Sensitivity , ecotoxicity and histopathological effects on neotropical fish exposed to glyphosate alone and associated to surfactant

1 University Center of Barretos Educational Fundation. Av. Prof. Roberto Frade Monte no 389, 14.783-226, Barretos (SP), Brazil. 2 Weed Science Environmental Research Studies of the College of Agricultural and Veterinary Science at Unesp, Via de Acesso Prof. Dr. Paulo Donato Castellane, Zona Rural, s/ n o , Jaboticabal, SP, Brazil. 3 Department of Morphology and Animal Physiology of the College of Agricultural and Veterinary Sciences at Unesp from Jaboticabal, SP, Brazil.


INTRODUCTION
Glyphosate (GFT) is the most used herbicide for weed control on several crops, due to its wide action spectrum, *Corresponding author.E-mail: cruzcl@yahoo.com.
The herbicide application on the aquatic environment needs a higher scientific base, due to the lack of knowledge about the molecules for non-target organisms, morph-physiological effects, and control effectiveness (Botelho et al., 2009) and residues on water.Another issue is the toxicity of the other formulation components, since the inert components, such as the surfactants, may be more toxic than the active ingredient (IA) for non-target organisms (Carraschi et al., 2011).For Amarante Jr et al. (2002), Tsui and Chui (2003), and Navarro and Martinez (2014), the surfactants from the glyphosate formulations are more toxic for fish than the molecule itself.
At this context, the acute and chronic toxicity assays may be performed to characterize the effects and to evaluate dangerousness and environmental risks of the formulations (Schmitt-Jansen et al., 2008).Thus, the organism selection is based on a series of criteria, such as sensibility, short reproduction cycle, operation easiness, and optimization costs (Cruz et al., 2008).
The neotropical species may be used as bioindicators for xenobiotics presence at the aquatic environment.In Brazil, the study about the application of this method for environmental monitoring is scarce (Glusczak et al., 2006;Shiogiri et al., 2012).
Thus, the aim of this study was to evaluate the sensitivity to a reference substance (potassium chloride) and the acute toxicity (LC 50 ;96 h) of the glyphosate and the Aterbane ® BR surfactant (alkylphenolpolyglycol ether) alone and in association with glyphosate (GFT + 0.5 and 1.0% surfactant) for the neotropical fish Piaractus mesopotamicus, Hyphessobrycon eques and Phallocerus caudimaculatus and the comparison with the standard species Brachydanio rerio and to evaluate possible histological effects caused by the glyphosate on gills, liver and kidney after acute exposure.

MATERIALS AND METHODS
The potassium chloride with 99.9% purity was used as reference substance, the Rodeo ® herbicide was the glyphosate source (480.0 g L -1 ) and Aterbane ® BR was used as the surfactant alkylphenolpolyglycol ether source (466.0 g L -1 ).

Sensitivity and acute toxicity tests
To perform the tests, the fish (P.mesopotamicus, H. eques, B. rerio and P. caudimaculatus) were acclimatized during seven days, under bioassay room conditions (25.0 ± 2°C temperature and 12 h photoperiod).
The fish were exposed to potassium chloride, glyphosate, surfactant and GFT + surfactant associations (Table 1), with three repetitions per treatment, three fish per repetition, at maximum density of 1.0 g L -1 , with 96 h duration (ABNT, 2011).
The substances concentrations on the definitive assays were previously adjusted by preliminary tests, to find the concentrations interval which caused zero and 100% mortality.

