Introduction

The recent years have brought on an increase in the number and use of new psychoactive substances (NPS). The number of NPS that reach the online markets every year is ever growing, with over 670 new drugs being currently monitored by the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA 2018). Synthetic phenylethylamines are one of the most prevalent NPS groups, as detailed by the 2018 World Drug Report (UNODC 2018), and within this group, psychoactive benzofurans or benzofury are among the most used drugs (Roque Bravo et al. 2019). 5-(2-Aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) have structural (Fig. 1) and pharmacological similarities to the well-known classic stimulants 3,4-methylenedioxyamphetamine (MDA) and 3,4-methylenedioxymethamphetamine (MDMA; ecstasy), acting as monoamine transporter inhibitors and as serotoninergic receptor agonists (Iversen et al. 2013; Rickli et al. 2015).

Fig. 1
figure 1

Chemical structures of psychoactive benzofurans and classical amphetamines. 5-APB, 5-(2-aminopropyl)benzofuran; 6-APB, 6-(2-aminopropyl)benzofuran; MDA, 3,4-methylenedioxyamphetamine; MDMA, 3,4-methylenedioxymethamphetamine

Full size image

A few clinical reports have disclosed benzofuran toxicity (Adamowicz et al. 2014; Chan et al. 2013; Helander et al. 2015; Hofer et al. 2017; Krpo et al. 2018; McIntyre et al. 2015; Seetohul and Pounder 2013; Turcant et al. 2017; Vallersnes et al. 2017), as well as lethality (Connelly et al. 2002). Some of the deleterious effects reported by users include agitation, mydriasis, tachycardia, seizures and hallucinations, similar to symptoms reported for MDA or MDMA (Jebadurai et al. 2013).

The liver is the organ where the majority of bioactivation and detoxifying metabolic reactions take place and, considering that oral ingestion is the most common route of administration for benzofurans (Barcelo et al. 2017), it becomes clear that there is a significant first-passage effect. As such, the liver is fully engaged in the biotransformation/elimination of these substances and of the resulting metabolites. Previous works have demonstrated that amphetamines and their metabolites can induce deleterious effects on liver cells by disturbing intracellular mechanisms, such as oxidative and mitochondrial homeostasis, and by activating programmed cell death pathways (Carvalho et al. 1996, 1997, 2001, 2013a; Dias da Silva et al. 2014a, b; Pontes et al. 2008a, b). Given the structural similarity, it becomes apparent that benzofurans may also have the potential to induce such effects on hepatocytes. One recent work proved that both 5-APB and its methylated analogue, 1-(benzofuran-5-yl)-N-methylpropan-2-amine (5-MAPB) have cytotoxic effects on freshly isolated rat hepatocytes through the production of reactive species and interference with mitochondrial membrane potential (Nakagawa et al. 2017). Notwithstanding, there is still a lack of research into the potential hepatotoxic effects of other benzofurans, in particular 6-APB. Noteworthy, despite being banned on countries such as the UK (ACDM 2013), 6-APB still remains one of the top selling NPS in Europe, being the most consumed benzofury (Roque Bravo et al. 2019).

Herein, we provide a comparison between the hepatotoxicity of 5-APB and its popular isomer 6-APB, which is particularly relevant in the clinical and forensic contexts, as it allows first-responders, clinicians and forensic experts, to better estimate the perils of the drug they are facing against, and determine the best treatment approach.

Materials and methods

Chemicals

Both 5-APB and 6-APB were acquired online at sensearomatics.net on March 2013. Chemical purity and identity of the drugs were verified by mass spectrometry, NMR and elemental analysis. Analytical data were consistent with the assigned structures with about 99% purity. Stock solutions of both drugs were made up in Hanks’ balanced salt solution (HBSS, no calcium or magnesium) and stored at – 20 °C. These solutions were thawed and diluted on the days that experiments were performed.

The cell culture reagents were purchased from Gibco (Thermo Fisher Scientific, Massachusetts, USA) and all other reagents were obtained from Sigma-Aldrich (Missouri, USA), unless otherwise stated.

Animals

Male Wistar Han rats with a body weight of 150–250 g were kept in sterile facilities under controlled temperature (20 ± 2 °C), humidity (40–60%), and lighting (12-h light/dark cycle) conditions and fed with sterile standard rat chow and tap water ad libitum. Isolation of hepatocytes was always conducted between 8:00 and 10:00 a.m. Surgical procedures were performed after rat anaesthesia induced by an i.p. injection of a combination of 20 mg/kg xylazine (Rompun® 2%, Bayer HealthCare, Germany) and 100 mg/kg ketamine (Clorketam® 1000, Vétoquinol, France), and maintained through inhalation of isoflurane vapour (IsoVet® 1000 mg/g, B. Braun VetCare, Germany). This study was performed at the highest standards of ethics after approval by the local Ethical Committee for the Welfare of Experimental Animals (University of Porto-ORBEA; project 158/2014) and by the national authority Direção Geral de Alimentação e Veterinária (DGAV). Housing and all experimental procedures were performed by investigators accredited for laboratory animal use in accordance with the Portuguese and European legislation (law DL 113/2013, Guide for Animal Care; Directives 86/609/EEC and 2010/63/UE).

Isolation of primary rat hepatocytes

Isolation of hepatocytes was performed using a modified two-step perfusion of the liver, as previously described by Dias da Silva et al. (2017) with some modifications. The liver was perfused in situ via the vena portae, with sterile EGTA-buffer at 37 °C, for approximately 10–15 min. The chelator promoted the irreversible cleavage of the desmosomes (junctional complexes) through calcium sequestration. The EGTA-buffer consisted of 248 mL glucose solution (9 g/L d-glucose), 40 mL Krebs–Henseleit buffer (60 g/L NaCl, 1.75 g/L KCl, and 1.6 g/L KH2PO4; adjusted to pH 7.4), 40 mL HEPES-buffer I (60 g/L HEPES; adjusted to pH 8.5), 30 mL MEM non-essential amino acid solution (100 ×), 30 mL MEM amino acids solution (10 ×), 4 mL glutamine solution (7 g/L l-glutamine), and 1.6 mL EGTA-solution (47.5 g/L EGTA; dissolved by addition of NaOH, adjusted to pH 7.6). Subsequently, hepatic collagen was hydrolysed by liver perfusion for 10–15 min with a sterile collagenase buffer supplemented with its cofactor calcium, at 37 °C. The collagenase buffer consisted of 155 mL glucose solution, 25 mL Krebs–Henseleit buffer, 25 mL HEPES-buffer I, 15 mL MEM non-essential amino acid solution (100 ×), 15 mL MEM amino acids solution (10 ×), 10 mL CaCl2 solution (19 g/L CaCl2·2H2O), 2.5 mL glutamine solution, and ~ 300 U/mL collagenase type IA from Clostridium histolyticum (dissolved immediately before use). After perfusion, the liver was dissected, removed from the animal, and the hepatic capsule gently disrupted in a sterile suspension buffer [124 mL glucose solution, 20 mL KH-buffer, 20 mL HEPES-buffer II (60 g/L HEPES; adjusted to pH 7.6), 15 mL MEM non-essential amino acid solution (100 ×), 15 mL MEM amino acids solution (10 ×), 2 mL glutamine solution, 1.6 mL CaCl2 solution, 0.8 mL MgSO4 solution (24.6 g/L MgSO4·7H2O), and 400 mg bovine serum albumin (BSA)]. The obtained suspension was purified by three low-speed centrifugations at 50g, for 2 min, at 4 °C. The viability of isolated hepatocytes at this stage was always above 90%, as assessed by the trypan blue exclusion method.

