Habitat differences affect the nuclear morphology of the erythrocytes and the hepatic melanin in Leptodactylus fuscus (Anura) in the Brazilian Cerrado savanna

The sensitivity of anuran to the effects of habitat destruction and contamination has led to a preoccupying global decline in their populations. Morphological biomarkers such as micronuclei and other erythrocyte nuclear abnormalities (ENAs), as well as the occurrence of hepatic melanin, can be used to evaluate the effects of habitat impacts. In the present study, these two parameters were combined for the in situ assessment of the effects of soybean cultivation on the grassfrog, Leptodactylus fuscus. Specimens were also collected from a protected area to provide a reference site (non-agricultural environment). The frequency of some of the nuclear abnormalities in the animals from the soybean plantation was much higher than that recorded at the reference site, in particular micronuclei, which were 3.6 times more frequent in the plantation, lobulated nuclei (3.4 times more frequent), and reniform nuclei, which were four times more common than at the reference site. The combined analysis of all the ENAs together also revealed a frequency approximately 1.4 times higher in the animals from the soybean plantation, in comparison with the protected area. Smaller areas of hepatic melanin were observed in the specimens from the soybean plantation. These results provide further evidence of the sensitivity of anurans to habitat impacts and indicate that animals found in soybean plantations are susceptible to systematic alterations of their cells.


Introduction
Anuran amphibians play an important role in both aquatic and terrestrial communities, where they may act as indicators of the effects of anthropogenic impacts (Böll et al. 2013;Green et al. 2019). Anurans may also play an important role in agroecosystems, given that they include some of the principal natural predators of many agricultural pests, worldwide (Arcaute et al. 2014). However, amphibians are also one of the planet's most threatened groups of animals, with populations declining rapidly in almost all parts of the world (Stuart et al. 2004;Ficetola and Maiorano 2016;Carvalho et al. 2019). This decline is due to the spread of disease (Vredenburg et al. 2010), habitat modifications (Blaustein et al. 2011), the introduction of exotic species (Collins 2010), increasing UV radiation , and the increasingly widespread use of pesticides (Arcaute et al. 2014;Pérez-Iglesias et al. 2016;Borges et al. 2019a;Carvalho et al. 2019;Benvindo-Souza et al. 2020 Multiple stressors for amphibians may arise simultaneously in agricultural environments, including habitat fragmentation (Prado and Rossa-Feres 2014) and the loss of vegetation cover from the margins of bodies of water (Silva et al. 2012a), which results in an increase in UV radiation (Lipinski et al. 2016), and the increasing distance between forest fragments and the ponds in which many amphibian species breed (Silva et al. 2012b). The application of pesticides contaminates not only the soil, but also bodies of water, either directly or indirectly (Lajmanovich et al. 2014), causing shifts in the physicalchemical parameters of the water (Babini et al. 2015). A number of studies have demonstrated the occurrence of mutagenic and genotoxic damage in anurans related to the environmental degradation caused by agricultural practices (Babini et al. 2015;Pollo et al. 2018;Gonçalves et al. 2019;Borges et al. 2019b).
The micronucleus (MN) test has become a standard procedure in ecotoxicological studies due to its simplicity, high sensitivity, and reliability, with the results becoming available rapidly. In addition to micronuclei, other erythrocyte nuclear abnormalities (ENAs) have been associated with the processes of cell division and death (Pollo et al. 2015). These abnormalities include lobed and notched nuclei, and binucleated cells (Pollo et al. 2015;Pérez-Iglesias et al. 2016;Mesak et al. 2018). Micronuclei are formed through the loss of chromosomal fragments or entire chromosomes, which form a second, smaller nucleus in the cell (Al-Sabti and Metcalfe 1995; Gregorio et al. 2019). An increase in the frequency of micronuclei in the cells of an organism indicates subcellular genotoxic effects and chromosomal damage (Bosch et al. 2011). A high frequency of ENAs is indicative of the interference of xenobiotic agents in the synthesis of DNA or the mutation in the nucleus (Pollo et al. 2017). These biomarkers are thus capable of detecting DNA damage caused by environmental impacts (Benvindo-Souza et al. 2020), which is why they were selected for the present study. I n a d d i t i o n t o t h e s e n u c l e a r a n o m a l i e s , melanomacrophages have been used as morphological biomarkers of the impact of environmental stressors such as UV radiation , benzo[a]pyrene , increasing temperature (Santos et al. 2014), and the intensification of agricultural activities (Franco-Belussi et al. 2020). Melanomacrophages are pigmented cells with phagocytic activity found in the hematopoietic organs, such as the liver and spleen, of ectothermic vertebrates (Agius 1980;Fishelson 2006;Franco-Belussi et al. 2013;Santos et al. 2014). These cells produce and store melanin, which in turn absorbs and neutralizes free radicals and other toxic agents (Zuasti et al. 1989;Fanali et al. 2018).
These biomarkers were used to investigate potential environmental impacts in the South American white-lipped grassfrog, Leptodactylus fuscus (Schneider 1799), a species native to the Brazilian Cerrado savanna biome. This medium-sized frog (males = 42.8±4.0 mm; females = 43.6 ±4.4 mm; Sugai et al. 2012) occurs throughout much of South America (Frost 2019) and is typical of open habitats (Wynn and Heyer 2001;Sugai et al. 2012). This frog was selected for the present study due to the fact that, despite its abundance in degraded habitats (Reynolds et al. 2004), little is known about the possible effects of agricultural stressors on these animals, given the paucity of studies on this species, in particular, under natural conditions. The results of this study provide new insights into the susceptibility of amphibians to the impacts of environments degraded by agricultural practices, and reconfirm the role of these activities in the decline of amphibian populations (Laurance et al. 2014), as confirmed in many different parts of the world (Arntzen et al. 2017;Swanson et al. 2018;Agostini et al. 2020). As agricultural frontiers are expanding throughout South America, but especially in Brazil, at an unprecedented rate (Dias et al. 2016), there is clearly an urgent need for in situ studies to determine the impacts of environmental factors related to the development of agricultural practices (Pollo et al. 2017), especially in key biomes, such as the Brazilian Cerrado. The data from the present study will also contribute to the reduction of knowledge gaps on the amphibian species (Benvindo-Souza et al. 2020) that occur in areas impacted by agriculture in Brazil.
In this context, the present study investigated the effects of soybean cultivation on the anuran species L. fuscus through the analysis of micronuclei and other ENAs, combined with the levels of hepatic melanin in the cells. It should be noted that this is the first in situ study to apply these two techniques specifically to the simultaneous monitoring of natural anuran populations in agricultural areas.

