WP4

Task 4.1 Interaction between radon and smoking in lung cancer

Leader: SURO, partners: IRSN, BfS, ISS, WHO (M1-M60)

Smoking is the main risk factor of lung cancer, and the assessment of its impact on the risk associated to radon is a major issue. This task will analyse the interaction between radon and smoking based on both updated data from cohorts of uranium miners (T4.1.1) and on updated data from residential studies (T4.1.2), using elaborated statistical approaches. Finally, trends in radon exposure and smoking habits in Europe will be studied and their impact on public health will be predicted (T4.1.3). Task 4.1 will be conducted in close collaboration with Task 4.6, and will contribute directly to WP8. The interaction between radon and smoking will be studied in several cohorts of uranium miners where, in addition to radon exposure, there is information on smoking (Czech, France, Colorado, New Mexico). Risk models for combined effect of smoking and radon will be developed, including modifying effects of age, time since exposure and exposure rate and the additive, multiplicative, or geometric mixture models. Analyses will include estimation of model parameters in separate cohorts, and in the joint cohort. Several epidemiological studies on residential radon and lung cancer collected information also on smoking habits. The results of these studies are consistent with both a multiplicative and a more-than-additive interaction between radon and smoking. These studies will be reviewed, and where possible updated, to summarize the evaluation of interaction between smoking and radon in residential studies and to compare such evaluation with results from the miner studies. The fraction of lung cancer attributable to radon in a population depends on the prevalence of smoking and on the interaction between radon and smoking. Changes in smoking prevalence may lead to a decrease in lung cancer rates in the future. A simulation study will be performed to assess the impact of trends in radon exposure and smoking habits on radon risk in several European countries, using different indicators: attributable fraction, lifetime risk and dose conversion coefficients.

The Task will be split into 2 subtasks:

Subtask 4.1.1: Interaction between radon and smoking in lung cancer in miner studies (in close collaboration with the PUMA project)

Subtask 4.1.2: Interaction between radon and smoking in lung cancer in indoor studies

Task 4.2 Radon-related risks other than lung cancer among adults

Leader: BfS, partners: IRSN, INSERM (M1-M60)

Given that absorbed doses from inhaled radon progeny to organs other than lung tend to be substantially lower than absorbed doses to the lung, it is expected that if there is an excess, the excess is small and large studies will be needed to detect such associations. Even a small excess risk, would have high implications for radiation protection, because large populations are exposed and current radon dose conversion factors would need revision. Data from the PUMA (pooled uranium miner cohort study) and French Constances cohort (www.constances.fr) will be used to investigate the radon-related risk.

The task will be split into 3 subtasks:

Subtask 4.2.1. Risk of death from cancers other than lung cancer in PUMA

Subtask 4.2.2. Risk of death from cardiovascular diseases in PUMA

Subtask 4.2.3. Risk of disease incidence in the Constances cohort

Task 4.3 Studies on the association of radon and childhood leukemia and brain cancer

Leader: STUK, other partners: TAU, IRSN, INSERM, NMBU, NIPH, DSA, ISPM, DCS (M1-M60)

In Task 4.3 epidemiological studies are conducted for risk assessment of residential radon, with case-control and cohort studies using radon measurements and/or predicted radon levels based on modelling housing and area characteristics. Finally, a pooled analysis of previous studies and studies to be conducted within the WP will be performed. This will contribute to better characterisation of cancer risk from radon exposure in childhood and risk of other cancers than lung.

Nationwide case-control studies in Finland with 500 cases of childhood leukemia (precursor B-cell ALL) diagnosed in 2000-2017 at ages 2-6 years and 650 childhood brain tumors at ages 2-15 years, with 1150 controls will be conducted with measurements of residential radon in the past dwellings of the participants. The mean radon concentration in Finland is 100 Bq/m3, and the relatively high exposure levels increase the statistical power.

Two nationwide case-control studies of childhood cancer have been established in France with individual questionnaire data. The ESCALE study includes 1466 cases (764 leukemias, 209 malignant CNS tumors) diagnosed in 2003-2004 and 1681 population controls. The ESTELLE study covers 1814 cases, including 740 leukemias, 339 CNS tumors diagnosed in 2010-2011 and 1423 controls. Natural background radiation (NBR) exposure levels (radon concentration and gamma radiation) in France are estimated on a fine scale from geological map information and radiation measurements. By pooling ESCALE and ESTELLE data, RadoNorm will consider the cumulative exposure to NBR from birth to diagnosis in relation to childhood leukemia and childhood CNS tumor risks.

