1 Introduction

Water contamination by pesticides is a global concern, with a particular focus on Brazil because it ranks among the leading consumers of pesticides in the world (Brovini et al., 2023; Panis et al., 2022). According to the latest data from the Brazilian Institute of the Environment and Renewable Natural Resources, the number of active ingredients sold in the country doubled in a decade, rising from 425 thousand tons in 2012 to 800 thousand tons in 2022 (IBAMA, 2012, 2022). The widespread use of pesticides in Brazil has been defended based on the favorable climate that promotes large-scale agricultural output centered on monocultures and the production of agrofuels (Brovini et al., 2021a; Gaboardi et al., 2023). Moreover, the country has permissive legislation that favors the production and registration of pesticides while neglecting inspection (Gaboardi et al., 2023; Souza et al., 2023a).

The herbicides glyphosate [N-(phosphonomethyl) glycine] and 2,4-dichlorophenoxyacetic acid (2,4-D) are the most widely used pesticides in Brazil and worldwide (Brovini et al., 2021a). Growers frequently combine these herbicides in tanks to achieve more efficient control of difficult-to-control weeds (Palma-Bautista et al., 2021; Takano et al., 2013). Heavily contaminated agricultural soils with glyphosate and 2,4-D lead to runoff and leaching (Carles et al., 2021; Lima et al., 2023), causing adverse effects on non-target organisms.

Considering this aspect, most of the knowledge on the toxic and genotoxic effects of glyphosate and 2,4-D on non-target organisms is derived from the effects of each compound separately. In this sense, both herbicides caused acute and chronic toxicity (Brovini et al., 2021a; Martins et al., 2021; Rodrigues et al., 2019), oxidative stress, and the induction of DNA damage (Arcaute et al., 2018; Gaaied et al., 2019; Nwani et al., 2013; Rodrigues et al., 2019) and micronuclei and chromosomal mutations (Marcato et al., 2017; Mercado & Caleño, 2020; Montero-Montoya et al., 2023; Truta et al., 2011; Vieira et al., 2024) in different organisms.

The International Agency for Research on Cancer categorizes glyphosate as probably carcinogen and 2,4-D (IARC, 2017) as possibly carcinogen to humans (IARC, 2018). In their review study, Panis et al. (2022) demonstrated the risks linked to widespread contamination of drinking water with proven, probable, or potentially carcinogenic pesticides in the second-largest grain-producing state in Brazil. In fact, in contrast to other countries with more restrictive legislation, pesticides with mutagenic, carcinogenic, and reproductive risks have been widely registered in Brazil, in an evident conflict of interests between the agribusiness industry and the scientific community (Rocha & Grisolia, 2019).

Since there is no safe exposure limit for genotoxic substances, and genotoxicity and mutagenicity are predictive factors for several human diseases, including cancer (Marchetti et al., 2023), the understanding the effects of 2,4-D and glyphosate in combination is particularly important because their interaction may result in unexpected genotoxic damage to non-target organisms.

However, there have been limited studies on the cytogenotoxic effects of mixtures containing 2,4-D and glyphosate. Additionally, most of these studies did not evaluate concentrations that reflect those found in the water, or only a limited number of mixtures were examined. For instance, Bernardi et al. (2022) only associated one high concentration for the aquatic environment (5 mg L −1 of 2,4-D + 10 mg L −1 of glyphosate) and reported increased DNA damage using the comet assay on fish. Finkler et al. (2022) assessed the highest concentrations applied in the agricultural fields (1.56 mg mL−1 for glyphosate and 0.28 mg mL−1 for 2,4-D) and the acute reference doses (11.66 mg mL−1 for glyphosate and 1.75 mg mL−1 for 2,4-D). This study found that these mixtures decreased the mitotic index and increased the frequency of chromosomal aberrations on Allium cepa.

Carvalho et al. (2020) evaluated the effects of several commercial formulations containing 2,4-D and glyphosate on fish. The mixtures consisted of 5% + 5% or 10% + 10% of the LC50 for each substance: 4.58 and 9.17 mg L−1 glyphosate, 0.023 and 0.046 mg L−1 2,4-D-acid, 33.90 and 67.80 mg L−1 2,4-D-amine and 0.13 and 0.26 mg L−1 2,4-D-ester. The results showed that the mixtures increased the genotoxic damage index, as measured by the comet assay.

Finally, Pavan et al. (2021) assessed five mixtures containing 2,4-D and glyphosate in concentrations representative of the aquatic environment. The lowest concentrations associated were 2,4-D at 4 µg L−1 + glyphosate at 65 µg L−1, which represent the maximum values allowed in Brazilian freshwater (Brasil, 2005; Table 1). The highest mixture assessed was 2,4-D at 74.5 µg L−1 + glyphosate at 1000 µg L−1. The mixtures increased the frequency of micronuclei and other erythrocyte nuclear abnormalities, as well as caused morphological and behavioral changes in tadpoles. It is important to note that this study only evaluated a limited number of mixtures. In addition, the authors did not compare the results of the mixtures regarding individual herbicide concentrations.