Study area and specimen collection
The present study was carried out in two areas of the Brazilian Cerrado biome, approximately 250 km apart, that have distinct histories of land use. The agricultural environment selected for the study was located in the municipality of Rio Verde, in southwestern Goiás, a state in the Brazilian Midwest. This municipality is considered to be the state's principal producer of soybean according to the Mauro Borges Institute of Statistics and Socioeconomic Studies and prior to 2017, it ranked eighth among all Brazilian municipalities according to the Brazilian Institute of Geography and Statistics. Rio Verde thus represents an environment with a high probability of genotoxic impacts on its resident organisms (Borges et al. 2019b). For the present study, 10 adult male Leptodactylus fuscus were collected by active searching on a single night in November 2018, during the soybean planting season. The animals were all collected in the vicinity of a single temporary pond of approximately 2m × 5m, and 15cm in depth, located in the middle of a soybean plantation (17°47.742′ S, 051°06.089′ W). There are scattered fragments of forest in the surrounding area, but no other bodies of water nearby, although there is some vegetation in the water. Data obtained from the NASA Prediction of Worldwide Energy Resources (NASA POWER) website (https://power.larc.nasa.gov/#dataaccess) show that the mean minimum temperature of this site in 2018 was 17.8°C, while the maximum temperature was 28.7°C, mean precipitation was 3.51 mm/day, and the mean incidence of sunlight was 5.11 kW/h/m 2 /day.
Another 10 adult male L. fuscus were collected by active nocturnal searching in the Emas National Park (ENP) in December 2018. The ENP is a federal conservation unit with an area of approximately 132,000 hectares located in southwestern Goiás, in the municipalities of Mineiros and Chapadão do Céu, and in the neighboring area of the state of Mato Grosso do Sul, in the municipality of Costa Rica (ICMBio 2019). This area is relatively flat and encompasses different Cerrado formations, such as grassland, shrubby savanna, and riparian forests (ICMBio 2019). The specimens were collected on a single night from a large area of wetland adjacent to a body of water located at least 10 km from the outer perimeter of the national park (18°06.990′ S, 052°5 5.024′ W). The collection site is dominated by typical shrubby Cerrado savanna vegetation, with sparse and poorly developed shrubs and herbaceous plants (ICMBio 2019). The mean minimum temperature recorded for the location in 2018 was 18.1°C, the mean maximum temperature was 29.2°C, the mean precipitation was 3.15 mm/day, and the mean insolation was 5.10 kW/h/m 2 /day (data obtained from the NASA POWER website). All the specimens were taken to the Animal Biology Laboratory of the Goiás Federal Institute in Rio Verde, where the analyses were conducted. The total mass of each specimen was determined using a precision analytical balance and the snout-vent length was recorded with a digital caliper (0.01mm precision). The characteristics of the matrix surrounding each sampling point are recorded within a radius of 1 km (Fig.  1). The ENP is covered (100%) in natural vegetation, with grassland and savanna typical of the Cerrado biome, while the soybean plantation is dominated by crops (92.59%), with only 7.41% of natural vegetation cover.

