Pesticide exposome reveals oxidative stress and immune response disruptions in occupationally exposed women with breast cancer

Abstract

Rural women are at a higher risk of developing breast cancer, but the molecular mechanisms involved remain poorly understood. This study investigates the impact of chronic pesticide exposure on systemic and tissue-level protein expressions in women diagnosed with breast cancer. Plasma and tumor samples were collected from 215 women diagnosed with breast cancer; among them, 128 patients were occupationally exposed (n = 128) and 87 patients were not occupationally exposed (87) were analyzed for further studies. High-throughput proteomic analyses of plasma revealed distinct protein expression profiles between the groups, primarily associated with inflammation, oxidative stress, and immune deregulation. Validation analysis showed that cytokines such as IL-1β, IL-12, IL-17A, and TNF-α were significantly reduced in both plasma and tumor samples from exposed patients compared to unexposed women. This reduction in cytokines, coupled with decreased tumor-infiltrating leukocytes (TILs) and CTLA-4 overexpression, suggests pesticide-induced immune compromise. Circulating IL-12 levels were significantly reduced in younger, eutrophic, and at menacme exposed patients. Pesticide exposure also influenced nitrosative stress production, leading to reduced NOx levels in plasma and impaired iNOS expression in TILs. These results suggest a potential link between pesticide exposure and the pathophysiology of breast cancer, offering new insights into the underlying molecular mechanisms.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-12870-2.

Introduction

Brazil ranks as one of the largest consumers of pesticides globally¹. Each year, the country uses over 500,000 tons of pesticides, especially glyphosate, atrazine, and 2,4-dichlorophenoxyacetic acid (2,4-D), reaching a per capita consumption of about 7 L annually². About 80% of the pesticides approved for sale in Brazil are prohibited in at least three European countries that are members of the Organization for Economic Co-operation and Development (OECD) due to its toxicity³. Consequently, the Brazilian population faces extensive exposure to these harmful substances, increasing the risk of long-term intoxication and the development of numerous chronic diseases^4–6.

Since the country stands as one of the largest food producers in the globe, family agriculture predominates in several regions. In this scenario, all family members are enrolled in the farming chain, including women. Despite the invisibility of their labor and the overlooked risks of pesticide exposure, they remain consistently exposed and are at heightened risk for developing serious chronic health conditions^7,8.

Pesticides have been linked to numerous deleterious long-term health effects^9–12. In women, they have been particularly associated with reproductive dysfunction^13,14 and tumor development^15,16, including breast cancer⁷. The majority of female breast cancer cases result from a combination of genetic, physical, chemical, and biological factors, underscoring the significant role of the exposome in the development and progression of the disease in about 90% of cases^17,18.

Approximately 25 pesticides significantly affect mammary cells, with 80% of them being implicated in the development of breast tumors¹⁹. Most evidence regarding the impact of pesticides on breast cells comes from in vitro studies, and include disruption of estrogen and progesterone signaling²⁰, impairment of mitochondrial function and oxidative stress generation²¹, enhanced cell growth and proliferation²², epigenetic changes²³, and DNA damage²⁴, among others. Animal data shows that prenatal exposure to atrazine delays the development of mammary glands and boosts steroid hormone production, which is thought to be a key factor in increasing hormone levels in adulthood²⁵. In breast cancer cells, even extremely low concentrations of glyphosate lead to DNA damage²⁴, while atrazine reduces protein expression and changes the activity of genes involved in antioxidant defense, which increases genomic instability driven by oxidative stress²⁶.

Despite these evidences, there is limited knowledge about the long-term impact of pesticide exposure on the critical mechanisms involved in breast cancer development and progression in humans. Additionally, its effects on the systemic and tumor-related biological profiles of breast cancer patients remain largely unexplored. To address this, the current study performed a proteomic screening of plasma to identify protein profile changes induced by pesticide exposure in a rural population of women diagnosed with breast cancer. Based on the key differentially expressed processes, we further validated our findings by analyzing specific proteins in both plasma and breast tumor samples, with the goal of identifying the primary processes affected by pesticide exposure in human breast cancer, focusing on oxidative stress, inflammation, and immune response-related markers, as well as underscoring its clinicopathological meaning.

Results

Study design, characterization of pesticide exposure, and patients data

This study was approved by the Institutional Human Research Ethics Committee and registered under CAAE number 35524814.4.0000.0107. It was conducted in the southwest region of Paraná, Brazil, an area dominated by family farming and intensive pesticide use. The study included a total of 215 women diagnosed with breast cancer from 27 municipalities who were treated at the Francisco Beltrão Cancer Hospital (Ceonc) in Francisco Beltrão, Paraná.

The study design is illustrated in Fig. 1. This region is marked by the widespread use of glyphosate. In this region, over 50% of the population is involved in agricultural activities, primarily centered on family farming. Most of the municipalites exhibit a high per capita consumption of pesticides concomitant to a high incidence of breast cancer (Fig. 1A).

Fig. 1.

Study design. This study included 215 breast cancer patients treated at Francisco Beltrão Cancer Hospital (Ceonc) in Francisco Beltrão. Paraná. Brazil. After providing informed consent. the patients were interviewed to gather information on their occupational pesticide exposure profiles. Based on the responses. 128 women were classified as pesticide-exposed and 87 as unexposed. (A) The map illustrates pesticide usage in tons and the number of cancer cases reported across 27 municipalities. based on the IBGE (https://www.ibge.gov.br/apps/basescartograficas)and SIAGRO (https://www.agricultura.pr.gov.br/Pagina/Siagro) databases. Darker shades indicate higher levels of pesticide trade. (B) Glyphosate-AMPA levels (ppm) were measured in urine samples. with the graph comparing results between unexposed and exposed patients. (C) An illustrative image shows urine samples collected from unexposed and exposed patients following five hours of pesticide handling.

After providing informed consent, participants were interviewed by using a specific questionnaire³ to gather detailed information about their occupational pesticide exposure profiles. The exposure criteria focused on continuous, unprotected, and direct pesticide handling. Rural women were classified as exposed if they reported a history of direct pesticide handling without using protective gloves during activities such as preparing and diluting pesticide solutions, applying pesticides, decontaminating personal protective equipment (PPE), and/or washing clothes worn during spraying. Additionally, they had to have spent at least 50% of their lives engaging in such activities at least twice a week throughout the year. In contrast, the unexposed group consisted of urban female workers with no past or current history of occupational pesticide exposure. Based on this, patients were categorized as unexposed (n = 87) or occupationally exposed (n = 128) women.

