Histopathological Biomarkers as Indicators of the Environmental Quality of Two Estuaries in Northeastern Brazil

ITALA Gabriela Sobral SANTOS (  italagsobral@hotmail.com ) UFRPE: Universidade Federal Rural de Pernambuco https://orcid.org/0000-0003-3952-9799 Alex Souza Lira Universidade Federal do Delta do Parnaiba Carolina da Silva Montes Universidade Federal do Pará: Universidade Federal do Para Flávia Lucena Frédou Universidade Federal Rural de Pernambuco Rossineide Martins Rocha Universidade Federal do Pará: Universidade Federal do Para


Introduction
Estuaries are transitional ecosystems characterized by ecological cycles that receive high inputs of organic and inorganic contaminants, such as pesticides and heavy metals (Salas et al. 2017) from the multiple coastal anthropogenic activities, which negatively affects the health of humans and biota (Dellamatrice and Monteiro 2014;Carr et al. 2017). These contaminants are discharged into the water and accumulate mainly in the sediment (Kameli et al. 2017;Willacker et al. 2017).
Discharges of domestic, industrial and agriculture e uents are among the activities with the greatest potential for contamination in tropical coastal ecosystems (Saldarriaga-Hernandez et al. 2020). The rst challenge is to identify chemicals that potentially present a risk to resident sh (Naidu et al. 2016).
Industrial activities are responsible for many impacts and the discharge of heavy metal, has become a worldwide concern (Boran and Altinok 2010; Ali et al. 2019). Industries of leather tanning, production of cement, chlor-alkali, mining, anticorrosive agent, and fertilizer production (Pacyna et al. 2010;Cipurkovic et al. 2014;Aslam and Yousafzai 2017;Becker et al. 2018;Chellaiah 2018;Vieira et al. 2019) are among the main industrial activities that threaten the balance of ecosystems. Pollutants, such as pesticides and herbicides, used in the cultivation of sugar cane, are also of great concern given their impact on estuaries, consequently on biota (Chavan and Muley 2014;Coelho et al. 2018).
Fish are considered good indicators for assessing the quality of aquatic environments since they bioconcentrate and bioaccumulate contaminants by direct absorption from water or indirectly by consumption of contaminated prey (Kroon et al. 2017), consequently reaching the upper trophic levels (Wang 2002). The choice of species to be used as an indicator of contamination is essential (Van der Oost et al. 2003), and those from different trophic guilds and with distinct trophic position are recommended because they use different absorption/accumulation routes, directly related to the type of food consumed (Terra et al. 2008).
Histopathological biomarkers are widely applied to investigate the effects of contamination on organisms, due to the possibility of identifying lesions in cells, tissues, and organs (Hook et al. 2014;Van der Oost et al. 2003;Viana et al. 2013). The mechanism to evaluate this biomarker is through histopathological changes in the tissue, as it is a direct response of sublethal effects, allowing the detection of the effect of pathogens in an e cient and easy-to-apply manner Fonseca et al. 2016). Among the organs, the liver is often used since it is essential for the physiological functioning of the animal, by detoxifying and excreting toxic substances, and bioaccumulate pollutants (Borges et al. 2018). Gills, owning to their fundamental role in gas exchange and osmotic regulation in shes, are similarly recommended for environmental biomonitoring (Ameur et al. 2015). This approach is widespread used to assess water quality in polluted environments (Dalzochio et al. 2016).
In this study, for the rst time in northeastern of Brazil, we evaluate the environment quality of the Santa Cruz Channel estuary (area 1) and Sirinhaém river estuary (area 2) using liver and gills as histopathological biomarkers for shes from different trophic levels. These areas are strongly impacted by the intense domestic and industrial e uent discharge, shrimp farming, and artisanal shing (Medeiros et al. 2001;David and Fontanetti 2009;Moura and Candeias 2009).

