Evidence of reproductive disturbance in Astyanax lacustris (Teleostei: Characiformes) from the Doce River after the collapse of the Fundão Dam in Mariana, Brazil

The Fundão Dam collapsed, on November 5th, 2015, dumping more than 50 million/m3 of iron ore tailings, enriched with metals, into the Doce River channel. The objective of this study was to evaluate the reproductive biology and histological damage in Astyanax lacustris specimens exposed to the metals from the dam collapse. The study was carried out at Doce River, in Espírito Santo State, Brazil. Monthly samplings were carried out for a year. Astyanax lacustris had multiple spawning: females’ reproductive peak was in September, October, November, and December; and males between September, October, January, and February. There was a latency in the formation of gonads. For male gonads, it was necessary a 6 cm growth for it to increase from 30 to 50% and 4 cm for female gonads to increase from 40 to 50%. There is a positive correlation between gonad’s concentration of Al and Fe and the rate of histological damage in females. Male gonads had a high rate of immature cells invading the cell lumen (47.36%) and female gonads showed a higher frequency of atresia (39.64%). Fish exposed to the contaminated water showed moderate-high gonad histological damage. The observed changes can directly influence the organism’s development and reproduction in the long run, thus affecting A. lacustris population present in the region.


Introduction
On November 5th, 2015, the Fundão dam collapsed in Mariana, a municipality located in Minas Gerais State, Southeast of Brazil, causing a flood of mud and mining waste (55-62 million m 3 ) into the Gualaxo do Norte River. The mud flowed into the Doce River and spread for 600 km until arriving at the Espírito Santo coast (Escobar 2015). Gomes et al. (2019) and Quadra et al. (2019) reported an increase in several metals in the Doce River shortly after the dam burst. Macêdo et al. (2020) observed high concentrations of metals, such as iron (Fe), aluminum (Al), manganese (Mn), and lead (Pb) in the Doce River, even 3 years after the dam broke. Xenobiotics, such as metals, can act in different pathways in organisms, involving different physiological processes (Tolussi et al. 2018), affecting sexual differentiation (Baroiller and Guiguen 2001) and gonadal development (Vested et al. 2014), and inducing vitellogenin synthesis in males (Vetillard and Bailhache 2006).
Unlike organic pollutants, metals do not suffer degradation or rapid elimination from the ecosystem, causing organisms exposed to these contaminants to accumulate them through the gills or by ingestion through the food chain (Merciai et al. 2015). When ingested, metals bind to the body's molecules such as water, proteins, and enzymes, forming stable biotoxic compounds and inactivating biomolecules (Duruibe et al. 2007). These biochemical changes, whether severe or prolonged, can cause structural changes (Van Dyk et al. 2009) and, thus, cause harmful reproductive effects (James 2011). Yancheva et al. (2016) state that one of the methodologies that have a direct connection with physiological functions, such as reproduction, is histology. Greenfield et al. (2008) affirm that histological changes in target tissues are sensitive Responsible Editor: Bruno Nunes biomarkers for the effects caused by exposure to xenobiotics. These changes occur before phenotypic ones and provide a more in-depth assessment of the effects of water pollution on communities, evaluating the incidence and prevalence of abnormalities in target tissues of exposed organisms. Yancheva et al. (2016) also state that histological changes are a reflection of the health of the entire population within the ecosystem studied. Due to the Doce River disaster, it is important to study a representative population of the affected ecosystem so we can evaluate the changes occasioned by it; therefore, the species population chosen to represent the ecosystem in this study was Astyanax lacustris.
Astyanax lacustris is native to the Doce River and has an early sexual maturity (maturation in 0.7-1.0 years and generations about 18 months old). In addition to its ecological importance, A. lacustris is indicated as a sentinel species for environmental investigations concerning aquatic contamination and experimental tests in the laboratory (Prado et al. 2011), as they have numerous offspring, short generations and are easy to reproduce and manage (Silva and Porto-Foresti 2020). Astyanax lacustris does not migrate for reproduction; it can reproduce in both lotic and lentic environments and has a long reproductive period with multiple peaks of active reproduction (Godinho et al. 2010;Weber et al. 2012).
The populations of A. lacustris from the Doce River have been continuously exposed to metals throughout their life cycle, due to the environmental disaster that occurred after the Mariana dam collapsed in 2015 (Passos et al. 2020). With the hypothesis that the enrichment of metals in water can negatively influence the reproduction of fish (James 2011), this work aimed to evaluate the reproductive biology of A. lacustris in the lower Doce River and to verify the deleterious histological effects of a population chronically exposed to tailings from the dam rupture in November 2015.

