Somatic DNA Damage Response and Homologous Repair Gene Alterations and Its Association With Tumor Variant Burden in Breast Cancer Patients With Occupational Exposure to Pesticides

Abstract

Homologous recombination is a crucial pathway that is specialized in repairing double-strand breaks; thus, alterations in genes of this pathway may lead to loss of genomic stability and cell growth suppression. Pesticide exposure potentially increases cancer risk through several mechanisms, such as the genotoxicity caused by chronic exposure, leading to gene alteration. To analyze this hypothesis, we investigated if breast cancer patients exposed to pesticides present a different mutational pattern in genes related to homologous recombination (BRCA1, BRCA2, PALB2, and RAD51D) and damage-response (TP53) concerning unexposed patients. We performed multiplex PCR-based assays and next-generation sequencing (NGS) of all coding regions and flanking splicing sites of BRCA1, BRCA2, PALB2, TP53, and RAD51D in 158 unpaired tumor samples from breast cancer patients on MiSeq (Illumina) platform. We found that exposed patients had tumors with more pathogenic and likely pathogenic variants than unexposed patients (p = 0.017). In general, tumors that harbored a pathogenic or likely pathogenic variant had a higher mutational burden (p < 0.001). We also observed that breast cancer patients exposed to pesticides had a higher mutational burden when diagnosed before 50 years old (p = 0.00978) and/or when carrying BRCA1 (p = 0.0138), BRCA2 (p = 0.0366), and/or PALB2 (p = 0.00058) variants, a result not found in the unexposed group. Our results show that pesticide exposure impacts the tumor mutational landscape and could be associated with the carcinogenesis process, therapy response, and disease progression. Further studies should increase the observation period in exposed patients to better evaluate the impact of these findings.

Keywords: breast cancer; mutational burden; occupational exposure; pesticides; somatic.

Figure 1

(A) Study design and map showing the sample distribution according to exposure. (B) Amplified map of Brazilian Paraná State with symbols representing exposed (pink triangle) and unexposed (gray circle) patients’ localities. (C) Detail of pesticide consumption in the study area; modified from Gaboardi et al. (24).

Figure 2

Genomic landscape and potential associations between mutations identified in tumor breast cancer samples. (A) Oncoplot showing the mutational profile of each gene in all tumor samples ordered by variant frequency. (B) Distribution of transitions and transversions in all tumor samples showing boxplot with overall summary of single-nucleotide variants (SNV) classified into six substitution classes, boxplot with distribution of SNVs classified into transitions (Ti) and transversions (Tv), and proportion of SNVs per sample classified into six substitution classes. (C) Co-occurrence or exclusive variant associations of the evaluated genes. (D) Panels with variant classification, type (SNP, single-nucleotide polymorphism; INS, insertion; DNP, double nucleotide polymorphism; DEL, deletion), SNV class, number of variants per sample, variant classification summary, and genes ordered by total number of variants. Color legend of panel A is the same for panels (B, D).

Figure 3

Genomic landscape and potential associations between mutations in tumor breast cancer samples grouped according to pesticide occupational exposure. (A, B) Oncoplots showing the mutational profile of each gene in groups exposed and unexposed to pesticides. (C, D) Mutational co-occurrence or exclusive associations between evaluated genes in (C) exposed and (D) unexposed groups. (E, F) Distribution of transitions and transversions in (E) exposed and (F) unexposed groups showing boxplot with overall summary of SNVs classified into six substitution classes, boxplot with distribution of SNVs classified into transitions (Ti) and transversions (Tv), and proportion of SNVs per sample classified into six substitution classes. Color legend of panels (A, B) is the same for panels (E, F), respectively. The symbol * means statistical significance (p < 0.05).

Figure 4

Distribution and frequency of pathogenic, likely pathogenic, and variant of uncertain significance (VUS) variants. (A) Graphical representation of variant classification detected in all samples. (B) Frequency and type of pathogenic, likely pathogenic, and VUS variant identified in each gene. (C) Proportion of pathogenic and likely pathogenic variants detected according to patient exposure status. (D) Frequency of variants classified as pathogenic, likely pathogenic, or VUS in exposed and unexposed groups. (E) Frequency of variants classified as missense, frameshift, nonsense, and splice site according to exposed and unexposed samples.

Figure 5

Tumor mutational burden (TMB) and clinicopathological variables according to pesticide exposure. (A) Frequency of high and low mutational burden levels in tumor samples with a pathogenic and/or likely pathogenic variant detected or not. Tumors harboring any deleterious variant present increase in mutational burden (p < 0.001). (B) TMB levels in samples according to pesticide exposure status and grouped by presence or absence of any predicted pathogenic variant. Tumors from unexposed patients present increased TMB when harboring any deleterious variant (p = 0.0214); this result is not observed in tumors from exposed patients. (C) TMB levels in exposed patient samples grouped according to disease onset. We found early-onset tumors (patients with <50 years old) with significantly higher TMB (p = 0.00978) in comparison to late-onset tumors (patients with ≥50 years old). (D–F) TMB levels in exposed patient samples grouped according to BRCA1, BRCA2, and PALB2 statuses, respectively. Tumors harboring a mutation in BRCA1 (p = 0.01308), BRCA2 (p = 0.0366), and/or PALB2 (p = 0.000588) presented higher TMB than wild-type tumors. The expanded form of “NS” means “No significance”.