Ethical and sampling statement
Licenses for the collection of specimens during the present study and animal experimentation were obtained from the Chico Mendes Institute for Biodiversity Conservation (ICMBio), under protocol 62687-1, and the Goiás Federal Institute Ethics Committee on the Use of Animals (CEUA/IFGoiano), under protocol number 6643030518.

Micronuclei and other erythrocyte nuclear abnormalities
After collection, the animals were euthanized by immersion in Benzocaine (5g/L). An abdominal incision was then made, from which the circulating blood cells of the abdominal cavity were obtained with a heparinized needle and syringe (25 mm × 0.7 mm). Two blood slides were prepared per animal, which were fixed in cold methanol for 20 min and then stained with a 5% solution of Giemsa in tap water for 12 min. A total of 2000 cells were analyzed per animal under an optical microscope (Laborana LAB-1001TB) which was attached to a digital camera (Laborana 3.0Mp), using ×100 magnification.
The criteria used to identify micronuclei were (a) staining intensity equal to that of the principal nucleus of the cell, but with a smaller diameter, (b) rounded shape unconnected to the principal nucleus, and (c) no overlap with the principal nucleus and located within the cytoplasm (Fenech 2000). The other ENAs quantified in the analyses were binucleated cells, nuclear buds, anucleated cells, lobed nuclei, notched nuclei, reniform (kidney-shaped) nuclei, and segmented nuclei, as described in our recent review (Benvindo-Souza et al. 2020). The data are presented as standard frequencies for each anomaly and for the whole set of ENAs.

Hepatic melanin
For the histological analyses, liver fragments were extracted from each L. fuscus specimen, weighed on a precision analytical balance, and fixed in metacarn (60% methanol, 30% chloroform, and 10% acetic acid) for 3 h at 4°C. These samples were then dehydrated in an increasing alcohol series and embedded in HistoResin (Leica-HistoResin embedding kit) for the extraction of 2-μm sections using a LeicaRM 2265 microtome (Switzerland). These sections were placed on slides, stained with hematoxylin and eosin (HE), and photographed under a Leica DM4B microscope attached to an image capture system (Leica DMC 4500), with a ×20 magnification. Hepatic pigmentation was quantified based on the color intensity, using the Image Pro-Plus Media-Cybernetics Inc. program (version 6.0). A total of 25 random histological fields were photographed and analyzed in each specimen. The procedures followed the protocols of Santos et al.