The objective of a third analysis in task 4.3 is to link the spatial-temporal development of childhood leukemia in Norway to situations where parents or children from all over Norway (MoBa, The Norwegian Mother and Child Cohort Study) have been exposed to a combination of stressors including radiation that potentially could trigger the disease. The work will focus on 3 aspects linking individual exposure agents to the disease, before the multiple stressor approach is applied: 1) linking the leukemia map to databases on the indoor/outdoor radon exposure, 2) linking the leukemia maps to the database on deposition of Chernobyl fallout in Norway, and 3) linking the leukemia map to the antioxidant status of newborn children. MoBa is a prospective population-based pregnancy cohort study, including 114,500 children and 95,200 mothers recruited from all over Norway in the period 1999 to 2008. Thus, a multiple stressor concept will be applied to assess the combined risk of leukemia.

A nationwide cohort study (Swiss National Cohort, SNC) including all children aged 0-15 years registered in national censuses in 1990, 2000, and annual censuses 2010-2017 was performed in Switzerland. Anticipated numbers of cases are 3,500 for any cancer, 1070 for leukaemia, and 770 for central nervous system (CNS) tumours. New geographic exposure models will be developed for indoor radon and terrestrial gamma radiation based on about 200,000 radon measurements and extensive air-borne gamma ray spectrometry covering a large part of the country.

A Pooled case-control study of childhood leukaemia and CNS tumours and exposure to radon and natural background radiation using nation-wide datasets from France, Finland, Norway, Denmark and Switzerland is planned. Study population comprises children aged 0-14 years included in previous and new (above) studies from each country. Outcomes will be all leukaemias, ALL, precursor B-cell ALL (and important cytogenetic subtypes including ETV6-RUNX1 and high hyperdiploidy), AML, central nervous system tumours with important subgroups. Exposures include indoor radon concentration, dose rates from terrestrial gamma and cosmic radiation, and cumulative RBM dose. Separate analyses will investigate exposures based on addresses at birth and at diagnosis and on full residential history.

The different epidemiological analysis of task 4.3 are separated into 5 subtasks:

Subtask 4.3.1 Register-based studies on the association of radon and childhood leukemia and brain cancers in Finland (A RadoNorm PhD project).

Subtask 4.3.2 Childhood cancer (leukemia and central nervous system (CNS) tumor) and natural radioactivity in France

Subtask 4.3.3 Linking childhood leukemia to multiple stressors and poor antioxidant status in Norway

Subtask 4.3.4 Association of radon and childhood leukemia and brain cancers in Switzerland

Subtask 4.3.5 Joint analysis of national datasets

Task 4.4 Uncertainties in radon risk assessment due to thoron

Leader: PHE, other partners: ISS, RIVM, IRSN, STUK (M1-M60)

The objective is to assess the uncertainty of radon risk assessment due to the interference of thoron (220Rn) with radon (222Rn) gas measurements. Typically, for indoor exposures, the activity concentration of thoron and its progeny are low compared with that of radon and its progeny. However, thoron exposure can be significant in some specific environments. Consequently, thoron can be an important source of radon gas measurement uncertainty if no distinction is made between thoron and radon, and if the radon detector is sensitive to thoron. Detectors that are sensitive to thoron can over-estimate radon exposure if thoron is present and will therefore result in an under-estimation of the risk. Radon gas detectors used in epidemiological studies will be assessed for their sensitivity of thoron. Dose calculations will also be performed to compare doses from radon and radon progeny with that from thoron and thoron progeny. Lung and organ doses such as to the brain and red bone marrow will be performed (Task 4.3). This will be used to assess the uncertainty of radon risk assessment due to thoron interference.

Task 4.5 Mechanisms of radiation action in lung cancer among never smokers

Leader: IGR, other partners: EORTC, IDIBAPS, IRSN, BfS, LUMC, PTB, EK (M1-M60)

This Task addresses molecular changes in cancer pathways of non-small lung cancers (NSCLC) among humans and rats exposed to radon and studies the effects of inhomogenous dose distribution using organotypic models of human bronchial epithelium. Radon is the leading cause of lung cancer in non-smokers, but the carcinogenesis mechanism remains unknown. Over the last years, several oncogenic molecular alterations have been described in non-smokers, comprising somatic mutations (EGFR, BRAF, etc.) and chromosomal rearrangements (ALK, ROS1, etc.), however, no specific alteration with a strong relation to radon exposure has been identified yet. The hypotheses for the study are: (i) the driver molecular alterations in NSCLC are associated with the cellular and molecular damage induced by radon, 2) radon gas can induce common genomic alterations in cancer pathways in rats and humans that can be identified in a radon-associated signature.