Sampling area
The Doce River Basin is divided into three physiographic regions. Both the upper and middle Doce River are located in Minas Gerais, while the lower Doce River is located in Espírito Santo, and it is responsible for approximately one-third of the state's water volume (Moretto 2001). Fish and water were sampled in the lower Doce River, downstream of the Mascarenhas Hydroelectric Power Plant (UHE), Baixo Guandu, Espírito Santo State (19°30′ 04.8″ S 40°53′ 23.2″ W), close to the mouth of the Mutum Preto stream. The climate presents pluviometric seasonality with greater rainfall occurring between October and March, with variations from 50 to 300 mm and an annual total of 1019 mm (Silva et al. 2010;ANA 2020). In addition, Sales et al. (2018) and ANA (2020) determined that the largest drought occurs from April to September, with an average rainfall of less than 50 mm per month. The data obtained at the National Water Agency (ANA 2020) come from a pluviometric station (station code: 01941003) located along the course of the river in the municipality of Baixo Guandu.

Fish sampling and biometric measurements
The project was carried out with the approval of the animal ethics committee (CEUA/UVV # 563-2018). Monthly samplings of A. lacustris were carried out for one year, from June 2018 to May 2019. The fish sampling was performed with a 12-mm mesh ring net. Between 10 and 20 individuals were sampled monthly (224 individuals at the end of the study) ( Table 1). The results of fish sampling are showed bimonthly.
The fish were anesthetized with Benzocaine solution (0.2 g L −1 ) and, afterward, euthanized by cervical section. For all specimens, sexing was performed based on the presence of spikes in the anal fin of the males and the shape of the body (Dos , with the counterproof being made with the analysis of the gonads. The sex ratio of the population was assessed using the chi-square test (X 2 ) with p < 0.05; furthermore, measurements of total length (TL), body weight (BW), and gonad weight (GW) were taken. Afterward, the following biological indices were calculated: gonadosomatic index (GSI = GW × 100/BW) and Fulton's condition factor (K = BW × 100/(TL^3)). Then, the gonads were removed; a part of them was immersed in glutaraldehyde 0.5%, where they were kept until histological analysis (Venturoti et al. 2019a(Venturoti et al. , 2019b; the other part was frozen at −80°C for later analysis of metals.

Water samplings and analysis
Monthly water samplings of Doce River (collection site) were carried out in triplicate for physicochemical and metal analysis. The following physicochemical parameters of the water were measured: dissolved oxygen (DO), temperature, conductivity, and pH with the Horiba U53 multiparameter (Tokyo, Japan); and hardness and alkalinity were measured by titration, according to APHA (2005). For metal analysis, the collected water was preserved in nitric acid (pH < 2) for later analysis. Total Al, Cr, Pb, Mn, and Fe were analyzed encompassing the metals with toxic potential and evidenced in previous analyses of the Doce River's water and sediment according to Gomes et al. (2017) and Macêdo et al. (2020).

Study of gonads
Presence and absence of gonads The probability of considering A. lacustris individuals capable of reproduction, at a certain body length, was obtained through gonads' presence (visible gonads) or absence (invisible gonads) (Blain and Sutton 2016). The collected fish were separated by size categories, and the number of mature and immature individuals in each category was described. Subsequently, the frequency of individuals of each sex with mature gonads was calculated. The body length estimated at the first maturation (L 50 ) represents the body length at which 50% of the fish are in reproductive capacity. The following formula was used to obtain the percentage of growth that the population needs to move from X% of mature individuals to 50% of them: Growth percentage (%) = (L50 − LX) * (100/maximum length of the individuals). For males, we used X = 30% and for females, X = 40%, since the smallest individuals collected presented 30% and 40% of the population already sexually matured for males and females, respectively.

Histological analysis of gonadal maturation
For histological analysis, the gonads were prepared according to Venturoti et al. (2019aVenturoti et al. ( , 2019b. Afterward, the stages of gonadal development were classified according to Prado et al. (2011) using a microscope (Leica Galen III model). The stages description of the gonadal development is detailed in the supplementary material (Table 2).