Water quality
Water samples were collected from both study sites (Rio Verde and the ENP) for physical-chemical analyses. The samples were collected in Rio Verde in November 2018 at approximately 20:00h and in the ENP in December 2018 at the same time and were then sent to the laboratory for analysis within 24 h of collection. One liter of water was collected from approximately 5 cm below the surface of a pond close to the study site and on the same day that the animals were collected at each site. The samples were stored in individual amber borosilicate glass flasks at a temperature of <4°C and sent to a private laboratory in Rio Verde, Goiás, Brazil, for quantification of carbamate, organochlorine, and organophosphate pesticides. The following substances were analyzed: 2. 4 D + 2. 4. 5 T, alachlor; aldicarb + aldicarb-sulfone + aldicarbsulfoxide, aldrin + dieldrin, atrazine, carbendazim + benomyl, carbofuran; chlordane (cis + trans), chlorpyrifos + chlorpyrifos-oxon, DDDT, diuron, endosulfan (alpha + beta + sulfate), endrin, glyphosate + ampa, lindane, mancozeb, metamidophos, metolachlor, molinate, methyl parathion, pendimethalin, permethrin, profenofos, simazine, tebuconazole, terbufos, and trifluralin. All of these compounds can be found in agricultural areas, either after their recent application or in residual form from previous applications. The analyses were conducted according to the procedures outlined in the Standard Methods for the Examination of Water and Wastewater, 23rd edition. The parameters were assessed based on the standards defined by CONAMA resolution number 357/2005-Class II and decree 1.745/1979.
The temperature, pH, dissolved oxygen, conductivity, resistivity, salinity, and total dissolved solids of the water of each study pond were measured in situ using a portable multi-parameter Bante900P apparatus. These measurements were conducted on the day of the collection of the specimens at each site, at approximately 20:00h.

Statistical analyses
The MN and ENA data and the melanin scores are presented as means ± standard deviation. The MN and ENA frequencies and the area of melanin were compared systematically between the two study areas (soybean plantation and the ENP). Prior to the analyses, the homogeneity of the variances was evaluated using Levene's test, and the data were (Log10) transformed to homogenize the variances, as appropriate. Student's t (Pollo et al. 2017) was applied to the parametric data (micronuclei, segmented nucleus, reniform nuclei, notched nuclei, anucleated cells, nuclear buds, binucleated cells, total ENAs, and the area of hepatic melanin), while Mann-Whitney's U was applied to the nonparametric data (lobed nuclei). A multiple Pearson correlation analysis was applied to verify the potential existence of correlations between the frequency of MNs, other ENAs, and the area of hepatic melanin in each environment. A probability of p < 0.05 was considered significant in all analyses.

Water quality
The presence of agrochemicals was not detected in the water samples collected from either study site (Supplementary Material 1). The water parameters recorded in situ at both sites are shown in Table 1. The ENP pond had a neutral pH, whereas the pH of the soybean plantations was acidic. The highest dissolved oxygen concentration was recorded in the ENP, while the highest values of all the other parameters were recorded in the soybean plantation.

Micronuclei and other erythrocyte nuclear abnormalities
In addition to micronucleated cells, other ENAs were found in the L. fuscus erythrocytes, such as reniform, lobed, notched, and segmented nuclei, anucleated cells, nuclear buds, and binucleated cells. The frequencies of MNs and other ENAs found in each environment are shown in Fig. 2, as a percentage per 2000 erythrocytes. A significantly higher frequency (t  Table 2) of MNs was recorded in the specimens collected in the soybean plantation in comparison with the ENP. When all the ENAs were considered together, a significantly higher frequency (t = 2.134; p = 0.046; Fig. 3) was also found in the soybean plantation in comparison with the ENP. There were also significantly more cells with lobed (U = 20.000; p = 0.023; Table 2) and reniform nuclei (t = 5.077; p = 0.000; Table 2) in the soybean plantation when these abnormalities were assessed separately.

Hepatic melanin
The area occupied by melanin in the liver tissue in the L. fuscus specimens was significantly smaller (t =−22.886; p = 0.000) in the specimens collected in the soybean plantation in comparison with the ENP (see Figs. 4 and 5). In the L. fuscus specimens collected in the ENP, melanin covered approximately 3.08% of the total area of the tissue, whereas it covered only 0.93% in the specimens from the soybean plantation.
No correlation was found between the frequency of MNs and the area of melanin in the specimens from either the ENP (r =−0.000; p = 0.998) or the soy plantation (r =−0.366; p = 0.298). A similar lack of correlation was found between the ENAs (all types combined) and the area of melanin in the specimens from the ENP (r = 0.222; p = 0.537) and the soybean plantation (r =−0.523; p = 0.120).