A retrospective study of archived pathology slides from the CEA’s experiments of radon exposure performed in rats between 1972 and 1992 for identifying the molecular phenotype of lung cancer induced by radon will be conducted. The activity will focus intially on the molecular assessment if the quality control of the samples is ensured. Then, based on the data generated, the actiovity will design a radon-associated signature using computational genomics in order to characterise the radon-related cancer pathways in rats, and to establish a hypothesis on radon carcinogenesis in humans.

In a second activity within the BioRadon study, a European multicenter observational prospective study, the correlation of the molecular phenotype with indoor radon exposure in NSCLC patients will be analysed (sponsor: EORTC). Overall, 1323 NSCLC patients will be enrolled. After the primary (mutational) endpoint analyses of the groups (driver-mutation group, driver-fusion group, WT-control group and other control group), correlations between molecular phenotype and radon concentrations will be assessed. In addition, we will compare the genetic alterations found with the molecular phenotype in animal (rats) and humans exposed at work (miners). Exploratory studies will also be carried out to assess the link between the germline, somatic DNA repair and others GAs and the indoor radon, in order to define a radon-associated signature in humans. IRSN will act as an expert regarding the reconstruction of cumulated radon exposure based on the alpha-track measurements and the European Radon Map (Joint Research Center, European Commission). This activity will be linked with Sherlock lung – Tracing Lung Cancer Mutational Processes in Never Smokers Study (by NCI-DCEG). This collaboration between Europe and US will foster knowledge in lung cancer etiology in never smokers. A joint meeting is planned to discuss the prospects for studies on radon as well as materials and data exchange with the European biobanks and EORTC.

A further activity is the retrospective assessment of the presence of driver molecular alterations in lung cancer in German uranium miners of the SAG/SDAG Wismut exposed to radon. Biological material originates from the histopathological archive of the former Institute of Pathology of the health service of the Wismut company consisting of biological material from 28,975 autopsy cases collected from 1957 to 1994. BfS has a biobanking with isolated DNA and RNA from FFPE blocks of more than 600 cases.

This subtask will characterise the molecular phenotype of a selected group of adenocarcinoma lung cancer specimens from this cohort, with smoking habit known and quality suitable for molecular testing. Overall, the population will be stratified in two subgroups: high exposure (if >300 WLM, n=15) vs. low exposure (if <50 WLM, n=15). The genetic alterations found will be compared with the molecular phenotype in animal (rats) and humans with indoor exposure (BioRadon Study). Inhalation of radon leads to an inhomogeneous exposure of the lungs in which the epithelial layer lining of bronchial bifurcations receives the highest dose. The aims of an experimental activity are (i) to test the hypothesis that locally high doses result in a transient increase of basal cell number and local accumulation of genotoxic damage, and (ii) to get detailed insight in the biological responses of cells present in a relevant complex tissue environment. For this purpose, cellular and tissue effects of homogeneous and heterogeneous exposures will be compared utilizing organotypic lung models derived from primary human bronchial epithelial cells (PBECs). Absorbed doses and dose distributions (homogeneous and corresponding to in vivo exposures) will be calculated in WP3 and will be delivered using well defined microbeam exposures performed at PTB. The potential cell type-specific radiosensitivity will be studied and the differential transcriptional responses of the individual cell types will be investigated using single cell RNAseq. This task is split into 4 subtasks:

Subtask 4.5.1 Molecular signature of lung cancers in rats exposed to radon

Subtask 4.5.2 Molecular characterization of lung cancer in patients exposed to radon (1920-LCG Bioradon prospective study) (A RadoNorm PostDoc project)

Subtask 4.5.3 Molecular characterization of lung cancer in miners exposed to radon (German Uranium mine’s cohort)

Subtask 4.5.4 Comparing the biological consequences of homogeneous and heterogeneous exposures using organotypic models of human bronchial epithelium

Task 4.6 Consideration of various sources of uncertainties in radon induced lung cancer risk inference

Leader: IRSN, other partners: BfS, HMGU, ISS, SURO, LMU, UPD, EK (M1-M60)

Aim of the task is to propose and fit Bayesian hierarchical models to refine the estimation of the risk of death by lung cancer due to low levels of chronic exposure to radon, by considering the cohorts of French, German and Czech uranium miners separately and jointly. As part of epidemiological studies, the treatment of both complex patterns of exposure measurement errors (shared/unshared, heteroscedastic, classical/Berkson) and other dosimetric uncertainties inherent to a simplified but plausible analytical expression of the human respiratory tract model will be improved in the estimation of the dose-response relationship of interest.