Culture of primary rat hepatocytes

A suspension of 5 × 105 viable cells/mL in complete culture medium was seeded onto the central 60 wells of 96-well plates (5 × 104 cells per well) or onto 6-well plates (1 × 106 cells per well) pre-coated with collagen. Complete cell culture medium consisted of William’s E medium (Sigma-Aldrich, Missouri, USA) supplemented with 10% heat-inactivated foetal bovine serum (FBS), 5 μg/mL insulin solution from bovine pancreas (Sigma-Aldrich, Missouri, USA), 50 nM dexamethasone (Sigma-Aldrich, Missouri, USA), 1% antibiotic solution (10,000 U/mL penicillin; 10,000 μg/mL streptomycin), 100 μg/mL gentamicin, and 250 ng/mL amphotericin B. After seeding, primary rat hepatocytes were left to adhere overnight at 37 °C, in an atmosphere of 5% CO2. On the next day, cells were exposed to the test drugs.

Human HepaRG cells

HepaRG hepatocyte cell line derives from a female patient with hepatocarcinoma. Upon reaching confluence, cells can transdifferentiate into two types of cells, one that is morphologically similar to human primary hepatocytes, and another that resembles bile canaliculus-like structures (Guillouzo et al. 2007). HepaRG cells retain the ability of expressing all major cytochrome P450 isoforms (with exception of CYP2E1 and CYP2D6) and many phase II enzymes, closely mimicking the in vivo situation (Guillouzo et al. 2007).

HepaRG cells were acquired from Life Technologies (Invitrogen, France) and were cultured as described before (Dias da Silva et al. 2017), at 37 °C, with an atmosphere of 5% CO2, in 75-cm2 flasks, using Williams’ E medium with l-glutamine, supplemented with 10% FBS, 1% antibiotic solution (10,000 U/mL penicillin; 10,000 μg/mL streptomycin), 5 μg/mL insulin (Sigma-Aldrich, Missouri, USA), and 50 μM hydrocortisone 21-hemisuccinate sodium salt (Sigma-Aldrich, Missouri, USA). The medium was changed every 2–3 days, and cells were sub-cultured by tripsinization at around 80% confluence, for a maximum of 10 passages. Upon reaching confluence, the differentiation was initiated by replacing the medium by fresh culture medium supplemented with 2% DMSO (Merck Millipore, Massachusetts, USA). The medium was then changed every 2–3 days, for 2 weeks, to promote differentiation into a 1:1 ratio of hepatic cells:biliary cells, as confirmed by microscopic observation. At the end of the differentiation period, cells were seeded onto 96-well plates at a density of 1.44 × 105 cells per well in a volume of 200 μL of complete medium and left to adhere overnight. They were used for cytotoxicity assays on the following day.

Human HepG2 cells

HepG2 cells, which derive from a male patient with hepatoblastoma, also express normal hepatocyte function. These cells secrete the majority of plasma proteins into the culture medium and retain the ability to produce and secrete bile acids (Bouma et al. 1989). These cells were previously used to successfully demonstrate the hepatotoxicity of other stimulants, namely amphetamines (Dias da Silva et al. 2013a, 2014a, b; Bouma et al. 1989) and piperazines (Dias da Silva et al. 2013b).

HepG2 cells were acquired from Life Technologies (Invitrogen, France) and cultured as previously described, with some modifications (Dias da Silva et al. 2013b). Briefly, the cells were kept in DMEM medium supplemented with 10% FBS and 1% antibiotic solution (10,000 U/mL penicillin; 10,000 μg/mL streptomycin), in 75-cm2 flasks, under a humidified 5% CO2 atmosphere at 37 °C, and subjected to regular medium changes (every 2–3 days). Whenever cells reached 80% confluence, they were sub-cultured by tripsinization for no more than 10 passages. For the cytotoxicity assays, cells were seeded in 96-well plates, at a density of 80 × 104 cells per well, in a volume of 200 μL of complete medium, and left to adhere overnight.

Drug exposures

On the day of the experiments, primary hepatocytes, HepaRG cells, and HepG2 cells were exposed to the test drugs for 24 h. The concentration range tested for each drug in the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay across the three different cell models is displayed in Table 1. MTT data were then used to calculate the concentrations of each drug to be tested in the following assays, which were conducted in rat primary hepatocytes only, due to the increased sensibility displayed by these cells. Accordingly, the concentrations of each drug producing 20%, 40%, 50% and 70% of the maximum cytotoxic effect in the MTT assay, i.e. EC20, EC40, EC50 and EC70, respectively, were estimated from the respective concentration versus response curves, using the parameters of the best fit regression model (see “Results” section). The selected concentrations are in line with those previously tested for amphetamine derivatives (Dias da Silva et al. 2014a).

Table 1 Concentrations (μM) of 5-(aminopropyl)benzofuran (5-APB) and 6-(aminopropyl)benzofuran (6-APB) tested for evaluating drug-induced alterations on cell viability, as assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay, after 24 h at 37 °C
Full size table

Cellular viability by the MTT reduction assay

The cytotoxicity of the drugs was evaluated using the MTT reduction assay, as described previously (Dias da Silva et al. 2015). The assay assesses cell viability through the indirect measurement of the activity of reductases that convert yellow soluble MTT into purple insoluble formazan salts. As MTT can only be reduced when these enzymes are active, the reaction is used as an indicator of the cell metabolic competence and, therefore, of viability. Briefly, after the exposure period, the culture medium was aspirated, followed by the addition of fresh culture medium containing 0.25 mg/L MTT in HBSS (no calcium, no magnesium). After a 30-min incubation at 37 °C in a humidified, 5% CO2 atmosphere, the formed intracellular crystals of formazan were dissolved in 100 μL DMSO. The absorbance was measured at 570 nm, using a multi-well plate reader (BioTek Instruments, Vermont, USA). Data were normalized to positive (1% Triton X-100) and negative controls (culture medium only).