Histological analysis after glyphosate exposure
After 96 h exposure, three representative animals from each treatment and each species (n=3) were euthanized through anesthesia (1.0 g L -1 benzocaine).Liver, kidney, and gills were collected.The samples were immersed in buffered formaldehyde solution (0.1 m; pH 7.3) for 24 h.After the fixation, the pieces were dehydrated in alcohol, diaphanized on xylene and included in Histosec ® (Merck).Then, the microtomy was carried out in an automatic microtome (RM2155 -Leica ® ), performing cuts ranging from 3.0 to 5.0 m thickness.The pieces were placed in glass slides, which were colored with hematoxylin-eosin and PAS (Schiff's periodic acid) (Behmer et al., 1976).The histological alterations were transformed into effect index, and then classified according to each tissue, as described in the results and discussion.For the gills, the following indexes were evaluated: 0 -similar to the control -normal; 1 -Hypertrophy and hyperplasia of primary and secondary lamellae; 2 -Epitelial lifting; 3 -Lamelar tip fusion; 4 -secondary lamella disarrangement; 5secondary lamella congestion; 6 -decrease of the interlamellar epithelium; and 7 -sub epithelial edema on the secondary lamella.For the liver, the indexes were as follows: 0 -similar to the controlnormal; 1 -capillaries congestion; 2 -hepatocytes hypertrophy with nucleus displacement to the periphery; 3 -hepatocytes fusion; 4disarrangement of cordonal structure; 5 -picnotic nucleus; and 6necrosis.And the indexes for the liver were the following: 0 -similar to the control -normal; 1 -Bowman's capsule release; 2epithelium disarrangement from the proximal tubules; and 3 -light increase of the proximal and distal tubules.

Fish sensitivity
The most sensitive fish to the potassium chloride (KCl) was P. caudimaculatus and the most tolerant was H. eques; however, the sensitivity between species was similar (Table 2).Thus, the potassium chloride was classified as moderately toxic for the fish (1.0<CL 50 <1.0mg L -1 ), except for P. caudimaculatus, which was highly toxic (0.1<CL 50 <1.0mg L -1 ) (Zucker, 1985).
Based upon the sensitivity and in comparison with the standard fish B. rerio, the tested species may be standardized for ecotoxicological assays in Brazil, except for P. mesopotamicus, because it does not fill the requirements for a test organism.This fish has low availability, because it is a rheophilic species, with only one spawning per year.The other species have a higher availability, due to a short reproduction cycle, as shown by Cruz et al., (2008) on the sensitivity test for H. eques exposure to potassium dichromate.
According to USEPA (2002), the advantage on the use of potassium chloride over potassium dichromate is the presence of hexavalent chromium on the potassium dichromate composition, which is highly toxic for the environment and may cause severe morphological and functional effects on neotropical fish as P. mesopotamicus (Castro et al., 2014).

