Over the past 2 decades, agricultural production in South America has been intensified because of world population growth and the high demand of the international market, leading to an increase in the use of agrochemicals (Barra et al. 2006; López et al. 2012; Zahid et al. 2016). These substances are often transported to superficial and underground aquatic environments through the discharge of wastewater and monsoon runoff on agricultural lands, eventually leading to changes in the quality of aquatic ecosystems (Barra et al. 2006; Lorenz et al. 2017). The toxic effects of agrochemicals in different nontarget animals could disrupt cellular processes and directly or indirectly interact with DNA, leading to genetic instability (Carrasco et al. 1990; Costa and Teixeira 2012; González et al. 2017; Benvindo-Souza et al. 2020). These effects can lead to harmful consequences for the fauna, including mutations, tumors, birth defects, and abnormal embryonic development (Brodeur et al. 2017). Ultimately, this could lead to population declines or even local extirpations (Gibbons et al. 2000; Hopkins 2000; González et al. 2017). In addition to the known risks that pesticide accumulation represents for biodiversity and human health, it is necessary to understand its consequences to trophic webs and broader implications for the ecosystem (Potter et al. 2004; Çördük et al. 2019; Marcelino et al. 2019).

The morphological analysis of blood cells’ nuclei, mainly erythrocytes, has been used as an important tool to evaluate the genotoxic effects of pollutants in wild species, such as fishes (Porto et al. 2005; Ghaffar et al. 2021), amphibians (Benvindo-Souza et al. 2020; Herek et al. 2021), lizards (Simonyan et al. 2018; Mestre et al. 2020), turtles (Çördük et al. 2019; De Oliveira et al. 2020), crocodilians (González et al. 2017), birds (Silveira et al. 2022; Rani et al. 2023), and mammals (Sandoval‐Herrera et al. 2021). The formation of micronuclei (MNs) is one of the anomalies caused by chemical pollutants, consisting of acentric chromosome and/or chromatid fragments unable to complete the mitotic process during the segregation in anaphase (Serrano‐García and Montero‐Montoya 2001). In addition, other nuclear abnormalities (e.g., binucleated cells, blebbed, notched, and lobed nucleus) may also be induced by genotoxic agents and be detected by microscopic analyses (Carrasco et al. 1990). The analysis of MNs and other erythrocytic nuclear abnormalities (ENAs) offers the advantages of requiring small sample sizes, being relatively low-cost and highly efficient in characterizing DNA damage (chromosomal abnormalities) caused by physical and chemical agents (Jha 2008; Hintzsche et al. 2017).

Contamination by agrochemicals is one of the main threats to freshwater aquatic ecosystems (Amoatey and Baawain 2019), affecting resident species, including fishes and aquatic reptiles (De Solla et al. 2007; Biaggini and Corti 2015; D’Costa et al. 2018; Odetti et al. 2024). Freshwater turtles are particularly good bioindicators for ecotoxicological studies because of their physiological and ecological characteristics, such as longevity and low generation turnover rate, and their role in different trophic levels (De Solla et al. 2007; Beau et al. 2019; Dos Santos et al. 2021). Also, because of their ectothermic condition, these animals exhibit periods of decreased metabolic activity that reduce the pollutant detoxification effectiveness and make them more susceptible to environmental stressors when compared to endothermic organisms (Hall 1980). Considering that some freshwater chelonians often use areas with intense agricultural activity as nesting and feeding sites (Saumure and Bider 1998; Rauschenberger et al. 2004; De Solla et al. 2014; Vogt et al. 2023), they face a higher risk of agrochemical exposure (Mingo et al. 2016).

The black spiny-necked turtle, Acanthochelys spixii (Duméril and Bibron 1835), can be found in central, southeast, and south regions of Brazil, occurring in freshwater and terrestrial habitats often associated with temporary swamps, slow-moving waters, heavily vegetated aquatic bodies, and small rivers, where it feeds on invertebrates, amphibians, and plants (Brandão et al. 2002; Brasil et al. 2011; Fraxe Neto et al. 2011). The species is not threatened in Brazil (Vogt et al. 2023), but it was on the list of endangered wildlife species of São Paulo State in the past (State Decree 42.838/1998) and currently is listed as insufficient data to assess its conservation status in the state (State Decree 63.853/2018). This freshwater turtle occurs in a region within a monoculture matrix of the Upper Paranapanema River basin (Fig. 1), southwest São Paulo State, together with 2 other chelonian species, Mesoclemmys vanderhaegei and Phrynops geoffroanus (Marques et al. 2013; Nehemy and São Pedro 2021). The possible impacts of agriculture activities on A. spixii are still unknown, which led the Chico Mendes Institute for Biodiversity Conservation (ICMBio) to recommend studies assessing the potential consequences of such activities on the species (Vogt et al. 2023).

