Introduction

Emerging environmental contaminants, such as pharmaceutical drugs and personal care products, often reach aquatic ecosystems (Archer et al. 2017), thereby deleteriously affecting aquatic biota (Archer et al. 2017). Pharmaceutical drugs constitute a diverse class of compounds with varied molecular structures, as well as different action mechanisms and metabolism in organisms (Fawell and Ong 2012). There are several sources of drugs entering the environment, such as human excretion through domestic effluents, animal excretion (poultry, livestock, aquaculture, veterinary use), leaching of sanitary landfills, and industrial effluents (Ternes 1998; Archer et al. 2017). Drugs are considered pseudo-persistent compounds considering that the amount that is naturally degraded equals the one that is discarded continuously into the environment (Oliveira et al. 2015). Anti-inflammatories, analgesics, antihypertensives, antidepressants, anxiolytics, antibiotics, and antiepileptics stand out among the most frequently detected pharmaceutical classes in the aquatic environment (Campanha et al. 2014; Ebele et al. 2017).

Pharmaceutical compounds are developed in order to possess high biological activity and, even at concentrations as low as ng L−1, are capable of exerting effects on aquatic biota exposed via the environment (Fabbri and Franzellitti 2016; Pereira et al. 2018; Matus et al. 2018). Although some drugs have relatively long half-lives, many of them are resistant to degradation and biotransformation (Daughton and Ternes 1999; Correia 2014). Drugs also have a varying degree of lipophilicity, which facilitates their absorption by plasma membrane of cells (Correia 2014). Once inside the cell, drugs act on metabolic, enzymatic, and cellular pathways, which are generally evolutionarily conserved (Fabbri and Franzellitti 2016). Therefore, it is necessary to better understand how these drugs act on non-target organisms and thus assess their impact on aquatic environments.

Acetaminophen is an analgesic and antipyretic drug, whose environmental presence has already been reported in effluents, and also in surface waters worldwide (Kolpin et al. 2002; Américo et al. 2012; Campanha et al. 2014). Studies indicate that in therapeutic doses, acetaminophen is considered safe; however, in overdosage, this drug can cause hepatotoxic effects (Xu et al. 2008), namely by causing oxidative stress (Ramos et al. 2014; Nunes et al. 2015).

Among the antihypertensive agents reported to occur in the aquatic environment, the most frequently detected are β-blockers, such as propranolol (Fabbri and Franzellitti 2016), a highly prescribed β-blocker used in the treatment of human hypertension, by blocking adrenoreceptors, thus reducing the frequency of heart beats (Owen et al. 2009). Propranolol was found in effluents and in surface waters worldwide (Andreozzi et al. 2003; Robert and Thomas 2006; Campanha et al. 2014; Thomas et al. 2014). Studies demonstrated that this pharmaceutical is capable of exerting effects on aquatic organisms, such as oxidative stress and neurotoxicity (Solé et al. 2009; Sun et al. 2015; Matus et al. 2018; Pereira et al. 2018).

Potential ecotoxicological effects of drugs on aquatic organisms may be assessed by evaluating sub-individual parameters (biomarkers), to detect physiological responses triggered by the presence of these chemicals in the environment (Nunes et al. 2004, 2008; Rodrigues et al. 2014; Oliveira et al. 2015). Among cellular biomarkers that can identify responses at the sub-individual level, heat shock proteins (HSPs) are considered valuable tools for the evaluation of the effects of chemical stressors on organisms. As proteins from the HSP70 family play the primary role of cell chaperones (Ashburner 1982; Garrido et al. 2001), the increase in their amount leads to a cytoprotective effect through the repair of altered proteins (Demeke and Tassew 2016), and consequently, enhanced amount of HSP70 can potentially prevent cell death. Thus, the immunolabeling of HSP70 coupled with an assessment of DNA fragmentation, which is a cell death marker, can provide an indication of the level of both cell stress and resulting cytotoxicity (Malaspina and Silva-Zacarin 2006).