Table 1 Maximum values allowed for 2,4-D and glyphosate herbicides (µg L−1) in fresh and potable water in Brazil and other countries
Full size table

To address these gaps, we assessed the cytogenotoxic potential of 16 mixtures containing realistic environmental concentrations set by Brazilian legislation to protect aquatic life and human health. The concentrations tested ranged from 0.4 to 400 μg L−1 for 2,4-D and 0.65 to 650 μg L−1 for glyphosate. The mixtures were formulated to replicate various scenarios, ranging from a less severe to the most extreme contamination scenario. Additionally, the study compared the effects of the mixtures concerning the individual herbicides. This is because chemical mixtures can result in either enhanced (additive or synergistic effect) or reduced (antagonistic effect) levels of toxicity compared to those of each herbicide when used in isolation (Ilyushina et al., 2020).

In this study, the cytogenotoxic assay was conducted on A. cepa root tip cells. This plant model is commonly used to evaluate environmental contaminants, including pesticides (Felisbino et al., 2018; Liman et al., 2015; Souza et al., 2016, 2022, 2023a, b; Türkoğlu, 2012; Verma & Srivastava, 2018). The A. cepa test has several advantages. Firstly, it assesses different endpoints such as cytotoxicity, genotoxicity, and mutagenicity. Secondly, it has a high correlation with carcinogenesis tests in rodents. Thirdly, it can detect chemical toxins even at low concentrations. Finally, it can indicate the mechanisms of action of the tested agents, whether pure substances or complex mixtures (Rank & Nielsen, 1994). Camilo-Cotrim et al. (2022) stated that the A. cepa bioassay is an alternative method to animal research based on the 3Rs (Replace Reduce Refine) principle due to its strong correlation with the carcinogenesis assay in rodents (Rank & Nielsen, 1994).

We hypothesize that the combination of 2,4-D and glyphosate may result in mixtures with increased cytogenotoxic effects, which are precursors of genetic instability.

2 Materials and Methods

2.1 Pesticides

Two commercial herbicide formulations were used in this study: (1) Roundup Original DI® Monsanto, containing 445 g L−1 (44.5% w/v) diammonium salt of [N-(phosphonomethyl) glycine], 370 g L−1 (37% w/v) acid equivalent of [N-(phosphonomethyl) glycine] and 751 g L−1 (75% w/v) other ingredients; and (2) U-46 BR, containing 806 g L−1 (80.6% w/v) 2,4- 2,4-dichloro phenoxy acetic acid (2,4-D amine) dimethylamine salt, 670 g L−1 (67% w/v) acid equivalent and 380.9 g L−1 (38.09% w/v) other ingredients.

The concentrations of 2,4-D and glyphosate were determined based on the acid equivalent concentration found in the commercial formulations. Stock solutions of 670 µg L−1 2,4-D and 740 µg L−1 glyphosate were prepared, as well as test solutions of 0.4, 4, 40 and 400 µg L−1 2,4-D and 0.65, 6.5, 65 and 650 µg L−1 glyphosate. The solutions were prepared using distilled water as the dilution medium.

The concentrations of individual herbicides were chosen based on the maximum values allowed by Resolution No. 357 of the National Council for the Environment (CONAMA) (Brasil, 2005). This resolution provides guidelines for the classification of water bodies. In class I waters (suitable for human consumption after simplified treatment and for the protection of aquatic communities, recreation, and irrigation) and class II waters (suitable for human consumption after conventional treatment and for the protection of aquatic communities, recreation, irrigation, sport, leisure, aquaculture, and fishing), the maximum permitted concentrations of 2,4-D and glyphosate are 4 and 65 µg L−1, respectively (Table 1).

The concentrations tested also included a) the maximum values allowed for class III waters (suitable for human consumption after conventional or advanced treatment and for recreation, irrigation, fishing and watering of animals) which can contain up to 30 µg L−1 2,4-D and 280 µg L−1 glyphosate (Brasil, 2005); b) the maximum values established by Ordinance No. 2914 of the Ministry of Health (Brasil, 2011) legislating potable water, which regulates the levels of pesticides in water intended for human consumption (30 µg L−1 for 2,4-D and 500 µg L−1 for glyphosate) (Table 1); and c) the values detected in tap water according to data from the Information System for the Surveillance of the Quality of Water for Human Consumption (SISAGUA), coordinated by the Ministry of Health: up to 30 µg L−1 for 2,4-D and up to 200 µg L−1 for glyphosate (Aranha & Rocha, 2019).

The concentrations tested in this study also included the concentrations of 2,4-D and glyphosate allowed by the legislation of other countries, as presented in Table 1.

The binary mixtures were prepared by combining each concentration of 2,4-D with each concentration of glyphosate. Table 2 provides a summary of the individual concentrations and mixtures tested.

Table 2 Concentrations of herbicides and binary mixtures
Full size table

2.2 Bioassay with A. Cepa

Seeds of A. cepa (variety Baia Periforme) that had not been treated with pesticides were used. Distilled water was used as a negative control. Initially, the seeds were germinated in Petri dishes lined with filter paper moistened with distilled water. When the roots reached approximately 1.5 cm, they were treated with the herbicides and mixtures for 24 h. Next, the roots were fixed in Carnoy I [ethanol + acetic acid (3:1, v/v)] and stored in a refrigerator. To prepare the slides, the roots were washed three times with distilled water for five minutes and then hydrolyzed in 5 N HCl for 20 min. The apical root meristem was sectioned, stained with 2% acetic orcein, and crushed between a slide and coverslip. The slides were analyzed under a light microscope (Leica DM750, Germany) equipped with an eyepiece camera (Leica ICC50, Germany) at × 400 magnification. A total of 5,000 cells per treatment (500 cells/10 slides) were analyzed.