Cytochrome P450 inhibition

To determine the influence of CYP metabolism in the toxicity induced by these drugs, different CYP isoforms of primary rat hepatocytes were inhibited prior and during exposure to the test drugs, as described before (Dias da Silva et al. 2013b). Briefly, the cells were incubated for 1 h with four CYP inhibitors: 500 μM metyrapone (CYP2E1 inhibitor), 10 μM quinidine (CYP2D6 inhibitor), 1 μM ketoconazole (CYP3A4 inhibitor) and 1 mM 1-aminobenzotriazole (ABT; general CYP inhibitor). Then, cells were co-incubated with these inhibitors and each test drug, during 24 h, at 37 °C. Afterwards, cell viability was evaluated using the MTT assay and data were normalized to positive (1% Triton X-100) and negative (cell culture medium only) controls. Additional controls with inhibitor only in cell culture medium were performed and compared to negative controls to ensure that no cytotoxicity derived from these treatments.

Intracellular reactive oxygen (ROS) and nitrogen (RNS) species

The assessment of intracellular reactive oxygen (ROS) and nitrogen (RNS) species was performed using the 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) fluorescence assay, as described by Dias da Silva et al. (2014a). Briefly, primary rat hepatocytes seeded onto 96-well plates were exposed for 30 min to 10 μM DCFH-DA, at 37 °C, protected from light. Afterwards, the cells were washed with HBSS (no calcium, no magnesium) and incubated with each tested drug, at 37 °C, for 24 h. The fluorescence was recorded using a fluorescence microplate reader (BioTek Instruments, Vermont, USA), set to 485 nm excitation and 530 nm emission. The acquired data were normalized to control conditions (only cell culture medium).

Total (tGSH), reduced (GSH), and oxidized (GSSG) glutathione

Primary rat hepatocytes seeded onto 6-well plates were exposed to each drug for 24 h, at 37 °C. After this time, cells were washed with HBSS (no calcium, no magnesium) and precipitated with 10% perchloric acid for 20 min, at 4 °C. Cells were then scrapped and the obtained suspension was centrifuged at 13,000g for 5 min, at 4 °C. The supernatant was collected and stored at − 80 °C until further determinations. The cell pellet was resuspended in 1 M NaOH and used to quantify protein by the Lowry assay (Lowry et al. 1951). After thawing, the supernatants were neutralized using 0.76 M KHCO3 (1:1) and centrifuged at 13,000g for 10 min, at 4 °C. Supernatants were then used to determine total glutathione (tGSH), oxidized glutathione (GSSG) and ATP.

The 5,5-dithio-bis(2-nitrobenzoic acid) (DTNB)-GSSG reductase-recycling assay was used to measure tGSH, as previously described (Dias da Silva et al. 2014a). Briefly, 100 μL of each neutralised sample, blank and standard were transferred into a 96-well plate, followed by the addition of 65 μL reagent solution (72 mM phosphate buffer, 0.69 mM NADPH, and 4 mM DTNB). The plates were incubated for 15 min, at 30 °C, in a multi-well plate reader (Power Wave X™, BioTek Instruments, Inc.) and then 40 μL of a 10 U/mL glutathione reductase solution was added. The absorbance was further read in kinetic mode at 415 nm to follow the formation of 5-thio-2-nitrobenzoic acid (TNB). The GSH content in the samples was interpolated from a standard curve, performed for every independent experiment. As 2-vinylpyridine blocks GSH, the same protocol was used for the quantification of GSSG, following sample incubation with 2-vinylpyridine for 1 h, at 4 °C. Experimental data were normalized to the protein content of each sample and the intracellular reduced glutathione (GSH) was calculated using the formula GSH = tGSH − 2 × GSSG.

Intracellular adenosine triphosphate (ATP)

A bioluminescence method was used for measuring intracellular ATP, as previously described Dias da Silva et al. (2013b). Samples were prepared as detailed above for the determination of glutathione. Afterwards, 75 μL of each sample neutralized supernatant, standard, or blank were transferred into a white 96-well plate. Then, 75 μL of a luciferin–luciferase solution (0.15 mM luciferin; 30,000 light units luciferase/mL; 10 mM MgSO4; 50 mM glycine; 1 mM Tris; 0.55 mM EDTA; 1% BSA) were added and the emitting light intensity was measured using a luminescence plate reader (BioTek Instruments, Vermont, USA). The ATP content in the samples was interpolated from a standard curve performed in every experiment and normalized to the protein content.

Mitochondrial membrane potential (Δѱm)

Mitochondrial integrity was evaluated using tetramethylrhodamine ethyl ester (TMRE), a positive-charged dye that solely targets active mitochondria. As described by Dias da Silva et al. (2013b), primary rat hepatocytes seeded onto 96-well plates were exposed to the drugs for 24 h at 37 °C. Following drug exposure, the medium was aspirated, and cells were washed with HBSS (no calcium, no magnesium) and incubated for 30 min with 2 μM TMRE prepared in fresh culture medium, protected from light. After the incubation period, cells were washed twice with HBSS (no calcium and no magnesium) and the fluorescence was quantified using a fluorescence microplate reader, set to 544 nm excitation and 590 nm emission. Data were normalized to negative controls (cell culture medium only).

Caspase-8, -9, and -3 activities

The activity of pro-apoptotic caspases-8, -9, and -3 was evaluated in primary rat hepatocytes seeded onto 6-well plates, following 24-h drug exposures at 37 °C. The incubation medium was discarded, the cells washed with HBSS (no calcium, no magnesium) and added of 150 μL lysis buffer (50 mM HEPES, 0.1 mM EDTA, 0.1% CHAPS, and 1 mM DTT; pH 7.4). The plates were then placed at 4 °C for 30 min, and cells scraped, collected into 2 mL microcentrifuge tubes, and centrifuged at 13,000g for 10 min, at 4 °C. Then, 50 μL of the supernatant were collected into a 96-well plate and added of 200 μL assay buffer (100 mM NaCl, 50 mM HEPES, 0.1 mM EDTA, 10% glycerol, 0.1% CHAPS, and 1 mM DTT). The reaction was started by adding 5 µL of caspase-3 (Ac-DEVD-pNA; 4 mM in DMSO), caspase-8 (Ac-IETD-pNA; 10 mM in DMSO), or caspase-9 (Ac-LEHD-pNA; 10 mM in DMSO) substrates. The plates were then sealed with parafilm, covered with aluminium foil, and incubated at 37 °C for 24 h. After this period, the absorbance was measured at 405 nm and data were normalized to the amount of protein of each sample. The protein content in the cytoplasmic fraction was quantified using the DC™ protein assay kit (Bio-Rad Laboratories, CA, USA), as described by the manufacturer. Results were expressed as percentage of negative controls.