In vertebrates, biotransformation processes of drugs occur mainly in the liver (Schinoni 2006; Correia 2014), potentially generating oxidative stress; formed free radicals (namely reactive oxygen species, ROS), may exert proteotoxic effects, through their reactivity with cellular proteins (Niforou et al. 2014). Consequently, this tissue is the most adequate for the evaluation of HSP70 levels in response to exposure to both acetaminophen and propranolol, coupled with an assessment of DNA fragmentation which may be indicative of cell death (TUNEL method). The biotransformation of both drugs, acetaminophen and propranolol, is likely to cause oxidative stress in aquatic organisms (Owen et al. 2007; Franzellitti et al. 2011; Ramos et al. 2014; Nunes et al. 2015), by increasing the levels of ROS. Consequently, the assessment of HSP70 levels may work as a proteotoxicity biomarker (Bierkens 2000), being important to be evaluated in non-target organisms exposed to pharmaceuticals. Increased levels of HSP70 after exposure to acetaminophen is a frequent response in vertebrates (Sumioka et al. 2004; Knockaert et al. 2011; Guiloski et al. 2017). Similarly, it has also been reported in aquatic animals exposed to propranolol (Sun et al. 2015) and other β-blocking drugs (Contardo-Jara et al. 2010).

Despite studies evaluating the levels of the HSP family in aquatic organisms (Iwama et al. 1998; Demeke and Tassew 2016), little is known about the effects of these two pharmaceuticals on proteotoxicity in fish. The response of organisms to chemicals can follow species-specific patterns (Owen et al. 2007; Nunes et al. 2014). Despite their importance to understand the establishment of sublethal effects of stressors in fish (Demeke and Tassew 2016), there are few studies focusing on hepatic HSP70 levels measured in neotropical native fish following exposure to these pharmaceutical drugs.

The species Phalloceros harpagos (Lucinda 2008), common name guaru, was selected as a test organism in the present study, as it is a representative of the autochthonous teleost fishes of South America, with a wide distribution in Brazil (Lucinda 2008; Mendonça et al. 2014; Yoshida and Uieda 2014).

The aim of this study was to evaluate the acute and chronic effects of the drugs acetaminophen and propranolol on the neotropical fish P. harpagos, based on the levels of the cell stress (HSP70) and cell death (DNA fragmentation) biomarkers, measured in liver, by means of immunofluorescence. With the present study, we intend to highlight the importance of using a conventional methodology for the evaluation of cellular biomarkers in a native species of tropical fish, since this new approach brings to the understanding of the potential toxicological effects of drugs in exposed fish that may affect their physiological traits and, consequently, their population in tropical aquatic ecosystems.

Materials and methods

Chemicals

The chemicals acetaminophen (CAS 103-90-2, purity 98–102%) and propranolol (CAS number 318-98-9, purity ≥ 99%), DNase I (RNase-free), and the monoclonal anti-heat shock protein 70 antibody produced in mouse were purchased from Sigma-Aldrich (Germany). IgG (H+L) Rabbit anti-Mouse, Alexa Fluor® 594, Superclonal™ secondary antibody (A27027), Fisher Scientific, Thermo Fisher Scientific (Finland). Dako Faramount Aqueous Mounting Medium (Agilent Technologies – USA).

Sample collection

P. harpagos was chosen for being a representative of the neotropical fauna. In addition, it also presents features that favor it as a test organism, such as its easy capture and handling. This species is also abundant in the environment and is well adapted to laboratory conditions, including to the exposure process and its maintenance in aquarias (Pereira et al. 2018; Matus et al. 2018). Belonging to the genus Phalloceros, this fish species has sexual dimorphism, females being larger than males. In addition, P. harpagos has small body size and is ovoviviparous with an incubation duration of 3 months (Machado et al. 2001). This species is omnivorous and feeds preferentially on aquatic and terrestrial insects, vegetables, algae and sediment (Lucinda 2008).

The collection of fish (P. harpagos) was carried out in a small river in the city of Cabreúva, in the Serra do Japi region, State of São Paulo, Brazil (23° 18′ 23.8″ S 47° 05′ 50.4″ W). This is an area of low environmental impact, with only few private properties, with agricultural and livestock activities (Yoshida and Uieda 2014; Prado and Novo 2015). This area is characterized by being an important fragment of semi-deciduous forest (Joly 1992), and is part of the Environmental Protection Areas (APA), created by Laws No. 4095/84 and No. 4023/84, regulated by the decree No. 43.284/98.

Specimens of P. harpagos were captured using a hand net, with a mesh of 0.3 and 0.5 mm. Animals were collected in April, May, October, and November 2016, and 150 individuals were collected in each of the first two sampling periods (April and May), and 100 in each of the last two sampling periods (October and November) in a total number of 500 animals. The collected fish were males (identified by the presence of gonopodia) and sexually immature females (size smaller than 2.5 cm, and identified by the absence of gonopodia). At the time of collection, individuals with an average size of 2 ± 0.5 cm were selected, with a mean weight of 100 ± 50 mg. Smaller or larger individuals, as well as other species, were immediately returned to the stream. The fish were then placed in transport boxes with constant aeration.