The effects of the herbicides and mixtures on cell cycle progression were evaluated using the phase index and the mitotic index. The phase index is determined by the frequency of cells in a particular phase of mitosis, while the mitotic index is characterized by the proportion of dividing cells. Any changes in these parameters compared to the control indicate cytotoxicity. The genotoxicity of the treatments was inferred by calculating the frequency of aberrant cells, while the mutagenic potential was assessed by determining the frequency of micronucleated cells (Leme & Marin-Morales, 2009). The following formulas have applied to calculate the mitotic index (MI), phases index (PI) and chromosomal aberrations (AC) and micronuclei (MN) frequencies:

$$MI (\%)=\frac{number\;of\;cells\;in\;division}{total\;number\;of\;cells\;observed} \times 100$$
$$PI (\%)=\frac{particular\;phase\;of\;mitosis}{number\;of\;cells\;in\;division}\;\times 100$$
$$AC(\%)=\frac{number\;of\;cells\;with\;mitotic\;and\;chromosomal\;abnormalities}{total\;number\;of\;cells\;observed}\times 100$$
$$MN (\%)=\frac{number\;of\;cells\;with\;micronuclei}{total\;number\;of\;cells\;observed}\;\times\;100$$

2.3 Statistical Analysis

The software used for the statistical analysis was GraphPad Prism version 9.4. The normality of the collected data was tested using the Shapiro‒Wilk test, and the homogeneity of variances was tested using the Brown–Forsythe test. As the data were not normally distributed and showed inequality of variance, comparisons between the experimental groups were performed using the Kruskal‒Wallis test followed by Dunn's test (p < 0.05). Comparisons were performed between a) the concentrations of each herbicide and the control group; b) the mixtures and the control group; and c) the mixture and each individual concentration of pesticide included in the mixture. The results are presented as the mean ± standard deviation.

3 Results and Discussion

3.1 Cytogenotoxic Effect of Isolated Herbicide: 2,4-D

Table 3 displays the results obtained for the 2,4-D which only significantly inhibited the mitotic index at a concentration of 400 µg L−1 (p = 0.0063, H = 28.23; Kruskal–Wallis test). The reduction in mitotic activity is a cytotoxicity marker for all living organisms. It can delay normal development and growth of exposed organisms (Camilo-Cotrim et al., 2022). Antimitotic agents can damage DNA and impair the functioning of essential enzymes involved in cell cycle progression, including polymerases and kinases (Türkoğlu, 2012). Consequently, the cell cycle is transiently blocked due the activation of checkpoints, providing time for damage to be repaired before the transition from G1 to S and from G2 to M (Fioresi et al., 2020; Miranda et al., 2023; Moreira et al., 2021; Souza et al., 2023b).

Table 3 Phase index, mitotic index, chromosomal aberrations and micronuclei (mean ± standard deviation) in meristematic cells of A. cepa treated with the isolated herbicides
Full size table

Considering the genotoxicity index, 2,4-D at 0.4, 4 and 40 µg L−1 significantly increased the frequency of aberrant cells compared to the control (p < 0.0001, p = 0.0247 and p = 0.0225, respectively, H = 29.40; Kruskal–Wallis test). According to Han et al. (2023), genotoxic agents can cause harm to DNA and cellular components responsible for maintaining the integrity of the genome, such as the spindle apparatus, topoisomerases, DNA repair systems, and DNA polymerases. Cells that are damaged and escape the repair processes activated during interphase contribute to an increased frequency of aberrant cells observed during mitosis (Fioresi et al., 2020).

On the contrary, the concentration of 400 µg L−1 did not induce any mitotic and chromosomal changes above basal levels (p > 0.9999, H = 29.40; Kruskal–Wallis test). The low genotoxic damage index observed at this concentration can be attributed to it strong mitodepressive action. According to Dhyèvre et al. (2014), the reduction in mitotic activity can prevent the observation of aberrant cells, i.e., the genotoxic effect can be overshadowed by the cytotoxic effect. The limited number of cells that completed mitotic division inherited damage induced in the parental cells. As a result, the frequency of micronucleated cells was higher in the group treated with 400 µg L−1 2,4-D than in the control group (p < 0.0001, H = 24.68; Kruskal–Wallis test). Micronuclei are among the main biomarkers of DNA damage and chromosomal instability. They arise from chromosome fragments (clastogenesis) or whole chromosomes (aneugenesis) that fail to segregate into daughter nuclei during mitosis (Fenech et al., 2020).

3.2 Cytogenotoxic Effect of Isolated Herbicide: Glyphosate

The results obtained for glyphosate are shown in Table 3. Glyphosate at 0.65 and 65 µg L−1 significantly increased the proportion of dividing cells (p = 0.0054 and p = 0.0068, respectively, H = 20.25; Kruskal–Wallis test). Additionally, both concentrations also increased the frequency of cells in prophase compared to the control (p < 0.0001 and p = 0.0076, respectively, H = 28.50; Kruskal–Wallis test). An increased mitotic index is also cytotoxicity marker and can lead to disordered cell proliferation (Leme & Marin-Morales, 2009). Prophase arrest indicates the interference of chemical agents in crucial events that ensure the correct segregation of genetic material during mitosis, such as chromatin condensation and the assembly of the mitotic spindle. Seong et al. (2002) reported that prophase-arrested mammalian cells exhibited randomly oriented spindle structures. Therefore, prophase delay may result in chromosomal abnormalities.