View Figure

• Download as Powerpoint
• Download Figure

The aim of this study was to quantify the number of nuclear abnormalities in erythrocytes to understand the impact of chemical contaminants related to agricultural activities on populations of A. spixii that inhabit ponds in monoculture matrices from a region of the Upper Paranapanema basin, São Paulo State, Brazil. The frequency of MNs and other ENAs between free-living individuals from contaminated environments and reference individuals from captivity were compared. A compilation of available studies on nuclear abnormalities in erythrocytes of freshwater chelonians is provided as evidence of the significant impact of agrochemicals on populations of A. spixii.

METHODS

Study Area. —

Sixteen aquatic environments including artificial and natural ponds and streams (Fig. 1) were sampled with the goal of capturing A. spixii. The study area comprises the municipalities of Angatuba, Buri, Campina do Monte Alegre, and Capão Bonito, localized in the Upper Paranapanema River basin, southwest of São Paulo State, Brazil. The region has a broad network of small lotic ecosystems, like streams, and lentic ecosystems, such as natural ponds and dams, and is mostly composed of mosaics of forest remnants within a monoculture matrix. Because of soil quality and water availability, the region presents a strong agricultural suitability, standing out for its grain production (e.g., corn, soy, beans, and wheat) through crop rotation in extensive areas (Favareto 2007). Twenty-seven kinds of agrochemicals (e.g., herbicides, insecticides, fungicides) have been detected in water bodies in the region, including some chemicals banned by Brazilian legislation, and permitted products that were found in concentrations above the recommended level, such as glyphosate (SISAGUA 2024). Despite the presence of an extensive agricultural matrix, the region hosts important native vegetation remnants of Cerrado and Atlantic Forest biomes (Kronka et al. 2005). The climate consists of type Cfb of Köppen, characterized by mild summers and the absence of a dry season, with annual rainfall rates varying between 1500 and 2000 mm (Peel et al. 2007).

The 16 aquatic environments were selected based on the accessibility and the presence of suitable characteristics for freshwater turtles, such as floating vegetation, rocks, and food availability. Most sampled aquatic environments (n = 11) were located inside monoculture matrices and were presumed to be contaminated with agrochemicals. The remaining sampled environments (n = 5) were located within Capão Bonito National Forest, a protected area created in 1944. These agrochemical-free areas were originally selected as reference environments, where the turtles were expected to be uncontaminated. However, we were unable to capture any individuals in these areas. As an alternative, captive individuals were sampled to serve as a reference group (see details below). Geographic coordinates and environment details on each sampled point can be found in Appendix 1.

Individual Capture and Sample Collection. —

Funnel traps were used to sample A. spixii in the aquatic environments between November 2018 and April 2019. Each point was sampled for at least 3 consecutive days (72 hr), and the cumulative sampling effort amounted to 10,920 hr of funnel traps. Each funnel trap consists of a cylinder on a metal frame (100 cm long) covered with nylon mesh, with a funnel opening of 40 cm diameter at both ends (Fig. 2). Funnel traps were installed near the margins of each sampled water body, partially submerged, lured with beef, and checked daily.

View Figure

• Download as Powerpoint
• Download Figure

It is important to note that, out of all the sampled aquatic environments, turtles were only captured in 1: a pond situated within a monoculture area. Because no individuals were captured in the presumed agrochemical-free area (e.g., the 5 aquatic environments in Capão Bonito National Forest), captive turtles from the Quinzinho de Barros Municipal Zoo, in Sorocaba, São Paulo State, were sampled to serve as reference individuals. According to the Zoo’s veterinarians, the turtles came mainly from apprehension of illegal trade and were all in good health. At the time of blood sampling, the individuals lived at the Zoo for at least 10 years (except for 1 individual that had been in captivity for 3 years), representing good control individuals for the study (Zapata et al. 2016). Furthermore, the captive turtles were maintained under controlled water conditions, ensuring a contaminant-free environment.

Measurements for all free-living individuals captured and those held in captivity at the Zoo were obtained using an analog caliper (0.05-mm precision). The length and maximum width of the carapace (LC and WC), length and maximum width of the plastron (LP and WP), and maximum height of the carapace (HC) were registered. The sex of each individual was determined based on the position of the cloaca and the shape of the plastron (Balestra et al. 2016). The free-living turtles received an individual marking by cutting the marginal shield of the carapace in “u” shapes using a saw, following the markings described in Balestra et al. (2016). Finally, peripheral blood samples (0.1 to 0.3 ml) were taken from the postoccipital venous sinus as described by Olson et al. (1975) using 1-ml sterile syringes. All these procedures were carried out within the laboratory. Free-living individuals were taken to the lab and released in the original pond on the same day after procedures.