To certify that female exposure was enough to ensure pesticide contamination, pesticide analysis was conducted in pooled urine samples collected after pesticide handling, aiming to detect glyphosate-AMPA levels (Fig. 1B and C) in occupationally unexposed and exposed women by LCMS²⁷. Glyphosate-AMPA levels were significantly higher in the pool of exposed women (black bar) when compared to the unexposed ones (grey bar).

Clinicopathological data of patients are presented in Table 1. Data obtention included patients characteristics (age at diagnosis, body mass index, menopausal status), tumor features (molecular subtypes, hormonal receptors status, human epidermal growth factor 2 receptor amplification (HER2)), and the presence of lymph nodal metastasis. Regardless of pesticide exposure status, most patients were diagnosed between the ages of 40 and 60, with 70.7% of cases classified as overweight or obese based on BMI. A majority of the patients were postmenopausal at the time of diagnosis (70.23%). The most common molecular subtype of breast cancer was Luminal B, accounting for 38.14% of cases. Additionally, an intermediate histological tumor grade was observed in 51.63% of cases, and 59.53% of patients had lymph nodal invasion at the time of diagnosis.

Table 1.

Clinicopathological characterization of patients.

Parameter	Subgroups	Total population (n = 215)	Exposed (n = 128)	Unexposed (n = 87)
Age at diagnosis (years)	< 40	10.7% (n = 23)	11.72% (n = 15)	9.2% (n = 8)
	40–60	50.23% (n = 108)	47.66% (n = 61)	54.02% (n = 47)
	> 60	39.07% (n = 84)	40.62% (n = 52)	36.78% (n = 32)
Body mass index (kg/2)	< 18.5	2.33% (n = 5)	1.56% (n = 2)	3.45% (n = 3)
	18.5–24.9	26.98% (n = 58)	27.34% (n = 35)	26.44% (n = 23)
	> 24.9	70.7% (n = 152)	71.09% (n = 91)	70.11% (n = 61)
Menopause at diagnosis	Yes	70.23% (n = 151)	70.31% (n = 90)	70.11% (n = 61)
	No	29.77% (n = 64)	29.69% (n = 38)	29.89% (n = 26)
Molecular subtype of breast cancer	Luminal A	29.77% (n = 64)	23.44% (n = 30)	39.08% (n = 34)
	Luminal B	38.14% (n = 82)	40.62% (n = 52)	34.48% (n = 30)
	HER2-amplified	15.81% (n = 34)	18.75% (n = 24)	11.49% (n = 10)
	Triple-negative	16.28% (n = 35)	17.19% (n = 22)	14.94% (n = 13)
Histopathological Grade	I	29.3% (n = 63)	28.91% (n = 37)	29.89% (n = 26)
	II	51.63% (n = 111)	55.47% (n = 71)	45.98% (n = 40)
	III	19.07% (n = 41)	15.62% (n = 20)	24.14% (n = 21)
Lymph node metastasis	Absence	40.47% (n = 87) 5	46.09% (n = 59) 5	32.18% (n = 28)
	Presence	9.53% (n = 128)	3.91% (n = 69)	67.82% (n = 59)

HER2 human epidermal growth receptor 2.

Pesticide exposure impacts the plasmatic proteomic profile of patients

To determine the impact of pesticide exposure on the differentially expressed proteins in plasma samples from breast cancer patients, we used label-free protein quantification by mass spectrometry²⁸. The proteomic analysis of plasma (Fig. 2) revealed a total of 223 proteins. The total number of identified proteins was categorized into unique proteins (exclusive to each condition) and differentially expressed proteins (Fig. 2A). A total of 142 proteins were upregulated by pesticide exposure, while 41 were downregulated. The proteins associated with these processes are shown in Supplementary Table 1.

Fig. 2.

In silico analysis of biological pathways and processes modulated in plasma from breast cancer patients exposed to pesticides compared to unexposed patients. (A) Upregulated. downregulated. Enrichment analysis of biological processes (B) and pathways (C) related to upregulated proteins in pesticide-exposed patients. emphasizing variations in oxidative stress and inflammatory response. (D) Up and (E) downregulated biological pathways. (F) Upregulated biological processes and (G) Downregulated biological processes. Data were processed using FunRich software. For D to G. the percentage of genes involved in each process is represented in blue. the statistical significance level (p-value) is indicated by red bars. and the orange line marks the reference for statistical significance (p = 0.05).

To identify the main signaling pathways and biological processes associated with differentially expressed proteins, we conducted separate in silico analyses for upregulated and downregulated proteins. In silico enrichment analysis (Fig. 2B and C) illustrates the main biological pathways (2B) and processes (2C) affected by pesticide exposure. The main pathways affected by pesticide exposure were apoptosis (downregulated), inflammatory responses (upregulated), and lipid transport (upregulated), while the main processes include downregulation of HIF-1 network and glutathione redox reactions, and upregulation of complement pathway, and peptide ligand-binding receptors.

Figure 2D to G shows the pesticide up and downregulated pathways and processes, respectively. Figure 2D highlights the biological pathways upregulated by pesticide exposure. There was a significant increase in the expression of proteins associated with the immune response (p = 0.007), innate immune system (p < 0.001), and lipid metabolism (p < 0.001). Pesticide-downregulated pathways (Fig. 2E) included the complement (p = 0.049) and clotting cascades (p = 0.003). The immune response was main biological processes differentially expressed in pesticide exposed patients (Fig. 2F, p < 0.001), while protein metabolism was significantly downregulated (Fig. 2G, p = 0.036). Based on these findings, our following step validated main components of immune, inflammation, and oxidative stress responses.

Pesticide exposure affected key components of the immune, inflammation, and oxidative stress responses

Given that the proteomic analysis indicated pesticide exposure influenced proteins related to immune response, inflammation, and oxidative stress, we further validated specific proteins and metabolites involved in these processes. Cytokine profiles (Fig. 3) from blood plasma and breast tumor samples were analyzed to assess the levels of interleukin 1 beta (IL-1β), interleukin 12 (IL-12), tumor necrosis factor alpha (TNF-α), and interleukin 17 A (IL-17A), in relation to the patients’ exposure profiles.

Fig. 3.

