy ions than the conventional model recommended by the NCRP. New probability distribution functions (PDF) to represent the uncertainties in new and existing data sets and knowledge of radiation quality effects, dose-rate modifiers, and the dependecies of risk on age at exposure are being considered from these analyses. Heart disease risk estimates were developed using a triple detriment approach based on US population data for coronary heart disease and stroke and recent meta-analysis of excess relative risks in exposed cohorts. Values of RBEs and dose-rate modifiers for heart disease are being investigated, and suggest a wide range of possibilities for the mortality risks from heart disease with results ranging from negligible to radiation cancer risks to approaching ½ of the cancer risk of males of above age 45 y at the time of exposure. A mathematical representation of the two-hit model of Alzheimer’s disease was developed. The model can be used to estimate a shift in the age of appearance of Alzheimer’s disease based on the number of neuronal cells with radiation induced hits.

Systems biology model developments included new molecular dynamic calculations of the Ku70/80 protein, RecA, and preliminary results for the RPA and Smad protein interactions with DNA. Computational tools to assess DNA repair foci and flow cytometry data for signal transduction phospho-proteins were developed. New flow based assays for dually staining proteins, cell population kinetics and telomere shortening were considered. A stochastic model of ligand-receptor binding to model TGFbeta activation was developed using a Green’s function approach.

A non-linear model of granulocytopoiesis following acute or chronic irradiation was compared to experimental data for several mammalian species and extended to human exposures including blood level kinetics following exposures to solar particle events. A probabilistic approach to solar particle event organ dose evaluation and shielding design was developed. These models will form the basis for ARRBOD V2.0

Models of cancer, acute and other risks developed by the Space Radiation Risk Assessment Project provide NASA with the ability to project risks and develop cost-effective mitigation approaches for future exploration missions.

The NASA Cancer risk projection model was further developed to estimate organ specific cancer risks. A review of NASA Space Radiation sponsored publications, or other data on RBE's for solid cancer and leukemia including surrogate markers is underway to formulate organ specific quality factors. The anatomical human geometry model, CAMERA was compared to spaceflight agreement with overall consistency of <15% found for Space Shuttle and ISS results. A new CT-based VOXEL model was compared to the CAMERA model and agreed very well (<5% difference) for GCR and hard SPE spectra. The effects of background cancer rates and survival curves were estimated through comparison of space radiation cancer risks for each of the 50 states using multiplicative and weighted multiplicative and additive risk transfer models. System biology models were developed for the ATM and TGF-beta pathways. Mechanisms that could lead to differences between signaling between low and high dose or LET were identified and will be further investigated. Of interest are the biochemical description of non-targeted effects through inter-cellular signaling. Also new computational tools to understand DNA damage repair foci and flow cytometry results with heavy ion beams were developed. Work on molecular dynamics simulations of the Ku and Rad51 proteins interacting with DNA were completed and identified likely binding regions on these proteins, and predictions of binding energies were made. ITC and FRET methods are being considered to verify computational predictions.

A Monte-Carlo track structure code, RITRACK was developed for relativistic heavy ion descriptions. The code includes the production and energy deposition of secondary electrons (delta-rays) with energies up to 10 MeV, and ions with energies from 0.1 MeV/u to 50 GeV/u. Track structure models of chromosomal aberrations using the random walk polymer models of the entire human genome were formulated and show good agreement with experiments.

Probabilistic risk assessment (PRA) tools were developed for SPE risks including the Acute radiation syndromes. A data base for energy spectra of all SPE's in the space era was developed and Hazard functions and sampling schemes formulated to consider shielding designs under PRA. This is the first PRA tool developed that considers SPE frequency of occurence, event size, and event spectral characteristics. The results will be very informative to the NASA Constellation program and the assessment of shielding approaches and mission architectures.

Models of cancer, acute and other risks developed by the Space Radiation Risk Assessment Project provide NASA with the ability to project risks and develop cost-effective mitigation approaches for future exploration missions.