Marc et al. (2004) demonstrated that glyphosate inhibits the activation of the CDK1/cyclin B, a regulator of the M-phase of the cell cycle, including the main events in prophase (Champion et al., 2017; Schatten, 2013). In addition, Díaz-Martín et al. (2021) reported that glyphosate decreased levels of acetylated α-tubulin and caused microtubule instability.

In this study, glyphosate at concentrations of 6.5, 65, and 650 µg L−1 significantly increased the frequency of aberrant cells compared to untreated cells (p = 0.0157, p = 0.0022 and p < 0.0001 respectively, H = 30.86; Kruskal–Wallis test). Regarding micronuclei, only the concentration of 650 µg L−1 induced a higher damage rate than the control (p = 0.0002, H = 30.05; Kruskal–Wallis test).

3.3 Cytogenotoxic Effect of Mixtures: 2,4-D + Glyphosate

The A. cepa test was also employed to assess the cytogenotoxic effects of the mixtures 2,4-D + glyphosate compared to the control group (Table 4). To better understand these results, we performed a comparative analysis of each mixture with respect to the effects of the individual herbicides (Table 5).

Table 4 Phase index, mitotic index, chromosomal aberrations, and micronuclei (mean ± standard deviation) in meristematic cells of A. cepa treated with mixtures of herbicides 2,4-D + glyphosate (µg L−1)
Full size table
Table 5 Cytogenotoxic effects of the mixtures in relation to that of the individual herbicides
Full size table

Six mixtures, M8, M10, M12, M13, M15 and M16, inhibited the mitotic index compared to the control (p = 0.0002, p = 0.0139, p = 0.0002, p = 0.0256, p = 0.0034 and p < 0.0001, respectively, H = 135.0; Kruskal Wallis-test) (Table 4). When comparing these effects with those of each herbicide, two distinct responses were obtained: (a) mixtures that showed a lower mitotic index than both the herbicides (M8, M10, and M12), which in turn did not differ statistically from each other (M8 vs. 4 µg L−1: p = 0.0175, M8 vs. 650 µg L−1: p < 0.0001, 4 µg L−1 vs. 650 µg L−1: p = 0.1263, H = 52.43; M10 vs. 40 µg L−1: p < 0.0001, M10 vs. 6.5 µg L−1: p = 0.0092, 40 µg L−1 vs. 6.5 µg L−1: p = 0.6394, H = 18.66; M12 vs. 40 µg L−1: p = 0.0006, M12 vs. 650 µg L−1: p = 0.0003, 40 µg L−1 vs. 650 µg L−1: p > 0.9999, H19.40; Kruskal–Wallis test (Table 5); and (b) mixtures that significantly inhibited the mitotic index compared to glyphosate and showed a higher mitotic index in relation the 2,4-D (M13, M15 and M16). In this case, the herbicides differed statistically from each other (M13 vs. 400 µg L−1: p = 0.0333, M13 vs. 0.65 µg L−1: p = 0.0333, 400 µg L−1 vs. 0.65 µg L−1: p < 0.0001, H = 25.81; M15 vs. 400 µg L−1: p = 0.0332, M15 vs. 65 µg L−1: p = 0.0332, 400 µg L−1 vs. 65 µg L−1: p < 0.0001, H = 25.81; M16 vs. 400 µg L−1: p = 0.0333, M16 vs. 650 µg L−1: p = 0.0333, 400 µg L−1 vs. 650 µg L−1: p < 0.0001, H = 25.81; Kruskal–Wallis test) (Table 5).

In summary, the cytotoxicity of mixtures M8, M10, and M12 was observed, while the individual herbicides did not show any cytotoxicity, indicating a synergistic effect. The mixtures M13 and M15 were found to be more cytotoxic than the isolated herbicides and showed additive or synergistic toxicities (Table 5). In general terms, an additive effect is the sum of the response of each substance. When the observed response exceeds the additive effect, this is assigned as a synergistic pattern (Carvalho et al., 2020; Finkler et al., 2022). The effect of M16 can be described as potentiation, in which the effect of only one of the constituents of the mixture is exerted (Felisbino et al., 2018).

Based on these results, it can be inferred that the cytotoxicity of the mixtures M8, M10, M12, M13, and M15 was due to the toxic action of the two herbicides, while only the M16 mixture (400 µg L−1 2,4-D + 650 µg L−1 glyphosate) exhibited an adverse effect attributed to 2,4-D.