Micronucleus Test and Other Erythrocyte Nuclear Abnormalities. —

Blood smears were prepared in duplicate for each individual, using a 5-μl aliquot of blood (Da Silva et al. 2012). The rest of the blood samples were stored in Eppendorf® tubes containing sodium heparin as an anticoagulant, which is recommended for reptile species, especially freshwater turtles (Yilmaz and Tosunoğlu 2010; Da Silva et al. 2012). The slides were air dried for 2 min and then stained with LB Panoptic stain, a rapid hematological staining (Laborclin Ltd., Pinhais, Paraná, Brazil). Briefly, the smears were submitted to the action of a fixative and 2 staining solutions through 5-sec immersions each time. Finally, the slides were washed with distilled water to remove the reagents and dried again at air temperature. This stain is both cost-effective and expeditious and additionally permits examination of slides using a conventional light microscope, maintaining their staining over time. The use of the LB Panoptic stain for blood analysis in reptiles is relatively recent, but studies have shown that this is an efficient method for observing erythrocytes (e.g., Garcia et al. 2021; Nóbrega et al. 2022).

Using an Olympus optical microscope (1000× magnification) a single person observed 3000 erythrocytes on each slide, totaling 6000 cells counted for each individual. Only nonoverlapping erythrocytes with intact cell and nuclear membranes were considered for analysis. To prevent duplication during the counting process, the slide was thoroughly inspected from one end to the other. The inclusion criteria for MNs comprised the absence of contact between the MNs and main nuclei, a size smaller than one-third of the main nuclei, and nonrefractivity, with the MNs maintaining consistent color and intensity in comparison to the main nuclei (Carrasco et al. 1990). Other ENAs were also identified and scored, which include binucleated erythrocytes (BEs), blebbed nucleus (BN), notched nucleus (NN), and lobed nucleus (LN), following the classification details of Carrasco et al. (1990). To enable comparative analysis with existing studies on nuclear abnormalities in erythrocytes of freshwater turtles, the frequency of each abnormality type was reported per 1000 cells in the results.

Statistical Analyses. —

Data of BEs, BNs, NNs, and LNs were pooled and presented as total ENAs. Frequencies of MNs and ENAs per 1000 cells for free-living individuals were compared to reference individuals from captivity. First, the normality of the data was assessed using the Shapiro-Wilk test. Since the data for MNs in captive individuals showed a nonnormal distribution (W = 0.45297, p = 4.136e-06), the Mann-Whitney U nonparametric test (also known as the Wilcoxon test) was used to assess significant differences between the groups. ENA data showed a normal distribution for both groups (free-living: W = 0.96099, p = 0.784; captive: W = 0.96423, p = 0.8541), so the Student t-test was employed to evaluate significant differences between them. To assess the effect of sex on the number of abnormalities, MN and ENA frequencies were compared between males and females. While MN data exhibited a nonnormal distribution (males: W = 0.78928, p = 0.03204; females: W = 0.75562, p = 0.002461), ENA data followed a normal distribution (males: W = 0.90331, p = 0.3515; females: W = 0.87268, p = 0.08392). Consequently, the differences between sexes were assessed using the Mann-Whitney U-test and the Student t-test respectively. A comparison of total carapace length between free-living and captive individuals was also conducted using the Mann-Whitney U-test, because of the nonnormal distribution of the data for captive individuals (W = 0.78801, p = 0.03111). A p value < 0.05 was considered as statistically significant. All analyses were performed using the R Studio 2022.07.1 software (R Core Team 2022).