Study area
The study area includes two estuarine systems, positioned along the coast of Pernambuco state, northeastern Brazil (Fig. 1). The Santa Cruz Channel Estuary (SCC, area 1) is the largest and highly productive estuary in Pernambuco coast (Macêdo et al. 2000) (Fig 1a). It is a 22 km long U-shaped channel with width ranging from 0.6 to 1.5 km, located within the Santa Cruz Environmental Preservation Area (APA Santa Cruz). A chloro-alkali industry was responsible for launching 22 to 35 tons of inorganic mercury in an important e uent of the SCC complex for approximately 24 years of operation (Meyer 1996). These impacts may have contributed for the reduction of up 10% of mangrove coverage in this area (Pelage et al. 2019).
The Sirinhaém River Estuary (SIR, area 2) is a small, shallow, coastal plain estuary, 9.5 km long, 350 m wide increasing up to 800 m at the river´s mouth and with maximum depth ranging from 1.2 to 4.5 m (Silva et al. 2011). Located between the Marine Protected Area of Guadalupe and the Marine Protected Area of Sirinhaém, SIR is characterized by a high density of mangrove, occupying 18 km 2 (Maia et al. 2006) (Fig. 1b). Although it is of extreme biological importance, multiple impacts are reported in the SIR estuary such as the sugar cane and domestic and industry e uent (CPRH 2011).
Since the sh sampled material arrived dead on board, the Brazilian animal ethics committee wasn't required given that any method of sacri ce inside the authorized shery activity wasn't demanded, therefore their approval wasn't needed. All specimens were stored on ice and transported to the laboratory. In the laboratory, all shes were measured (Standard Length -cm) and weighed (g). Posteriorly, an abdominal incision was made to assess the liver.

Histology
The liver was weighted, and the median portion of this organ and the second branchial arch right were removed. Subsequently the collected tissues were xed in 10% formalin for 24 hours and then preserved in 70% ethanol for histopathological analysis and subjected to routine histological processing for para n embedding (Prophet et al. 1995). Samples were subsequently sectioned into 5-μm-thick sections using a RM2245 microtome (Leica Microsystems, Germany). The liver was stained with hematoxylin and eosin (HE), hematoxylin and eosin-phloxine B (HE-P), Mallory Trichrome. The gills were stained with (HE) and (HE-P). Stained sections were analyzed, and photomicrographs were obtained using an Eclipse Ci-S light microscope (Nikon, Japan) connected to a Nikon DS-Ri1 digital camera.

Histopathological indicators
2.4.1. Histopathological Index of the liver (HIL) Liver changes were semi-quantitatively assessed according to the adapted protocol described by Bernet et al. (1999). The Histopathological Index of the Liver (HIL) was calculated for everyone by multiplying the importance factor by the occurrence value of each alteration (see Table 1-List of alterations).
According to this index, Bernet et al. (1999) an occurrence value was attributed to each change by taking into consideration the following tissue injury levels: (1) minimum -easily reversible change; (2) moderate -reversible if the stressor is neutralized; and (3) severe -often irreversible, thus leading to decreased liver function. It was possible to assign an occurrence value to the changes after they were identi ed. This value corresponds to the level and extent of the tissue changes, namely: (0) unchanged; (2) occasional; (4) moderate and (6) severe (diffuse lesion).
High HIL values indicate more severe injuries in the affected organs. These indices are indicative of the extent and intensity of histological alterations in tissue morphology expressed into a quantitative value.
Because of this, the method recommends to use the frequency of the histological alterations, calculated from the occurrence of an alteration divided by the total HIL of the sample and multiplied by 100, in all analyzed sh and for each site.
The Hepatosomatic Index (HSI) (Eq.1) (Slooff et al. 1983) and Condition Factor (K)(Eq.2) (Smolders et al. 2002) were also used to assess the status of the sh health, since the variation in the weight of the organ in relation to body weight is a response of the individual when exposed to contaminants (da Mata Pavione et al. 2019). The HSI was calculated as: While the condition factor calculated by Eq. 2: Where W = body weight (g) and L = total length of sh (cm).

Fish health Index (FHI)
The Fish Histopathological Index (FHI) (Zimmerli et al. 2007) (was used to classify liver injury severity. This index is divided into ve classes as follows: Class I (< 10), normal/healthy tissue structure without impairments or pathological alterations; Class II (11 -20), normal tissue structure with slight histological alterations; Class III (21 -30), moderate modi cations of normal tissue and morphology; Class IV (31 -40), pronounced histological changes (in the liver); and Class V (>40), severe histological alterations of normal tissue and morphology.

