Comprehensive computational analysis via Adverse Outcome Pathways and Aggregate Exposure Pathways in exploring synergistic effects from radon and tobacco smoke on lung cancer

Abstract

Lung cancer remains the leading cause of cancer mortality worldwide, with tobacco smoke and radon exposure being the primary risk factors. The interaction between these two factors has been described as sub-multiplicative, but a better understanding is needed of how they jointly contribute to lung carcinogenesis. In this context, a comprehensive analysis of current knowledge regarding the effects of radon and tobacco smoke on lung cancer was conducted using a computational approach. Information on this co-exposure was extracted and clustered from databases, particularly the literature, using the text mining tool AOP-helpFinder and other artificial intelligence (AI) resources. The collected information was then organized into Aggregate Exposure Pathway (AEP) and Adverse Outcome Pathways (AOP) models. AEPs and AOPs represent analytical concepts useful for assessing the potential risks associated with exposure to various stressors. AOPs provide a structured framework to organize knowledge of essential Key Events (KEs) from a Molecular Initiating Event (MIE) to an Adverse Outcome (AO) at an organism or population level, while AEPs model exposures from the initial source of the stressor to the internal exposure site within the target organism, situated upstream of the AOP. Combining these frameworks offered an integrated method for knowledge consolidation of radon and tobacco smoke, detailing the association from the environment to a mechanistic level, and highlighting specific differences between the two stressors in DNA damage, mutational profiles, and histological types. This approach also identified gaps in understanding joint exposure, particularly the lack of mechanistic studies on the precise role of certain KEs such as inflammation, as well as the need for studies that more closely replicate real-world exposure conditions. In conclusion, this study demonstrates the potential of AI and machine learning tools in developing alternative toxicological models. It highlights the complex interaction between radon and tobacco smoke and encourages collaboration among scientific communities to conduct future studies aiming to fully understand the mechanisms associated with this co-exposure.

Keywords: AOP-helpFinder; Adverse Outcome Pathways (AOP); Aggregate Exposure Pathway (AEP); computational toxicology; lung cancer; radon; text mining; tobacco smoke.

Figure 1

Overview of the study progression. The development of the AOPN and AEP is based on information derived from scientific literature and exposure data, obtained through databases and bioinformatics tools. The identification of gaps and the overview of the literature following clustering involve examining and categorizing the available data and types of data. This includes assessing the number of mechanistic vs. epidemiological studies, or identifying which KEs are most and least studied in the context of radon-tobacco co-exposure. Studies focusing on exposure will be preferentially used for the development of the AEP, while studies focusing on mechanistic relationships will be used for the development of the AOPN. * The term “Event” in the Stressor-Event search corresponds to the 238 biological keywords associated with lung cancer, while in the Event-Event search, it refers exclusively to the final components (MIE, KE, AO) included in the AOPN for evaluating KERs.

Figure 2

Clustering of the 378 articles assessing the combined impact of radon and tobacco on lung cancer. This classification was carried out into 4 clusters using the k-means method. Each cluster is detailed by the types of studies conducted (including human, animal model, and cellular research) and displays the type and predominant biological events identified within each cluster.

Figure 3

PRISMA diagram of the specific automatic process leading to the acquisition of a relevant article pool for the study of co-exposure to radon and tobacco smoke in the context of lung cancer.

Figure 4

Aggregate Exposure Pathways (AEP) illustrating the interconnected risk arising from simultaneous exposure to tobacco smoke and radon. The red areas indicate points of interaction. Radon decay products can attach to tobacco smoke particles in the air, resulting in higher concentrations of radon (and radon products) in smoking areas. In heavy smokers, changes in lung function lead to reduced elimination of radon and tobacco smoke from the lungs. This results in greater accumulation of radon and smoke particles in lung tissues, increasing the potential for molecular damage.

Figure 5

AOPN triggered by combined exposure to radon (physical agent) and tobacco smoke (chemical agent). The AOPN illustrates distinct MIEs initiated by radon (“deposition of energy”), tobacco smoke (“uptake of carcinogens by CYPs”), or both (“Reactive oxygen species”), each leading to different types of DNA damage that converge on a shared pathway via the KE “Inadequate DNA repair.” The AO is labeled as “Lung cancer,” reflecting a generic term encompassing increased risk across multiple subtypes. This terminology is consistent with AOP-Wiki conventions, thereby supporting interoperability and linkage across AOPs.