Systemic and tissue cytokine profile of women with breast cancer exposed to or not exposed to pesticides. (A and E) Interleukin 1 beta (IL-1β); (B and F) Interleukin 12 (IL-12); (C and G) Tumor Necrosis Factor-alpha (TNF-α); (D and H) Interleukin 17A (IL-17A). Blood cytokines were measured by enzyme immunoassay. tumor cytokines were determined by CBA. excepting for IL-17A that was assessed by PCR *p < 0.05. ****p < 0.01. ns  not significant.

Pesticide exposure affected negatively the circulating levels of IL-1β (36.49 pg/mL to exposed and 94.58 pg/mL to the unexposed group, p < 0.0001, Fig. 3A), IL-12 (21.67 pg/mL to exposed and. 76.24 pg/mL to the unexposed group, p < 0.0001, Fig. 3B), TNF-α (78.49 pg/mL to exposed and 136.2 pg/mL to the unexposed group, p < 0.0001, Fig. 3C), and IL-17A (47.45 pg/mL to exposed and 149.0 pg/mL to the unexposed group, p < 0.0001, Fig. 3D). Cytokine levels in breast tumors were significantly affected, showing reduced IL-12 (0.32 pg/mg tissue to exposed and 0.59 pg/mg tissue to unexposed, p = 0.0130, Fig. 3F) and TNF-α (0.23 pg/mg tissue to exposed and 0.42 pg/mg tissue to unexposed, p = 0.0249, Fig. 3G). No significant changes were observed in IL-1β levels (Fig. 3E). IL-17A levels, measured by enzyme immunoassay and CBA analysis, as well as PCR analysis, showed no significant differences between the groups (Fig. 3H).

Histological analysis of tumor-infiltrating lymphocytes (TILs) (Fig. 4A) revealed significant differences between the groups. The mean lymphocyte infiltration score was higher in unexposed patients (1.75 ± 0.25) compared to those exposed to pesticides (1.00 ± 0.20, p = 0.0537), suggesting a reduction in lymphocyte infiltration among exposed patients. Given the low cytokine levels detected in exposed patients, along with reduced TIL presence, we conducted a biopsy analysis of cytotoxic T-lymphocyte antigen 4 (CTLA-4, Fig. 4B). Notably, all breast cancer biopsies from pesticide-exposed patients exhibited diffuse CTLA-4 labeling in TILs, whereas in the unexposed group, CTLA-4 expression was observed in approximately 30% of samples.

Fig. 4.

Tumor-Infiltrating Leukocytes (TILs) analysis in breast cancer biopsies from pesticide-exposed and unexposed patients. (A) TILs (blue dots) were assessed using optical microscopy in hematoxylin–eosin stained biopsies. with scoring classified as 1 + (< 25%). 2 + (26–60%). and 3 + (> 60%). (B) Immunofluorescence labeling for CTLA-4. The images presented are representative of each group (Magnification: 20x).

Regarding the plasmatic oxidative stress profile (Fig. 5), nitric oxide metabolites levels (NOx, Fig. 5A) were reduced in pesticide-exposed patients (28.35 μM to exposed and 45.67 μM to unexposed, p < 0.0001). No significant changes were observed in systemic antioxidant capacity (Fig. 5B) or lipid peroxidation profile (Fig. 5C). Although breast tumor analysis was conducted, the detected levels were too low to be considered reliable, and therefore, these data were excluded from the study. The expression of inducible nitric oxide synthase (iNOS, Fig. 5B), an enzyme involved in NO synthesis, was also assessed in tumor tissue samples. iNOS expression was significantly lower in tumor-infiltrating lymphocytes (TILs) of pesticide-exposed patients, with labeling observed in 20% of samples from the exposed group compared to 40% in the unexposed group (p < 0.001).

Fig. 5.

Oxidative stress profile in breast cancer patients exposed or not to pesticides. (A) Nitric oxide metabolite levels (NOx); (B) Total plasma antioxidant capacity (TRAP); (C) Lipoperoxidation levels measured by high-sensitivity chemiluminescence (D) Immunofluorescence labeling for inducible nitric oxide synthase (iNOS). The images presented are representative of each group (20 × magnification). ). **** p < 0.001. ns not significant.

Since significant variations were observed in IL-12, TNF-α, and NOx systemic levels among exposed patients, we further assessed their circulating levels based on clinicopathological parameters (Table 2). Notably, IL-12 levels were significantly reduced in exposed breast cancer patients compared to unexposed patients concerning age at diagnosis (p = 0.02 for the 40–60 years age group), BMI ( p = 0.04 for eutrophic patients), and menopausal status (p = 0.05 for patients in menacme). No significant differences were observed for IL-12, TNF-α, and NOx in other comparisons.

Table 2.

Systemic IL-12. TNF-a. and NOx levels according to clinicopathological data.