Acute toxicity for glyphosate, surfactant and the associations
The lethal concentration 50% (LC 50 ; 96 h) of glyphosate for the tested species was >975.0 mg L -1 (Table 3).The mortality ranged from 10 to 40%, regardless of the tested concentration.The lethal concentration was not estimated precisely, due to a lack of pattern regarding the ratio concentration/mortality.
The most sensitive species to the surfactant exposure was P. caudimaculatus, although the LC 50 ;96 h was similar for all tested species (Table 2).No mortality occurred for B. rerio under 7.0 mg L -1 exposure; however, 100% mortality occurred with 11.0 mg L -1 .For P. mesopotamicus, no mortality occurred with 8.0 mg L -1 , and 100% mortality occurred with 12.0 mg L -1 surfactant exposure.For H. eques, no mortality occurred with 4.0 mg L -1 and 100% mortality was achieved with 14.0 mg L -1 .
For P. caudimaculatus no mortality occurred at 3.0 mg L -1 and 100% mortality was obtained with 8.0 mg L -1 surfactant exposure.
The glyphosate toxicity observed in this study was similar to the values found for Salmo gairdneri and Oncorhynchus tshawytscha (LC 50 ;96 h = 1070.0 to 1440.0 mg L -1 ) (Mitchell al., 1987); Hybognathus amarus and P. promelas (LC 50 ;96 h > 1000.0 mg L -1 ) (Beyers, 1995) and P. promelas and Polonichthys macrolepidoyus (LC 50 ;96 h = 1154.0 to 1132.0 mg L -1 ) (Riley and Finlayson, 2004).The commercial formulations type (inert compounds and surfactant) has an important role for the glyphosate toxicity characterization.The lower toxicity of the formulation studied in this paper (Rodeo ® ) may be due to the absence of surfactants.Thus, the knowledge about the formulations toxicity is fundamental at the moment of decision making, concerning the monitoring of possible glyphosate environmental effects.
The surfactant presence in any concentration has increased the formulation toxicity in about 50 % for H. eques and P. mesopotamicus, indicating that the surfactant addition interfere on the glyphosate toxicity, regarding some non-target organisms.According to Giesy et al. (2000), the polyethoxylate amine surfactant (POEA) may be responsible for making the Roundup ® formulation more toxic for aquatic organisms.Riley and Finlayson (2004) observed that the LC 50 ;96 h for H. transpacificus was 5.5 and 3.9 mg L -1 for P. promelas, after the exposure to Rodeo ® formulation with the addition of R-11 ® surfactant, while the LC 50 of the commercial formulation without the surfactant was 270.0 and 1132.0 mg L -1 for the fish, respectively, showing the connection between the surfactant presence and glyphosate toxicity.
Based on the classification by Zucker (1985) and Giesy et al. (2000), the glyphosate and the associations with 0.5 and 1.0% surfactant (recommended doses for herbicide application) can be considered practically non-toxic and the alkyl phenolpolyglycol ether is classified as moderately nontoxic (1>LC 50 <10 mg L -1 ).

Gills
The pacu (P.mesopotamicus) gills consist of four branchial arches, supported by two rows of primary lamellae, which are covered with stratified epithelium.The secondary lamellae consist of epithelial, lining, pillar, chloride, and mucous cells (Figure 1A), as described by Severi et al., (2000).The observed histological changes are described below according to each concentration (Figure 1), on Table 4.
The epithelium increase and the blood congestion at the secondary lamellae also occurred on O. mykiss exposed to methiocarb insecticide (Altonok et al., 2006) and on Oreochromis niloticus exposed to diquat herbicide (Henares et al., 2008).The epithelium increase works as a barrier to decrease the xenobiotics absorption (Mallatt, 1985), similar as described for Cyprinus carpio exposed to GFT (Neskovic et al., 1996).
The presence of sub epithelial edema and the disarrangement of the secondary lamellae was reported for C. carpio (Neskovic et al., 1996) and for O. niloticus exposed to GFT (Jiraungkoorskul et al., 2002).The blood  congestion caused by the blood flow increase leads to the sub epithelial edema and the lamellae disarrangement.According to Shiogiri et al. (2012), P. mesopotamicus exposed under acute conditions to Roundup Ready ® also presented pillar cells hyperplasia and lamellar epithelium hypertrophy.Although, such alterations are reversible (Henares et al., 2008;Shiogiri et al., 2012) and after some days of exposure, the tissue can remake its common histomorphology, different from paraquat herbicide, which caused cells disarrangement, uncommon regeneration of the epithelial cells and deformation of the branchial cartilage cells from Colossoma macropomum, exposed to 10.0 mg L -1 (Salazar-Lugo et al., 2011).
H. eques was the most sensitive bioindicator to GFT, with several histological changes at the gills.Despite the changes not being specific to the GFT toxicity, this kind of biomarkers indicates the presence of xenobiotics on the  and 6  2 and 2  9 50  2, 3 and 4  4 and 6  2 and 3  975  2, 3 and 4  4 and 6  2 and 3 environment.This fact is important to environmental monitoring and identification of contaminants on the aquatic environment.P. caudimaculatus displayed branchial structure similar as observed for P. mesopotamicus and H. eques on the control and under 900.0 mg L -1 exposure.The sub epithelial edema on the secondary lamellae and the increase of the interlamellar epithelium was similar to a report with O. niloticus exposed to 36.0 mg L -1 glyphosate as the Roundup ® formulation (Jiraungkoorskul et al., 2002).

