research indicates that exposure to pesticides can significantly influence the mutational landscape within tumors, potentially playing a role in the development of cancer, the response to treatments, and the progression of the disease.
further studies should extend the observation period for individuals exposed to pesticides to more accurately evaluate the implications of these findings. this research focused on breast cancer, occupational exposure to pesticides, mutational burden, and somatic mutations.
brazil stands as a major consumer of agricultural pesticides worldwide, a practice that raises considerable concerns about human health.
while pesticide use has been widespread in brazil for decades, data regarding consumption, particularly among small-scale farmers, have been limited. exposure to pesticides can increase the risk of cancer through various mechanisms, including the generation of oxidative stress, alterations in adhesion molecules, the inhibition of acetylcholinesterase, endocrine disruption, and the contribution to genomic instability, which are all recognized hallmarks of cancer.
the role of dna repair in cancer development
double-strand breaks in dna can pose a significant threat to genomic stability.
the body employs multiple dna repair mechanisms to counteract these harmful effects, as failure to repair these breaks can result in chromosome aberrations, apoptosis, and the development of cancer. homologous recombination is a critical pathway involved in repairing double-strand breaks that occur primarily during dna replication and due to cell damage.
genes crucial for homologous recombination, and any alterations within them, are closely linked to carcinogenic features, such as a high mutational burden, which arises from the accumulation of unrepaired dna damage. the loss of function of tumor suppressor genes, such as the tp53 gene, reveals that dna damage response pathways must be downregulated for cells to proliferate and avoid apoptosis.
therefore, tumor cells frequently exhibit alterations in genes responsible for these dna damage response pathways, potentially leading to a loss of genomic stability and the suppression of cell growth. these factors emphasize the significant impact of lifestyle and environmental conditions on breast cancer risk.
mutational patterns and environmental exposures
the examination of mutational patterns related to dna damage and repair processes in cancer has revealed several signatures that can be linked to exposures to carcinogens and defects in dna maintenance pathways, such as specific base transversions often seen in smoking-related lung cancer, which is caused by tobacco exposure.
while factors like obesity, physical activity, and the use of tobacco and alcohol are known to be associated with breast cancer risk, the potential role of environmental exposure to pesticides in breast cancer development remains relatively unclear.
despite the limited high-quality evidence and the statistical challenges in inferring causality, there is a trend toward increased risk, even if few studies are statistically significant.
although inconsistent results are found in the literature regarding pesticide exposure and breast cancer risk, it is suggested that hormone-positive breast cancer could have an increased association with pesticide exposure. a study was conducted involving patients between january and september, managed at the francisco beltrão cancer hospital (ceonc) in paraná state, brazil, which assists 27 municipalities.
study methodology and patient population
after providing informed consent, tumor samples were collected during diagnostic breast cancer biopsy surgery, without selection bias.
the study was reviewed and approved by the ethics committees of the state university of west paraná. all enrolled patients were invited to complete a questionnaire containing 61 questions about their occupational history. of the women diagnosed with breast cancer, a specific number were eligible for this study.
based on their responses, the study population was categorized as either occupationally exposed or not exposed to pesticides.
the study comprised 91 patients exposed to pesticides and 67 unexposed to pesticides. all clinicopathological variables were collected from medical records. multiplex pcr-based assays were conducted on unpaired tumor samples, designed to cover the coding regions and flanking splicing sites of brca1, brca2, palb2, tp53, and rad51d.
tumor dna sequencing was prepared using the nextera xt dna kit (illumina, san diego, ca, usa) and performed in three independent runs, using paired-end methodology on a miseq genome analyzer (illumina) with a reading length of up to bp.
variant analysis and data interpretation
base quality score recalibration, indel realignment, and variant calling were performed following the genome analysis toolkit v4.
somatic variants were called by mutect2 from gatk, using default parameters for unpaired tumor samples. somatic variants were functionally annotated using gatk funcotator and the ensembl variant effect predictor (vep). to explore the mutational landscape based on pesticide exposure, somatic variant data were processed and analyzed using the r programming language version 4.
tumor mutational burden (tmb) was assessed by calculating the total number of somatic non-synonymous variants in the target region divided by the total size of the target region per mega-base.
breast cancer patients were subsequently divided into low- and high-tmb groups using the median value. to assess correlations with clinicopathological features, the mutational burden data were combined with corresponding clinical information. all nonsense and frameshift mutations not registered in the clinvar, catalogue of somatic mutations in cancer (cosmic), or international agency for research on cancer (iarc) database were considered pathogenic, according to the american college of medical genetics and genomics (acmg) standard terminology.
novel variants in a splice site that were not yet described were determined as likely pathogenic if there were known pathogenic variants in the same splice site already registered in clinvar.
statistical models included subjects with complete data on the specific environmental variable of interest and the adjustment variables.
