Glyphosate commercial formulation effects on preoptic area and hypothalamus of Cardinal Neon Paracheirodon axelrodi (Characiformes: Characidae)

In Colombia the use of glyphosate commercial formulations (Roundup™) for spraying have left deleterious effects on animals and humans. Much of this spraying takes place at the Orinoco basin, habitat of one of the most exported ornamental fish in Colombia, Cardinal neon. To evaluate the effect of Roundup Activo™ four experimental treatments were carried out with 0 mg/L (T1), 0.1 mg/L (T2), 1 mg/L (T3) and 5 mg/L (T4) during 30 days of exposure. The fishes were processed for high-resolution optical microscopy. The main finding of Roundup Activo™ exposure was an increase in mast cells number in brain blood vessels and some neuronal nuclei of the preoptic and posterior diencephalic areas, including hypothalamus. A correlation between concentrations and mast cells number was observed, with the largest mast cells number in T4 treatment. Mast cells presence is a stress benchmark, suggesting the beginning of allergic, inflammatory and apoptotic events. Presence of mast cells in these brain areas may lead to alterations on reproduction, visual and olfactory information integration among other processes. These alterations may result in diminished survival, affecting the conservation of this species in its natural habitat.


Introduction
Glyphosate is an acidic compound used as a not specific herbicide. Its IUPAC name is N-(phosphonomethyl) glicine, and is commonly used as an isopropilamine salt with a surfactant. Polioxiethilen-amine (POEA) and Cosmoflux 411F are the most common surfactants included in commercial formulations (Ramírez, Rondón, 2003). Glyphosate remanence time in water depends on sediment chemistry, staying up to 14 days after spraying, at a concentration of 0.6 mg/kg in sediment, and up to 400 days after aspersion in lentic water bodies. In general Roundup™ may stay in water one to three weeks (WHO, 1994). This herbicide in used in legal agriculture in order to control weeds, grain drying and spraying of transgenic crops. However, in Colombia, glyphosate has been also widely used for illegal crop spraying (Ramírez et al., 2008). In September 2015, Colombian government banned glyphosate for illegal crops spraying, but a 2014 United Nations Office on Drugs and Crime (UNODC) shows that illegal crop areas had increased on about 40% in Colombian Meta and Guaviare departments, Cardinal neon Paracheirodon axelrodi (Schultz, 1956) 2 e190025 [2] habitat; opening a discussion on glyphosate interdiction, pointing out to use continuous manual spraying, as a new illegal crops control alternative.
Studies on fish exposure to glyphosate and glyphosate commercial formulations have been made on sub lethal doses, imitating what is most probably found at natural ecosystems (Glusczak et al., 2007;Kreutz et al., 2011). According to Cattaneo et al. (2011), acetylcholinesterase (AChE) activity in Cyprinus carpio Linnaeus, 1758 brain is progressively reduced after a 96h glyphosate treatment with 0.5 to 10 mg/L concentration. Sandrini et al. (2013) have shown a diminished (50%) AChE activity associated to 1.165 mg/L and 1.52 mg/L glyphosate exposure for Danio rerio (Hamilton, 1822) and 1.77 mg/L and 2.47 mg/L for Jenynsia lineata (Jenyns, 1842). Sinhorin et al. (2014) showed that siluriform fish Pseudoplatystoma sp. suffers from liver, brain and muscle oxidative stress after commercial glyphosate exposure at 2.25, 4.5, 7.5 and 15 mg/L concentrations. Hued et al. (2012) have reported that 5, 10, 20, 35, 60 and 100 mg/L Roundup Max™ concentrations reduce clotting time, female courtship and coupling success in J. lineata exposed for 28 days. Mesnagea (2012) found that surfactant to be more toxic to fish that glyphosate, POEA exposed may lead to loss of cell membranes in human germinal, liver and placental cells at 1 to 3 mg/L concentration. Roundup™ levels up to 200 mg/L are reported to produce damage in fishes, while half surfactant concentrations are able to lead to erratic swimming, and behavioral alterations in Oncorhynchus mykiss (Walbaum, 1792) and Oreochromis spp. (Cox, 2000).
For the above, in this article we want to identify the effect of a commercial presentation of glyphosate on some areas of the encephalon in order to generate data that allow us to support protection plans for conservation of this species and additionally inform about the consequences of the excessive use of glyphosate.

