The in vitro toxicity of nitrile and epithionitrile derivatives of glucosinolates from rutabaga in human and bovine liver cells

Previous evidence suggests that select nitrile and epithionitrile derivatives of glucosinolates can cause liver disease in cows grazing on brassica forage crops. A toxic incidence in New Zealand in cattle grazing brassica led us to investigate the direct in vitro hepatotoxicity and possible inhibition of the ABCG2 transporter of five nitrile compounds. In this study, we investigated 1-cyano-2-hydroxy-3-butene (CHB, epithionitrile derivative of progoitrin), 1-cyano-2-hydroxy-3,4-epithiobutane (CHEB, nitrile derivative of progoitrin), 3-butenenitrile (nitrile from sinigrin), 4-pentenenitrile (nitrile from gluconapin), and 5-hexenenitrile (nitrile from glucobrassicanapin). Cell viability was assessed following 24- and 72-hr treatments with the 5 different compounds using the MTT assay (HepG2 cells and bovine primary liver cells). Additionally, ABCG2 transporter function was assessed. The results showed that none of the tested compounds caused cytotoxicity at concentrations up to 2 mM for 24hr. Over 72-hr the maximum concentration was 20 μM but no reduction in cell viability was observed. No inhibition of the ABCG2 transporter occured at concentrations up to 1 mM. Overall this study suggests that direct or secondary toxicity due to selected nitrile or epithionitrile derivatives of these glucosinolates was not the cause of the toxic event in cattle.


Conflict of interest and funding statement:
This work was funded by a grant from AGMARDT NZ. We have secured copyright to use all images in this work.
Previous evidence suggests that select nitrile and epithionitrile derivatives of glucosinolates can cause liver disease in cows grazing on brassica forage crops. A toxic incidence in New Zealand in cattle grazing brassica led us to investigate the direct in vitro hepatotoxicity and possible inhibition of the ABCG2 transporter of five nitrile compounds. In this study, we investigated 1-cyano-2-hydroxy-3-butene (CHB, epithionitrile derivative of progoitrin), 1cyano-2-hydroxy-3,4-epithiobutane (CHEB, nitrile derivative of progoitrin), 3-butenenitrile (nitrile from sinigrin), 4-pentenenitrile (nitrile from gluconapin), and 5-hexenenitrile (nitrile from glucobrassicanapin). Cell viability was assessed following 24-and 72-hr treatments with the 5 different compounds using the MTT assay (HepG2 cells and bovine primary liver cells). Additionally, ABCG2 transporter function was assessed. The results showed that none of the tested compounds caused cytotoxicity at concentrations up to 2 mM for 24hr. Over 72hr the maximum concentration was 20 M but no reduction in cell viability was observed. No inhibition of the ABCG2 transporter occured at concentrations up to 1 mM. Overall this study suggests that direct or secondary toxicity due to selected nitrile or epithionitrile derivatives of these glucosinolates was not the cause of the toxic event in cattle. 7 1 1 Introduction 2 Background 3 In the spring of 2014 in the Southland and South Otago regions of New Zealand there was a 4 large and unprecedented outbreak of sudden deaths, photosensitization, reduced body 5 condition, increased incidence of metabolic disease, and reproductive problems in dairy cattle 6 grazing rutabaga (Brassica napus ssp. napobrassica, swedes)(1). These crops were virtually 7 weed-free and the rutabaga plants were well-grown, leafy with long stems, and were starting 8 to flower. Daily access by cattle to the crops was restricted (break feeding) according to time 9 or calculated consumption per animal. This was an unusual poisoning scenario as NZ cattle 10 routinely feed on this crop with no observed toxicity. Photosensitization was the most 11 outstanding clinical presentation of many of the cows and was secondary to liver disease as 12 indicated by elevated serum liver enzyme activities. Histopathological lesions of the liver in 13 many of the cows that died or that were euthanized showed distinctive but subtle lesions in 14 small interlobular bile ducts, variable portal fibrosis and bile duct hyperplasia, as well as mild 15 fatty change or patchy necrosis in the parenchyma (2). These lesions closely resembled those 16 seen in bulb turnip (B. rapa) photosensitization (3). Kidney lesions were variable and 17 comprised tubular dilation, cast formation, and scattered tubules showing epithelial necrosis 18 (2). Following the 2014 outbreak, an epidemiological investigation was carried out by 19 DairyNZ. Samples of flower, leaf, stem and bulb were analysed for 21 different 20 glucosinolates known to be found in swedes, turnips and rape crops. The concentration of one 21 glucosinolate, progoitrin (25 µmol/g dry matter), was 10-50 times higher than any of the 22 other 20 (<2 µmol/g dry matter) (1). 10 In addition, the pH of the medium where this degradation occurs is important; breakdown at a 11 low pH (<4) produces predominantly nitrile metabolites and at higher pH isothiocyanates (5).
12 This complexity means that it has been difficult to determine the full spectrum of biological 13 actions of the glucosinolate metabolites (6). 14 15 It has been hypothesized that nitrile and/or epithionitrile derivatives of glucosinolate 16 compounds from turnip (B. rapa), and rape (B. napus ssp. biennis) forage crops cause 17 hepatotoxicity or cholangiotoxicity in cattle (7). When crambe (Crambe abyssinica) and 18 rapeseed meals were fed to rats, bile duct and liver and renal tubular epithelial cell damage 19 resulted and this was attributed to the nitriles (8). The same lesions were found in rats fed the 20 epithionitrile CHEB from epi-progoitrin (9). Crambe seed meals have been reported as 21 having high concentrations of the parent glucosinolate epi-progoitrin; which formed the 22 nitrile 1-cyano-2-hydroxy-3-butene (CHB) and two diastereomeric isomers of the 23 epithionitrile, 1-cyano-2-hydroxy-3,4-epithiobutane (CHEB) at low pH (8). Rapeseed (B. 24 napus) meal reportedly have high concentrations of progoitrin and hydrolysis produced the 25 same daughter compounds as crambe seed but with the (R) configuration (8).