Histological damage to the gonad
For both sexes, 8 gonad sections from each individual were read microscopically. A qualitative histological analysis was performed using a microscope. After this procedure, the results were transformed in a semi-quantitative way using a protocol as said by Agbohessi et al. (2015). Also, according to Agbohessi et al. (2015), the score and the importance factor for each alteration identified were multiplied and then added to obtain the gonad damage index of each individual. As stated by Bernet et al. (1999) and Agbohessi et al. (2015), the formula used to calculate these indices was I org = ∑ rp ∑ alt (α org rp alt × ω org rp alt ), where org = organ (constant), rp = reaction pattern, alt = alteration, α = score, and ω = importance factor.
For females, the histological analysis and damage index determination were also carried out quantitatively. For this, a reading of all cells and all cuts was made. The numerical data obtained were later transformed into a percentage of damage (Prado et al. 2011).
These indices were used to compare the severity of each histological alteration found in the collected fish. To classify the severity of the damage index in the testicles (It) and the damage index in the ovaries (Io), the results were evaluated based on a classification system provided by Agbohessi et al. (2015): Class 1 (index <10): tissue with normal structure and few histological changes; Class 2 (index 10-20): tissue with normal structure and moderate histological changes; Class 3 (index 20-30): clear changes in organ tissue; Class 4 (index > 30): severe changes in organ tissue. In order to determine the severity index of each individual, the changes described in Table 3 were accounted for.
In addition to the indices calculated by score and pathological importance of the lesions, another point of interest was used for histopathological characteristics (Agbohessi et al. 2015). The prevalence of each change was calculated according to the percentage of occurrence among the collected fish. Table 2 Microscopic characteristics of the gonadal maturation stages of Astyanax lacustris organisms exposed to Rio Doce mining tailings

Males Females
Resting (1) Seminiferous tubules with only sperm and occluded lumen.
Presence of early and advanced perinuclear follicles and oogonia nests.

Maturation
(2) Seminiferous tubules with sperm cell cysts and lumen full of sperm and acidophilic secretion.
Parcial spawning (3) Presence of few sperm cell cysts and partially empty tubular lumen.
Total spawning (4) Seminiferous tubules containing spermatogonia and empty tubular lumen with only a few remnants of sperm.
Presence of perinuclear follicles and post-ovulatory follicles. Table 3 List of histological changes detected microscopically in the gonads of Astyanax lacustris exposed to mining waste from Rio Doce. F) The formula used was: Prevalence of histological changes = (Number of fish with the change/Total number of fish) × 100 (Agbohessi et al. 2015).

Metal analysis
The concentrations of metals in water and gonads were measured. For digestion of the gonads, ultrapure water, ultrapure nitric acid, and hydrogen peroxide (5:2:1), and a microwave energy digester (Ethos UP from Millestone) were used. We followed the protocol of the device for digestion: 30 min at 200°C and maximum power of 1000 W. After being digested, the samples were quantified for Al, Cr, Pb, Mn, and Fe in a graphite oven, in the Atomic Absorption spectrophotometer (Thermo ICE3500). The quality assurance and quality control (QA/QC) tests of the analyses were carried out to monitor and control the reliability of the analysis methodology. For each evaluated metal, a calibration curve was prepared and for Fe and Mn, a certified reference material (ERM® -BB422) was used ( Table 4). The bioconcentration factor of each metal was also calculated, relating the concentration of metals in the water and the gonad of each fish, using the following formula: BCF = Cg/Cw, where Cg is the concentration of the metal in the gonad of each organism and Cw is the metal concentration in the collected water (Nenciu et al. 2016

Statistical analysis
The results of metal concentration and the water quality parameters were grouped every 2 months and presented as mean and standard error. The results of metal concentration in the gonad were also grouped every 2 months and presented as mean and standard deviation. The concentration of the different metals in the gonads was compared-between sexes and bimonthly-by a two-way ANOVA (p < 0.05), followed by a Tukey test (p < 0.05). The values were logarithmic (Log 10) in order to normalize the results. The GSI and FC were compared between bimesters by ANOVA (p < 0.05), followed by a Tukey test (p < 0.05).
To determine the correlation between the histological damage index and quantification of metals in the gonad, a Pearson correlation was performed (p < 0.05).