Experimental design

Upon arrival at the laboratory facilities, the fish were acclimated for 15 days (quarantine, depuration, and acclimation period) in glass aquaria (3 aquaria with 43 L of capacity, and 75 fish per aquarium, 50 × 25 × 35 cm), with Klarina® mineral water (for human consumption), a controlled temperature of 23 °C (± 1 °C), photoperiod of 12-h light/12-h dark, continuous aeration, being fed daily, ad libitum, with flocculated ration for tropical fish Tetra Min®. During this period, two inspections were carried out per day and dead animals were immediately removed from the aquarium.

For each test (acute and chronic), fish were individually placed in plastic containers with 250 ml of water. Fish were divided into 6 experimental groups, and each of the 6 experimental groups was composed of 15 fish (90 fish in total per test, as described by Nunes et al. 2008). Each exposure vessel contained one fish (individually exposed), and was constantly aerated. The photoperiod was set at 12-h light/12-h dark, at a temperature of 23 °C (± 1 °C) for all acute and chronic trials of both pharmaceutical compounds. The plastic containers used for both exposures corresponded to polyethylene bottles, 2 L capacity, previously used for human water consumption. They were duly washed with distilled water and air-dried.

Acute exposure

The experimental design for the acute exposure was based on the standards issued by the Organization for Economic Cooperation and Development (OECD) guideline 203 (OCDE 1992). The experiment lasted 96 h with medium changes and drug replacement every 48 h and daily visual inspections. Fish were not fed during the exposure period. Six groups were defined: the control group (Klarina® water only) and the groups exposed to five drug concentrations (acetaminophen 8, 80, 800, 8000, 80,000 μg L−1; propranolol 0.1, 1, 10, 100, 1000 μg L−1). These concentrations were defined according to two criteria: (1) the highest concentration of each drug should be around the LC50/10, i.e., below to lethal concentrations (LC50 values of both pharmaceutical drugs described for the fish species Oryzias latipes, which was the most phylogenetically similar species to P. harpagos, given the lack of LC50 values for this species in the literature; according to Yamamoto et al. (2007) the LC50 value was of 800,000 μg L−1 for acetaminophen; according to Kim et al. (2009), the LC50 was of 11,400 μg L−1 for propranolol); (2) the lowest concentrations were chosen according to concentrations already reported in the environment, for acetaminophen (6 μg L−1, Ternes 1998; 10 μg L−1, Kolpin et al. 2002; 5.529–6.957 μg L−1, Robert and Thomas 2006) and for propranolol (0.1–0.09 μg L−1, Andreozzi et al. 2003; 0.026–1.900 μg L−1 Huggett et al. 2003; 0.016–0.135 μg L−1, Zhou et al. 2009). To attain the defined test concentrations, it was necessary to prepare two stock solutions, of both acetaminophen and propranolol. The concentrations for the tests were obtained by diluting the respective pharmaceuticals on Klarina® mineral water to prepare the stock solutions at the following concentrations: paracetamol 400 mg L−1 (400,000 μg L−1) and for propranolol 10 mg L−1 (10,000 μg L−1). Because these were semi-static regimes of exposure, and also in order to maintain constant drug concentrations throughout the exposure, the exposure media were fully replaced every 48 h and the distinct volumes of the drug solutions were added at this time.

Fifteen fish per experimental group were placed in individual plastic containers, with a total volume of 250 mL of media each. At the end of the exposure period, five fish from each treatment were randomly collected and assigned to the analysis of the cellular biomarker (HSP70).