Considering the mixtures that did not cause a change in the mitotic index compared to the control, the findings obtained were as follows: (a) the response of M1, M3, M4, and M6 was equivalent to that of the individual herbicides (M1 vs 0.4 µg L−1: p = 0.0545, M1 vs. 0.65 µg L−1: p > 0.9999, 0.4 µg L−1 vs 0.65 µg L−1: p = 0.1303, H = 6.52; M3 vs. 0.4 µg L−1: p > 0.9999, M3 vs. 65 µg L−1: p = 0.1383, 0.4 µg L−1 vs. 65 µg L−1: p = 0.1752, H = 5.05; M4 vs. 0.4 µg L−1: p = 0.8238, M4 vs. 650 µg L−1: p = 0.0582, 0.4 µg L−1 vs. 650 µg L−1: p = 0.6394, H = 5.47; M6 vs. 4 µg L−1: p = 0.3115, M6 vs. 6.5 µg L−1: p = 0.1801, 4 µg L−1vs. 6.5 µg L−1: p > 0.9999, H = 4.165; Kruskal–Wallis test); b) M2 showed a higher mitotic index than both the herbicides (M2 vs 0.4 µg L−1: p = 0.0162, M2 vs. 6.5 µg L−1: p > 0.0001, 0.4 µg L−1 vs 6.5 µg L−1: p = 0.4427, H = 18.49; Kruskal–Wallis test); (c) M5 and M7 exhibited equivalent mitotic index to that 2,4-D, but significantly lower than glyphosate (M5 vs. 4 µg L−1: p = 0.1114, M5 vs. 0.65 µg L−1: p > 0.0001, 4 µg L−1 vs. 0.65 µg L−1: p = 0.0622, H = 19.35; M7 vs. 4 µg L−1: p = 0.1910, M7 vs. 65 µg L−1: p > 0.0001, 4 µg L−1 vs 65 µg L−1: p = 0.0412, H = 18.77; Kruskal–Wallis test); (d) M9 and M11 showed a lower mitotic index than the both 2,4-D and glyphosate (M9 vs 40 µg L−1: p = 0.0092, M9 vs. 0.65 µg L−1: p = 0.0002, 40 µg L−1 vs. 0.65 µg L−1: p = 0.9286, H = 17.07; M11 vs. 40 µg L−1: p = 0.0156, M11 vs. 65 µg L−1: p = 0.0005, 40 µg L−1 vs 65 µg L−1: p > 0.999, H = 15.25; Kruskal–Wallis test); and (e) M14 exhibited equivalent mitotic index to that glyphosate, but significantly higher than 2,4-D (M14 vs. 400 µg L−1: p = 0.0175, M14 vs. 6.5 µg L−1: p = 0.1049, 400 µg L−1 vs. 6.5 µg L−1: p < 0.0001, H = 23.81; Kruskal–Wallis test) (Table 5).

In relation to non-cytotoxic mixtures, M5 and M7 showed a mitotic index equivalent to that of 2,4-D, but significantly lower than glyphosate (individually cytotoxic). Conversely, other hand, the cytotoxicity of 2,4-D was reduced in M14. As the toxicity of the herbicides was nullified in these mixtures, an antagonistic effect is suggested. According to Carvalho et al. (2020), antagonism occurs when the toxicity of the mixture is smaller than the effect caused by any of the components alone. Although M1 and M3 are similar to 2,4-D and glyphosate, they showed an antagonistic effect. Similarly, M9 and M11 were considered antagonist mixtures when compared to the separate herbicides. On the other hand, there were no significant interactions observed in the herbicide mixtures M2, M4, and M6 that would result in a discernible alteration in the mitotic index.

Regarding the chromosomal aberration index, M1, M2, M4, M5, M6, M9, M13 and M16 (half of the mixtures) induced greater damage than the control (p < 0.0001, p < 0.0001, p = 0.0036, p = 0.0026, p < 0.001, p = 0.0004, p = 0.0072 and p = 0.0130, respectively, H = 52.43; Kruskal–Wallis test) (Table 4) and yielded the following results: a) mixture whose genotoxicity is equivalent to 2,4-D (M1) (M1 vs. 0.4 µg L−1: p = 0.9452, M1 vs. 0.65 µg L−1: p = 0.0342, 0.4 µg L−1 vs. 0.65 µg L−1: p = 0.0012, H = 13.27; Kruskal–Wallis test) (Table 5); b) mixture whose genotoxicity is equivalent to glyphosate (M13) (M13 vs. 400 µg L−1: p = 0.0246, M13 vs. 0.65 µg L−1: p > 0.9999, 400 µg L−1 vs. 0.65 µg L−1: p = 0.1881, H = 7.38; Kruskal–Wallis test) (Table 5); and mixtures whose genotoxicity is equivalent to both herbicides (M2, M4, M5, M6, M9 and M16) (M2 vs 0.4 µg L−1: p = 0.3706, M2 vs. 6.5 µg L−1: p = 0.2940, 0.4 µg L−1 vs. 6.5 µg L−1: p = 0.0042, H = 10.21; M4 vs 0.4 µg L−1: p = 0.2057, M4 vs. 650 µg L−1: p = 0.0573, 0.4 µg L−1 vs. 650 µg L−1: p > 0.9999, H = 6.06; M5 vs. 4 µg L−1: p > 0.9999, M5 vs. 0.65 µg L−1: p = 0.6198, 4 µg L−1 vs. 0.65 µg L−1: p = 0.2995, H = 2.97; M6 vs. 4 µg L−1: p > 0.9999, M6 vs. 6.5 µg L−1: p = 0.2617, 4 µg L−1 vs. 6.5 µg L−1: p = 0.6472, H = 3.12; M9 vs 40 µg L−1: p = 0.9609, M9 vs. 0.65 µg L−1: p = 0.3788, 40 µg L−1 vs. 0.65 µg L−1: p > 0.9999, H = 2.41; M16 vs. 400 µg L−1: p = 0.1597, M16 vs. 650 µg L−1: p = 0.2882, 400 µg L−1 vs. 650 µg L−1: p = 0.0010, H = 12.96; Kruskal–Wallis test) (Table 5).