The environmental and biological responses of the anurans
The results of the present study indicate a clear effect of agriculture on anurans (Fig. 1), as indicated by the reduced coverage of native vegetation within a radius of 1 km of the pond located  within the soybean plantation. L. fuscus is considered to be a generalist species well adapted to changes in habitat, and is known to be a colonizer of degraded habitats (Sugai et al. 2012). Despite this, the environmental stressors found in the agricultural landscape appeared to have provoked significant changes in the blood parameters and liver melanin of the frog specimens in comparison with the natural environment (the ENP). The natural, relatively open conditions of the Cerrado savanna and the absence of extensive forest cover probably determined the similar climatic conditions recorded in the two areas during the study period (2018), although the ENP had much greater vegetation cover and more extensive wetlands. The pH and temperature of the water are known to influence the toxicity of many pollutants (Hoffman et al. 2010;Pollo et al. 2017), while for amphibians, the ideal pH for healthy development ranges between 6.3 and 7.7 (García and Fontúrbel 2003;Babini et al. 2015;Pollo et al. 2017). The pH recorded in the ENP (7.41) would thus have been adequate for amphibians, whereas that recorded in the soybean plantation (5.38) was excessively acidic. The higher conductivity, total dissolved solids, and salinity recorded in the soybean plantation may have been the result of a process of mineralization provoked by the agricultural practices implemented at this site (Gatica et al. 2012;Babini et al. 2015). High salinity, in particular, may be especially deleterious for organisms dependent on the freshwater environment, given the need to maintain an internal osmotic balance (Pollo et al. 2017). High salinity can cause osmoregulatory mechanisms to collapse, which may result in cell damage or even death (Cañedo et al. 2013;Pollo et al. 2017).
Only 0.1% of pesticides reach the target organism (Pimentel 1995), while farmland soils may be contaminated extensively by these chemicals (Leeb et al. 2020). Pignati et al. (2017) alerted that the planting of soybean in Brazil is the agricultural activity that most uses pesticides and that the municipality of Rio Verde is one of the 10 municipalities that most consume pesticides in Brazil, with a mean of 17.7 L of agrochemicals being applied per hectare. In the context of the present study, as the adult frogs are post-metamorphic animals, the specimens collected in the plantation may have been exposed to contaminated soil for long periods of time, which may lead to the progressive cutaneous absorption of pesticides from contaminated soil (Van Meter et al. 2015).