Extracellular lactate dehydrogenase (LDH)

Cell lactate dehydrogenase (LDH) leakage occurs upon rupture of the cytoplasmic membrane. After a 24-h exposure to each drug, 10 μL of the incubation medium of primary rat hepatocytes seeded in 96-well plates were added of 40 μL of 50 mM potassium phosphate buffer and 200 μL of 0.2 mM β-NADH (in 50 mM potassium phosphate buffer), as described by Dias da Silva (2019). Then, the addition of 25 μL of 22.72 mM sodium pyruvate (in 50 mM potassium phosphate buffer) initiated the oxidation of β-NADH into β-NAD+, which was monitored by reading the absorbance decay at 340 nm, every 16 s, for 3 min, using an automatic plate reader Power Wave X™ (BioTek Instruments, Inc.). Data were normalized to positive (1% Triton X-100) and negative (cell culture medium only) controls.

Hoechst 33342/propidium iodide (PI) fluorescent staining

Hoechst 33342 is a cell-permanent counterstain that binds to DNA and emits a blue fluorescence; and propidium iodide (PI) is a nuclear membrane impermeant dye that issues red fluorescence solely in dead cells. In order to identify hepatocytes undergoing apoptosis or necrosis, Hoechst 33342 and PI were used as staining probes, as previously described by Valente et al. (2016). Briefly, after washing hepatocytes seeded onto 6-well plates with HBSS (no calcium, no magnesium), cells were incubated with 50 μM PI for 15 min, under light protection, rinsed twice, and further incubated with 5 μg/mL Hoechst 33342 for 5 min, previous to observation under an inverted fluorescence microscope Nikon Eclipse Ti (Nikon, Amsterdam, Netherlands), at an original magnification of 200 × .

Acridine orange staining for acidic vesicular organelles (AVO)

Acidic vesicular organelles (AVO) formation was assessed through acridine orange staining, as described in Valente et al. (2017). This lysosomotropic dye issues green/yellow/orange/red fluorescence in a pH-dependent manner: at neutral pH it is a hydrophobic green molecule; inside acidic organelles it is converted into its protonated form, building aggregates that issue yellow-to-red fluorescence. Briefly, following exposure, hepatocytes were incubated with 5 μg/mL acridine orange for 20 min, at 37 °C, to detect AVOs formation. Following incubation with acridine orange, cells were rinsed three times with HBSS (no calcium, no magnesium) and observed under an inverted fluorescence microscope Nikon Eclipse Ti (Nikon, Amsterdam, Netherlands), at an original magnification of 200 × .

Statistical analysis

Data obtained in the MTT assay in four independent experiments were fitted to the Logit model, chosen based on a goodness-to-fit principle (best fit dosimetric model) (Dias da Silva et al. 2013c). Comparisons between concentration–response curves were performed using the extra sum-of-squares F test. All other assays were performed at least in three independent experiments and data are presented as mean ± standard error of the mean (SEM). The normality of data distribution was analysed using the Kolmogorov–Smirnov test, and statistical comparisons between groups were done using one-way analysis of variance (ANOVA), followed by Dunn’s multiple comparison test, when the results followed a Gaussian distribution; or otherwise by Kruskal–Wallis test. All statistical analyses were made using GraphPad Prism® software, version 6.07 (San Diego, CA, USA).

Results

5-APB was more hepatotoxic than 6-APB in all of the tested hepatocyte models

The in vitro cytotoxic effects observed for the tested drugs in the different hepatocyte models are presented in Fig. 2. Our data show that the drug-induced decrease in cell viability is concentration-dependent for both drugs in all tested cell models. As observed by the respective potencies of each drug (Fig. 2, Table 2), 5-APB elicited greater cytotoxicity than 6-APB, in all the tested in vitro models (p < 0.0001; F test). It is also clear that primary rat hepatocytes display higher sensitivity to the hepatotoxicity elicited by 5-APB and 6-APB, than human immortalized hepatocytes (p < 0.0001; F test). In addition, HepG2 cells revealed to be more resistant to the toxic effects of these drugs than HepaRG cells (p < 0.0001; F test). Accordingly, the curves obtained for HepG2 cells were shifted to the right and the EC50 values were higher, when compared to those obtained for HepaRG cells (Fig. 2, Table 2). Based on the relative sensitivities of these in vitro models to benzofuran toxicity, further assays were carried out in primary rat hepatocytes at concentrations that elicited 20, 40, 50, and 70% of cell death (Table 3), as interpolated from the best-fits attained in the MTT assay.

Fig. 2
figure 2

Toxicity elicited in primary rat hepatocytes, HepaRG differentiated cells, and HepG2 cells after 24-h exposure to 5-(2-aminopropyl)benzofuran (5-APB, purple solid line) and 6-(2-aminopropyl)benzofuran (6-APB, black solid line). Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay and data are presented as percentage of cell viability relative to the negative controls (n = 4). Curves were fitted to the dosimetric Logit model (parameters displayed in Table 2). The dashed lines are the upper and lower limits of the 95% confidence interval of the best estimate of mean responses. The dotted horizontal lines represent 50% and 100% effect

Full size image
Table 2 Parameters derived from nonlinear fits (asymmetric Logit function) of concentration–mortality data of 5-(aminopropyl)benzofuran (5-APB) and 6-(aminopropyl)benzofuran (6-APB)
Full size table
Table 3 Concentrations (μM) tested for evaluating mechanisms underlying hepatotoxicity of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) when primary rat hepatocytes were exposed to the drugs for 24 h at 37 °C
Full size table

Noteworthy, we also observed the surfacing of intracellular black-brown pigments, following drug incubations at the highest concentrations tested.