Physicochemical and metal analysis in water
The results of the water physicochemical parameters and the concentrations of Fe, Mn, Al, Cr, and Pb in water are described in Table 5. The water quality parameters did not present differences along the bimesters (p< 0.05 for all parameters). The dissolved oxygen average was 7.71 ± 0.34 mg L −1 , with a higher concentration in May and June. Temperature had an average of 27.09 ± 1.54°C, with higher values in January and February, and, lower, in July and August. We obtained the same results for conductivity and alkalinity, with higher results in May and June (101.00 ± 0.008 and 80.9 ± 3.75 mg L −1 , respectively). The average pH was 6.59 ± 0.29 and the average hardness was 44.21 ± 1.26 mg L −1 , with higher values in Jan-Feb and Jul-Aug, respectively. Concerning metals, the ones with the highest concentration were Al and Fe, with averages of 543.78 ± 125.58 μg L −1 and 491.59 ± 97.36 μg L −1 , respectively. Al was found in greater concentration in January and February, with 986.83 ± 237.08 μg L −1 ; and Fe was found in greater concentration in November and December, with 900.88 ± 131.78 μg L −1 , both in the period of greatest rainfall.

Biological indices and sexual maturation
A total of 105 males and 109 females were collected, and no significant differences were found in the sex ratio of the population (X 2 = 0.075; gl = 1; p = 0.78). The average male weight Table 4 Percentage of recovery of analytes present in the certified reference material (ERM-BB422), limits of quantification and detection practicable (mg kg −1 wet weight), limits of detection (LD), and limit of quantification (LQ) of analytes was 9.14 ± 3.81 g, and the average length was 8.46 ± 1.44 cm.
The average weight of the females was 11.89 ± 6.44 g, and the average length was 9.27 ± 1.75 cm.
Males obtained 30% gonadal maturation with 5 cm in length and 50% gonadal maturation with 11 cm in length. Between obtaining 30% of matured organisms and 50% of matured organisms, male fish had to grow 35.29% in relation to their maximum size. Females obtained 40% of gonadal maturation at 7 cm in length and 50% of gonadal maturation at 11 cm in length. Thus, between 30 and 50% of matured organisms, females needed to grow 22.22% (Table 6).
Males had a lower number of gonads at peak maturation (stage 2) in May-Jun and Jul-Aug, with 0% and 40%, respectively. Jan-Feb showed 100% of gonads at peak maturation, being considered the main months of male reproduction. Females had a higher reproduction rate in Sep-Oct and Nov-Dec, with a greater number of gonads at peak maturation (65% and 45%, respectively), being Sep-Oct considered the main months of female reproduction (Fig. 1a).
The GSI of males and females were proportional over the months. Both had a GSI peak in Sep-Oct and lower levels of GSI in Mar-Apr. The GSI of the males showed a significant difference only between Mar-Apr and Nov-Dec (p = 0.02). For females, there was a significant difference when comparing the months of Nov-Dec with Mar-Apr (p< 0.001) and with May-Jun (p = 0.003). Sep-Oct also showed a significant 0.002 ± 0.001 a 0.0002 ± 0.0001 a 0.0012 ± 0.001 a 0.00001 ± 0.0001 a 0.0001 ± 0.00001 a 0.001 ± 0.001 a Table 6 Sexual maturation of Astyanax lacustris exposed to the Doce River mining tailings in different length classes ( difference when compared to Mar-Apr (p< 0.001) and May-Jun (p = 0.002) (Fig. 1b). The Fulton condition factor (K), determinant of body condition or wellbeing of the organism, remained constant in both sexes, with K ranging from 1.3 to 1.6 for males and from 1.2 to 1.6 for females (Fig. 1c).