These findings indicate that M1 exhibited a potentiation effect due to 2,4-D. The effect of M5, M9 and M16 also suggests potentiation since only one herbicide in the combinations exerted genotoxic activity. The genotoxicity of the M2, M4, and M6 mixtures was similar for both herbicides, indicating no additive or synergistic effect. Finally, the response of M13 suggests a synergistic effect.

Regarding the nongenotoxic mixtures in relation to the control, the combination 2,4-D + glyphosate yielded: a) mixtures equivalent to the both the herbicides (M7, M10, M11, M14 and M15) (M7 vs. 4 µg L−1: p = 0.0910, M7 vs. 65 µg L−1: p = 0.0774, 4 µg L−1 vs. 65 µg L−1: p = 0.9999, H = 6.44; M10 vs. 40 µg L−1: p > 0.9999, M10 vs. 6.5 µg L−1: p = 0.9601, 40 µg L−1 vs. 6.5 µg L−1: p > 0.9999, H = 1.37; M11 vs. 40 µg L−1: p > 0.9999, M11 vs. 65 µg L−1: p = 0.1001, 40 µg L−1 vs. 65 µg L−1: p = 0.6496, H = 4.57; M14 vs. 400 µg L−1: p = 0.1818, M14 vs. 6.5 µg L−1: p > 0.9999, 400 µg L−1 vs. 6.5 µg L−1: p = 0.0283, H = 7.19; M15 vs. 400 µg L−1: p = 0.2320, M15 vs. 65 µg L−1: p = 0.1959, 400 µg L−1 vs. 65 µg L−1: p = 0.0009, H = 13.03; Kruskal–Wallis test) (Table 5); b) mixture statistically different to the 2,4-D (M3) (M3 vs. 0.4 µg L−1: p = 0.0021, M3 vs. 65 µg L−1: p = 0.496, 0.4 µg L−1 vs. 65 µg L−1: p = 0.1372, H = 11.58; Kruskal–Wallis test) (Table 5); c) mixture statistically different to the glyphosate (M8 and M12) (M8 vs. 4 µg L−1: p = 0.1040, M8 vs. 650 µg L−1: p = 0.0004, 4 µg L−1 vs. 650 µg L−1: p = 0.2506, H = 14.81; M12 vs. 40 µg L−1: p = 0.7882, M12 vs. 65 µg L−1: p = 0.0016, 40 µg L−1 vs. 650 µg L−1: p = 0.0576, H = 12.48; Kruskal–Wallis test) (Table 5). The ability of one or both herbicides to induce genotoxic damage was reduced in all these mixtures, characterizing an antagonistic effect. (Table 5).

Finally, no mixture increased the frequency of cells with micronuclei compared to the control (p > 0.05, H = 25.42; Kruskal–Wallis test) (Table 4). Therefore, it suggested that 2,4-D at 400 µg L−1 lost its mutagenic effect when combined with varying concentrations of glyphosate. The same occurred for glyphosate at 650 µg L−1 when mixed with different concentrations of 2,4-D, resulting in an antagonistic effect for M4, M8, M12, M13, M14, M15 and M16) (M4 vs. 0.4 µg L−1: p = 0.7515, M4 vs. 650 µg L−1: p = 0.5037, 0.4 µg L−1 vs. 650 µg L−1: p = 0.0344, H = 6.41; M8 vs. 4 µg L−1: p > 0.9999, M8 vs. 650 µg L−1: p = 0.0145, 4 µg L−1 vs. 650 µg L−1: p = 0.0933, H = 8.69; M12 vs. 40 µg L−1: p > 0.9999, M12 vs. 650 µg L−1: p = 0.3969, 40 µg L−1 vs. 650 µg L−1: p = 0.0608, H = 5.55; M13 vs. 400 µg L−1: p = 0.0079, M13 vs. 0.65 µg L−1: p > 0.9999, 400 µg L−1 vs. 0.65 µg L−1: p = 0.0043, H = 12.81; M14 vs. 400 µg L−1: p = 0.0015, M14 vs. 6.5 µg L−1: p = 0.1923, 400 µg L−1 vs. 6.5 µg L−1: p < 0.0001, H = 22.87; M15 vs. 400 µg L−1: p = 0.0084, M15 vs. 65 µg L−1: p =  > 0.99999, 400 µg L−1 vs. 65 µg L−1: p = 0.0002, H = 16.96; M16 vs. 400 µg L−1: p = 0.0072, M16 vs. 650 µg L−1: p = 0.7227, 400 µg L−1 vs. 650 µg L−1: p = 0.1882, H = 9.36; Kruskal–Wallis test) (Table 5).