Progress was made in developing systems biology models of the double strand break repair (DSB). In mammalian cells there are two mechanisms of DNA double strand break repair: Non-homologous end-joining (NHEJ) and homologous recombination (HR). The error-prone NHEJ is the main mechanism in resting cells and the G1 phase of the cell cycle. A systems biology model (Cucinotta et al., in press) of the NHEJ pathway was developed and used to make quantitative descriptions of the gamma-H2AX foci and DSB rejoining experiments. The model includes a kinetics description of several DNA repair proteins including the Ku70/80 hetero-dimer, the catalytic sub-unit of the DNA-PK repair complex denoted DNA-PKcs, and the Ligase-IV/XRCC4 complex. The regulation of DNA-PKcs by autophosphorylation for simple and complex DSB was described. The model is being extended to consider the radiation quality dependence of the relative fraction of simple and complex DSB, rejoining and associated repair defects, and the kinetics of various radiation induced foci (RIF). In further work, the addition of ATM and the MRN complex to the model was achieved and the role of processing of damaged ends by the Artemis and WRN proteins are being modeled

For biophysical understanding of sub-tissue structures a description of cell size and shape, and geometry of multiple cell lineages is needed. A “tissue box” model (Ponomarev and Cucinotta, 2006) was developed to represent heterogeneous tissue architecture and to score HZE tracks and nuclear reactions in 2D and 3D tissues. Accurate segmentation of imaging data from 2D or 3D from a variety of imaging methodologies is possible. The tissue box model will be applied to represent experimental models used by SRP funded investigators. Nuclear reactions were shown to be rare in a small tissue sample (<100 cubic-micron), however the possibility of a large imparted dose at such sites is a continuing concern.

A random walk polymer model of whole chromosomes was extended to include description of all human chromosomes within a typical cellular volume and to predict the role of DNA loops and attachment points on the spatial distributions of DSB (Ponomarev and Cucinotta, 2006) To overcome the background DSB inherent in experimental methodologies a subtraction technique was formulated and applied to several data sets in collaboration with the Tufts U. NSCOR (PI. L. Hlatky) (Ponomarev et al., 2007). Also these models were used in an imaging approach to study DNA repair foci (Costes et al., 2007) in collaboration with the LBNL NSCOR (PI. M. Barcellos-Hoff). A more detailed model of DNA damage and mutation using Monte-Carlo scoring of energy deposition in atomistic models of DNA and chromosomes is also under development (Nikjoo et al. 2007).

Work on the transmission of specific chromosomal aberrations in subsequent cell cycles was initiated using basic cytogenetic theories. A modeling project to predict the initial yield of terminal and interstitial deletions was begun. Differences between normal cells and specific DNA repair defects are also under study and the role of longer times for open breaks in ATM and MRN deficient cells will be studied to consider the increases in overall and specific types of aberrations and the potential impacts on transmission frequencies.

New statistical models of the probability of SPE frequency and size were developed using our data base of historical SPE and solar cycles (Kim et al., 2006, 2007). The model was extended back to the 15th century for the >30 MeV proton fluences using ice-core data on nitrate concentrations normalized to modern events as reported by McCracken and collaborators. The time dependence of dose-rate for the 30 largest SPE’s was also evaluated. These studies indicate that acute radiation risks will only occur with a realistic probability under EVA conditions, and therefore the focus of acute risk models should be on the so-called prodromal risks (nausea, vomiting, fatigue, etc.) that may occur during EVA in deep-space or on the lunar surface. The DoD based, RIPD model was adapted for a description of risks from SPE. A computer code of the model was developed ab-initio at JSC. This model uses a logistic scale to assign performance degradation probabilities and time-courses for the acute risks. Using the BRYNTRN code and initial estimates of RBEs from the scientific literature, application of the model to the 1972 SPE for EVA conditions was made. (Hu et al., 2007 and in preparation). Work in collaboration with Dr. Smirnova of Moscow State U. was begun to study the probabilities of mortality from one or more, large SPE’s, including the role of an adaptive response due to simulation of granulocytes by a first SPE, leading to protection against a second SPE within the ensuing next few months.