Material and Methods
Experimental design. Fishes from 2 to 2.8 cm total length were kept in 40 L aquaria for 30 days with low light incidence. Four different treatments were used, T1: 0 mg/L, T2: 0.1 mg/L, T3: 1.0 mg/L, and T4: 5.0 mg/L, with four replicas/ treatment, thirty fishes/aquarium. These concentrations were adjusted based on the active ingredient of the commercial product. Fishes were kept in semi static systems with pH of 6.0 to 6.5, and temperature of 25 to 26°C. Animals were feed with Tetracolor® 47.5% crude protein, at 6% of total biomass, three times/day at 8:00, 12:00 and 16:00 hours (Anjos, Anjos, 2006). The individuals were obtained in stores specializing in ornamental aquatic species. The species is identified in the Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Colleción de Peces, Villa de Leyva, Colombia with reference code IAvH-P 10312.
Tissue sampling and processing. Fishes were euthanized by benzocaíne (0.5 g/L) treatment, followed by spinal cord section at cervical level, according to ethical fish management procedures (AVMA, 2013). After a 30 days exposure, six animals per aquarium were euthanized. Three of them were used for high-resolution optical microscopy, the other three were kept for future ultrastructural studies. Brains were dissected and fixed in a modified Karnovsky solution (2% formaldehyde and 2.5% glutaraldehyde) in phosphate buffer pH 7.2 for two hours. Tissues were rinsed three times with phosphate buffer, followed by postfixation with 2% osmium tetroxide in phosphate buffer. Brain were rinsed again with buffer and dehydrated with ethanol 70%, 90%, 95% and 100%. After dehydration, tissues were embedded in Poly/ Bed®812-propylen oxide 1:2 and 1:1, with a final 100% Poli/Bed®812. Samples were polymerized at 60°C (Obando et al., 2013;Rincón et al., 2017). Sections were obtained with Microm Slee Cut 40650 rotatory microtome and stained with toluidine blue (Obando et al., 2013;Rincón et al., 2017). Photographs were examined in the microscope Zeiss Axio Scope A1 coupled to a Zeiss Axiocam ICM 1 camera. Diagrams of cuts were made using Adobe Photoshop CC 2015 software. A description of mastocyte distribution and number per field was made using these digital images. The preoptic area (± 72 µm) and the posterior diencephalic region (± 516 µm) were analyzed (Obando et al., 2013). The cell count MC was performed by direct counting supported by the analysis of images made in the ImageJ software (https:// imagej.nih.gov/ij/) with a space margin of 12 µm between each cut analyzed to avoid repeating any cell previously counted. In total, ± 588 µm were analyzed per replica.

Statistics analysis.
A Shapiro-Wilk test was used to evaluate normality. ANOVA tests were used for detection of significant differences among treatments at a p<0.05 level with R 3.2.3 software.

Results
Preoptic Area (AP). The cuts were made in two areas of the brain (preoptic area and posterior diencephalic region) ( Fig.  1). MC cells were observed at T2, T3 and T4, associate with blood vessels adjacent to brain tissue (Figs. 2-3). However, some MC cells were found within brain parenchyma in 3 e190025[3] some neuronal nuclei, like mid area of dorsal telencephalon (MD), lateral area of dorsal telencephalon (LD) and anterior parvocellular preoptic nucleus (PPa).  Posterior diencephalic area and hypothalamus. MC cells distribution was similar for T2, T3 and T4 treatments (Fig. 3). These cells were observed at Optic Tectum (TeO), Torus semicircularis (TS), Grey periventricular area of optic tectum (PGZ), Torus lateralis (TLa), diencephalic ventricle (DiV) and ventral area periventricular hypothalamus (Hv). The MC cells were observed in blood vessels adjacent to PDR. However some MC cells were also found within cerebral parenchyma. No signs of MC degranulation were observed. Figure 4 shows significant statistical differences in the number of MC cells for the different treatments. These results show that the number of MC cells is directly proportional to the increase in the concentration of Active Roundup™ in the treatments. For the treatments with 1 and 5 mg/L of Roundup Activo™ MC cells were found in some neuronal nuclei of both the preoptic area and the posterior diencephalic region, being more abundant in mid area of dorsal telencephalum (MD), anterior parvocellular preoptic nucleus (PPa), periventricular gray area of optic tectum (PGZ), optic tectum (TeO) and Torus lateralis (TLa) (Tab. 1). The MCs cells were also found in the blood vessels adjacent to the brain, treatments with 1 and 5 mg/L of Roundup Activo™ were the ones with the highest number of these cells (Tab. 2).  No morphological alteration of AP and HT was observed after Roundup Activo™ treatments. AP, PDR and HT staining and distribution are similar to other teleost species (Rincón et al., 2017).