Chronic exposure

Based on the OECD standard guideline 215 (OCDE 2000), the chronic test lasted for 28 days, and the medium was changed every 48 h. Fish were fed with flocculated ration for tropical fish (Tetra Min®) every 2 days, ad libitum. Six experimental groups with 15 fish per group were adopted. Fish were individually kept in plastic containers with a volume of 250 mL per vessel: a control group (exposed to Klarina® mineral water, for human consumption) and 5 groups exposed to the different concentrations of both drugs, selected on the basis of ecological relevance criteria. The here-adopted concentrations were close (same magnitude) to the concentrations already reported to occur in the environment. For acetaminophen, the tested concentrations were as follows: 5, 10, 20, 40, and 80 μg L−1 (acetaminophen has been shown to occur in levels of 6 μg L−1, Ternes 1998; 10 μg L−1, Kolpin et al. 2002; 5.529–6.957 μg L−1, Robert and Thomas 2006). For propranolol exposure, the concentrations were as follows: 0.0625, 0.125, 0.25, 0.5, and 1 μg L−1 (propranolol has been already found in the wild in levels from 0.1 to 0.09 μg L−1, Andreozzi et al. 2003; 1.900 μg L−1, Huggett et al. 2003; 0.016–0.135 μg L−1, Zhou et al. 2009). Similarly, to what was done for the acute exposure, two stock solutions were prepared for the chronic trials, and both had concentrations of 10,000 μg L−1 for the two drugs. Similarly, to what was done for the acute exposure, stock solutions were prepared with Klarina® water, with the following concentrations: paracetamol 400 mg L−1 (400,000 μg L−1) and for propranolol 10 mg L−1 (10,000 μg L−1). The exposure media were fully replaced every 48 h and different volumes of the pharmaceutical’s solutions were added at this time.

At the end of the exposure periods, five fish of each exposed group were randomly assigned for the analysis of cellular biomarker (HSP70).

Euthanasia and tissue processing

After the acute and chronic exposure periods, the previously exposed fish were submitted to anesthesia by hypothermia (2 to 3 min) and euthanasia by decapitation. Livers used for the detection of HSP70 (stress cell biomarker) were dissected under stereomicroscope at room temperature and immersed in a fixative solution (4% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.4). Livers were fixed during 24 h, at 4 °C.

Immunofluorescence analysis

Fixed livers (N = 5 per experimental group) were processed for inclusion in paraffin according to the methodology described by Silva-Zacarin et al. (2012); i.e., after 24 h immersed in fixative solution (4% paraformaldehyde in 0.1 M sodium phosphate buffer pH 7.4), these organs were washed in sodium phosphate buffer (0.1 M, pH 7.4) and submitted to low-temperature slow dehydration in solutions containing increasing concentrations of ethanol. The dehydrated material was diaphanized in xylene and embedded in liquid paraffin for further inclusion in ultra-pure paraffin (Paraplast). The paraffin-embedded material was submitted to microtomy. Histological sections of 6 μm, obtained with a microtome (LEICA RM2255), were extended on salinized glass slides (Star Frost). After drying, the slides containing the histological sections at different depths of the organ were used for HSP70 immunostaining and DNA fragmentation immunodetection, which is indicative of cell death. Four slides were prepared per individual from each experimental group, with 5 individuals per group.

Immunolabeling of HSP70

For HSP70 analysis, the slides containing the histological sections were routinely deparaffinized. Then, the rehydrated histological sections were subjected to antigen retrieval using 1 M sodium citrate buffer, pH 6.0, for 40 min at 90 °C. After antigen recovery, the slides were maintained in phosphate buffer saline (PBS) in a humid chamber. Histological sections were permeabilized with 0.5% Triton X-100 for 10 min and thereafter non-specific sites were inhibited with 3% bovine albumin solution for 30 min. After being washed 3 times with PBS buffer, the histological sections were incubated for 1 h with the monoclonal primary antibody (anti-HSP70, Sigma, produced in mouse), with dilution of 1:100, in the absence of light, at 37 °C, and in a humid chamber. After incubation with the primary antibody, the histological sections were washed 3 times in PBS and were subsequently incubated with the rabbit-raised Alexa 594-conjugated secondary antibody, at the dilution of 1:300, for 1 h at room temperature, in the absence of light, in a humid chamber. Finally, the histological sections were washed 3 times with PBS, in the absence of light in the humid chamber, and the slides were assembled in specific aqueous medium and held at 4 °C until the analysis. The negative control of the reaction consisted of the omission of the primary antibody in the incubation phase, which was replaced by PBS.

Immunofluorescence analysis for the detection of HSP70 in P. harpagos liver histological sections was performed using confocal laser scanning microscopy (LEICA® TCS-SP8). The analysis was performed using the laser OPSL 552 nm (excitation 580 nm, emission 650 nm; LEICA® TCS-SP8) at 3% intensity, with 1.00 AU of pinhole opening (laser opening), gain 800 (brightness), and offset (− 10%). The analysis was performed with the × 40 objective and × 3 optical zoom. The three-dimensional reconstruction was done by scanning 5 μm of depth (z-axis scanning) for each liver section, totaling 11 optical sections (frames) spaced in 0.50 μm, in order to obtain a 3D image from each liver section, which was generated at the speed of 100 fps (frames per second).