all statistical analyses and visualizations were performed in the r programming language version 4. the shapiro—wilk test was applied to assess the normality of the data. for continuous variables and group comparisons, the mann—whitney u test was used.
clinical and demographic characteristics of women included in this study were observed. the median age of diagnosis was. tumor samples were sequenced with high depth, with a mean coverage of samples of × to 1,×.
mutational landscape findings
after processing and filtering steps for somatic variant identification, of samples harbored variants classified as missense, frameshift, nonsense, or splice-site variants, distributed in the brca1, brca2, palb2, tp53, and rad51d genes.
the proportion of missense variants was the highest among other mutation types. the distribution of variants in all samples showed brca2 as the most affected gene. in general, missense variants were the most frequent type. the genomic landscape and potential associations between mutations identified in tumor breast cancer samples were analyzed.
oncoplots were used to show the mutational profile of each gene in all tumor samples, ordered by variant frequency.
the distribution of transitions and transversions in all tumor samples was also analyzed, showing boxplots with the overall summary of single-nucleotide variants (snvs) classified into six substitution classes, boxplots with the distribution of snvs classified into transitions (ti) and transversions (tv), and the proportion of snvs per sample classified into six substitution classes.
the co-occurrence or exclusive variant associations of the evaluated genes were also analyzed. among breast cancer samples with detected variants, the most mutated genes in the exposed group were brca2 and palb2, while in the unexposed group, they were brca2 and brca1. concerning co-interaction analysis, differing correlation results were observed between the exposed and unexposed groups.
genomic landscapes and potential associations between mutations in tumor breast cancer samples were grouped according to pesticide occupational exposure.
oncoplots were used to show the mutational profile of each gene in groups exposed and unexposed to pesticides. mutational co-occurrence or exclusive associations between evaluated genes were analyzed in both exposed and unexposed groups. the distribution of transitions and transversions was shown in both exposed and unexposed groups, with boxplots providing an overall summary of snvs classified into six substitution classes, the distribution of snvs classified into transitions (ti) and transversions (tv), and the proportion of snvs per sample classified into six substitution classes.
a total of 28 pathogenic, 10 likely pathogenic, and 12 vus (variant of uncertain significance) variants were identified.
all variants were identified in 47 tumor samples from different patients. several variants were detected in more than one sample: the pathogenic variants (p. c61g in brca1, p. hr in tp53), and the vus (p. s46c in rad51d). regarding tp53, 26 variants were classified as pathogenic and likely pathogenic on clinvar, 21 of which were also predicted as pathogenic on cosmic.
six brca1 variants are classified as pathogenic on clinvar, 21 on the cosmic database, and only one ranked as pathogenic in both databases (brca1 c.).
four brca2 variants are pathogenic on clinvar and six on cosmic, with no concordance between the two databases. eight palb2 variants were classified as pathogenic on clinvar, 12 on cosmic, and only one classified as pathogenic in both databases (palb2 c.).
two rad51d variants were classified as vus on clinvar, and one was predicted as pathogenic in cosmic (rad51d c.).
further analysis was performed on the distribution and frequency of pathogenic, likely pathogenic, and vus variants. graphical representations of variant classifications were presented, along with the frequency and type of each variant identified in each gene.
the proportion of pathogenic and likely pathogenic variants detected according to patient exposure status was also examined. in addition, the frequency of variants classified as missense, frameshift, nonsense, and splice site was analyzed according to the exposed and unexposed samples.
the mutational burden and its association with various clinicopathological variables and pesticide exposure were also studied.
mutational burden analysis
in general, samples that presented a pathogenic or likely pathogenic variant showed a higher mutational burden.
it was observed that breast cancer patients exposed to pesticides showed no difference in mutational burden in the presence or absence of pathogenic variants in the tumor. mutational burden was found to be increased in tumors that harbor a tp53 pathogenic or likely pathogenic variant only in unexposed patients.
the tumor mutational burden (tmb) and clinicopathological variables were analyzed according to pesticide exposure. the tmb levels in samples were analyzed according to pesticide exposure status, grouped by the presence or absence of any predicted pathogenic variant.
tmb levels were also analyzed in exposed patient samples grouped according to disease onset and brca1, brca2, and palb2 statuses.
breast cancer tumors and other neoplasias tend to present tp53 as the most mutated gene. however, this study found brca2.
it is important to highlight that somatic variants of clinical relevance or interest were analyzed. the study also showed the co-occurrence of brca1 and tp53 mutations in exposed patients, a condition already observed in germline brca1 mutation carriers and found in colorectal cancer tumors, where the co-occurrence of brca1 and tp53 mutations led to a poorer prognosis.
it must be noted that there are discrepancies between variant annotations in clinvar, a large public archive composed mainly of germline variants and the phenotype consequence, and cosmic, the largest public resource of somatic mutations in human cancer. it has been demonstrated that environmental exposure to certain toxins can induce haploinsufficiency, which has been proposed to contribute to breast cancer development, especially in brca2 cells bearing heterozygous mutations.