At posterior diencephalic area, neurons form ventral hypothalamic (Hv) area, located around diencephalic ventricle (Div) do show round and strongly stained nuclei, with scarce cytoplasm. Optic Tectum (TeO) neurons have weakly stained nuclei, while at grey periventricular area from optic tectum (PGZ) neurons around Torus Semicircularis (TS) have with scarce cytoplasm, but strongly stained AP and HT do have a lot of parenchymal and peripheral blood vessels, the latter being of larger size (Rincón et al., 2017).
Histopathological alterations. Appearance of MC cells after T1, T2 and T3 treatments was the main effect observed on preoptic and posterior diencephalic areas, including hypothalamus. As MC cell granules contain heparin sulfate, these cells can be visualized with a metachromatic stain like toluidine blue, leading to a deep purple staining (Reite, Evensen, 2006;Nelissen et al., 2013). The MC cells or granular eosinophilic cells (EGCs) are homologous to mammal mast cells, where they play a role in innate and adaptative immune response (Nelissen et al., 2013), promoting an inflammatory reaction triggered by an allergenic or toxic agent, stress or trauma (Reite, Evensen, 2006), mediated by the release of pro-inflammatory and vasoactive substances like histamine, serotonin, cytokinins and proteolytic enzymes (Dezfuli et al., 2000;Theoharides, Cochran, 2004). These cells also present substances like nerve growth factor (NGF), acid and alkaline phosphatases, antimicrobials (piscidins), tumor necrosis factor (TNF) among others (Theoharides, Cochran, 2004;Wilhelm et al., 2005;Mulero et al., 2007).

e190025[5]
Based on MC cells role in immunity in teleost fishes, our results may indicate that the exposure to Roundup Activo™ is related to an increase in MC cells due to a stress reaction, leading to the activation of P. axelrodi immune system. This interpretation is supported by the observation of a MC cells infiltration in peripheral brain blood vessels and in some neuronal nuclei like AP and PDR (Figs. 2-3), specifically at mid area of dorsal telencephalon (MD), anterior preoptic parvocelular nucleus (PPa) and lateral area of dorsal telencephalon (LD).
Statistical analysis showed significant differences (P<0.05) between control (T1) and T3 and T4 for MD, while at PPa, the larger number of MC cells was observed after T3 treatment. However, no significant differences were detected at LD between control (T1) and exposed animals (Tab. 1). Lack of effect after exposure was also observed at TeO, but at PGZ, with a 5 to 8 range of MC cells, T4 treatment exhibited a larger number of cells. At TS and TLa, even with a low number of MC cells, significant differences were found between control and T3 and T4. For all other neuronal nuclei no statistical differences were observed (Tab. 2).
The presence and number of MC cells is directly related to glyphosate concentration in Roundup Activo™, for example the MC cells number in brain parenchyma was higher in T4 than T1 (17±1 and 0.66±0.57). Finally, no significant differences were detected between T2 and T3 treatments (Fig. 4). These results could be explained as a Roundup Activo™ triggered activation of immune system in fishes like P. axelrodi in spite of the low concentrations used. Our results also increase the possibility of even greater effects, such as increased number of mast cells, degranulation, decreased AChE activity, increased vascular permeability, acute inflammation and cell death (Reite, Evensen, 2006;Meshkini et al., 2018). These brains did show high MC cells numbers after exposure to 7.5 mg/L up to 120 mg/L glyphosate in Roundup Ultra™, where even degranulation was observed (Eslava et al., 2007).
MC cells are produced from stem cells at hematopoietic organs, and do migrate via blood vessels to affected areas, where they do differentiate (Reite, Evensen, 2006;Sfacteria et al., 2015). It seems to be the case in our observations as the larger number of MC cells were detected nearby brain peripheral blood vessels. Arrival of larger number of MC cells to brain depends on Roundup Activo™ concentration, as more cells are observed after T4 exposure (Tab. 2), supporting the idea that Roundup Activo™ exposure may cause tissue damage and/or immune system activation in P. axelrodi brain.