RESULTS

Eleven free-living individuals of A. spixii were captured in only 1 aquatic environment located in an agricultural matrix of rotational grain cultivation (Fig. 1). Seven turtles from the Zoo were obtained as reference individuals. All analyzed individuals (n = 18) were adults according to Bager et al. (2016). The mean total carapace length was 133.4 ± 11.6 mm for the free-living individuals and 146.9 ± 10.9 mm for the Zoo individuals (Appendix 2), with captive individuals exhibiting significantly larger sizes (W = 61, p = 0.04609). The frequency of MNs and other ENAs was significantly greater in free-living individuals (MNs: W = 69, p = 0.004; ENAs: t = 5.168, df = 12.987, p = 0.0002) compared to the reference group (Table 1; Fig. 3). MN and ENA frequencies were not different between males and females (MNs: W = 42.5, p = 0.7401; ENAs: t = −0.63971, df = 14.249, p = 0.5325). A representative image of each type of nuclear abnormality is shown in Fig. 4. The MN frequency was 104 times higher in free-living individuals than that in captivity, while the ENA frequency (pooled data of BEs, BNs, NNs, and LNs) was 4.7 times higher (Table 2). Additionally, the magnitude of differences found between free-living and captive individuals of A. spixii is much higher than that reported in similar studies with other freshwater chelonians species (104 times vs. 15.6 times in MNs; Table 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Number of micronuclei (MNs) and other erythrocyte nuclear abnormalities (ENAs) in the black spiny-necked turtle Acanthochelys spixii from free-living and captive individuals. Data are expressed as frequency per 1000 cells, depicting the mean ± standard deviation (SD) of each group.^a

View Fullscreen

Table 2. Available studies on nuclear abnormalities in erythrocytes of freshwater chelonians. Frequency of micronuclei (MNs) and/or other erythrocyte nuclear abnormalities (ENAs) for every 1000 cells are presented for both reference and contaminated individuals, followed by the difference between groups, the source of contamination, and its respective reference.

View Fullscreen

DISCUSSION

This study revealed for the first time a possible impact of agrochemicals on the health of free-living black spiny-necked turtle A. spixii. Individuals captured within a monoculture matrix exhibited a significantly higher frequency of MNs and ENAs than those from captivity (Table 1; Fig. 3). Erythrocytic nuclear abnormalities in reptiles can occur spontaneously, with their frequency varying according to factors such as species, sex, age, and health status (Zúñiga-González et al. 2000, 2001; Latorre et al. 2015; De Oliveira et al. 2020). In the present study, the frequency of nuclear abnormalities did not vary between sexes. Regarding age-related effects, the frequency of nuclear abnormalities varies in different ways according to taxa (Zuñiga-González et al. 2001). In ectothermic animals, excluding the neonatal phase where higher micronuclei rates may be observed (Schaumburg et al. 2014), comparisons between young and adult individuals of various ages may either show stable abnormality rates (Poletta et al. 2008) or an increase in older individuals (Romanova et al. 2021). No previous study in reptiles has reported a pattern similar to that observed in the present study, where larger, presumably older, adults (captive) exhibited fewer abnormalities than smaller, free-living adults, thus minimizing the potential age effect on our results. Moreover, all animals analyzed in our study were apparently healthy, with no external damage or signs of disease. Therefore, we believe the differences observed between free-living and captive individuals are unlikely to be explained by factors related to the spontaneous occurrence of erythrocytic nuclear abnormalities.

Comparing our results with previously established reference values would be ideal (Latorre et al. 2015); however, this is the first study to investigate erythrocyte nuclear abnormalities in A. spixii, and thus no reference values are available. Nevertheless, the comparison with other freshwater turtles can be used as a proxy of the variation expected between healthy individuals and those exposed to some environmental disturbance. In this context, the disparity observed between free-living and captive A. spixii specimens is considerably greater than that reported in similar studies involving other freshwater turtles (Matson et al. 2005; Frossard et al. 2013; Zapata et al. 2016; Çördük et al. 2019; De Oliveira et al. 2020). This notable difference (104 times vs. 15.6 times in MNs; see Table 2) strongly suggests that free-living A. spixii may have been exposed to a significant environmental stressor. The main environmental impacts associated with the emergence of nuclear abnormalities in reptile erythrocytes are related to extreme climatic conditions (e.g., temperature or radiation) and presence of pollutants, such as heavy metals, industrial waste, and pesticides (Odetti et al. 2024). The black spiny-necked turtles were found in a natural pond with no marginal vegetation, located within an agricultural area of rotational grain cultivation. No other obvious source of pollution, such as industries or mining activities, are present in the vicinity of the pond. According to the farm owner, the land had been used for the annual rotation of monocultures (mainly corn, soy, and wheat) for over 20 yr, with intense exposure to agrochemicals. The main active compounds used in grain monocultures in Brazil are herbicides (45%), such as glyphosate, atrazine, and lactofen (Pignati et al. 2017). Some agrochemicals are known to be genotoxic agents, resulting in increased occurrences of erythrocyte nuclear abnormalities in bee (Battisti et al. 2021), fish (Ghaffar et al. 2021), tadpole (Herek et al. 2021), lizard (Simonyan et al. 2018; Mestre et al. 2020), crocodile (González et al. 2017), turtle (De Oliveira et al. 2020), bird (Rani et al. 2023), and bat (Sandoval‐Herrera et al. 2021). Hence, it is plausible that the high rates of MNs and ENAs found in free-living A. spixii have been caused by agrochemicals through a significant amount and/or prolonged duration of exposure.