Histopathological Index of the gill (HIG)
The gill histopathological alterations were calculated from the adaptation of the Poleksic and Mitrovic-Tutundzic (1994) classi cation. The type and location of the damaged gill tissue are divided into ve groups: a. hypertrophy and hyperplasia of gill epithelia and related changes; b. changes in the mucous cells; c. gill parasites; d. blood vessel changes and e. last terminal stages. After that, the method determined stage of each group of alteration (Table 1), The HIG was calculated as Eq.3: Where I, II, and III are the number of stages for lesions. An overall score is used to classify the organ 0-10: functioning normally, 11-20: slightly to moderately damaged, 21-50: moderately to heavily damaged and > 100 irreparably damaged.

Statistical analysis
Firstly, the normality and homoscedasticity for HIL and HIG data were tested using Shapiro and Bartlett tests respectively. However, the assumptions of parametric tests were not reached, hence the Kruskal-Wallis non-parametric test, followed by Nemenyi test for multiples comparisons (post-hoc) was performed to compare, the HIL and HIG values difference between species, guild and site. In addition, for comparisons between the lesion classes HIL (4 stages) and HIG (3 stages) for each species.
To assess the degree of resemblance in the histopathological alterations among species for liver and gills, multidimensional scaling (MDS) based on a Euclidian distances matrix was applied using the frequency of occurrence of histopathological alterations (%) as input, with the specimens considered as the sampling unit. Wisconsin double standardization was applied to improve the gradient detection ability of the dissimilarity indices (Bray et al. 1957). To compare the whole set of histopathological alterations among sites and sh species, a two-way PERMANOVA was applied (Wang et al. 2009). All analyses were performed in the R environment (Core Team R 2020), using vegan (Oksanen et al. 2017) and rrcov packages (Todorov and Filzmoser 2009). For all analyses, a signi cance level of 0.05% was considered.

Fish analysis
A total of 146 individuals were analyzed. Gobionellus stomatus (3.82) and Bairdiella ronchus (1.13) in Santa Cruz Channel Estuary had the highest HSI (Table 2). In terms of condition factor (CF), the highest values were observed in Caranx latus (2.68) and B. ronchus (1.97), and the lowest in G. stomatus (1.01).