Five individuals from each experimental group were evaluated and five different liver sections per individual were used to obtain 3D images, summing to twenty-five 3D images per group. For each image, five fields (designated ROI) in hepatic parenchyma were used, so that only hepatocytes were selected for the quantification of emitted fluorescence intensity (positive immunolabeling for HSP70). Each ROI was 20 μm wide and 20 μm high, with an approximate area of 315 μm2. Thus, 25 ROI were analyzed per individual (N = 5 per group) and, consequently, 125 ROIs per experimental group. The fluorescence intensity in each ROI was measured by means of the software Leica Application Suite AF 2.6.0 (LAS-AF). Measurements from ROI were submitted to statistical analysis.

Immunodetection of cell death by the TUNEL method

Histological sections of the paraffin-embedded livers were deparaffinized and hydrated. Slides containing the fish liver sections were then submitted to the TUNEL (Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling) method (Gavrieli et al. 1992). In order to detect single- and double-stranded DNA strand breaks generated during apoptosis, the In Situ Cell Death Detection Kit (ISCDDK) was used in fish liver sections, according to the manufacturer’s instructions (Roche, Sigma-Aldrich).

After application of the TUNEL reaction mixture, which contains TdT and fluorescein-dUTP, slides containing liver sections were washed with PBS, and mounted with aqueous medium. A slide with histological sections was previously incubated in a buffer solution containing recombinant DNase I, which induces DNA strand breaks, representing the positive control of the TUNEL reaction (Fig. 1c). Analysis were performed by means of a laser scanning confocal microscopy, using a laser with wavelength of 488 nm (excitation 450 nm, emission 565 nm) in order to detect positive nuclei labeled with fluorescein.

Fig. 1
figure 1

Liver of P. harpagos. Exposure to acetaminophen. Immunofluorescence for HSP70 (red pseudo-color) marking and analysis by confocal laser scanning microscopy. a Acute exposure to acetaminophen, control group. b Acute exposure to acetaminophen, group 8 μg L−1. c Chronic exposure to acetaminophen, control group. d Chronic exposure to acetaminophen group 80 μg L−1

Full size image

The analysis was performed using 10% laser intensity, with 0.50 AU pinhole aperture (laser aperture), gain 1243.9 (brightness), and offset − 17.5% (background). The analysis was performed with the × 40 objective and × 2 optical zoom. The three-dimensional reconstruction was done with a total thickness of 5 μm, spacing 0.50 μm, totaling 11 optical sections in each 3D image generated at the speed of 100 fps.

Five individuals from each experimental group were evaluated, and 5 fields (ROI) per section in five different histological sections were analyzed per individual from each experimental group, in order to detect negative- and/or positive nuclei for DNA fragmentation.

Statistics

The data obtained concerning the quantification of the immunofluorescence labeling of HSP70 family proteins were submitted to the normality test (Shapiro-Wilk) and to the variance homogeneity test (Bartlett), which characterized the data as non-parametric. Therefore, the statistical test used was Kruskal-Wallis, followed by the Dunn test for comparison between the control group and the groups exposed to the drugs. All statistical analyses were performed using GraphPad Prism 6.0 software, with p < 0.05.

Results

Evaluation of the HSP70 by immunolabeling

Immunofluorescence indicative of HSP70-positive marking in livers of fish that were exposed to acetaminophen (Fig. 1), as evidenced by the immunolabeling of P. harpagos hepatocytes in the control group (Fig. 1a and c) and in the exposed groups (Fig. 1b and d). It is possible to observe that, after the acute exposure, there was a more evident marking of tissues of fish from the group exposed to 8 μg L−1, when compared to those from the control treatment, corroborated by results of the statistical analysis (Fig. 3). Similarly, data obtained for the chronic test showed a visible decrease of the marking that occurred in animals of the group exposed to 80 μg L−1, in relation to the control.

Figure 2 shows immunofluorescence indicative of HSP70-positive labeling in fish livers following exposures to propranolol, as evidenced by the immunoblot of P. harpagos hepatocytes in the control group (Fig. 2a and c), and in the exposed groups (Fig. 2b and d). After both acute and chronic exposures to this drug, it was possible to observe the increase in the intensity of the immunostaining of the fish from exposed groups in relation to the control, in agreement with the results of the statistical analysis (Fig. 3).