Micronuclei and other nuclear abnormalities
A significantly higher frequency of micronuclei (MNs) was recorded in the L. fuscus specimens from the soybean   (Fenech 2000;Pollo et al. 2015) or as a product of the removal of excess material from the principal nucleus, leading to genomic instability (Prieto- García et al. 2007;Pollo et al. 2015). A similar pattern has been recorded in previous in situ studies of anurans, both larvae and adults, in habitats disturbed by farming (Pollo et al. 2015;Babini et al. 2015;Cruz-Esquivel 2017;Borges et al. 2019b). These findings highlight the importance of this type of study for the assessment of the health of the populations that inhabit disturbed environments (Hoffman et al. 2003). As in situ studies are potentially influenced by the interaction of multiple stressors (Pollo et al. 2017), it is important to note that the higher temperatures typically found in open farmland may increase the toxic effects of agrochemicals (Hoffman et al. 2003). In the present study, the temperature of the water recorded in the soybean plantation was approximately 12°C higher than that recorded in the ENP, which may have contributed to the high frequency of MNs recorded in this environment. It is also important to note here that the micronucleus test detects late adverse effects (Marquis et al. 2009), which implies that the MNs and other ENAs recorded in the specimens from the soybean plantation may reflect long-term genotoxic effects. As pesticides are applied during the planting of crops, the failure to detect these substances in the water samples may reflect the diluting effects of the rainfall in the region (Dores and De-Lamonica-Freire 2001). A number of experimental studies have also demonstrated the formation of MNs in larval anurans in response to exposure to agrochemicals (Mesak et al. 2018;Pérez-Iglesias et al. 2019). Experimental studies of the effects of pesticides on the formation of MNs in adult anurans are less common, however, although Pérez-Iglesias et al. (2016) found evidence of the influence of Glyphosate, one of the world's most widely used herbicides, on the frequency of MNs and hepatic melanomacrophages in anurans. An increase in the frequency of MNs and other ENAs caused by mutagenic compounds may indicate an increase in recessive mutations and genomic instability, which favors the decline of populations of sensitive animals (Ossana and Salibián 2013). This type of analysis thus provides an early warning of damage at the genetic level that may lead, eventually, to damage in the cells, tissues, the entire body and, ultimately, the population as a whole (Mouchet and Gauthier 2013).
In the present study, a higher frequency of ENAs was also found in the specimens from the soybean plantation. While the processes that lead to the formation of these anomalies are not yet fully understood (Benvindo-Souza et al. 2020), it does appear to reflect adverse reactions in the cells or the control mechanisms used to eliminate cells with damaged DNA (Fijan 2002). Lobed and reniform nuclei do appear to be precursors of micronuclei and binucleated cells, however (Araújo et al. 2019). Lobed nuclei have been observed in response to exposure to environmental contaminants in a number of anuran studies (Lajmanovich et al. 2014;Pollo et al. 2015;Babini et al. 2016;Pérez-Iglesias et al. 2018). Borges et al. (2019b) detected an increase in the frequency of reniform nuclei in tadpoles collected from an agricultural environment in comparison with those collected in the ENP, which is consistent with our results. Binucleated cells, like MNs, are related to cell division (Baršienė et al. 2010;Pollo et al. 2015), while nuclear budding is related to the amplification or polyploidization of the DNA, with excess genetic material being removed through the subsequent formation of MNs (Prieto- García et al. 2007;Pollo et al. 2015). Anucleated cells are considered to be a response to stressful situations, such as changes in the diet, pathologies or metabolic damage (Lajmanovich et al. 2014). Together, these anomalies and the other ENAs, such as notched and segmented nuclei, are thought to be provoked in anurans by pesticides (Lajmanovich et al. 2014;Babini et al. 2016;Mesak et al. 2018;Pérez-Iglesias et al. 2018. In the case of in situ studies, these ENAs have been observed in anurans in urban areas (Pollo et al. 2015), soybean plantations (Pollo et al. 2015;Borges et al. 2019b), areas of silviculture and ranching (Cruz-Esquivel et al. 2017), and areas associated with mining operations (Pollo et al. 2017). The results of the present study thus provide further evidence that agricultural environments contribute to an increase in the frequency of ENAs.
It is important to note here that L. fuscus, which is a relatively resistant and adaptable species, tolerant of habitat disturbance, may present different responses when compared to more sensitive species that may be affected even more intensively. This is why many Brazilian anurans are threatened by habitat loss and alteration, due primarily to the expansion of . The data are presented as means (circles) and standard deviations (vertical bars). The asterisk indicates a significant difference in the area of melanomacrophages between the two environments according to Student's t (data Log10 transformed) farmland, as defined by their IUCN status. These species include Physalaemus cuvieri (Borges et al. 2019b) and Leptodacylus latrans (Josende et al. 2015), for example, which have already been shown to undergo an increase in MNs and ENAs when exposed to agricultural activities or pesticides. The situation faced by these species is even more preoccupying than that of L. fuscus, which is not considered to be threatened currently by habitat change (Reynolds et al. 2004). Despite the greater diversity of anurans in the tropical region in comparison with more temperate environments, there are relatively much fewer studies, and most studies based on the MN test have focused on the exotic, North American species, Rana catesbeiana, rather than tropical amphibians (Benvindo-Souza et al. 2020).
The results of previous studies of MNs and other ENAs in leptodactylids exposed to environmental stressors are presented in Table S2