Histopathology of liver
We observed 13 types of alterations in the liver distributed in the classes: circulatory (1), in ammatory (3), progressive (1) and regressive (8), summarized by the Histopathological Index of Liver (HIL - Fig. 2). These alterations are divided into four categories given the reaction pattern, such as (1) circulatory disturbances (pathological condition of blood and tissue ow), including hemorrhage and aneurysm, (2) regressive changes (causes reduction or loss or the organ function) which involve cytoplasm alteration, hepatic steatosis, hepatic parenchyma, necrosis, nuclear alterations, and vacuolar degeneration, (3) progressive changes (increased activity of cells or tissues) represented by hypertrophy and hyperplasia in hepatic parenchyma and (4) in ammation (consequence of other reaction patterns) such as in ltration and parasites. The more relevant alterations are regressive changes. Six alterations were the most representative: hemorrhage (97.26%), structural alterations in liver tissue (100%), steatosis hepatic (92.46%), necrosis (72.60%), vacuolar degeneration (95.89%), and in ltration (94.52%). All studied sh species showed these changes but with different magnitudes.
Considering the areas, area 1 was signi cant different to area 2, and C. undecimalis in area 2 and G. stomatus in area 1 had the most severe changes. The HIL index of G. stomatus in area 1 were signi cant higher (66.47 ± 3.32) than those observed in B. ronchus (55.21 ± 5.85) (Fig. 2-HIL total). In area 2, the HIL was signi cantly different between species C. undecimalis (67.08 ± 5.97), C. latus (45.4 ± 3.44) and C. parallelus (52.10 ± 4.26). Based on Fish Health Index (FHI), all species were considered with severe alterations given the morphology and hepatic parenchyma.
The rank of the histopathological index of liver (HIL), in severity order, was Centropomus undecimalis > Gobionellus stomatus > Bairdiella ronchus > Centropomus parallelus > Caranx latus. Although C. undecimalis and G. stomatus are more impacted in absolute values of HIL, G. stomatus had severe histological changes, such as fatty degeneration. In relation to the trophic guild, HIL values were signi cant different between piscivore and piscivore/zoobenthivore, respectively C. undecimalis and C. latus, and detritivore versus piscivore/zoobenthivore, corresponding to G. stomatus and C. latus. In general, piscivores (C. undecimalis and B. ronchus) and detritivores (G. stomatus) were more affected by local contamination.
In the MDS plot based on liver alterations, it is possible to observe a clear separation of G. stomatus from the other species (Fig. 3a). There were signi cant alterations in the whole histopathological alterations between G. stomatus and other species (pseudo-F = 9.89; p < 0.001), and there was difference between areas (pseudo-F = 21.84; p < 0.001), mainly due to the high occurrence of four changes (Fig. 3b), the vacuolar degeneration (Fig.4g), architectural and structural alteration, hyperplasia, and hepatic steatosis, in comparison with the other species. It was observed a high frequency of some alterations by species: such as necrosis in C. undecimalis (Fig. 4e), in ammation by leukocyte in ltrate in C. latus (Fig. 4d) and parasites in B. ronchus.
Hepatic steatosis (Fig. 4b-c) was noted with a varying degree of intensity, indicating accumulation of lipids. However, particularly to G. stomatus, it was observed the liver parenchyma completely lled with fat accumulation (Fig. 4g).
We found parasites in 12.32% of the individuals of all species, except in G. stomatus. The most affected species were B. ronchus and C. latus, with 29.16 and 17.14% of livers infected (Fig. 3b). Parasites were identi ed as metacercariae of digenetic trematodes (Fig. 5b-d) and myxozoan parasite (Fig. 5a). Metacercaries were found within the exocrine pancreas and hepatic parenchyma. Its presence has a clear separation within the blood congestion (Fig. 5b) and presents a central portion of inde nite shape, with a circular tendency in pink and branches in yellow. These pathogens were surrounded by a capsule of connective tissue (Fig. 5c). Myxozoan paratise (Fig. 5a) were observed in the bile ducts of B. ronchus which presented granulomas with in ammatory response, the parasite's morphology is mainly composed of collagen bers, blue in color stained with mallory of trichrome. It occupies the entire bile duct and multiplies within it, increasing the size of this in ammatory capsule.

Histopathology of gills
There were 18 types of alterations in the gills, distributed in 3 stages: stage I (10), stage II (7) and stage III (1) (Fig 6). The most abundant changes were: lifting of epithelial cells (65.06%) and fusion of the several secondary lamellae (33.56%) in stage I, lamellar aneurysm (63.01%), vessel blood congestion (46.57%) and rupture of the lamellar epithelium (56.16%) in stage II, and necrosis (34.93%) in stage III. The necrosis alteration is an irreversible damage.
The histopathological index of gill (HIG) (Fig. 6) in the area 1 was signi cant different than area 2. In area 1, it was signi cantly higher in Gobionellus stomatus (93.55 ± 10.36) than in Bairdiella ronchus (66.95 ± 12.21). In area 2, the HIG did not signi cantly varied among species. In terms of severity, the rank of HIG were: Gobionellus stomatus > Bairdiella ronchus > Centropomus undecimalis > Centropomus parallelus > Caranx latus. Although all species were considered to have moderate to heavy damaged gills, C. undecimalis in area 2 and G. stomatus in area 1 were more impacted. Moreover, clearly, the species caught in the area 1 is more contaminated in relation to the area 2 species. Thus, the anthropized environment were more decisive in relation to the severity of the lesions than the species trophic level.
The frequency of changes was generally divided into four classes, 6.16% of sh with normally functioning gills (HIG= 0-10), and 13.1% slightly to moderately damaged gills . The more representative third class (21-50), with 45.89% of the analyzed species, corresponds to moderately to heavily damaged gills, with more severe changes that can affect normal functioning. In the latter class (33.56%) (HIG values > 100), it is observed irreparably damaged gills.
In the MDS plot based on gills (Fig. 7a), it was observed spatial differentiation of G. stomatus and B. ronchus compared to other species (pseudo-F = 2.23; p < 0.05), showing signi cant different between areas (pseudo-F = 8.06; p < 0.001). There is a predominance of speci c alterations in some species, i.e., G. stomatus and B. ronchus with necrosis, C. parallelus presenting hemorrhage with rupture of epithelium, and C. undecimalis with a complete fusion of all the secondary lamellae (Fig. 8c). In general, changes such as necrosis, aneurysm lamellar, rupture of the lamellar epithelium and hemorrhages with rupture of epithelium were present in all studied sh in signi cant proportions ( Fig. 7b and 8).