Impact on physiological functions. The possibility that Roundup Activo™ could trigger an immune response, may lead to eventual brain physiological dysfunction. Reports on glyphosate effects in mammals range from alteration on mitochondrial dynamics, to increase in neuronal apoptosis, aromatase diminished activity, suppression of cytochrome P450, hippocampus oxidative stress, carcinogenesis and increase in blood brain barrier permeability. All these reports are useful in order to understand the effect of pure and commercial formulations of glyphosate, not only in mammals but in teleost fishes, due to the great similarity in many cellular, cerebral and metabolic processes (George et al., 2010;Gui et al., 2012;Mesnagea et al., 2012;Samsel, Seneff, 2013;Cattani et al., 2014).
The lateral (LD) and medial (MD) areas of C. auratus dorsal telencephalon are involved in recognition, learning and spatial orientation (Saito, Watanabe, 2004). These nuclei and Torus semicircularis do receive lateral line and preglomerular lateral nucleus sensitive and mechanoreceptor projections (Striedter, 1991). Dorsal telencephalon is innervated by olfactory projections from olfactory bulbs (Nieuwenhuys, Meek, 1990). Finally, dorsal telecephalon is reported to play a role in some mnemonic processes (Portavella et al., 2004).
Anterior preoptic parvocelular nucleus (PPa) and PDR ventral area periventricular hypothalamus (Hv) do play a neuroendocrine role, secreting isotocin an homologue of mammal oxytocin, involved in reproduction and behavior regulation. However, isotocin function is not clear yet (Gozdowska et al., 2006;Bernier et al., 2009). Arginin-Vasotocine, Cholecystokinin, Corticotrophin Releasing Factor, Galanin, Gastrin Releasing Peptide, Gonadotrophin Releasing Hormone, Growth Hormone Releasing Hormone, Neuropeptide Y, Somatostatin and Tyrotrophin Releasing Hormone are also released. These molecules are important in neuroendocrine and metabolism regulation, control of cardiovascular rhythm and blood pressure, release of gonadotrophin, corticotrophin, growth hormone and gastric acid, regulation adenohypophysary secretion, inhibition of growth hormone, gonadal maturation, activation of reproductive dynamics and osmoregulation (Bernier et al., 2009;Cerdá-Reverter, Canosa, 2009;Martins et al., 2014). An increased presence of MC cells in some neuronal nuclei, may trigger apoptosis events, by means of microgia activation as reported by Zhang et al. (2016), leading to diminished functions related to reproduction, adaptation and survival of P. axelrodi within the ecosystem.
MC cells were also found in PDR optic tectum (TeO) and periventricular gray area (PGZ) peripheral blood vessels. These nuclei de play a role in food recognition, optomotor response, light intensity response, movement detection, color vision and integration of visual information (Marachlian et al., 2018). However, TeO also serves as an integration center for sensitive and mechanosensitive information coming from Torus semicircularis (TS) (Striedter, 1991). Feedback from TeO is important for Isthmi nucleus role in food recognition and predator avoidance. In addition, TeO do receives input from neuropeptide and GnRH neurons involved in reproductive stimuli, visual response to courtship and reproduction sites (Robles et al., 2011;Cooper et al., 2015;Umatani et al., 2015).
Torus lateralis (TLa) is considered a neuronal nucleus exclusive for Actinopterygian fishes (Butler, Hodos, 2005). It has been reported that Oncorhynchus mykiss TLa is involved in processing taste information (Folgueira et al., 2005). Paracheirodon axelrodi dwells on dark waters where any loss of visual or taste information may reduce feeding, depredator avoidance, shelter finding, and reproductive success.
The increase in the number of MC cells by exposure Roundup Active™ could also affect the physiological functions of other brain nuclei because the neuronal nuclei of the CNS have reciprocal connections with other nuclei. For example, it is known that AP is connected with nuclei as the periaqueductal gray area of the midbrain, the postrema of the caudal medulla, and the central gray area of the spinal cord (Nieuwenhuys et al., 1988). Paracheirodon axelrodi may be also used as an environmental biomarker of industrial and agriculture pollutants effect on aquatic ecosystems, as this species inhabit basins already used for extensive rice and palm oil plantations aspersed with Roundup Activo™. This fish may be used as a public health biomarker (Marchand et al., 2009;Barišića et al., 2015;Javed et al., 2015), as high glyphosate levels have been reported in waters of the Meta river basin (González et al., 2014), used for human consumption, which is a possible source of health problems.
In our study the use of Round Activo™ caused an increase in the MC cells number in the preoptic and posterior region area of P. axelrodi brain, the increments may affect directly CNS, causing deleterious performance in the species. In addition, this effect is dependent of the concentration, suggesting that glyphosate levels reported in nature may lead to central nervous system alterations, affecting P. axelrodi in its environment.