Inhibition of CYP isoforms impacts the toxic potential of benzofurans in rat hepatocytes

Cytotoxicity resulting from drug exposure following specific inhibition of CYP2E1, CYP2D6, CYP3A4 or P450 in general is presented in Figs. 3 and 4. These data indicate that inhibition of the different isoforms of P450 led to a decline of the cytotoxicity elicited by 5-APB and 6-APB, as supported by the shift of the concentration–response curves towards the right. By comparing the cytotoxic potencies of the drugs in the presence or absence of the different P450 inhibitors (Table 2), it is also evident that the activities of CYP3A4 and P450, in general, present higher impact on the toxification of both drugs (p < 0.01), compared to the effect of CYP2D6 and CYP2E1.

Fig. 3
figure 3

Toxicity elicited in primary rat hepatocytes after 24-h exposure to 5-(2-aminopropyl)benzofuran (5-APB), following inhibition of CYP2E1, CYP2D6, CYP3A4 or P450 in general. Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay and data from incubations with (grey solid line) or without (black solid line) inhibitor are presented as percentage of cell viability relative to the negative controls (n = 4). Curves were fitted to the dosimetric Logit model (parameters displayed in Table 2). The dotted lines represent 50% and 100% effect

Full size image
Fig. 4
figure 4

Toxicity elicited in primary rat hepatocytes after 24-h exposure to 6-(2-aminopropyl)benzofuran (6-APB), following inhibition of CYP2E1, CYP2D6, CYP3A4 or P450 in general. Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay and data from incubations with (grey solid line) or without (black solid line) inhibitor are presented as percentage of cell viability relative to the negative controls (n = 4). Curves were fitted to the dosimetric Logit model (parameters displayed in Table 2). The dotted lines represent 50% and 100% effect

Full size image

Benzofurans induce oxidative stress in primary rat hepatocytes

Induction of oxidative stress is one of the cornerstone mechanisms of toxicity elicited by classic psychoactive stimulants, such as amphetamines (Dias da Silva et al. 2014a) and cocaine (Martins et al. 2018), but also by more recently emerged club drugs, such as piperazines (Dias da Silva et al. 2015, 2017). We thus determined the role of oxidative stress in the hepatotoxicity elicited by the tested benzofurans by evaluating the formation of ROS/RNS in rat hepatocytes, 24 h after drug exposure. Figure 5 shows that both drugs induced concentration-dependent production of reactive species, and this result was significantly different from controls at concentrations higher than the respective EC40 (p < 0.05; ANOVA followed by Dunn’s test).

Fig. 5
figure 5

Production of reactive oxygen (ROS) and nitrogen (RNS) species in primary rat hepatocytes, after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Fluorescence data from the 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) assay were normalized to negative controls (no drug exposure, set to 1) and were from five independent experiments. *p < 0.05, **p < 0.01, **p < 0.001; compared to controls. Statistical analysis of data was performed using one-way ANOVA followed by Dunn’s multiple comparison test

Full size image

In addition, since GSH is a pivotal first-defence against free radicals, we determined the intracellular levels of GSH and GSSG (Fig. 6). All tested concentrations of 5-APB and 6-APB induced a concentration-dependent depletion of GSH, that reached statistical significance at concentrations above the EC50 for 5-APB (p < 0.001; Kruskal–Wallis test) and above the EC40 for 6-APB (p < 0.05; Kruskal–Wallis test). The GSSG levels also increased at the EC40 and EC50 for both benzofurans (p < 0.01; Kruskal–Wallis test).

Fig. 6
figure 6

Intracellular levels of total (tGSH), reduced (GSH; back bars) and oxidized (GSSG; white bars) glutathione in primary rat hepatocytes, after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Results were from three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; compared to GSH in control. &p < 0.05, &&p < 0.01; compared to GSSG in control. Statistical analysis of data was performed using Kruskal–Wallis test

Full size image

Benzofurans disrupted mitochondrial homeostasis and functioning in primary hepatocytes

Mitochondrial homeostasis was evaluated by measuring the incorporation of the selective probe TMRE in the organelle. Results in Fig. 7 show an increase in Δѱm for both drugs (p < 0.05; ANOVA followed by Dunn’s test), when compared to control conditions, that was solely reverted at the highest concentration tested (EC70).

Fig. 7
figure 7

Mitochondrial membrane potential (ΔΨm) measured by TMRE inclusion assay in primary rat hepatocytes, after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Data from five independent experiments were normalized to control conditions (no drug exposure, set to 1). *p < 0.05, **p < 0.01, ****p < 0.0001; compared to control. Statistical analysis of data was performed using one-way ANOVA followed by Dunn’s multiple comparison test

Full size image

Since Δѱm is a vital component of the electrochemical proton motive force that drives energy generation by mitochondria, we also determined the intracellular ATP levels. ATP is essential for cells to successfully carry out their energy-dependent functions and, consequently, to survive. Under our experimental conditions, 5-APB and 6-APB significantly impaired cell energetics for all concentrations tested (p < 0.05; ANOVA followed by Dunn’s test), as depicted in Fig. 8.

Fig. 8
figure 8

Intracellular adenosine triphosphate (ATP) levels in primary rat hepatocytes, after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Results were from three independent experiments. *p < 0.05, ***p < 0.001, ****p < 0.0001; compared to control. Statistical analysis of data was performed using one-way ANOVA followed by Dunn’s multiple comparison test

Full size image

Benzofurans activated upstream and downstream apoptotic events that ultimately may render primary hepatocytes to necrosis

The activation of caspase cysteine–aspartic proteases is a hallmark of apoptotic programmed cell death. Figure 9 displays results obtained for the activation of initiator caspase-8 and -9 and executioner caspase-3, following 24-h exposure to 5-APB and 6-APB. Our data show that all treatments led to the activation of upstream and downstream apoptotic pathways, as evidenced by activation of caspases-8, -9, and -3 (p < 0.05; ANOVA followed by Dunn’s test); the exception was 6-APB at EC20 for caspase-3. All the effects were more exacerbated for 5-APB, as compared with 6-APB.

Fig. 9
figure 9

Caspase-8, -9 and -3 activation in primary rat hepatocytes, after exposure to the EC20, and EC50 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Results from four independent experiments were normalized to control (no drug exposure, set to 1). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; compared to control. Statistical analysis of data was performed using one-way ANOVA followed by Dunn’s multiple comparison test

Full size image

To further clarify the cell death mode, either by necrosis and/or apoptosis, we performed the Hoechst 33342/PI staining of primary rat hepatocytes exposed to 5-APB and 6-APB for a period of 24 h at 37 °C. As depicted in Fig. 10, nuclei in control cells (Fig. 10a) display regular contours and have a large and round size. Early apoptotic events (pyknotic nuclei lacking PI labelling, marked with green arrowheads) can be identified in 6-APB EC20 and EC40 (Fig. 10c, e) and 5-APB EC40 and EC70 (Fig. 10d, h). Late apoptotic cells (red condensed nuclei, marked with orange arrowheads) are present in images from all treatments, and necrotic cells (red large nuclei, marked with red arrowheads) in significant numbers in Fig. 10h (5-APB EC70) and g, i (6-APB EC50 and EC70, respectively).