this haploinsufficiency causes the reduction of brca2 function, which sensitizes cells to dna damage and compromises dna repair, and under prolonged exposure to these toxins, it can even promote brca2 protein depletion in wild-type cells.
the researchers speculate that pesticide exposure could cause the same effect, as it tends to occur during a long lifetime, not only for tumors bearing deleterious mutations in dna damage response genes but also in tumors with both functional alleles. moreover, several studies suggest the endocrine-disrupting potential of certain pesticides.
although the current evidence is not fully conclusive of their mechanism of action, exposure to these compounds has been associated with an increased risk of breast cancer. interestingly, it has also been shown that farmers under pesticide exposure who presented chromosomal abnormalities also had lower expression of brca2, which further reinforces the essential role of brca2 in dna stability.
variant-specific findings and conclusions
since whole-exome sequencing (wes) was not performed, it is impossible to associate the results with any mutational signature at this time.
a higher tp53 missense variant frequency was identified. two novel splicing variants were identified in canonical splice sites, one in intron 3 (c.). the missense tp53 (c. hr) variant found in two patients (one exposed and one unexposed) in this study was already documented in lung cancer tumors.
the nonsense variant (c.). in summary, all variants were clustered within exons 5—8, the evolutionary conserved dna-binding region of the tp53 protein, which is considered a hotspot area by the iarc database. this indicates that patients under pesticide exposure are more prone to dna damage compared to unexposed individuals.
palb2 presented the second-highest proportion of pathogenic or likely pathogenic variants in the study.
palb2 plays an important role in cancer development and progression, as even heterozygous mutations appear to contribute to early events of oncogenesis. two frameshift variants were identified in palb2 in seven patients of the exposed group, and one nonsense variant (c.
ffs) found in the same sample, localized at the evolutionary conserved chromatin-associated motif (cham) domain, which is responsible for palb2 chromatin association and dna repair function. it is postulated that variants with strong evidence for pathogenicity in palb2 are commonly located in the coiled-coil (cc) motif or the wd40 domain.
the recurrent frameshift variant found in six patients exposed to pesticides occurs in the palb2 (c. mfs) and is localized at the cc motif on the amino-terminal region that mediates palb2 interaction with brca1 and rad51d. all brca1 pathogenic variants were found in exposed patients, except for the (c.
c61g) observed in two samples (one from an unexposed patient). this missense mutation occurs in the brca1 ring domain, decreasing brca1 availability at dna damage sites and hindering dna repair.
it has also been demonstrated that brca1 c61g mammary tumors develop resistance to cisplatin platinum therapy, a drug that induces oxidative damage in cells.
as for brca2 pathogenic variants, three were in exposed patients, and only one was in an unexposed patient (c.
afs), which was already detected in breast, ovary, and lung cancer patients by another study. mutations in these genes hinder dna repair, leaving cells more vulnerable to dna damage and prone to therapy-induced lethality and are also related to therapy resistance mechanisms.
somatic mutations may change over time due to selective pressure derived from therapy and genetic instability, and the genotoxicity effect of pesticide exposure may also impact this mutational landscape, as several toxic substance exposures produce a characteristic mutational pattern that impaired dna repair capacity.
deleterious variants in dna damage response genes, mainly in lung cancer, are associated with a higher mutational burden, as observed in breast carcinomas with dna damage repair gene variants and in these tumor samples with brca1, brca2, and palb2 variants.
a high mutational burden was found in patients carrying a pathogenic, likely pathogenic, or vus variant in both analyzed groups, with the highest frequency of truncating and likely deleterious variants in exposed patients compared to unexposed patients.
evidence suggests that the variation of high mutational burden in cancer types could also be related to chronic mutagenic exposure. indeed, individuals exposed to several pesticides have increased dna damage, including dna strand breaks, a consequence of the direct exposure and the oxidative stress generated from it, which could be the mechanism behind these findings in the exposed group.
interestingly, tumors from unexposed patients harboring tp53 mutations presented significantly higher tmb than those with wild-type tumors; however, this result was not observed in the exposed group. usually, tp53 mutations are found in tumors with a high mutational burden.
for example, in lung cancer, high tmb and tp53 mutations are frequently observed in tumors with the sbs4 signature, which is associated with tobacco smoking. this indicates that tp53 mutations are related, at least to some extent, to carcinogen exposure.
this study highlights the impact of pesticide exposure on the mutational landscape in breast cancer, particularly in francisco beltrão, brazil.
the findings suggest that exposure to pesticides may be linked to an increased risk of breast cancer and could affect how tumors respond to treatment. further investigation is necessary to fully understand these relationships and develop effective strategies to protect public health.