Fig. 2
figure 2

Liver of P. harpagos. Exposure to propranolol. Immunofluorescence for the labeling of HSP70 (pseudo-red color) and analysis by confocal laser scanning microscopy. a Acute exposure to propranolol, control group. b Acute exposure to propranolol, group 1000 μg L−1. c Chronic exposure to propranolol, control group. d Chronic exposure to propranolol. Group 0.0625 μg L−1

Full size image
Fig. 3
figure 3

Data obtained in the quantification of immunofluorescence (in pixel) indicative of positive immunolabeling for HSP70 family proteins. ROI immunolabeling intensity mean values (125 per group)

Full size image

For fish acutely exposed to acetaminophen (Fig. 1a), the immunofluorescence indicative of HSP70 in the hepatic parenchyma had a significant increase in all groups (F5,750 = 394.7; p < 0.0001); on the contrary, the chronic exposure of P. harpagos to acetaminophen (Fig. 1b), resulted in a significant decrease of this immunofluorescence in the liver of animals exposed to all concentrations tested (F5,750 = 353.3; p < 0.0001).

For animals acutely and chronically exposed to propranolol (Fig. 2c and d), there was an increase in immunofluorescence for HSP70. In acutely exposed fish, the concentrations that were responsible for significant increases were 1, 10, and 1000 μg L−1 (F5, 750 = 50.84, p < 0.0001), and for the chronic assay, the concentrations that induced significant increases of immunofluorescence were 0.0625, 0.125, 0.25, and 0.5 μg L−1 (F5,750 = 438.3, p < 0.0001).

Evaluation of cell death by immunolabeling of DNA fragmentation

The results for cell death were analyzed by a laser confocal microscope, and sporadic hepatocytes with positive nuclei for DNA fragmentation were found in all experimental groups (data not shown), so that cell death was observed in basal level, which is the normal condition during the cell turnover in liver. Both treated and control animals showed the same pattern concerning the occurrence of hepatocytes with positive nuclei for DNA fragmentation, which demonstrates that there was no induction of DNA fragmentation in nucleus of hepatic cells after both acute and chronic exposures to acetaminophen and propranolol, in comparison to animals from the control groups.

Discussion

Acetaminophen is a compound that can generate oxidative stress in cells of a large array of distinct species, causing oxidation of protein thiol groups (Nunes et al. 2015; Oliveira et al. 2015), which may require the stimulation of HSP production to protect cellular proteins from stress (Ahn and Thiele 2003). The increase in the level of HSP70 in hepatocytes of vertebrates exposed to acetaminophen is a response frequently associated with the oxidative stress generated during the biotransformation process of this drug (Sumioka et al. 2004; Knockaert et al. 2011; Guiloski et al. 2017). Increased levels of HSP gene expression were also observed in liver mouse after being exposed to this drug (Salminen et al. 1998), probably as a consequence of oxidative stress resulting from acetaminophen metabolism (Salminen et al. 1998; Xu et al. 2008; Knockaert et al. 2011). One of the effects of ROS in cells is precisely the reactivity with cellular proteins (Scheibmeir et al. 2005), with alteration of their three-dimensional structure. Increased levels of damaged proteins may activate the mechanism of HSP70 transcription and de novo synthesis of these chaperones in order to protect cells from this toxic effect (Iwama et al. 1998; Castro et al. 2014). The cytoprotective effect exerted by HSP70 is crucial for limiting the increase of acetaminophen-induced injury in distinct tissues (Tolson et al. 2006). These oxidative effects on proteins may explain the increased immunolabeling of HSP70 in hepatocytes of P. harpagos in response to oxidative stress resulting from acute exposure to acetaminophen. The pathway of activation of HSP70 in eukaryotes is evolutionarily conserved, and involves the signaling by partially denatured proteins (or with changes in their three-dimensional conformation) to activate the UPR (unfolded protein response), leading to an increase in the level of transcription of the HSP genes by the specific transcription factors (Fujimoto and Nakai 2010; Garbuz 2017), such as HSF1 (heat shock factor 1). This effect occurs via redox-dependent activation (Ahn and Thiele 2003), leading to the intracellular increase of the amount of HSP70 proteins in the cell compartments. Considering that the increase in the intracellular amount of HSP70 proteins after acute exposure to acetaminophen is a frequent response in vertebrates (Sumioka et al. 2004; Knockaert et al. 2011; Guiloski et al. 2017), and that the genes of the HSP70 family are generally evolutionarily conserved, it is possible to suggest that this same response occurred in the P. harpagos individuals exposed in our study.