Fig. 10
figure 10

Representative fluorescence microphotographs of Hoechst 33342/propidium iodide staining of primary rat hepatocytes after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) (b, d, f, and h, respectively) and 6-(2-aminopropyl)benzofuran (6-APB) (c, e, g and i, respectively) for 24 h, at 37 °C. Microphotograph a represents control conditions. Red arrows indicate necrotic cells, green arrows indicate early apoptotic cells and orange arrows indicate late apoptotic cells. Original magnification, 200 ×

Full size image

Intact cell membrane is also a classical feature of apoptotic cells, while membrane leakage is often considered a typical necrotic feature. Results of LDH leakage in primary rat hepatocytes, shown in Fig. 11, evidenced that both drugs were capable of inducing concentration-dependent damage to cytoplasmic membrane integrity with consequent LDH leakage, in accordance with the Hoechst 33342/PI staining data.

Fig. 11
figure 11

LDH leakage in primary rat hepatocytes, after exposure to the EC20, EC40, EC50, and EC70 of 5-(2-aminopropyl)benzofuran (5-APB) and 6-(2-aminopropyl)benzofuran (6-APB) for 24 h, at 37 °C. Results from four independent experiments were normalized to negative and positive controls (set to 0% and 100%, respectively). *p < 0.05, ***p < 0.001, ****p < 0.0001; compared to control. Statistical analysis of data was performed using one-way ANOVA followed by Dunn’s multiple comparison test

Full size image

Benzofurans induce autophagy in primary rat hepatocytes

To date there are no reports of autophagy induced by psychoactive benzofurans. Exposure of primary rat hepatocytes to different effect concentrations of 5-APB and 6-APB for 24 h, at 37 °C, induced morphological changes, as observed by phase contrast microscopy, which included a marked increase in cytoplasmic vacuolization (a typical feature of autophagy). Acridine orange was further employed to substantiate the presence of AVO. As expected, acridine orange labelling evidenced the presence of a larger amount of acidic vesicles following exposure to EC40 of 5-APB and 6-APB (Fig. 12c, e); these are located around the nuclei, as highlighted by pink arrowheads in phase contrast images (Fig. 12d, f).

Fig. 12
figure 12

Representative phase contrast (b, d, f) and fluorescence (a, c, e) microphotographs of primary hepatocytes stained with acridine orange. Cells treated with EC40 of 5-APB (c, d) and EC40 of 6-APB (e, f) for 24 h, at 37 °C, present acidic vesicular organelles stained in orange to red (pink arrow), which are absent in control cells (a, b). Blue arrows indicate cell nuclei. Original magnification, 200 ×

Full size image

Discussion

The benzofurans 5-APB and 6-APB, currently unscheduled in several European and US countries, are psychoactive substances with stimulant and entactogenic properties similar to those of amphetamines and, based on the results herein presented, also display similar toxicological mechanisms (Carvalho et al. 1996; Dias da Silva et al. 2014a; Brown and Yamamoto 2003). These mechanisms are barely elucidated, with little or no information pertaining to the toxicological pathways of these NPS towards such an important organ as the liver. Unfortunately, there is a significant lack of descriptive literature on clinical intoxications with these drugs. Nonetheless, clinical hepatotoxicity is a common outcome of other non-psychoactive benzofuran derivatives (Babany et al. 1987; Hautekeete et al. 1995; Jaiswal et al. 2018; Ratz Bravo et al. 2005; Tsuda et al. 2018; Verhovez et al. 2011), which comes in line with our results. Our study is the first comparing the in vitro hepatotoxicity induced by 5-APB and 6-APB. One previous study has already shed some light upon the toxic potential of benzofurans (Nakagawa et al. 2017) by demonstrating that 4 mM 5-APB induced hepatotoxicity by triggering loss of mitochondrial membrane potential, hampering ATP production and increasing ROS to a greater level than that produced by MDMA. Herein, we further contribute towards the knowledge on benzofury toxicity by showing similar effects for 6-APB, but with much lower potency, and by unveiling the involvement of apoptosis, necrosis and autophagy in the hepatotoxicity elicited by 5-APB and 6-APB.

Although the concentrations tested here are higher than the levels found in blood of intoxicated patients (from 0.395 nM to 31.96 μM) (Adamowicz et al. 2014; Chan et al. 2013; Krpo et al. 2018; McIntyre et al. 2015; Barcelo et al. 2017), it is well-known that the liver is exposed to concentrations of amphetamine-like stimulants far higher than the concentrations achieved in blood (De Letter et al. 2006; Garcia-Repetto et al. 2003). Considering the aforementioned structural and functional similarities between these drugs, it is reasonable to consider that the same occurs with the benzofurans tested in this work.

Overall, our study demonstrated lower cytotoxicity in immortalized HepG2 cells, compared to HepaRG cells. HepG2 cells are a cell line that expresses very low levels of essential drug-metabolizing enzymes, such as CYP2B6, CYP1A2 and 3A4, which have been described to be involved in the metabolism of amphetamines and other NPS (Aninat et al. 2006; Carvalho et al. 2012; Gerets et al. 2012; Meyer et al. 2012). Although displaying an enzymatic profile more representative of the hepatocyte in vivo, HepaRG cells were also established from a donor that poorly expressed CYP2D6 and CYP2E1 (Gerets et al. 2012), and the putative involvement of these isoforms in the metabolism of the studied compounds is not to be ignored. Therefore, the use of highly metabolically competent primary rat hepatocytes, which display great levels of P450 isoforms absent or present in low amounts in these human-derived cells, for evaluating the toxicity of 5-APB and 6-APB was therefore considered essential to address the flaws presented by the immortalized models. For both substances, the EC50 in primary cells was over 2.7-fold lower, supporting the potential relevance of metabolism by P450, as toxicity appears to be magnified with the increased expression/activity of the drug-metabolizing enzymes. This might signify that the hepatotoxic effects may be due not only to the drugs itself, but also to their indirect action through toxic metabolites. Further investigation is required to establish the toxicological profile of benzofury metabolites and their comparative toxicity towards the parent drugs.