Biologically-based concepts of lung carcinogenesis are implemented into mechanistic risk models. This subtask integrates results from WP4 tasks concerned with molecular analysis to improve causality assessment and risk quantification for lung cancer caused by exposure to radon. Newly developed model will inspired by the results of genetic alterations after radon exposure. They are tested with both historic rat data sets (CEA) and human cohorts (French and Czech Uranium miners). The approach to implement distinct evolutionary lung cancer pathways into mechanistic risk models will be linked to the framework of Adverse Outcome Pathways (AOPs) which has been developed in a toxicological setting. The preferred model will be used in the comparison of dose-response relationships with regular regression models. Should robust molecular radon signatures be identified, they will strongly influence retrospective risk assessment.

Aim of a third activity is to explore the impact on lung cancer lifetime risk estimates of different scenarios of exposure (duration, age at exposure, time since exposure…), populations (European, Asian, American), models and risk coefficients (uranium miner models, indoor radon models).

The Task is split into 3 Subtasks:

Subtask 4.6.1 Establishing Bayesian hierarchical models for lung cancer risk analysis in human and animal

Subtask 4.6.2 Biologically based models of lung carcinogenesis:

Subtask 4.6.3 Analysis of lung cancer lifetime risk estimates:

Task 4.7 Risks from radon and NORM in drinking water

Leader: STUK, other partners: NMBU, DSA, UEF (M1-M60)

Few studies have evaluated the risks from ingestion of naturally occurring radionuclides such as radon, uranium and radium. Very high concentration can occur in groundwater, but radionuclides are largely eliminated from tap water in the processing at waterworks. Meaningful exposures occur, nevertheless, in private wells drilled in the bedrock, as they rarely employ sophisticated removal systems. For radon, activity concentrations can reach several hundred Bq/l, but substantially lower for other radionuclides (<1 Bq/l for radium and <0.1 Bq/l for uranium isotopes. Small case-cohort studies in Finland analysed water concentrations of major radionuclides in subjects using water from private wells drilled in bedrock with cumulative organ doses of the order of 5 mSv (0.2-0.4 mSv per year) failed to show increases in risks of leukemia, or stomach, bladder or kidney cancers (Auvinen 2002, 2005, Kurttio 2006). A small Canadian study showed non- conclusive results from bone cancer in children and adolescents (Finkelstein). Several ecological studies with poor exposure assessment have shown conflicting results, but they are not highly informative for risk assessment. Further studies addressing the health risks from ingested radon and NORM and combined effects of radiation and chemical toxicity are needed. These studies should come with reliable exposure assessment and cover a wide range of exposure levels. PhD project on Radon in drinking water (NMBU, DSA)

Task 4.8 Effects and mechanisms of action of combined exposures to radon or NORM and other stressors relevant of true exposure situations of humans and biota

Leader: NMBU, other partners: SU, IRSN, DSA, NIPH, EK (M1-M60)

The objective of Task 4.8 is to improve knowledge on transfer and effects, including underlying mechanisms, of combined exposures to radon or NORM and other stressors. Relevant exposure situations for humans are exposure to radon combined with tobacco smoke or nanoparticles via inhalation, however, the mechanisms are not well understood. For organisms exposed to NORM combined with metals and particles, ecosystem transfer (uptake, accumulation) depends on exposure routes and interactions between stressors (toxicokinetics), while effects depend on competition between stressors to interact with key biological targets (toxicodynamics). This task includes the linking of exposure to uptake, internal distributions and responses to understand why combined exposures could depart from additivity along the source – adverse outcome pathway.