Pars meter	Subgroups	Exposed	Unexposed	p-value	Exposed	Unexposed	p-value	Exposed	Unexposed	p-VHlue
Age at diagnosis (years)	< 40	26.84 (19.5–38.64	24.6 (5.67–42.79	0.76	89.3 (33.9–67.9)	120.22 (19.33–203.57)	0.61	36.76 (26.33 –118.72)	41.76 (26.33–75.89)	0.83
	40–60	20.45 (9.28– 143.52)	29.36 (13.75 (190.86)	0.02*	87.9(0.0–40.71)	93.57 (16.4–83.57)	0.6	32.41 (24.54–76.11)	31.1 (7.2–160.69)	0.76
	>60	32.58 (5.14– 142.25)	24.92 (13.43–133.95)	0.95	117.69 (37.18–307.86)	119.28 (59.97–172.1)	0.78	31.0 (22.2–88.07)	31.33 (8.72–129.15)	0.77
Body mass index (Kg/m2)	< 18.5	27.79 (27.79–27.79)	26.2 (26.2–26.2)	1	131.45 (95.0–167.9)	22.85 (22.65–22.85)	3.67	36.76 (34.8–38.72)	27.3 (25.02–29.59)	0.09
	18.5–24.9	20.77 (10.88–38.96)	24.6 (14.39–190.86)	0.04*	88.6 (63.6–07.86)	99.27 (19.36–203.57)	0.68	30.68 (24.54–76.11)	32.63 (7.63–79.8)	0.85
	> 24.9	22.65 (5.14 -143.32)	27.15 (5.67–159.59)	0.46	96.4 (0.0–63.6)	104.26 (16.4–283.57)	0.23	33.24 (22.2–118.72)	31.54 (7.2–160 89)	0.66
Menopause at diagnosis	Yes	22.0 (5.14–143.32)	26.84 (13.43–190.86)	0.41	99.74 (0.0–07.86)	110.7 (16.4–172.1)	0.74	31.98 (22.2–88 07)	31.87 (7.2–160.89)	0.56
	No	25.72 (10.88–51.0)	27.15 (5.67–159.59)	0.05*	86.45 (12.21 –167.9)	93.57 (19.36–283.57)	0.56	33.7 (25.63–118.72)	29.8 (7.63–75.89)	0.71
Molecular subtype	Lumina A	22.0 (10.88–143.52)	41.38 (13.43–190.86)	0.15	97.86 (0.0–63.6)	106.4 (36.4–46.4)	0.87	32.2 (25.24–66.96)	29.48 (7.2–67.59)	0.24
	Luminal Li	22.69 (10.88–142.25)	27.15 (5.67–159.59)	0.18	93.6 (12.21–67.9)	99.04(19.33–283.57)	0.48	34.76 (22.2–76.11)	34.15 (7.63–160.89)	0.53
	HER2-amplified	20.77 (15.67–23.64)	19.81 (19.5–20.13)	0.94	118.56 (73.57–307.86)	119.3 (93.57–203.57)	0.74	30.46 (25.63–70.67)	30.24 (26.11 –75.89)	0.24
	Triple-negative	27.8 (5.14–55.56)	19.18 (14.71 –37.0)	0.5	89.3 (33.6–137.86)	97.9 (16.4–52.1)	0.85	32.63 (24.76–118.72)	33.61 (25.67–102.2)	0.99
Tumor grade	I	22.0 (5.14141.5)	28.43 (5.67–190.86)	0.08	100.73 (0.0–33.57)	135.95 (39.45–203.57)	0.19	34.04 (24.54–88.07)	28.82 (7.2–160.89)	0.74
	II	26.84 (10.88–143.52)	27.6 (13.43133.95)	0.45	81.45 (21.87–307.86)	100.19 (39.3–283.57)	0.54	31.98 (22.2–118.72)	37.03 (10.02–129 15)	0.21
	III	27.79 (19.82 –38.64)	26.52 (14.71 -159.59)	0.5	136.4 (12.21–167.9)	96.82 (16.4–52.1)	0.55	34.35 (27.35–73.5)	30.02 (24.37 -102.2)	0.65
Lymph node metastasis	No	25.24 (5.14–142.25)	26.36 (5.67–137.78)	0.46	81.45 (12.21–67.9)	133.1 (22.85–283.57)	0.1	31.11 (22.2–65.89)	30.24 (7.63–129.15)	0.43
	Yes	21.09 [9.28–143 52)	27.15 (13.43–190.86)	0.14	103.6 (0.0–307.86)	97.58 (16.4–03.57)	0.48	33.26 (24.54–118.72)	32 2 (7.2 160.89)	0.77

*Statistical significance

Discussion

Using high-throughput proteomics, we demonstrated that pesticide exposure negatively regulates the immune, inflammatory, and oxidative stress responses in breast cancer patients, impairing the body’s ability to produce cytokines and oxidative stress mediators essential for effective antitumor responses. These changes in the inflammatory landscape may profoundly influence tumor progression and disease outcomes.

The simultaneous high incidence of breast cancer and intensive pesticide use areas in the southwest region of Paraná has sparked interest in exploring potential correlations between regional environmental factors and the development of this disease, particularly in cases involving more aggressive tumors. Data from our group reveal that more than 50% of breast cancer patients living in this area reported direct or indirect contact with pesticides during their lifetime⁷. This finding is particularly significant, as breast cancer remains the leading cause of death among women worldwide and continues to exhibit high incidence and mortality rates in this agricultural region.

Given that our study population is primarily exposed to pesticides through equipment decontamination and clothes washing, we analyzed glyphosate-AMPA residues in pooled urine samples from exposed and unexposed patients. This analysis revealed significantly higher levels of these compounds in the exposed group, emphasizing that female contamination occurs predominantly during their work on pesticide handling and decontamination activities. Notably, glyphosate-AMPA residues were also detected in the urine pool of the unexposed group, underscoring the widespread pesticide contamination, which can occur through non-occupational exposure via the ingestion of contaminated food and water. These findings align with our previous study that detected pesticide residues in urine samples from women with breast cancer in southwest Paraná after clothes washing and pesticide handling⁷. Urinary aminomethylphosphonic acid (AMPA), a metabolite of glyphosate, has been linked to breast cancer development in women²⁹. Proposed carcinogenic mechanisms for AMPA include alterations in DNA methylation and disruptions in the estrogen signaling pathway, processes that can occur even at environmentally relevant concentrations²³.

Despite advances in understanding the mechanisms of pesticide action and its relationship with cancer, limited knowledge exists regarding the systemic biological components affected by long-term pesticide exposure, particularly in the context of breast cancer. To address this gap, we conducted a high-throughput proteomic analysis to identify potential differences between occupationally exposed and unexposed women with breast cancer. Our findings revealed that the majority of differentially expressed processes linked to pesticide exposure were associated with inflammation, oxidative stress, and immune responses deregulation. Consequently, we focused on validating these results by quantifying cytokines and oxidative stress metabolites in both plasma and tumor samples.

Immune responses play a crucial role in the prognosis of diseases like breast cancer, where the balance and regulation of immune activity can profoundly impact tumor development, progression, and aggressiveness^30–32. In the present study, we demonstrated that prolonged and consistent occupational exposure to pesticides significantly affects both circulating and tissue cytokine levels in breast cancer patients. Pesticides exert immunotoxic effects through various mechanisms, disrupting both innate and adaptive immune responses, which are crucial for tumor surveillance and control³³. Prolonged exposure to these substances can lead to chronic antigenic stimulation, resulting in a phenomenon known as immune compromise³⁴. This disruption in cytokine production has been linked to the carcinogenic process, directly influencing tumor control and progression.