However, chronic exposure to acetaminophen yielded a significant decrease in the immunolabeling of HSP70 in hepatocytes of exposed fish, which may be a direct deleterious consequence of oxidative stress inducing cytotoxicity in hepatocytes after chronic exposure, compromising the de novo synthesis of the inducible members of the HSP70 family. In this case, already existing chaperone protein molecules may not be enough to supply cellular needs, and such scenario may occur in case of cytotoxicity (Soti et al. 2003). Study with organic compounds showed that some endocrinal disruptors cause a decrease of HSP70 gene expression in copepod Tigriopus japonicas, and the authors suggest that may be a certain selectivity of these compounds to HSP70 genes, regulating their expression (Rhee et al. 2009). Chronic exposure to acetaminophen-induced cytotoxicity due to oxidative stress, evidenced by lipid peroxidation, in Oncorhynchus mykiss (Ramos et al. 2014). Changes in liver histology of the fish species Danio rerio, chronically exposed to acetaminophen at concentrations of 0.5 and 10 μg L−1, induced signs of cytotoxicity in hepatocytes, such as loss and vacuolization of their cytoplasm (Galus et al. 2013). In the present study, the lowest concentration of acetaminophen used in the chronic exposure (5 μg L−1) is in the same range of concentrations to which D. rerio were exposed in the cited study. It is thus possible to suggest that the main cytotoxic mechanisms that have been already described for other fish species may be also accounted for the cytotoxic effect that was here observed.

Although the chronic exposure to acetaminophen induced a decrease in the immunolabeling of HSP70 in hepatocytes of exposed fish, the negative immunolabeling for TUNEL method in these cells could suggest acetaminophen-induced cell death without DNA fragmentation. Absence of oligonucleossomol DNA fragmentation during apoptosis was observed in different cells, namely in neuroblastoma cells (Yuste et al. 2001), and also in human Jurkat cells exposed to curcumin pigment (Sikora et al. 2006). This result showed that, despite the chemical stress caused in hepatocytes of P. harpagos after exposure chronically to acetaminophen, fish were able to face the challenge of such exposure. Another possibility to explain the negative immunolabeling for TUNEL method in hepatocytes can be the induction of autophagy by acetaminophen because this drug inhibits mTOR activity in primary hepatocytes (Ni et al. 2012), by means of increased generation of mitochondrial ROS and decreased cellular ATP levels. This effect helps cells to relieve oxidative stress induced by acetaminophen. As the Akt/mTOR signaling pathway is an important regulator of autophagy, its inhibition may contribute to increased autophagy in the liver. Thus, the here-obtained data can suggest increased autophagy in the liver of fish acutely exposed to acetaminophen, in order to favor cell survival; for these reasons, we did not find increased cell death in this experimental group in relation to the control group. Autophagy counteracts apoptotic cell death via the cell survival pathway, so that the autophagic response occurs earlier than apoptotic cell death, placing autophagy upstream of apoptosis (Wang 2015).

Both acute and chronic exposures to propranolol caused a significant increase in the intensity of HSP70 labeling in P. harpagos hepatocytes. This effect indicates a change in cellular homeostasis in response to the stress generated by drug exposure as a compensatory response, but not a cytotoxic response as that observed for acetaminophen. Studies relating changes in HSP70 levels after exposure to propranolol in aquatic organisms are scarce. In the study by Sun et al. (2015), the authors observed that an acute exposure (96 h) of D. rerio to propranolol was capable of increasing the expression of HSP70 family genes, which, according to the authors, indicates a protective mechanism against proteotoxic damage. Increased expression of the HSP70 family genes also occurred in gills and digestive glands of mussels of the species Dreissena polymorpha exposed to the β-blocker metaprolol, indicating a need to activate mechanisms of repair, transport, or even protection of proteins (Contardo-Jara et al. 2010).

Level of members of HSP70 family, such as HSP72, is influenced by beta-adrenergic-receptor intermediates, such as cAMP and cAMP-dependent protein kinase (Choi et al. 1991; Pizurki and Polla 1994). Fabbri and Moon (2016) demonstrated that β-2-adrenergic receptors signaling systems existed in the fish liver. In fact, the structure and function of these receptors seem to be largely conserved in vertebrates including mammals, and the fish hepatic adrenergic receptor is a cAMP-dependent signaling transduction pathway. As demonstrated by Sun et al. (2014), β-blockers are able to induce upregulation of transcript abundance of HSP70 in embryos and larvae of zebrafish exposed to propranolol and metoprolol.