The aim of a first activity is to carry out in vitro experiments using normal bronchial epithelial cells on the mechanisms and levels of interaction between radon, cigarette smoking and nanoparticles. Experiments will be carried out with an air liquid interface (ALI) exposure system, where airborne nanoparticles (carbon, uranium, etc), cigarette smoke and radon gas can be directly deposited on cell cultures. Cells will be exposed to single factors and various combinations. Dose response relationships will be measured allowing estimation of the mode of interaction by the envelope of additivity approach. The following endpoints will be analysed: cell viability (MTT assay and trypan blue counting), clonogenic survival assay, apoptosis (cleaved PARP/Caspase 3) and senescence (beta-galactosidase staining). Inflammation can increase the risk of adverse health effects, therefore proinflammatory markers IFN, IL-6 and IL-1β will be examined. The possible presence of DNA double strand breaks will be assayed by γH2AX analysis (immunofluorescence) and/or by alkaline comet assay and cytogenetic methods (chromosomal aberrations, micronuclei). Epigenetic modifications commonly appear prior to mutations in the process of carcinogenesis and histone modifications and promoter-specific methylation of target genes will be assessed.

The objective of a second activity is to investigate how multiple stressor exposure could affect toxicokinetics and toxicodynamics, and why combined exposures could depart from additivity. A key issue is to compare exposure routes via dietary intake and via lungs (inhalations/intratracheal instillatio), and to identify differences in accumulations and uneven internal distributions. Four different laboratory experiments will be performed using 3 different test organisms, (i) mixed NORM (U, Ra, Po-210) and NORM with metals (Cd) and nanoparticles (silicate, iron) exposures of terrestrial organisms with the reporter C. elegans, (ii) mixed exposure of aquatic organisms: with reporter Daphnia Magna, (iii) mixed NORM exposure of reporter mice via drinking water, and (iv) mixed NORM exposure of reporter mice via lungs.

The task is split into 2 subtasks:

• Subtask 4.8.1 Effects of combined exposure of human bronchial epithelial cells to radon, tobacco smoke and nanoparticles

The aim of the PhD project is to identify biomarkers of exposure to radon in combination with other stressors.
Experiments on human cells will focus on normal bronchial epithelial BEAS-2B cells exposed using an air liquid interface (ALI) exposure system to radon gas, cigarette smoke and various naturally and occupationally relevant nanoparticles. Experiments with non-human cells will be carried out on the effect of radon and nanoparticles in collaboration with the CERAD CoE in Norway. Dosimetric information will be derived from modelling carried out in WP 3. DNA damage induced by various combination of the stressors will be assessed by cytogenetic methods, gamma H2AX focus assay and level of DNA responsive gene expression. Qualitative and quantitative analysis of the endpoint will show if different stressor combinations induce a fingerprint of exposure that can be used as a future biomarker. Subtask 4.8.1 will include a RadoNorm PhD project: Biomarkers of exposure to multiple stressor in human and non-human species

• Subtask 4.8.2 Effects of multiple stressor exposures from NORM (U, Ra, Po-210), metals (Cd) and nanoparticles (silicate, iron) of selected organisms.

Task 4.9 Assessment of combined toxicity and cumulative risk

Leader: IRSN, other partners: NMBU, SU, IRSN, UPD, SCK•CEN (M1-M60)

The aim is to characterise the impact of radiation on humans and non-human biota in combination with other stressors under different environmental conditions, and to assess the cumulative risk. The main objectives are to: 1) establish a thoroughly evaluated set of predictive computational models for combined effects and cumulative risk assessment of multiple stressors ranging from the molecular initiating events to the adverse outcomes; 2) apply these models on existing data as well as novel knowledge generated within this project to relevant stressors, exposure scenarios, effect endpoints and species to characterise the impact of multiple stressors having similar and dissimilar modes of action; 3) assess and reduce uncertainty in integrative risk assessment. Among the approaches planned to be investigated, the Adverse Outcome Pathway (AOP) concept allows the description of toxic effects from the molecular initiating events (MIE) to the adverse outcomes (AO), through linear chains of biological key events (KE) on single organisms and populations. Such models that relate effects at several levels (molecular, cellular, tissue, individual and population) to adverse effects have not been undertaken with respect to radionuclides exposures. This task will develop novel AOP and AOP networks related to radon and NORM exposure with the aim to identify the critical events leading to disease status, and to address the common pathways and potential interactions with other stressors.

The task is split into 2 subtasks:

Subtask 4.8.1 Effects of combined exposure of human bronchial epithelial cells to radon, tobacco smoke and nanoparticles

Subtask 4.9.2 Evaluate of cumulative risks on non-human biotas exposed to uranium and chemicals