Concentration of metals in the gonads
Both males and females showed higher concentrations of Al and Fe in their gonads, in relation to Cr, Mn, and Pb (Fig. 2). Concerning Al, Cr, Fe, and Pb, males bioaccumulated significantly higher concentrations than females. When comparing the concentrations between sexes, only Mn showed no significant difference in Sep-Oct. When comparing metal concentrations between bimesters in males and females, separately, Cr and Pb showed no significant difference. Al showed a significant difference for both sexes between May-Jun and Sep-Oct, with p = 0.006 for males and p = 0.016, for females. Fe and Mn presented a difference only in relation to males between Jul-Aug and Sep-Oct (p = 0.034) and Nov-Dec compared to May-Jun and Jul-Aug (p = 0.044 and p = 0.032), respectively. Al presented a BCF with potential bioaccumulative effects by REACH and TSCA only in Jul-Aug (BCF = 2167.68). On Fig. 1 Annual distribution of (a) relative frequency of different stages of ovarian and testicular maturation; (b) gonadosomatic index; and (c) Fulton condition factor of Astyanax lacustris exposed to the Doce River mining tailings. Lowercase letters indicate a significant difference between bimonths in relation to males and uppercase letters indicate a significant difference between bimonths and females. Values are represented in mean and standard error the other hand, Cr had potentially bioaccumulative BCF in all months, except for May-Jun, with BCF = 120. Jan-Feb months showed highly bioaccumulative effects, with BCF = 10,027.23 (Table 7).

Histological damage to the gonads
Six types of changes were identified in the fish collected in the Doce River. In males, the greatest amount of damage was found in Jul-Aug. The invasion of immature cells had the highest prevalence among the collected fish. The invasion of immature cells and the detection of desynchronized maturation showed higher prevalence in the months of Jan-Feb and May-Jun (100% both). As for the detection of intersex, we found it present in Jul-Aug and Nov-Dec, with 20% and 16.66% prevalence, respectively, in the collected organisms.
In females, atresia had the highest prevalence among the collected fish, with only Mar-Apr and May-Jun showing a prevalence under 100%. Hyperplasia and membrane detachment had 100% prevalence in Jan-Feb and no prevalence in the months of May-Jun (Table 8). Fig. 2 Concentration of metals in the gonads of individuals from Astyanax lacustris exposed to mining waste from the Doce River. Values are represented in mean and standard error. Capital letters represent a significant difference in the bioaccumulation of metals in males in relation to periods. The lowercase letters represent a significant difference in the bioaccumulation of metals in females in relation to the periods. Asterisks (*) represent a significant difference in the bioaccumulation of metals between males and females Table 7 Bioconcentration factor (BCF) of Astyanax lacustris exposed to the mining tailings from the Doce River water along the bimesters (-) indicates that the formula could not be applied due to water values < LQ On semi-quantitative analyses, both sexes obtained, on average, moderate levels (class 2) of histological damage (Table 8). Pearson correlation (p < 0.05), for semiquantitative analysis, did not present a significant correlation between histological damage and metals in the gonad (Table 9). The average rate of damage to the female gonads on quantitative analysis was 23.41, thus classifying the female gonads as Class 3 of histological damage. Pearson correlation between metals in the gonad and quantitative damage analysis, showed a positive correlation for Al and Fe (p = 0.04, CP = 0.37 and p = 0.03, CP = 0.41, respectively) ( Table 9).

Physicochemical and metal analysis in water
Previous studies, such as the one by Escobar (2015) and Gomes et al. (2019), reported that, due to anthropological activities on the banks, the Doce River was already polluted before the disaster. Despite this, Macêdo et al. (2020) found that the metals with the highest concentrations in the Doce River's water 32 months after the disaster were Al and Fe, corroborating the results obtained in the present study. Even 4 years after the disaster, the results show that Al remains above the maximum limits allowed by Brazilian law (National Environmental Council, Resolution 357/2005) and by USEPA (1998). The results obtained in the analysis of the gonads corroborated, partially, the ones obtained in the water analysis, with Al being the metal with the highest concentration.