3.4 Representative Micrographs of Chromosomal Abnormalities and Micronuclei Observed

Figure 1 presents illustrative micrographs showing the main chromosomal abnormalities and micronuclei identified after exposure to the isolated herbicides and mixtures.

Fig. 1
figure 1

Illustrative micrographs depicting the normal cell cycle of A. cepa (a-e) and main chromosomal aberrations and micronucleus identified following exposure to the isolated herbicides and the mixtures (f-j). (a) Interphase. (b) Prophase. (c) Metaphase. (d) Anaphase. (e) Telophase. (f) Micronucleus (arrow). (g) Metaphase with chromosomal loss (arrow). (h) Anaphase with chromosome adhesion. (i) Anaphase with chromosome bridge (arrow). Telophase with bridge (bigger arrow) and chromosomal fragments (smaller arrows). Bar scale: 20 µm

Full size image

3.5 Final Considerations

Water contamination by pesticides has been studied by several researchers (Rocha & Grisolia, 2019; Daam et al., 2019; Moutinho et al., 2020; Brovini et al., 2021a; b; Mesnage et al., 2021; Panis et al., 2022; Gaboardi et al., 2023) due to its negative impact on human and environmental well-being.

Several studies have reported the presence of glyphosate and 2,4-D in Brazilian waters. However, it is important to note that the recorded concentrations vary depending on the study. In a broad review, Brovini et al. (2023) reported that the highest concentration of glyphosate found in freshwaters was 500 µg L−1, with an average of 56.7 µg L−1. Glyphosate-AMPA was also found at concentrations ranging from 51.88 to 117.07 and 23.46 to 41.25 μg L−1, respectively (Gomes et al., 2022). In another study, Pires et al. (2023) found groundwater samples were contaminated with glyphosate (up to 1.5868 μg L−1) and AMPA (up to 0.2751 μg L−1), while surface water samples contained glyphosate (up to 0.0236 μg L−1). According to Brovini et al. (2023), the highest concentration of 2,4-D was found on freshwaters up to 30 µg L−1 (average = 1.1 µg L−1). Finally, Lima-Junior et al. (2024) recorded an average value of 0.122 µg L−1 for 2,4-D, and Pinheiro et al. (2010) reported 2,4-D at 74.50 µg L−1. According to data from SISAGUA, up to 30 µg L−1 for 2,4-D and up to 200 µg L−1 for glyphosate were recorded in drinking water from 2014 to 2017 (Aranha & Rocha, 2019).

The concentrations above cited are in accordance with the maximum concentration allowed by Brazilian laws. However, Brovini et al. (2021a) reviewed the surface freshwater concentrations of bestseller pesticides in Brazil and discussed the implications for the aquatic organisms. The highest value reported for glyphosate was 110 μg L−1 (median = 7.5 μg L−1). Toxicity analyses were performed on Daphnia magna, Oncorhynchus mykiss, and Raphidocelis subcapitata. The authors estimated that glyphosate concentrations below 0.1 μg L−1 pose a minimal risk, while concentrations exceeding 1 μg L−1 represent a high risk to aquatic organisms.

The highest median value reported for 2,4-D was 10 μg L−1, and the median concentration in Brazilian surface freshwater was 0.05 μg L−1. Concentrations of 2,4-D lower than 1 μg L−1 pose a low risk, while concentrations higher than 10 μg L−1 could cause a high risk. Based on these findings, Brovini et al. (2021a) suggested that the permissible levels of glyphosate and 2,4-D by Brazilian legislation in freshwater ecosystems should be revised.

When we examined the pesticides separately, it was observed that the levels permitted in water by Brazilian laws showed cytotoxic and/or genotoxic effects on A. cepa. Our data are consistent with previous studies that assessed glyphosate and 2,4-D at concentrations allowed in water by Brazilian legislation using the A. cepa test. Bordin et al. (2023) reported that glyphosate at 65 µg L−1 induced chromosomal aberrations and nuclear abnormalities in A. cepa roots. In addition, Miranda et al. (2023) showed that glyphosate at concentrations of 50, 65, and 280 µg L−1 and 2,4-D at concentrations of 4, 10, and 30 µg L−1 were cytotoxic or genotoxic to onion. As A. cepa is a bioindicator of water quality, these studies support the previous findings of Brovini et al. (2021a) regarding the need to review the maximum allowable concentrations of glyphosate and 2,4-D in water.

Other studies have also reported the cytogenotoxic effects of the glyphosate on non-target organisms. However, it is important to note that the most studies conducted with plants assessed concentrations higher than those examined in this study, ranging from 0.36 to 7.2 mg/ml (Truta et al., 2011), 5 to 30 mg L−1 (Mercado & Caleño, 2020), 1.56 to 11.66 mg/ml (Finkler et al., 2022), and 1.34 to 13.40 mg L−1 (Vieira et al., 2024). Regarding aquatic organisms, Rodrigues et al. (2019) showed that glyphosate (1.7 to 100 mg L−1), its primary degradation product aminoethyl phosphonic acid—AMPA (1.7 to 100 mg L−1), the commercial formulation ATN (1.7 to 50 mg L−1), and the surfactant polioxietil amina (0.4 to 4 mg L−1) induced DNA damage in zebrafish larvae. In another study, Hong et al. (2018) found that glyphosate (0.35 – 5.60 mg L−1) induced oxidative stress, DNA damage, and micronuclei in freshwater shrimp.