The immune system operates through a complex network of cytokines that mediate and regulate specific responses to internal and external stimuli. Dysfunctions in this system may result in immunodeficiencies, either primary or secondary in nature^33,34. Among the environmental factors contributing to such dysfunctions, chronic exposure to pesticides stands out as a significant risk factor, with the potential to impair tumor immunosurveillance, given the interdependent nature of the immune system and other physiological systems. Despite growing concern regarding the immunotoxic effects of pesticides, the underlying mechanisms remain poorly understood due to the heterogeneity of the compounds involved and the diversity of immune pathways affected. Several studies have reported immunological alterations associated with pesticide exposure. Notable findings include an increase in CD8⁺ T lymphocytes in individuals consuming contaminated water³⁵; a higher prevalence of atopic conditions and autoimmune diseases³⁶; and recurrent respiratory infections in chronically exposed Workers³⁷. Other studies have observed enhanced proliferative responses of T lymphocytes and increased interleukin-2 (IL-2) production³⁸; reduced expression of CD25⁺ cells³⁹; as well as decreased leukocyte counts and serum immunoglobulin G (IgG) levels⁴⁰.

The antitumor immune response is regulated by a network of cytokines, whose production and signaling must remain within homeostatic patterns of T helper cells to effectively eliminate tumors⁴¹. In our study, the Th1 response, represented by cytokines such as IL-1β, IL-12, and TNF-α, was negatively impacted by pesticide exposure, which was associated with more aggressive clinicopathological features. One of the key findings was the depletion of IL-12, a critical cytokine in the antineoplastic response. IL-12 stimulates a Th1-type immune response, promoting tumor rejection and contributing to survival^42,43. A reduction in IL-12, even in the early stages of cancer, can hinder the immune system’s capacity to control tumor progression, enabling the growth of larger tumor masses⁴⁴. As a result, this may facilitate the development of more aggressive tumor behaviors, including an increased incidence of metastases, as observed in patients exposed to pesticides.

Clinicopathological data analysis revealed significant reductions in IL-12 levels in exposed breast cancer patients compared to unexposed individuals, particularly concerning age at diagnosis, BMI, and menopausal status. A previous study reported that pesticide exposure leads to lower IL-12 levels in breast cancer patients with early-onset disease⁴⁵. Given that IL-12 is associated with type I immune responses, which are crucial for cancer defense, its reduction may contribute to disease progression and increased aggressiveness, especially in young rural women. Furthermore, reduced IL-12 levels could be detrimental even in clinically favorable conditions, such as a normal BMI. Additionally, the combination of low IL-12 levels and high estrogen levels observed in premenopausal women may have significant implications for tumor development and progression. These findings support our previous studies, suggesting that the IL-12 axis is a target of pesticide exposure in breast cancer and is associated with poor clinical outcomes^46–48.

We also observed a significant reduction in IL-1β levels in patients exposed to pesticides, further supporting the idea that pesticide exposure can induce immune dysfunction, as reported in other studies involving breast cancer patients exposed to these compounds⁴⁹. The immune response mediated by the Th1 pathway is well-established for its antitumor function, and it is associated with better prognoses and longer life expectancy in cancer patients⁵⁰. Therefore, the decrease in Th1 cytokines induced by pesticide exposure, as observed in this study, suggests a significant impairment of the immune response, potentially hindering the body’s ability to control and combat breast cancer. These findings support the hypothesis that pesticides may have a direct toxic effect on immune cells, which could explain the depletion of Th1 cytokines in breast cancer patients with poorer prognosis. While similar results have been reported by Santos⁴⁶ the available data in the literature remains limited, underscoring the importance and need for the research presented in this study.

Additionally, we explored the hypothesis that exposure to pesticides significantly alters circulating and tissue levels of IL-17, a cytokine associated with the Th17 response. The Th17 axis is essential in both innate and adaptive immunity and plays a role in modulating breast cancer progression through cytokine production⁵¹. In our study, we observed a significant reduction in circulating IL-17 levels in patients occupationally exposed to pesticides compared to the unexposed group. In the tissue of these patients, IL-17 did not show a significant difference between the evaluated groups. In summary, these data suggest a possible immunological impairment. Although we did not find specific data on IL-17 dysregulation in the context of pesticide exposure and breast cancer, studies in non-cancerous conditions suggest that IL-17 levels do not show significant variation among workers exposed to pesticides⁵². However, research in both in vitro and in vivo models indicates that IL-17 plays a role in processes related to carcinogenesis^51,53,54.

Therefore, decreased levels of these cytokines may contribute to the development of substantial tumor masses over time, potentially leading to more aggressive tumor behavior, including metastasis. In our study, patients exhibited commitment of circulating cytokines, reflecting an inability to produce adequate cytokines. This depletion of cytokines in the tissues of patients exposed to pesticides was corroborated, resulting in the commitment of naïve T cells. T cell commitment is associated with reduced cytokine production, leading to decreased cytotoxic effector activity of T cells and reduced leukocyte infiltration. This phenomenon is complex and raises multiple questions³⁴, as different exhaustion profiles may emerge, including irreversibly exhausted T cells.

Although a reduction in IL-12 and IL-17A levels was observed, it is important to emphasize that this trend does not extend to all immunological markers evaluated, since IL-1β levels in tumor tissue did not show a statistically significant difference. These results support the hypothesis that pesticides may exert a direct exclusion effect on immune system cells, contributing to the depletion of cytokines associated with Th1 and Th17 profiles, especially in patients with an unfavorable clinical prognosis. The scarcity of human data in this context reinforces the relevance of the present investigation and highlights the need for additional studies to deepen the understanding of these mechanisms.

In this context, we assessed the TILs content and the expression of CTLA-4 in biopsies of these patients. Reduced TILs content was observed in pesticide-exposed patients, reflecting the reduced cytokine levels observed in plasma and breast tumor samples of these patients. Furthermore, high expression of CTLA-4 was observed in pesticide-exposure biopsies. This molecule is involved in immune checkpoint mechanisms and plays a crucial role in T cell inactivation. The interaction between CTLA-4 and CD80 inhibits positive costimulatory signaling mediated by CD28, thereby providing inhibitory signals to T cells, which can suppress cytokine production⁵⁵. Pesticides are capable of activating the aryl hydrocarbon receptor (AhR) pathway, which regulates the differentiation of regulatory T cells and promotes CTLA-4 expression, thereby favoring a functionally exhausted phenotype in CD8⁺ T cells^56,57. These combined mechanisms may contribute to the immunosuppressive profile observed in exposed patients, potentially compromising the immune response.