Biological indices and sexual maturation
According to Vicentini and Araújo (2003), the sex ratio between individuals is basic information for the reproductive potential, providing important data about the dynamics and population structure of a species. Súarez et al. (2017), studying A. lacustris in the Pantanal, Brazil, reported higher proportions of females and reported that the imbalance in the proportion of males and females is unfavorable for natural selection in the environment. Contrary to the findings by Súarez et al. (2017) with A. lacustris, the present study revealed a balanced proportion between males and females in the population of the lower Doce River, which is a beneficial factor for the reproduction of this population; furthermore, the onset of sexual maturity is also an important component in the population dynamics of fish (Súarez et al. 2017). When analyzing the data obtained in the present study, we found a latency in gonad maturation when compared to the individuals of other populations of A. lacustris (Table 10).
Studying a smaller population of A. lacustris (maximum size = 7 cm for males and females), Súarez et al. (2017) reported male individuals with 30% gonadal maturation at 2.1 cm in length and 50% at 3.1 cm in length. The difference between 30 and 50% of matured individuals was only 1 cm (growth = 14.28%). On the other hand, the males in the present study needed to grow 6 cm (35.29% of the maximum size), so that 30% of individuals with gonads passed to 50%. Regarding females, the population of A. lacustris in the Pantanal increased 0.3 cm (4.28% of the maximum size) to pass from 40% of individuals with gonads to 50% (Súarez et al. 2017). In the present study, A. lacustris females needed to grow 22.22% to increase from 40% of individuals with gonads to 50%. Only one female was collected with a length between 14 and 15 cm, and it did not have evident gonads. Studies carried out by Souza et al. (2015), Dala-Corte and Azevedo (2010), and Oliveira et al. (2019) with species of the genus Astyanax observed a much lower need for proportional growth (1.11-5%). This latency in the formation of gonads is a strong indication of a disruption in the reproductive biology of the population of A. lacustris from the lower Doce River and deserves to be the target of future investigations. Previous studies have reported that chronic exposure to metals, such as Al, can cause a delay in the sexual maturation of fish, as well as increased spermatogonia and spermatocytes and a significant decrease in spermatids and spermatozoa (Paschoalini et al. 2019). Another factor observed is that Table 8 Prevalence (%) of the histopathological changes observed in the Astyanax lacustris gonads exposed to the Doce River mining tailings. The importance factor is indicated in parentheses in each change. DC, deleterious changes; IS, intersex A. lacustris individuals from the lower Doce River showed a size greater than 60% of the maximum size for 50% of the population to have gonads (male = 64% and female = 61%). The population of A. lacustris of the Paraguay River has 50% of individuals with gonads with 44.28% of the maximum size, for both males and females. Species of the Astyanax genus usually have a seasonal breeding strategy, which increases in the rainy season, between spring and summer (Dala-Corte and Azevedo 2010). The present study corroborates this affirmation since the reproduction peak found was in Sep-Oct for females and Jan-Feb for males, during the rainy season in the region. Despite the higher percentage of gonads with reproductive capacity in the rainy season, the other periods also showed matured gonads in a lower percentage, a fact that indicates multiple reproduction. Souza et al. (2015) suggest that this fragmented reproduction strategy may be associated with the unpredictable nature of the reproduction region's conditions, a reality observed at UHE Mascarenhas due to the high volume of water released.

Concentration of metals in the gonads
Through a study done with Barbus grypus, Barbus sharpeyi, and Cyprinus carpio, Alhashemi et al. (2012) concluded that sex is an important factor that can interfere with the bioaccumulation of metals in fish. C. carpio males showed greater bioaccumulation of metals when compared to females, corroborating the results obtained in the present study.
According to Passos et al. (2020), the lower Doce River is a region that has frequent resuspension events during the  rainfall season (October to March). Due to the high amount of rain, the water flow released by UHE Mascarenhas also increases (in the dry season, the water flow is 183.82 m 3 s −1 and in the rainy season, it increases to 387.25 m 3 s −1 (ANA 2020)). Besides that, it is known that the UHE Mascarenhas has a periodic opening of its floodgates twice a day, according to ANA (2020). Those events can make elements, such as metals previously sedimented, available again in the water column. This corroborates the result found in the present study, which showed a higher concentration of Al and Fe in the gonads during the rainfall months. In freshwater fish, Al is known to affect their reproduction, due to a directly proportional reduction in vitellogenesis (Hwang et al. 2000). In addition, acidified water, a result of high concentrations of Al (Hwang et al. 2000;Correia et al. 2010), is known to impair fish reproduction by affecting fertility, egg viability, spawning success, gonadal development, and production of gametes (Correia et al. 2010). Passos et al. (2020) state that, in addition to Al, Fe also causes deleterious effects in fish exposed to the metal, which can lead to increased mortality and histopathological changes in liver cells. When compared to the results of the present study, it is concluded that, in addition to histopathologies in liver cells, Fe may also be responsible for histopathologies in gonadal cells. Although Al and Fe showed higher concentrations in the gonads of the organisms, Cr presented a greater bioaccumulative effect in relation to other metals when BCF was calculated. Despite being an essential metal, Bakshi and Panigrahi (2018) revealed that Cr can affect the behavioral, histological, biochemical, genetic, and immunological conditions of organisms, leading to changes in the fish exposed to it.