Liver
The evaluated fish liver (Control) showed cordonal organization of the hepatocytes in direct contact with the capillary sinusoids.The hepatocytes showed a hexagonal shape, with central nucleus, basophil, decondensed chromatin, a visible nucleolus and rosy cytoplasm, indicating high acidophilia (Figure 2A).
The main histological changes occurred after acute exposition were as follows: capillary sinusoids congestion (hyperemia), hypertrophy and fusion of the hepatocytes and cordonal structure disarrangement of the hepatocytes, with 900.0 and 925.0 mg L -1 exposure (Table 5), followed by tissue necrosis with 925.0 to 975.0 mg L -1 exposure for P. caudimaculatus (Table 5) (Figure 2C, D, E, F).
The cellular nucleus displacement to the periphery and the hepatocytes hypertrophy was also observed for O. niloticus exposed to GFT (Jiraungkoorskul et al., 2002).The displacement indicates the increase of the organelles and of the enzymes which are responsible for the xenobiotics metabolism, whereas the blood congestion may be a signal of the pressure decrease of the hepatic circulation, similar to O. mykiss exposed to 5.0 mg L -1 GFT (Topal et al., 2015).The cordonal disarrangement also occurred with Spaurus aurata exposed to imazapyr, terbutrin and triasulfuron (Arufe et al., 2004) and with Trichogaster trichopterus exposed to paraquat (Banaee et al., 2013).According to Shiogiri et al. (2012), the P. mesopotamicus liver showed hepatocytes vacuolization and cells hypertrophy when exposed to 3.0 or 4.0 mg L -1 glyphosate as Roundup Ready ® .The vacuolization may work as storage of compounds which are difficult to be metabolized.
The liver, used as GFT exposure biomarker, was the tissue which exhibited the highest histological index, reaching tissue necrosis during P. caudimaculatus exposure to Rodeo ® .This alteration was also described by Ayoola (2008) on C. gariepinus, same for the fibrosis which occurred on O. mykiss exposed in a chronic manner to 5.0 or 10.0 mg L -1 GFT (Topal et al., 2015).The necrosis probably occurred due to an excessive work performed by the hepatocytes, in attempt to eliminate the herbicide during the detoxification process.

Kidneys
The fish kidneys are formed by a renal corpuscle containing the glomerulus constituted by capillaries and the Bowman's capsule.Around the renal corpuscle are also found the hematopoietic tissue with basophilic cells, proximal and distal tubules, as observed on the control for all studied species on this research (Figure 3A).
The alterations were the shrinkage of the Bowman's capsule and disarrangement of the proximal tubules epithelium in all assessed concentration for P. caudimaculatus and H. eques, whereas did not occurred alterations of the tissue structure on P. mesopotamicus in any tested concentration (Table 6 and Figure 3B).
The glomerulus capsule release, the disarrangement of the proximal tubules epithelium and the proximal and distal tubule light increase was also observed on O. mykiss exposed to linuron herbicide (Oulmi et al., 1995) and on O. niloticus exposed to Roundup ® (Jiraungkoorskul et al., 2002).The Bowman's capsule space increase, or the capsule release, and the disarrangement of the tubules cells was also observed on C. gariepinus exposed to GFT, which leads to the kidneys physiological functions loss (Ayoola, 2008).

Conclusions
The fish species H. eques, P. caudimaculatus and P. mesopotamicus present great potential to be used as standard organisms for ecotoxicological assays for herbicides monitoring, for they present an excellent answer to the tested reference substance.The addition of

Table 1 .
Concentrations of the substances used on the toxicity assays.

Table 4 .
Histological alterations on fish gills after glyphosate exposure.

Table 5 .
Histological alterations on the fish liver after glyphosate exposure.

Table 6 .
Histological alterations on fish kidneys after glyphosate exposure.