Our results suggest that this immunological molecule could be impairing leukocyte infiltrates and cytokine production, supporting findings from others⁴⁷. It suggests that women with breast cancer exposed to pesticides experience immune compromise at both systemic and tissue levels. This scenario involves multiple aspects of immune dysregulation, including reduced tumor-infiltrating lymphocyte (TIL) content, increased CTLA-4 expression by tumor cells, and decreased IL-12 levels both systemically and within the tissue, suggesting an environment favorable to immune evasion by the tumor. However, it is important to emphasize that these represent only a subset of the complex immunological landscape involved in tumor progression. Numerous other immune markers and pathways may also be implicated in this process, underscoring the need for further studies to achieve a more comprehensive and accurate understanding of the underlying mechanisms.

Although the literature does not yet provide specific data on the effects of pesticide exposure on the immunological profile of breast TILs, several studies demonstrate that lymphocytic toxicity induced by these substances occurs through different mechanisms, including a decrease in the Th1 cell population^47,58, which corroborate our findings. These results indicate that exposure to pesticides interferes with the pattern of TILs and compromises cytokine production in occupationally exposed breast cancer patients. The analysis of these markers could contribute to a more accurate stratification of disease prognosis in breast cancer patients subjected to chronic exposure to pesticides.

Immune responses play a crucial role in the modulation of systemic oxidative stress, a key event in the development and progression of breast cancer. However, the interaction between oxidative stress and pesticide exposure in this context is still not fully understood. Breast cancer is influenced by various intracellular and extracellular factors, including reactive oxygen and nitrogen species (RONS), which generate an increased production of ROS compared to normal cells, thus facilitating tumor proliferation and metastasis⁵⁹. While no changes were observed in ROS/antioxidant levels in our study, pesticide exposure led to a reduction in circulating NOx levels. This metabolite is directly linked to the immune response, and its decrease may result from the impaired cytokine production observed in the context of pesticide exposure. The cellular accumulation of reactive nitrogen species (RNS) is implicated in cancer initiation and progression, causing alterations in gene expression profiles, dysregulating signaling pathways, and inducing abnormal modifications in proteins⁶⁰. On the other hand, the reduction of elevated intracellular RNS concentrations controls reversible modifications of regulatory proteins caused by oxidative/nitrosative stress, acting as redox switches that regulate the activities of intracellular effectors downstream of different cellular signaling pathways through ROS/RNS signaling. RNS have also been implicated in disrupted cellular homeostasis, leading to deleterious consequences through deregulation of signaling pathways and the induction of protein modifications⁶⁰.

Reduced iNOS expression was observed in TILs from pesticide-exposed biopsies, indicating that breast tumor tissue was also impacted in relation to the NO axis. Evidence suggests a relevant role for iNOS in the biology of tumor infiltrating lymphocytes (TILs), with iNOS expression being a potential marker of immune response, providing a cytotoxic activation in immune effector cells^61–63. Tumor-associated iNOS overexpression has been also correlated with a favorable prognosis in breast cancer⁶⁴. In contrast, reduced iNOS expression is frequently associated with tumor recurrence and shorter survival in other cancer types, such as ovarian cancer⁶⁵. We observed that iNOS expression in TILs occurred in 20% of cases in the exposed group, compared to 40% in the unexposed group. It evidenced the negative influence of pesticide exposure in modulating iNOS expression, probably reflecting on the observed reduced systemic NOx levels.

In light of the evidence presented, it becomes increasingly clear that chronic exposure to pesticides poses a tangible threat to the integrity of mammary tissue, regardless of whether it is in a healthy or neoplastic state. Although still limited, previous studies have indicated that non-tumoral mammary cells exposed to such compounds may undergo early alterations compatible with carcinogenic processes⁶⁶ These changes include DNA damage, dysregulation of hormonal and proliferative pathways⁶⁷, reduced activity of antioxidant mechanisms, increased levels of inflammatory markers such as TNF-α, elevated expression of PPAR-γ, and insufficient activation of classical immune pathways such as IL-12 and NF-κB⁶⁸, as well as relevant epigenetic modifications²³. Collectively, these findings suggest a disruption in the homeostatic balance between cellular defense mechanisms and environmental aggressors, which, under conditions of continuous exposure, may favor the malignant transformation of mammary tissue.

This study have some limitations, including: (i) only selected cytokines were validated; (ii) exposure was self-reported and not individually quantified for all participants; (iii) cytokine and proteomic data reflect measurements from a single point in time, limiting insight into dynamic immune or oxidative responses; (iv) potential confounding factors such as genetic susceptibility, diet, other environmental exposures and comorbidities were not fully controlled for; (v) modest sample size; (vi) single-point cytokine measurement and the absence of pesticide measurements in all samples; (vii) the rural population studied may have unique genetic, dietary, or environmental factors not generalizable to all breast cancer populations; and (viii) validation in independent cohorts and longitudinal studies would be required before these markers could be considered clinically relevant.

Despite this, it demonstrated that pesticide exposure can disrupt inflammatory components, leading to a state of immune compromise and IL-12 deficiency, especially in patients with specific clinical features. This disruption affects the body’s ability to generate oxidative stress mediators, contributing to the generation of nitrosative stress. Also, our data highlight the complex relationship between pesticide exposure and the development of breast cancer, showing that pesticides suggest a potential contribution to immune modulation by negatively regulating the immune, inflammatory and oxidative profiles of breast cancer patients, which may result in a worse prognosis and poor clinical outcomes.

Methods

Study design

This research was approved by the Institutional Ethics Board with the reference number CAAE 35,524,814.4.0000.0107. The study involved 215 individuals diagnosed with breast cancer who lives in the 27 municipalities that compose the 8th Health Care Regional, located at Francisco Beltrão, Paraná, Brazil, and assisted by the Francisco Beltrão Cancer Hospital (Ceonc – Associação Beneficente Deus Menino), from May 2015 to December 2023. Each participant provided informed consent, and the study adhered to the ethical guidelines outlined in the Declaration of Helsinki for research involving human subjects. The selection of patients, execution of assays, and analysis of data adhered to the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) criteria.

Pesticide exposure assessment

To assess pesticide exposure, participants completed a comprehensive 61-question survey covering their past and present occupational experiences, previously developed a validated by our team³. Based on their responses, we classified them as either occupationally exposed or unexposed to pesticides. Women in the exposed group (n = 128) reported spending at least half of their lives working with pesticides and having direct contact with these chemicals at least once a week. Their exposure included handling pesticide-contaminated clothing and personal protective equipment (PPE) worn by family members, preparing and diluting concentrated pesticides, and assisting in spraying diluted pesticides on crops. These tasks were typically performed for 4–8 h per day over 2–3 consecutive days, every 1–2 weeks. In contrast, women in the unexposed group had no occupational history involving pesticides. They did not engage in rural work requiring pesticide use, had no past or present contact with these chemicals, and never handled or washed pesticide-contaminated clothing or PPE. The lack of an occupational exposure history qualified them as the unexposed group (n = 87).