Similarly, 2,4-D at concentrations of 25.2 and 50.4 mg L−1 caused DNA strand breaks in fish (Arcaute et al., 2018), as did 2,4-D at concentrations of 10, 20 and 40 mg L−1 (Zafra-Lemos et al., 2021). Tadpoles exposed at concentrations higher than 52.5 μg L−1 died within 24 h of the start of the experiment. In addition, the occurrence of mouth and intestinal malformations (at 100 μg L−1), erythrocyte nuclear abnormalities (at 30.0, 52.5, 75.0 and 100 μg L−1) and micronuclei (at 52.5, 75.0, and 100 μg L−1) were also recorded (Santos et al., 2024).

In studies on mammals, there is often conflicting evidence regarding the genotoxic effects of glyphosate. In the review study, Kier and Kirkland (2013) reported that the genotoxic properties of glyphosate are primarily attributed to the impact of "inert ingredients" present in commercial formulations rather than the active ingredient itself. In this sense, Chaufan et al. (2014) showed that the glyphosate formulation produced an increase in reactive oxygen species, nitrotyrosine formation, superoxide dismutase activity, and glutathione (GSH) levels in the HepG2 cell line. In addition, the commercial formulation induced apoptosis. On the other hand, no effects were found with acid glyphosate and AMPA treatments.

Mesnage et al. (2022) demonstrated that Roundup formulations induce more biological changes associated with carcinogenesis than glyphosate. In another study, glyphosate, Roundup 360 PLUS, and AMPA induced DNA damage and caused oxidation of purines and pyrimidines in human peripheral blood mononuclear cells. Notably, the commercial formulation induced DNA damage at a concentration as low as 5 μM, while glyphosate and AMPA caused DNA lesions at concentrations of 250 μM and 500 μM, respectively (Woźniak et al., 2018). Additionally, Montero-Montoya et al. (2023) reported that glyphosate induce micronuclei, chromatin buds, nucleoplasmic bridges, and extranuclear fragments in peripheral lymphocytes.

Regarding 2,4-D, both the active ingredient at a concentration of 10 μg ml−1 and the formulation Dedalo Elite at a concentration of 2 μg ml−1 increased the frequency of micronuclei in Chinese hamster ovary cells. In addition, both compounds induced DNA damage and cytotoxicity (Laborde et al., 2020). Mesnage et al. (2022) evaluated the carcinogenic properties of 2,4-D. The researchers found that the herbicide is a strong inducer of oxidative stress and the unfolded protein response, while having relatively minimal effects on DNA damage.

As previously stated, glyphosate has been classified as a probable carcinogen (IARC, 2017), while 2,4-D is classified as a possible carcinogen to humans (IARC, 2018). Continuous exposure to cytogenotoxic and mutagenic agents can have an impact on human health, as noted by Fenech et al. (2020) and Marchetti et al. (2023). Therefore, the contamination of drinking water by mixtures containing proven, probable, or potential carcinogenic pesticides is a cause for concern, as highlighted by Panis et al. (2022).

Previous studies have reported the genotoxic activity of mixtures containing 2,4-D and glyphosate on various organisms (Bernardi et al., 2022; Carvalho et al., 2020; Finkler et al., 2022; Pavan et al., 2021). Our results are consistent with these findings. In this study, we prepared mixtures to replicate a range of contamination scenarios. Variations in cytogenotoxic responses may be due to differences in combined concentrations and analyzed parameters, such as mitotic index, chromosomal aberrations, and micronuclei. The results obtained for mixtures that exhibited additive/synergistic or potentiated effects are highlighted. The detection of these interactions warns of potential risks to exposed organisms. Therefore, the findings could assist pesticide regulators and environmental managers in improving pesticide legislation. However, there is currently no national discussion on the issue of pesticide mixtures in water.

In contrast, some countries have taken steps to decrease their use of pesticides. For example, the European Union has implemented some of the strictest legislative limits globally, as demonstrated in Table 1. Additionally, Directive 2020/2184 of the European Union adopts the precautionary principle and establishes a maximum limit of 5 µg L−1 for the sum of all pesticides in drinking water.

To the best of our knowledge, this is the first report with the large number of binary mixtures 2,4-D + glyphosate analyzed. The data obtained in this study help reveal the adverse effects of co-exposure to these herbicides when relevant concentrations to the aquatic environment are considered.

4 Conclusion

This study investigated the cytogenotoxicity of the herbicides 2,4-D and glyphosate, as well as their binary mixtures, using the A. cepa test. The results showed that even concentrations allowed by Brazilian legislation were cytotoxic and increased the frequency of chromosomal aberrations, indicating a potential risk to exposed organisms. Additionally, certain mixtures of 2,4-D and glyphosate were found to be cytotoxic and genotoxic, exhibiting additive, synergistic, or potentiated effects. The research emphasizes the importance of formulating pesticides that are less harmful to non-target organisms. Additionally, the findings may aid pesticide regulators and environmental managers in enhancing pesticide regulations and facilitating discussions on pesticide mixtures in water. Further research should prioritize investigating the toxic mechanisms of 2,4-D + glyphosate mixtures using different assays and biomarkers. Studies on both active ingredients and co-formulations are also necessary.