The lowest concentration of acetaminophen that elicited the increase of HSP70 level, following the acute exposure, is similar to the levels found in rivers throughout the globe (10 μg L−1, Kolpin et al. 2002; 13 μg L−1, Américo et al. 2012; 30.421 μg L−1, Campanha et al. 2014). For propranolol, the range of concentrations that induced increased levels of HSP70 after the acute exposure was already found in specific locations of the aquatic environment (0.1–0.09 μg L−1, Andreozzi et al. 2003; 0.026–1.900 μg L−1, Huggett et al. 2003). Similarly, levels of propranolol to which fish were exposed were also reported to occur in the wild (0.1–0.09 μg L−1, Andreozzi et al. 2003; 0.1–1.900 μg L−1, Huggett et al. 2003; 0.016–0.135 μg L−1, Zhou et al. 2009; 0.0152, Campanha et al. 2014; 0.0260 μg L−1, Thomas et al. 2014). Thus, acute exposures to both pharmaceuticals (acetaminophen and propranolol) and chronic exposure (namely to propranolol) are likely to occur in the wild. Consequently, the here-observed modifications in fish hepatocytes (showing the effort of the organism to prevent possible cytotoxic damages in liver resulting from drug-induced proteotoxicity under realistic concentrations of both pharmaceuticals) may occur under real scenarios of contamination. On the other hand, the decreased HSP70 level observed after chronic exposure of P. harpagos to acetaminophen has occurred for concentrations similar to those levels already found in rivers (6 μg L−1, Ternes 1998; 10 μg L−1, Kolpin et al. 2002; 13 μg L−1, Américo et al. 2012; 30.421 μg L−1, Campanha et al. 2014; 5.529–6.957 μg L−1, Robert and Thomas 2006). Thus, decreased HSP70 level can become the hepatocytes more sensible to additional xenobiotics from the aquatic environment, at realistic conditions, that could cause liver injury.

The here-obtained results showed that HSP70 is a sensible cell stress biomarker to measure the effects of drugs exposure on neotropical fish species, because their normal pattern of expression was significantly changed in animals that were exposed to the two tested drugs. Immunofluorescence is an excellent method for in situ localization and assessment of cell biomarkers in liver fish (Ansoar-Rodríguez et al. 2016). The innovation that our study brought to the current state of the art on cell stress biomarker and/or cell death biomarker in fish exposed to xenobiotics was the evaluation of these biomarkers in a species of neotropical fish exposed to pharmaceutical drugs.

However, the absence of increased level of DNA fragmentation in hepatocytes from drugs-exposed fish could be a consequence of the cytoprotective effect exerted by HSP70 family proteins, and/or by exposure to environmentally relevant concentrations that did not promote cytotoxicity. The concentrations used herein are considered to be environmentally relevant. On the other hand, in the environment, the pharmaceutical compounds are associated with other contaminants, including other drugs that act through the same pharmacological and toxicological routes (Daughton and Ternes 1999; Fawell and Ong 2012), which may potentiate these effects in non-target organisms, such as fish. In addition, drugs are generally considered to be pseudo-persistent, with continuous entry into the aquatic environment (Ebele et al. 2017), which exceeds the ability of the environment to degrade them. This pseudo-persistent nature of pharmaceutical drugs causes aquatic organisms to be continually exposed to relatively high levels of these compounds, and therefore, organisms exposed via environment may be continuously under the toxic action of combination of these chemical compounds in waters.

Finally, in a real scenario on field conditions, a considerable set of anthropogenic compounds may exist in the aquatic environment, and they can exert similar effects to these obtained in vitro conditions and, thereby, compromise the fish ability to deal with other environmental stressors.

Conclusion

In conclusion, the increased intensity of HSP70 immunolabeling in fish acutely exposed to acetaminophen and propranolol, as well as chronic exposure to propranolol, indicated a cytoprotective response induced by these two pharmaceutical substances on P. harpagos liver. On the contrary, chronic exposure to acetaminophen probably induced the response to a chemical injury in hepatocytes that have reached the threshold of cytoprotection by HSP70. This scenario has probably led to the decreased levels of HSP70 in hepatocytes, but without induction of DNA fragmentation as hallmark of typical apoptotic cell death.

The importance of this study lies in the fact that the results obtained are ecologically relevant, since the concentrations tested were already reported in the environment. Additionally, there is few data in literature about sublethal effects of drugs on fishes which makes it difficult to compare the here-obtained results, reinforcing the relevance of this ecotoxicological study about the evaluation of HSP70 and cell death after acute and chronic exposures to acetaminophen and propranolol in neotropical species P. harpargos.