Discussion
The use of histopathological biomarkers is widely used to assess the health status of sh in anthropized environments, since it allows the detection of sublethal effects, being a low-cost and easy-to-apply tool. In accordance with our ndings obtained with the use of liver and gills and biomarkers, the species evaluated were effective for the diagnosis of environmental impact in the study areas, based on the degree of cellular damage in individuals.
Fish species with different trophic guilds were chosen in this study, which implies that they have different contamination routes (Viana et al. 2013), and then a close correlation with the contamination because of biomagni cation process. The trophic guild piscivores (C. undecimalis and B. ronchus) and detritivores (G. stomatus) were more affected, and hence may act as better bioindicators of impacted areas. Piscivorous species tend to have the greatest contamination effect due to the natural synergistic process of biomagni cation. However, detritivorous species such as Gobionellus stomatus from polluted environments can be considered more sensitive to environmental variations, being considered a good indicator of the environmental quality of the estuary, similar to that obtained by (Do Prado et al. 2020).
It was observed high values of hepatosomatic index (HSI) in all species. The liver is an organ commonly used to assess contaminant effects and anthropic impact in sh, mainly through morphometric indices such as HSI and condition factor (K). Several studies have already evaluated the positive correlation between liver enlargement and the exposure to chemical pollutants, which can be a metabolic activity during the process of detoxifying contaminants (Schmitt and Dethloff, 2000;Van der Oost et al. 2003).
The size of the liver can be used as a tool to monitor the health of sh in impacted aquatic environments (Van der Oost et al. 2003). Higher HSI values are associated to the presence of toxic compounds and, the increase of liver size is a result of hypertrophy and hyperplasia processes, as well as the increasing ability to metabolize xenobiotics (Araújo et al. 2018;da Mata Pavione et al. 2019). This relationship between the high HSI and accumulation of lipids was also reported by Fåhraeus-Van Ree and Spurrell (2003) for Limanda ferruginea from Witless Bay, Newfoundland and Labrador.
As for the HSI, it was observed in G. stomatus the lowest K value among the analyzed species. The K is often associated to water quality and food availability, hence indicator of the sh health ( Van der Oost et al. 2003;Rossi et al. 2020). This index may be reduced by the presence of pollutants given the increased metabolic activity and demand of energy for detoxi cation process, altering the sh feeding rate (Heath 1995). All species in the studied estuaries had high values of histopathological alteration index of the liver (HIL) (HIL=45.44 -67.8). Several authors have proven the e ciency of histopathological biomarkers for environmental monitoring through the HIL, in order to evaluate the health condition of sh (Viana et al. 2013;Kostić et al. 2017;Borges et al. 2018). The observed values were higher than the values recorded for species in other estuarine and coastal areas submitted to anthropic impacts: Solea senegalensis (HIL<10) in contaminated sediments from south of the Gulf of Gabes, Tunisia (Ghribi et al. 2019); Solea senegalensis (HIL<12) in sediment from Sado estuary (Costa et al. 2009); and Solea spp. (HIL<12) in Bilbao estuary (Bay of Biscay) (Briaudeau et al. 2019). In Brazilian estuaries, our indices were much higher those reported for Atherinella brasiliensis (HIL<30) in the Estuarine-Lagoon Complex of Iguape-Cananéia, São Paulo (Salgado et al. 2018) and Cathorops spixii (HIL<12) in Paranaguá Bay, Brazil (dos Santos et al. 2014).
According Zimmerli et al. (2007), severe histological changes in the tissue (HIL>40) may be categorized as irreparable, like necrosis, that can lead to mortality due to the loss of vital tissue function or increase of susceptibility to disease. The HLI and the HSI are directly related, and possibly correlated -high HSI values lead to high HIL, and vice-versa. High HSI is commonly associated to sublethal impacts of e uents, such as retention zones for contaminants, exposure time and bioavailability of contaminants (da Mata Pavione et al. 2019).
The liver changes found for the studied species ranged from mild to severe, with loss of tissue function, and liver tissue and bile duct necrosis were the most serious injury (72.60%). Santos et al. (2014) observed necrosis alteration in 100% of liver of Cathorops spixii when evaluating the effects of bioaccumulation of butyltins in Paranaguá Bay. Khoshnood (2017) reported that different sources of pollution can result in moderate to severe damage to liver cells. Our ndings were similar to the obtained by Oliva et al. (2013) with Solea senegalensis in Huelva, Spain; Cathorops spixii in Paranaguá Bay (dos Santos et al. 2014), and Arius thalassinus in Hodeida, Yemen Republic (Saleh and Marie 2016). In general, all the studies presented above reported alterations as observed in our study, the liver had an increased vacuolar degeneration, necrotic foci, hemorrhage, leukocyte in ltration, granulomas and parasites when submitted to an anthropo ed environment. Among these alterations, the granuloma has a complexity due to the origin of its formation.
Granulomas observed for all species in this study were associated with in ammatory processes, which are capsules surrounded by multiple brotic layers involving an in ammatory response (Fricke et al. 2012). This alteration may be a response to environmental stress, or a process initiated by parasites (Araújo et al. 2019), as observed in our study. Parasites affect structure of tissue, increase granulomas, and can cause irreversible damage ). These reaction patterns were similarly observed in all sh species of our study, associated with the presence of parasites. Highly anthropized environments directly affect the immunity of sh that inhabiting this ecosystem, making them more susceptible to parasite infestations (de Melo Souza et al. 2019;Falkenberg et al. 2019;Byers 2020).