To validate the hypothesis that significant disparities concerning hepatotoxicity among the employed models may rely on the activity and/or expression of their P450 isoforms, further experiments employing enzymatic inhibitors were carried out in primary rat hepatocytes, which is the most relevant model in what concerns P450 content. The most important alterations in drug potency occurred following specific inhibition of CYP3A4 and unspecific inhibition of P450, but whether the latter results from CYP3A4 inhibition by the non-specific inhibitor, requires further confirmation. Metabolism of benzofury through CYP3A4 may involve an epoxide formation, as furan toxicity has been associated with bioactivation through this enzyme into a highly reactive epoxide, followed by ring opening into an γ-keto aldehyde (Meyer et al. 2012). In accordance, Connelly et al. (2002) reported that 2,3-benzofuran converts to 2-(2-hydroxyphenyl)ethanal via ring cleavage, in Sprague–Dawley rats. Such toxic γ-keto aldehyde intermediates are able to form adducts with cell biomolecules, previous to their urinary excretion (Smith 2011). Previous works pertaining to unveil the metabolism of both 5-APB and 6-APB corroborate this hypothesis by showing that the first step of their metabolic pathway is indeed furan oxidative metabolism through ring cleavage (Welter et al. 2015a, b). Alternatively, the epoxide can ultimately give rise to a lactone that can also form adducts, including with CYP3A4. In fact, a number of furan-containing compounds were described to act as mechanism-based inhibitors of CYP3A4 due to the formation of covalently bonded adducts (Meyer et al. 2012). One previous study with the well-known benzofuran amiodarone and its main metabolite, desethylamiodarone, established that the metabolite caused greater loss of HepG2 viability than the parent compound, the metabolite formation was greatly dependent on CYP3A4, and the abolishment of desethylamiodarone cytotoxicity occurred upon co-exposure of amiodarone with a CYP3A4 inhibitor (Wu et al. 2016). As such, the hypothesis of 5-APB and 6-APB metabolites being more toxic than the parent drugs should not be discarded.

Both CYP2E1 and CYP2D6 inhibitors also decreased the sensitivity of hepatocytes to benzofurans, but to a lesser extent. Dinger et al. (2016) previously identified 6-APB and 5-APB as substrates of CYP2D6, but other studies attempting to establish which CYP isoforms were involved in the metabolic pathways of both drugs did not succeed due to the low rate of metabolite formation (Connelly et al. 2002; Smith 2011). In line with our data, other studies revealed that CYP2D6 inhibition reduced the hepatotoxic potential of some other synthetic stimulants, namely cathinones (Valente et al. 2016; Pedersen et al. 2013), and piperazines (Dias da Silva et al. 2015).

One important finding of our work was the formation of intracellular brown-black pigments following exposure of primary hepatocytes to the highest drug concentrations tested. These pigments had maximal absorbance in the same wavelength range of the formazans and could, therefore, influence the absorbance readings in the MTT viability assay. However, this was not problematic since their formation only occurred at concentrations for which the maximum effect in the assay had already been achieved (higher than 10 mM for 6-APB and 5 mM for 5-APB). Similar findings had already been observed for amphetamines (Dias da Silva et al. 2014b) and were attributed to the production of aminochromes, whose subsequent oxidation leads to the production of melanin-type polymers. These aminochromes are formed through catechol intermediates that oxidize into the corresponding o-quinones, which are highly redox active molecules that react with GSH, free cysteine and cysteinyl residues on proteins, resulting in significant modification of macromolecules, including proteins, lipids and deoxyribonucleic acid (DNA). Also, o-quinones can enter in oxidative cycling, increasing pro-oxidant burst and depleting cell from antioxidant defences (Dias da Silva et al. 2014a). In agreement with these facts, significant oxidative stress arose from the exposure of hepatocytes to 5-APB and 6-APB, which is also in line with the epoxide formation hypothesis. Previous studies employing non-psychoactive benzofurans such as amiodarone and benzbromarone, developed as treatment for arrhythmia and gout, respectively, showed the benzofuran ability to induce an increase in ROS, as did Nakagawa and co-workers when fresh rat hepatocytes were exposed to 5-APB and its N-methylated congener 5-MAPB (Nakagawa et al. 2017; Kaufmann et al. 2005; Waldhauser et al. 2006). Although no catechol or o-quinone metabolites have been described for 5-APB and its 6-congener, the increased oxidative stress observed following the exposure of cells to these drugs suggests that the formation of these products should not be excluded. Noteworthy, amiodarone and benzbromarone reportedly led to quinone and catechol formation through hepatic metabolism (McDonald and Rettie 2007; Parmar et al. 2016; Ramesh Varkhede et al. 2014). Importantly, the benzofuran moiety was already identified as liable for the formation of reactive quinones (Parmar et al. 2016). There is one report describing the activation of melanogenesis following exposure to a benzofuran compound, mecinarone, thus supporting the possibility of formation of melanin-like polymers following benzofury exposure (Bailly et al. 1981). To further identify the source of these polymers, additional studies should be carried out by isolating the metabolites of 5-APB and 6-APB, exposing primary hepatocytes to each of them, and ultimately characterizing the polymers and adducts formed.

The most noticeable cell defence against reactive species and electrophilic xenobiotics is GSH. The role of this pivotal antioxidant becomes even more important in the liver, a primary target for damage induced by pernicious drugs, much due to the relevance of the drug-metabolizing processes in this organ (Forman et al. 2009; Yuan and Kaplowitz 2013). Of note, amiodarone and usnic acid, another benzofuran-containing molecule, are reported in the literature to cause a diminution of glutathione in mouse hepatocytes (Han et al. 2004; Takai et al. 2016). The significant GSH depletion promoted by 5-APB and 6-APB in the present work is also corroborated by previous experiments in rat hepatocytes with 5-APB and 5-MAPB (Nakagawa et al. 2017) and with amphetamine-related drugs (Dias da Silva et al. 2014a, 2015, 2017), but this effect was not always accompanied by the complementary expected increase in GSSG levels. This can result from the efflux of GSSG into the cell culture medium, which is favoured to defend cells against the disproportionate oxidative stress (Carvalho et al. 1996; Dias da Silva et al. 2014a) but it may also be explained by the formation of conjugates of GSH with reactive metabolites, resembling what occurs with amphetamines (Carvalho et al. 1996; Hiramatsu et al. 1990; Monks et al. 2004).