Proteomic study

For the proteomic analysis, patients representing all molecular subtypes were included. Plasma samples were grouped according to pesticide exposure status (exposed vs. non-exposed), using pooled plasma samples from individuals classified based on their exposure history, in accordance with previously described methodologies⁶⁹. The protein concentrations in clarified supernatants were gauged using the Qubit™ Protein Assay Kit from Life Technologies. Subsequently, samples underwent concentration and exchange with 50 mM ammonium bicarbonate employing a 3-kD ultra-filtration device by Millipore. Protein extracts (50 µg) were subjected to denaturation (0.1% RapiGEST at 60 °C for 15 min, Waters, Milford, USA), reduction (10 mM dithiothreitol at 60 °C for 30 min), alkylation (10 mM iodoacetamide for 30 min at room temperature in the dark), and enzymatic digestion with trypsin at a 1:50 (w/w) enzyme to protein ratio. The digestion process was halted by the addition of 10µL of 5% trifluoroacetic acid (TFA)^69,70. Peptides were desalted using C18 micro columns (Harvard apparatus), dried via vacuum centrifugation, resuspended in 0.1% formic acid, quantified by Qubit protein assay, and subjected to analysis via label-free methodologies. Two micrograms of digested peptides underwent triplicate technical analyses following a 3-h gradient (5% to 40% B/167 min; 40% to 95% B/5 min; 95% B/8 min). The solvent composition for Easy-nanoLC1000 (Thermo Fisher) was as follows: solvent A (95% H2O/5% acetonitrile (ACN)/0.1% formic acid) and solvent B (95% ACN/5% H2O/0.1% formic acid). The trap-column employed was Easy column C18, 2 cm × 100 μm i. d. × 5 μm, 120 Å, while the analytical column measured 25 cm in length with an internal diameter of 75 μm (3 μm spheres, Reprosil Pur C18). Label-free quantification was conducted utilizing an Easy-nLC 1000 (Thermo Scientific) coupled to a QExactive Plus operating in FullScan-DDA MS2 mode with dynamic exclusion set at 45 s and a spray voltage of 2.70 kV. Full scan parameters included a resolution of 70,000 at m/z 200, with a m/z range of 350−2000, AGC of 1 × 106, and injection time of 50 ms. The 15 most intense ions for HCD fragmentation were selected with a normalized collision energy of 30, precursor isolation window of m/z 1.2 and 0.5 offset, resolution of 17 500 at m/z 200, AGC at 5 × 105, and injection time of 100 ms. For data analysis, all replicate samples were processed using Proteome Discoverer 2.1 software employing the human database UniProt (V. Nov 2018-https://www.uniprot.org/). The parameters included full-tryptic search space, up to two missed cleavages allowed for trypsin, precursor mass tolerance of 10 ppm, and fragment mass tolerance of 0.05 Da. Carbamidomethylation of cysteine was set as a fixed modification, while methionine oxidation and protein N-terminal acetylation were considered as dynamic modifications for label-free quantification. Spectra analyses utilized a target-decoy strategy with a maximum delta CN of 0.05, all available peptide-spectrum matches, and a target false discovery rate (FDR) of 0.01 (strict). Peptide and protein filtering criteria were established for high confidence. Statistical analysis was performed using the Perseus Computational Platform v 1.6.10.50, involving Log2 transformation, median normalization, and Student t-test (p-value < 0.05). Volcano plot presentation illustrated the distribution of log p-value versus log2 fold differences.

Functional in silico analyses were conducted utilizing FunRich: Functional Enrichment Analysis Tool, incorporating integrated databases such as the Human Protein Reference Database (HPRD), Entrez Gene, and Uniprot for biological pathways and processes. For protein–protein interactions, databases such as BioGRID, Intact, and Human Proteinpedia were accessed. Information on localization, expression, and signaling pathways were derived from databases including the Human Protein Atlas, Human Proteome Browser, Human Proteome Map, Proteomics DB, Reactome, NCI, Cell map, and HumanCyc. Transcription factor consensus sequences prediction data were obtained via automated search in 29 mammalian databases using the FunRich tool, followed by Bonferroni statistical analysis.

Cytokine determination

Based on the high-throughput proteomic screening and in silico analyses, we investigated circulating and tumor immune, inflammatory, and oxidative stress-related mediators. Cytokine analyses were conducted by using commercial antibody-specific enzyme-linked immunosorbent assay (ELISA) kits in plasma and Cytometric Bead Arrays (CBA) kits in breast tumors (BD Biosciences, USA). Results were quantified as pg/mL for plasma samples and pg/mg tissue for biopsies, by fitting the data to a standard curve obtained using human recombinant cytokines.

Oxidative stress measurement

Lipid peroxidation, total antioxidant capacity (TRAP) and nitric oxide metabolites (NOx) levels were assessed to evaluate the oxidative stress profile, as previously published⁷¹. To assess lipid peroxidation levels, plasma samples were diluted in phosphate buffer containing tert-butyl hydroperoxide. Real-time high sensitivity chemiluminescent reactions were monitored using a Glomax 20/20 luminometer (Promega, USA) for 60 min. The obtained results were expressed as relative light units (RLU). For TRAP determination, plasma samples were diluted and mixed with a solution containing luminol and 2,2 azobis. Chemiluminescence profiles were measured in a luminometer (Promega, USA) for 30 min. The obtained results were expressed as nM of Trolox, a soluble vitamin E analogue. For NOx content, plasma aliquots underwent deproteinization, centrifugation, recovery of the supernatant, and dilution in glycine buffer, followed by incubation with activated cadmium granules. Subsequently, aliquots of the recovered supernatant were mixed with Griess reagent in equal proportions. Absorbance was measured at 550 nm using a standard microplate reader, and results were expressed as μM nitrite.

real-time PCR: RNA extraction and IL-17A analysis by quantitative polymerase chain reaction (qPCR)

Following the quantification of IL-17A by ELISA and CBA methods, an additional analysis was performed using PCR. Total RNA was extracted from both cultured cell lines (HeLa, THP-1, and RAW 264.7) and homogenized breast tumor tissue samples using the Quick-RNA™ Miniprep Kit (Zymo Research Corporation), following the manufacturer’s protocol. Tumor tissue was snap-frozen in liquid nitrogen and stored at − 80 °C, while cells were maintained in DMEM supplemented with 10% fetal bovine serum (FBS) under standard conditions (37 °C, 5% CO₂).