Hepatic melanin
A significantly smaller area of melanin was detected in the liver tissue of the L. fuscus specimens collected in the soybean plantation in comparison with those from the ENP. This type of cell plays a role in detoxification ) and other functions related to the immune system, as well as neutralizing free radicals to protect the tissue from oxidative stress (Franco-Belussi et al. 2013). A number of previous studies have demonstrated the harmful effects of pesticides through the analysis of melanomacrophages in the liver of anurans (Paetow et al. 2012;Çakici 2015;Oliveira et al. 2016;Huespe et al. 2016;Pérez-Iglesias et al. 2016. In addition to pesticides, the action of agents such as benzo[a]pyrene , UV radiation , fluorine , and cadmium (Wu et al. 2017) has already been demonstrated in the anuran liver. De  also found that the agricultural matrix can alter the area of melanomacrophages in anurans, with the exact level of response varying among species. Drugs such as cyclophosphamide, flutamide, and testosterone cypionate may also influence the area of liver melanomacrophages . Regnault et al. (2014) and Fanali et al. (2017Fanali et al. ( , 2018 demonstrated a significant reduction in the area of the melonamacrophages in anurans exposed to benzo[a]pyrene, a hydrocarbon considered to be a high-risk contaminant capable of inducing the formation of MNs (Fortin et al. 2015). This decrease in the area of melanomacrophages may be related to liver stress and hepatocyte apoptosis (Regnault et al. 2014). In an in situ study, Rohr et al. (2008) found that Atrazine and phosphate in the water caused a reduction in the area of melanomacrophages in anurans. In this same study, an experiment demonstrated that animals exposed to atrazine suffered a significant reduction in melanomacrophages in comparison with the control group, which indicates that atrazine supp r e s s e s t h e i m m u n e r e s p o n s e e x e r t e d b y t h e melanomacrophages in anurans.
However, other studies of pesticides have found an increase in the pigmented area of the liver of anurans exposed to these contaminants Pérez-Iglesias et al. 2016. Pérez-Iglesias et al. (2019) attributed this to a possible initial reaction to the exposure to xenobiotics in the immune system, with the phagocytic properties of the melanomacrophages contributing to the inflammatory process. These results reinforce the function of the organ and its physiological plasticity. Recent studies have shown that factors such as the hepatosomatic index and liver pigmentation vary under different types of land use (Gondim et al. 2020;Franco-Belussi et al. 2020). In a comparison of Leptodactylus macrosternum specimens from farmland and natural vegetation in Brazil's semi-arid zone, Gondim et al. (2020) found considerable variation in the hepatosomatic indices among sites, but no systematic difference between agricultural and non-agricultural areas. In agricultural environments in the Cerrado, Franco-Belussi et al. (2020) found that the amount of melanin and substances originating from hepatic cellular catabolism (lipofuscin and hemosiderin) varied among five anuran species, as well as between populations from areas with different types of land use. In the same study, L. fuscus had the smallest amount of melanin, and the abundance of the species was correlated negatively with the percentage of pasture planted in the landscape (Franco-Belussi et al. 2020). The extremely low percentage of melanin recorded in L. fuscus in the present study is consistent with these findings. Overall, few data are available on the effects of agriculture on anuran melanomacrophages, not least because of the confounding effects of the multiple stressors found in these environments. Given this, there is a clear need for further research to better understand the response of these cells in anurans.

Conclusions
Based on the in situ analysis of biomarkers, the present study demonstrated the effects of soybean cultivation on the physiology of the anuran species Leptodactylus fuscus, in terms of the increase in the frequency of nuclear abnormalities observed in the erythrocytes and the reduction in liver melanin. The analysis of micronuclei and other nuclear erythrocyte abnormalities revealed an increase in genotoxic damage in the specimens from the soybean plantation in comparison with those from the natural habitat (the reference site). The analysis of the hepatic melanin also demonstrated a reduction in the pigmented area in the specimens from the soybean plantation in comparison with the reference site. The evidence thus indicates that agriculture causes systemic effects in L. fuscus and that this process should be investigated further through the analysis of other parameters of the liver tissue, including cytochemical tests, such as the analysis of hemosiderin and lipofuscin, and enzymatic assays, to provide more conclusive insights into the phenomenon. Future studies may also expand on the data presented here, including new analyses of a larger number of populations representing different types of farmland and impact.

Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s11356-021-14974-4. Author contribution RAA: conceptualization, methods, validation, formal analysis, investigation, writing-original draft, visualization. WRR: methods, validation, investigation. CGAS: methods, validation, investigation. MBS: methods, formal analysis, writing-review and editing, visualization. NPLA: methods, validation, investigation. REB: conceptualization, methods. LFB: writing-review and editing, visualization. CDO: writing-review and editing. LRSS: conceptualization, methods, validation, resources, writing-review and editing, visualization, supervision, project administration, funding acquisition. All authors have read and approved the final manuscript. Availability of data and materials The data sets used and/or analyzed during the present study are available from the corresponding author on reasonable request.

Declarations
Ethics approval and consent to participate For this study, licenses for the collection of specimens and animal experimentation were obtained from the Chico Mendes Institute for Biodiversity Conservation (ICMBio), under protocol 62687-1, and the Goiano Federal Institute Ethics Committee on the Use of Animals (CEUA/IFGoiano) under protocol number 6643030518.

Consent for publication Not applicable.
Competing interests The authors declare no competing interests.