Furthermore, our data also substantiate the ability of benzofurans to interfere with mitochondrial homeostasis, by impacting Δѱm and cell energetics. Hinder of ATP production had previously been reported for non-psychoactive benzofurans (Kaufmann et al. 2005; Waldhauser et al. 2006). Although the ATP decline confirms results from previous works with MDMA (Dias da Silva et al. 2014a) and synthetic cathinones (Valente et al. 2016; Dias da Silva et al. 2019), the apparent mitochondria hyperpolarization was unexpected, since a decrease of mitochondrial membrane potential often occurred for other related stimulants, as well as for 5-APB in a previous work (Nakagawa et al. 2017). However, Nakagawa and colleagues only assessed alterations to the Δѱm for 4 mM 5-APB; at this concentration level, our results also suggest a decrease of Δѱm for 5-APB. Hyperpolarization of the mitochondrial membrane was previously registered after cardiomyocyte exposure to cocaine and ethanol (Martins et al. 2018), as well as in neurons exposed to piperazines (Arbo et al. 2016). It may be possible to explain this occurrence when considering that the TMRE assay is susceptible to changes brought upon by dysregulation of intracellular ionic charges (such as calcium release from mitochondrial and endoplasmic reticulum stores) which, in turn, can cause an increase in Δѱm (Perry et al. 2011). Further assays to evaluate intracellular calcium levels of the hepatocyte would be essential to confirm whether this is also the case for benzofurans.

Mitochondrial hyperpolarization was previously described to induce caspase activation and chromatin condensation (Griffiths and Rutter 2009). In this line, mitochondria are also dynamic organelles in regulating the intracellular signalling programme that mediates apoptosis. All pathways of apoptosis converge upon the activation of proteases, such as caspase-3, which orchestrate the efficient cell dismantling. Our experimental data evidences that 5-APB and 6-APB trigger the two main pathways leading to caspase activation: the extrinsic route initiated by cell surface receptors leading directly to caspase 8 activation, and the intrinsic pathway that is regulated by mitochondria through activation of caspase-9. Both extrinsic and intrinsic apoptotic pathways converge to activate caspase-3, whose proteolytic activity leads to a cascade of events that culminate in the execution of programmed cell death. Activation of the initiator caspases elicited by both benzofurans is greater than that observed for caspase-3 activity, and 5-APB seems to more efficiently impact the activity of these proteases than 6-APB. Similarly, caspase-3 activation has been previously demonstrated for MDMA (Dias da Silva et al. 2013a; Montiel-Duarte et al. 2002), along with caspase-8 and -9 activation for cathinones (Valente et al. 2016; Dias da Silva et al. 2019) and piperazines (Dias da Silva et al. 2015) in rat hepatocytes. Apoptosis triggered by activation of initiator and effector caspases is also corroborated by the results of the Hoechst/PI staining, which show condensation of nuclei chromatin (Ormerod et al. 1993). Apoptosis (Kaufmann et al. 2005; Chen et al. 2018; Liu et al. 2016), as well as necrosis (Waldhauser et al. 2006; Han et al. 2004; McMurtry and Mitchell 1977) have been reported following benzofuran exposure in hepatic cell lines (HepG2), isolated rat and mouse hepatocytes and human chondrosarcoma cells.

The concentration-dependent LDH leakage observed for 5-APB and 6-APB had been also previously described for D-amphetamine, MDMA and a few synthetic cathinones in rat hepatocytes (Valente et al. (2016); Beitia et al. 1999; El-Tawil et al. 2011), and can be related with the progression of apoptotic events through a phenomenon designated secondary necrosis (Zhang et al. 2018). Accordingly, during apoptosis, if hepatocytes are not scavenged, they progress to a programmed lytic phase, in which extrinsic pathway-activated caspase-3 may cleave gasdermin D-related protein DFNA5, generating necrotic DFNA5 N-terminal fragment that forms oligomers and transits to the plasmatic membrane wherein it triggers the formation of non-selective pores and consequently leads to the rupture of the membrane integrity. In the case of 5-APB, the observed LDH leakage is compatible with the significant increase in activation of initiator caspase-8 and effector caspase-3, and with cell morphological features indicative of necrosis. For 6-APB, the activation of effector caspase occurs at higher levels than those verified for 5-APB, however marked necrosis in league with the LDH leakage is also at place. While a great number of early and late apoptotic cells was observed at the lower tested concentrations, necrotic events prevail at the concentrations higher than 2.15 mM for 5-APB and 2.74 mM for 6-APB, substantiating drug-induced bimodal cell death, depending on the severity of the injury.

Autophagy is an engulfing process during which obsolete and/or altered cell organelles and proteins are phagocytized and recycled, to maintain the cellular environment in good condition. The process is sparked by various stimuli such as oxidative stress (Navarro-Yepes et al. 2014) and controlled by different signalling pathways. Excessive activation of autophagy can induce apoptosis or necrosis, as well as precipitate cell function shutdown (Galluzzi et al. 2008). Autophagy was previously described following exposure to non-psychoactive benzofurans (Chen et al. 2014; Geng et al. 2018; Hsieh et al. 2015; Lin et al. 2015), but our findings represent the first report of autophagy induced by 5-APB and 6-APB. This homeostatic process was already described in neuronal cells exposed to methamphetamine (Chandramani Shivalingappa et al. 2012; Huang et al. 2017), as well as in liver exposed to alcohol (Schneider and Cuervo 2014). The activation of autophagy described herein also correlates with the observed increase in ROS and RNS, as previously described for other synthetic stimulants (Valente et al. 2017; Chandramani Shivalingappa et al. 2012; Huang et al. 2017).

In conclusion, psychoactive benzofurans are marketed without previous testing for human safety and fallaciously mentioned to be safer than classical stimulants. Notwithstanding, we demonstrated for the first time the concentration-dependent hepatotoxic effects of two benzofurans, 6-APB and 5-APB, in three in vitro models of the hepatocyte. For all models, 5-APB proved to be the most hepatotoxic drug and primary rat hepatocytes displayed highest sensitivity to the toxicity of both drugs. It also became evident that metabolism plays an important part in the toxicity of benzofury, and we hypothesize the formation of intermediate epoxide and o-quinones that may help explain the marked oxidative stress and dark-brown pigment formation in cell cultures exposed to both drugs. Disruption of redox homeostasis may have interfered with mitochondrial function, resulting in accentuated exhaustion of ATP supplies and in the enhanced activation of cell death pathways, such as caspase-induced apoptosis, necrosis and autophagy. All these toxic mechanisms overlap those of amphetamines and structurally related NPS, including synthetic piperazines and cathinones. The present findings assume great relevance from a scientific perspective, as they advance the knowledge on NPS by clarifying the toxicological mechanisms of other benzofury that may yet reach the market; and also from a clinical and forensic perspective as to assist health professionals in the event of intoxicated patients who present themselves to emergency room services after consuming these compounds.