All sh specimens evaluated in both areas presented around 95.89% vacuolization in the liver. It usually occurs due to the storage of lipids or glycogen as an energy source when found in low proportion or minimums, but it is possible to increase if subjected to some stressors like contaminant exposure and nutritional status of the species (Wolf et al. 2015). Oliva et al. (2013) found 100% vacuolization in the Solea senegalensis, from a polluted estuary in Huelva, Spain.
Gobionellus stomatus had a steatosis in high degree of intensity when compared to other species. The steatosis or fatty degeneration found in all species in this study may be derived from the toxic exposure to pollutants, such as organic compound and heavy metals (Wolf and Wolfe 2005) . The severity of this pathology is due to the direct impact on liver functions, enabling an increase in the individual's susceptibility to poison by contaminants (Roberts and Rodger 2012). This pathology in G. stomatus may be associated with a speci c response of this species to contaminants, similar results were observed the degenerative processes associated with excessive fat storage in sh from heavily polluted river in Zoarces viviparus of the Baltic Sea (Fricke et al. 2012).
In general, the studied shes had moderately to heavily damaged degree of gill damage. Some individuals had permanent lesions, such as necrosis, resulting in irreparable damage, indicating that these sh are being subjected to long exposure to toxic agents. The presence of these lesions indicates that a complex mixture of contaminants is present in natural environments affecting the shes and then health condition (da Silva Montes et al. 2020). Mucous and chloride cells are the cells mostly affected by toxic substances, as they act in the defense and transport mechanism of chloride ions, respectively. These stress-inducing factors can induce pathological changes (David and Fontanetti 2009;Nimet et al. 2020).
In this study, the histopathological index of gills (HIG) ranged from 36.68 to 93.55. Fishes as histopathological bioindicators to assess environmental contamination are being used by several authors. Carvalho et al. (2020) evaluated Micropogonias furnieri and Menticirrhus americanus (HIG = 100) from the Bay of Sepetiba-Rio de Janeiro, an area known for the disposal of materials and contaminants, and recorded indexes higher than those reported in the present study. Centropomus undecimalis, with an average HIG of 66.45, is frequently used as a bioindicator of water quality in impacted areas. Several authors in Maranhão (Brazil) (Cantanhêde et al. 2014(Cantanhêde et al. , 2016(Cantanhêde et al. , 2018Santos et al. 2014) found values lower than the observed in our study (HIG = 19.11 to 53).
In both estuaries, all sh species had several alterations. In area 1, mainly G. stomatus was more impacted, presented severe lesions, such as necrosis. In general, the histological changes observed in the liver and gills are not of speci c origin, being associated with their response when exposed to toxic agents, which can vary from heavy metals, and vinasse to pesticides (Chavan and Muley 2014;Coelho et al. 2018). Even though the studied areas are impacted by different anthropic sources and different severities, such as the launch of 22 to 35 ton Hg in the area 1 for 24 years and the sugar industry established in area 2 since of the 19th century (Meyer 1996;Pelage et al. 2019), it is di cult to determine the main source that causes histopathological changes in the liver and gills. Heavy metals are responsible for several deleterious effects in sh tissues leading to changes in metabolism, such as hypertrophy (Cr), hyperplasia, lamellar fusion, rupture of epithelium lamellar (Zn), hypertrophy and hyperplasia of primary and secondary lamellae; and liver hepatocytes vacuolations (Cd) (Labarrère et al. 2012;Naïja et al. 2016;Aslam and Yousafzai 2017). Moreover, the release of vinasse in the estuary can favor the appearance of lesions, such as rupture and hemorrhage in the gills (Correia et al. 2017). The pollution of these areas has already been studied by Lopes (2018) which evaluated the pollution in sediments and biota, nding Hg values above the level 2 established by o cial agency for the (CONAMA 2012), which means a high probability of toxic effect to biota. This presence of this metal resulted in alteration in neurological physiology, measured by inhibition of AChE activity in C. undecimalis, used the as the bioindicator species.
Alterations were more severe for the sh collected in area 1, and correlated to the trophic guild, specially piscivorous and detritivores, which were mostly affected. Gobionellus stomatus had fatty degeneration with a magnitude of severity, totally different from other species. This species and C. undecimalis (or of similar genus) are already considered keystone species in these estuaries (Lira et al. 2018) and considering our ndings, can also certainly be considered as e cient bioindicators in tropical estuaries. The combination of methodologies pioneering in the Northeast of Brazil, using different species and trophic guilds, was e cient in diagnosing the health status of the area using sh as a bioindicator. There is no study of how these potential damages, evident by the histological alterations, can impact human health, mainly by trophic biomagni cation, since the studied species are of great importance to the local community, through a source of food and income. This study serves as a warning to the population that feeds on sh of impacted estuarine areas, such as the present case, and public policies and constant environmental monitoring are needed to investigate the effects of aquatic pollution and consequently, assure safe food to the population.