Histological damage to the gonads
The release of immature cells into the lumen is indicative of a deficiency in spermatogenesis and, possibly, testicular functionality (Agbohessi et al. 2015;Da Cuña et al. 2011). In the present study, it was also observed the release of immature cells into the lumen in the testicles of A. lacustris in greater quantity (47.36%) when compared to other histological damages found. Intersex was also observed, and despite being found in a smaller quantity, it is of great concern in the reproductive biology of fish since it is considered a signature of exposure to xenobiotics and it can directly impact the development and reproduction of offspring (Prado et al. 2011). Although the genotype of intersex fish had not been determined, the macro and microscopic aspects of their gonads (Prado et al. 2011) indicated that the individuals were male.
Atresia is a degenerative and resorption process that occurs both naturally and under conditions of environmental contamination (Weber et al. 2003). Studies with Danio rerio and Chalcalburnus tarichi (Kaptaner and Ünal 2011;Luzio et al. 2016) found a significant number of atretic oocytes when fish were exposed to xenobiotics, corroborating the present study and highlighting that exposure to xenobiotics can delay the development of oocytes, inducing them to death and potentially reducing the individual's reproductive success. Cardoso et al. (2018) performed an experiment with Danio rerio females in order to confirm the applicability of the semiquantitative method of counting in female gonads. and concluded that, if the objective was to obtain a general classification of maturation or some differences between treatments, the semi-quantitative method is considered appropriate; however, if the objective of the work is to explore in detail the structural components of the gonads (such as primary, cortical, vitellogenic, and pre-vitellogenic oocytes), the recommended method is the quantitative one. The results of the present study are in agreement with this statement since the stages of gonadal maturation of males and females were successfully determined by the semi-quantitative method; nonetheless, when a more detailed analysis of female gonads was made using the quantitative method, the correlation between histological damage and the concentration Al and Fe was significantly positive. Moreover, when the results were classified into levels of histological damage, the semi-quantitative method classified females as Class 2 of histological damage with moderate changes, and the quantitative methodology classified females as Class 3 of histological damage with clear alterations in the tissue of the organ. Thus, it is concluded that the method of quantitative determination of histological damage in female gonads from A. lacustris is more recommended than the semi-quantitative method.

Conclusions
The environmental disaster caused by the rupture of the Fundão dam raised the concentration of metals, such as Al and Fe, in the waters of the affected region, a scenario that persists until today, as observed in the present study and previous ones. The population of A. lacustris from the lower Doce River shows obvious deleterious effects on the gonads, namely: (1) latency formation; (2) high concentrations of Al and Fe; and (3) clear histological alterations, such as the presence of atresia, hyperplasia, and the invasion of immature cells into the lumen. In females, there is a positive correlation between the concentration of Al in the gonads and the histological damage index estimated by the quantitative method. Our results allow us to predict that the permanence of high concentrations of metals (ex: Cr, Fe, and Al) in the water of the lower Doce River will continue to subject the population of A. lacustris to harmful effects on the gonads. This long-term scenario may further compromise the reproductive success of this population and, consequently, of other species that are at a higher level in the trophic chain.
Availability of data and materials All data generated or analyzed during this study are included in this published article and its supplementary information files.
Author contribution JM contributed to the study conceptualization, the formal analysis, the investigation, methodology, project administration, supervision, visualization, and to both writings-original draft and review & editing; The investigation and the project administration were also performed by DSC and BCT; TMP, AB, CV, and SN performed the investigation and LCG contributed with all roles, except for the investigation and the writing of original draft.
Funding Research and Innovation Fundation of Espírito Santo (FAPES) financially supported (grant #19/2017) and granted a scholarship to JM, DSC, BCT, and AB.
The Brazilian Coordination for the Qualification of Higher Level Personnel (CAPES) granted a scholarship to TMP (Finance code 001).
The Brazilian National Council for Scientific and Technological Development (CNPq) granted a scholarship to LCG and SN.

Declarations
Ethics approval The project was carried out with the approval of the animal ethics committee (CEUA/UVV # 563-2018).

Consent to participate Not applicable.
Consent to publish Not applicable.

Competing interests
The authors declare no competing interests.