RNA purity and concentration were assessed using a NanoDrop™ One UV–Vis Spectrophotometer, ensuring A260/280 ratios between 1.8 and 2.0. cDNA synthesis was performed using the SuperScript™ IV Reverse Transcriptase Kit (Invitrogen, cat. no. 18090010) in a 20 µL reaction containing 5X RT buffer, 100 mM DTT, dNTP mix (10 mM each), gene-specific primers (4 µM), RNaseOUT™, and 200 U/µL SuperScript™ IV.

qPCR was conducted on a QuantStudio™ Real-Time PCR System (Applied Biosystems) using iTaq Universal SYBR^® Green Supermix (Bio-Rad) in a 20 µL reaction mixture containing 10 µL of SYBR Green mix, 400 nM of each primer, and 2 µL of cDNA, with nuclease-free water completing the final volume. Reactions were performed in a 96-well plate, with technical triplicates for each target gene and a no-template control (NTC) included to monitor contamination.

Thermal cycling conditions were as follows: initial denaturation at 95 °C for 30 s, followed by 40 cycles of 95 °C for 30 s, 55–58 °C for 30 s (annealing, optimized per primer), and 72 °C for 60 s, concluding with a melt curve analysis from 60 °C to 95 °C to confirm primer specificity and detect primer-dimer formation.

Gene-specific primers were used as follows: IL-17A: Forward 5′-GGACTGTGATGGTCAACCTGA-3′, Reverse 5′-TCATGTGGTAGTCCACGTTCC-3′, GAPDH: Forward 5′-TGGGAACCTGGTCAATCACC-3′, Reverse 5′-GCATCACCCCATTTGATGTT-3′, HuPO: Forward 5′-GCAATGTTGCCAGTGTCTGT-3′, Reverse 5′-GCCTTGACCTTTTCAGCAAG-3′.

Primer specificity was confirmed through melt curve analysis, ensuring a single peak without primer-dimer formation. Relative gene expression was determined using the Comparative Ct (ΔΔCt) method, normalizing IL-17A Ct values against the geometric mean of GAPDH and HuPO reference genes (ΔCt). Expression levels were calculated using the formula 2^(-ΔΔCt), comparing results to control samples such as untreated cells or normal tissue. Data analysis was performed using QuantStudio™ Design & Analysis Software.

TILs estimation

The density of TILs on hematoxylin and eosin (H-E) stained breast tumor biopsy sections was determined based on the recommendation by the International TILs Working Group⁷². Tumor area was first selected at low magnification and the percentage of the area that was filled with mononuclear cells assessed at high magnification (× 100). All mononuclear cells were defined as TILs, excluding granulocytes and other polymorphonuclear leukocytes. For evaluating the density of TILs, we used a semi-quantitative score, classifying as 1 + (< 25%), 2 + (26–60%), or 3 + (> 60%).

Immunofluorescence labeling for CTLA-4 and iNOS

Paraffin-embedded tissue Sects. (3 μm thick) were first deparaffinized using xylene and then rehydrated through a series of ethanol gradients. Antigen retrieval was performed by heating the tissue sections in a 1 mM citrate buffer (pH 6.0) at 100 °C for 15 min. To block endogenous peroxidase activity, the slides were immersed in a 3% hydrogen peroxide (H₂O₂) solution for 15 min. Next, a blocking solution containing 5% fetal bovine serum and 0.3% Triton X was applied, and the sections were incubated in a humidified chamber at 37 °C for 15 min to prevent nonspecific antigen binding. The sections were then incubated overnight at 4 °C with a mouse anti-human CTLA-4 IgG monoclonal antibody (1:1000 dilution, BNI3 clone, Thermo Fisher Scientific) or a mouse anti-human iNOS monoclonal antibody (1:1000 dilution, Santa Cruz Biotechnologies, catalog number sc-7271). Following this, the slides were rinsed five times in 0.9% saline for 5 min each, kept protected from light, and incubated at 37 °C for 60 min with a secondary antibody (Alexa Fluor 488 goat anti-mouse IgG, 1:1000 dilution, Thermo Fisher Scientific). The sections were then washed twice in 0.9% saline for 5 min each before being incubated with 4’,6-diamidino-2-phenylindole (DAPI, Sigma Aldrich) at a concentration of 5 mg/mL for 30 min. Finally, the slides were mounted using glycerol and stored under light-protected conditions for fluorescence microscopy analysis. Negative controls were prepared by incubating sections with the antibody diluent without the primary antibody. Immunoreactivity intensity and localization were assessed using the Motic BA 410E fluorescence microscope and MOTICAM ProS5 Plus camera. BC slides were independently evaluated by two researchers blinded to the patients’ clinical and prognostic data. Immunofluorescence-positive cells were categorized on a scale from 0 to 4: negative (0), < 1% (1), 1–10% (2), 11–33% (3), 34–66% (4), and > 60%. Quantification in positive cells was performed using the ImageJ software.

Data analysis

Statistical assumptions of normality (Shapiro–Wilk test) and homoscedasticity (Levene’s test) were assessed for all variables. Duplicate dataset analyses were represented using box plots (min–max), with the average value indicated by + . Outliers were detected using the Grubbs test, with none observed in the study. Results were compared using Student’s t-test or the Mann–Whitney test, considering p ≤ 0.05 as statistically significant. GraphPad Prism 9.0 software package (GraphPad Software, San Diego, CA, USA) was utilized for statistical analyses.

Author contributions

All authors equally contributed to data collection and analysis and critically revised and approved the final manuscript. CP and JMBMV designed the study, collected and analyzed data, wrote the first draft of the manuscript and revised its final version.

Funding

This work was supported by Brazilian National Council for Scientific and Technological Development (CNPq; Grants 402364/2021-0, 441017/2023-1, and 305335/2021-9), Secretariat of Science, Technology, and Higher Education of the State of Paraná (SETI, grant 220/2024), Araucária Foundation (grant 078/2022), and Itaipu Binacional (grant 4500007500/2024).

Data availability

Proteomics data are listed as supplementary Table 1. Further data will be available upon a reasonable request. Please get in touch with Dr. Panis.

Declarations

Competing interests

The authors declare no competing interests.