DNA damage caused by exposure to contaminated water and soil by agrochemicals can affect both adults and neonates of freshwater reptiles (Matson et al. 2005; Frossard et al. 2013; Zapata et al. 2016; González et al. 2017; Çördük et al. 2019; De Oliveira et al. 2020; Odetti et al. 2024). For example, De Oliveira et al. (2020) observed a high frequency of nuclear abnormalities and various morphological changes in the shape and size of erythrocyte in the Amazon River turtle (Podocnemis expansa) neonates from eggs artificially incubated in substrate at different concentrations of glyphosate, atrazine, and fipronil, as well as in concentrations permissible by the Brazilian government. A similar result was observed by González et al. (2017) when they evaluated the frequency of MNs and other ENAs in embryos of broad-snouted caiman (Caiman latirostris) exposed to 4 mixtures of pesticides. Adults of the Colombian slider turtle (Trachemys callirostris) that inhabit contaminated areas with mercury from agricultural fungicides and pesticides also showed a significant increase in the frequency of MNs and the extent of DNA damage (Zapata et al. 2016). In fact, agrochemical contamination in soil and water is one of the most well-known causes of DNA damage in freshwater reptiles (Odetti et al. 2024).

Chronic exposure to agrochemicals could affect different biological processes, varying according to the chemical nature of each substance (De Oliveira et al. 2020). Glyphosate, for example, is considered an aneugenic compound, which means that it promotes changes in cell division processes, stimulating the whole chromatids generation outside the main nucleus (MNs) or in the formation of multinucleated cells (ENAs) (Serrano‐García and Montero‐Montoya 2001). The presence of MNs and ENAs can also be associated with the excessive production of reactive oxygen species (ROS) caused by the exposure to agrochemicals, which can damage proteins and the formation and functionality of essential structures for the proper functioning of cells, such as microtubules, microfilaments, and/or intermediate filaments (Dorval and Hontela 2003; Dornelles and Oliveira 2014; Burella et al. 2018; Zhang et al. 2019). Thus, chronic exposure to agrochemicals represents a serious concern for the health and integrity of living organisms, negatively influencing their homeostasis and ability to carry out vital cellular processes.

The effects of agrochemicals on reptiles are not restricted to cellular abnormalities. It can also lead to other biological damage, such as alterations in gastrointestinal microbiota (Kittle et al. 2018), reduction in immune responses (Siroski et al. 2016; Mestre et al. 2020), and oxidative stress (Héritier et al. 2017). Changes in embryonic development have also been associated with agrochemicals in substrates near the nesting sites of some reptile species (Rich and Talent 2009; Allan et al. 2017). These substances have affected their morphology and physiology (Latorre et al. 2015; Hirano et al. 2019; De Oliveira et al. 2020; Mestre et al. 2020). Additional research on the physiology and behavior of A. spixii in agricultural landscapes would be useful to understand the actual impacts of agrochemicals in this species.

Individuals of black spiny-necked turtles were found in only 1 of the 16 sampled environments. Although this study is the first carried out specifically with A. spixii in the Upper Paranapanema River basin, it was unexpected not to find this species in other sampled locations, considering that the majority of the sites displayed similar characteristics and were effectively sampled. According to the latest Brazilian extinction risk assessment, A. spixii is not easily found in the wild throughout its distribution (Vogt et al. 2023). In fact, the species appears to be naturally rare, with very few records in herpetological studies in the region since 2017 (V. São Pedro, pers. comm.).

In summary, our results show strong evidence of the genotoxic effects of agrochemicals in a free-living population of black spiny-necked turtle, A. spixii, from a natural pond inserted in an agricultural matrix of rotational grain cultivation. These effects may represent an underestimated impact on the species conservation, which had not yet been effectively detected (Vogt et al. 2023). On the other hand, recognizing specific constraints in the present study is essential, yet these limitations can serve as a foundation for future investigations. Additional genotoxic data of populations from other preserved and impacted areas, or even from controlled laboratory tests, can generate a better overview of this impact and establish reference values for nuclear abnormalities in our target species. Analyzing the chemical composition of natural environments—such as identifying and quantifying agrochemicals—and correlating these findings with ENAs and MNs can significantly advance our understanding of their impact on A. spixii populations.