Declarations Acknowledgments
We are grateful to BIOIMPACT Laboratory of the Federal Rural University of Pernambuco for eld assistance, Aquaculture Technology Laboratory and the Fisheries Oceanography Laboratory for providing the laboratory structure needed for the analysis (UFRPE). We are also grateful to the Cellular Ultrastructure Laboratories -Institute of Biological Sciences of the Federal University of Pará (UFPA), for the help with histological analyses. This research was funded by PRPPG/UFRPE 015/2018 and is a contribution to LMI TAPIOCA. The Coordination for the Improvement of Higher Education Personnel (CAPES) nanced a student grant to the rst author, and the National Council for Scienti c and Technological Development (CNPq) nanced a productivity research grant Flávia Lucena-Frédou and Rossineide Martins da Rocha (308554/2019-1, 307688/2019-4, respectively).

Declaration of Ethics approval and consent to participate
Not applicable, because of the sh sampled material arrived dead on board, the Brazilian animal ethics committee wasn't required given that any method of sacri ce inside the authorized shery activity wasn't demanded, therefore their approval wasn't needed Declaration of Consent for publication "Not applicable"

Declaration of Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Histopathological index of gills (mean± SD) by studied species. Letters indicate the results of Nemenyi tests among species. Bai.ron-Bairdiella ronchus, Gob.sto-Gobionellus stomatus, Car.lat-Caranx latus, Cen.par-Centropomus parallelus, Cen.und-Centropomus undecimalis.