Geographic variation of mutagenic exposures in kidney cancer genomes

Substantial geographical variations in adult cancer incidence rates exist, often unexplained by known risk factors. This suggests the presence of unidentified environmental or lifestyle causes. While traditional epidemiological studies have identified several important risk factors, their success in recent decades has been limited. This necessitates alternative approaches to identify further contributors to cancer development. Analyzing mutational signatures within cancer genomes offers a complementary strategy. Cancer genomes accumulate somatic mutations throughout an individual's life, resulting from both endogenous processes (e.g., imperfect DNA replication and repair) and exogenous exposures (e.g., UV radiation, cigarette smoke). Mutational signatures are patterns of somatic mutations reflecting individual mutational processes. Kidney cancer exhibits particularly high incidence rates in Central and Northern Europe, with increasing incidence in high-income countries. Although obesity, hypertension, and tobacco smoking are known risk factors for clear cell renal cell carcinoma (ccRCC), they account for less than half the global burden and don't explain geographical or temporal trends. Previous studies had limitations in sample size and geographical diversity. To address these gaps, this study analyzed a large international ccRCC dataset, combining epidemiological and whole-genome sequencing data to identify unknown carcinogens and investigate the mechanisms of action of established risk factors.

The literature review section is implicitly woven into the Introduction and Discussion sections of the paper. The introduction cites previous research demonstrating unexplained geographical variations in cancer incidence and the limitations of traditional epidemiological methods. It then introduces mutational signature analysis as a complementary approach and discusses previous research on ccRCC risk factors and genomic studies, highlighting the need for a larger, more geographically diverse study. The discussion section references studies on aristolochic acid and its mutational signature, compares the findings to a previous study on esophageal squamous cell carcinoma, and cites literature regarding the contribution of both mutagenic and non-mutagenic processes to cancer development. It also references various studies analyzing mutational signatures in relation to lifestyle and environmental exposures.

This study included 962 ccRCC cases from 11 countries across four continents, representing a wide range of ccRCC incidence rates. Epidemiological data (sex, age at diagnosis, BMI, hypertension, smoking status) were collected. DNA was extracted from ccRCC tumors and matched blood samples, then subjected to whole-genome sequencing (average coverage 54-fold for tumors and 31-fold for blood). Somatic mutation burdens were analyzed for single base substitutions (SBS), doublet base substitutions (DBS), and indels. Mutational signatures were extracted using SigProfilerExtractor and mSigHdp, algorithms employing non-negative matrix factorization and Bayesian hierarchical Dirichlet processes. De novo signatures were extracted and compared to the COSMIC database. Signature activities were attributed to each sample using MSA. Driver mutations were identified using a dNdS approach, and their mutational spectra were analyzed. Evolutionary analysis (using DPClust) was performed to determine the timing of mutational processes. Associations between mutational signatures and risk factors were explored using regression analysis. Untargeted metabolomics (UHPLC-QTOF-MS) was conducted on plasma samples (n=901) to identify potential mutagenic agents. Geospatial analysis was performed on residential history data for specific signatures. Targeted metabolomics analyses also investigated circulating PFAS compounds and cystatin C levels. Polygenic risk scores were used to complement observational data.

Geographical variation in somatic mutation loads and patterns was observed in ccRCC. The study identified several key mutational signatures with varying geographical distributions. SBS22 (renamed SBS22a), associated with aristolochic acid exposure, was highly prevalent in Romania, Serbia, and Thailand. Three new aristolochic acid-related signatures (SBS22b, DBS20, ID23) were identified and correlated with SBS22a. SBS12, a signature of unknown cause, was found in 72% of Japanese ccRCCs but was rare elsewhere. SBS40 (decomposed into SBS40a, SBS40b, and SBS40c) was ubiquitous but showed higher mutation burdens in countries with higher ccRCC incidence. SBS4 and DBS2, known tobacco smoking signatures, correlated with tobacco consumption. No mutational signatures associated with obesity or hypertension were found. SBS40b mutation burdens strongly correlated with country-specific ccRCC incidence rates and with markers of impaired kidney function. Untargeted metabolomics identified N,N,N-trimethyl-L-alanyl-L-proline betaine (TMAP), a marker of reduced kidney function, as associated with SBS40b. Evolutionary analysis indicated that signatures from putative exogenous exposures were primarily present in early stages of cancer development. No association was found between several established risk factors for ccRCC (obesity, hypertension, diabetes) and mutational burden except for tobacco smoking and the associated mutational signatures.

This large-scale, international study demonstrates the existence of multiple, geographically variable mutagenic processes contributing to ccRCC mutation loads. This contrasts with findings from a study of esophageal squamous cell carcinoma, which showed no geographical differences in mutation burdens or signatures. The results emphasize the contribution of both geographically variable mutagenic and non-mutagenic carcinogenic exposures to global cancer incidence. The high prevalence of aristolochic acid-related signatures in southeastern Europe suggests a much wider extent of exposure than previously thought. The identification of SBS12 and SBS40b highlights previously unsuspected mutagenic exposures. The ubiquitous nature of SBS40b, and its strong correlation with ccRCC incidence and impaired kidney function, warrants further investigation into its etiology. The absence of association between obesity, hypertension, and mutation burden supports a model of cancer development where mutations are essential, but additional factors impact clone expansion and cancer progression. Future research needs to focus on understanding both mutagenic causes and the promotion of mutated clones by non-mutagenic processes.

This study reveals significant geographical variations in mutational signatures associated with ccRCC, implicating multiple widespread mutagenic processes. Aristolochic acid exposure is far more widespread than previously appreciated. Novel mutagenic exposures (SBS12 and SBS40b) are identified, with SBS40b potentially significantly contributing to global ccRCC burden. Further research should focus on identifying the causal agents of these signatures, assessing the extent of population exposure, and exploring mitigation strategies. Large-scale whole-genome sequencing studies across diverse populations are needed to uncover additional cancer-causing agents.

The study relied on retrospective data collection across multiple centers, with potential variations in data collection protocols. The number of cases from some countries was limited, potentially affecting the generalizability of findings. Untargeted metabolomics identified associations with some signatures, but the causative agents for many signatures remain unknown. The study did not consider all potential environmental exposures or lifestyle factors that could potentially affect the risk of ccRCC.