Photosensitization
10 Clinical reports from the poisoning incidence highlighted a significant degree of 11 photosensitization in poisoned animals (2). This was determined to be a secondary feature of 12 the observed liver damage. Certain types of liver damage, especially when bile ducts are 13 involved, leads to the disruption in the normal biliary excretory pathway of dietary 14 chlorophyll breakdown pigments. These pigments include pheophorbide a and 15 phytoporphyrin (phylloerythrin) (10). In the liver, the adenosine triphosphate (ATP)-binding 16 cassette (ABC) transporter G2 (ABCG2), also known as the breast cancer resistance protein 17 (BCRP), actively transports both of these pigments into the bile (11,12). This implies a 18 potential role for ABCG2 in the pathogenesis of the phototoxicity. We therefore hypothesized 19 that one or more of these nitriles could inhibit ABCG2 and therefore prevent the normal 20 biliary excretion of chlorophyll derivatives leading to photosensitization. This aspect of the 21 liver metabolism of nitriles has not been previously investigated. Bovine liver primary cells protocol 16 The protocol was a revised version of previously published protocols (21). Liver samples 17 were minced with trauma shears and washed with 20 ml antimicrobial ATB solution. Tissue 18 was homogenized in 50 ml HBSS until no large pieces of tissue remained and the sample had 19 the consistency of a thick slurry. 25 µl of collagenase II 1 mg/ml was added before the 20 mixture was moved into a cell culture hood and stirred gently for 12 min. The cells were then 21 filtered through cheesecloth and centrifuged at 1080 rpm for 5 min at 4 °C. The supernatant 22 was discarded carefully so as not to disturb the pellet and 20 ml of EGTA-PBS solution was 23 added to the tube. The tube was centrifuged again at the same settings and time. The 24 supernatant was discarded again and another 20 ml of ATB was added for the final spin. The 25 remaining pellet was re-suspended in DMEM media and a sample was taken for trypan blue 1 exclusion assay for cell viability (15). All methods were carried out in accordance with the 2 relevant guidelines and regulations. Ethical approval was not required for these studies as the 3 sample were sourced from a commercial abattoir, however the University of Otago Animal 4 Ethics Committee was informed of the work and the protocols being used in this study. 5 6 The best cell culture results followed two days of incubation allowing time for viable cells to 7 attach to the flask. Twenty-four hours after initial culture the media was discarded and fresh 8 media was put on, cells were then left for two days before seeding for MTT. Trypsin and re-9 seeding of flasks were found to reduce the number of viable cells and so flasks were seeded 10 at an initial concentration that eliminated the need for new flasks within the test period.

Cell culture and cytotoxicity assay
11 Seeding densities were determined during resuspension in DMEM with viable cell number 12 determined by the trypan blue assay. 13 14 ABCG2 vesicle transporter assay 15 As bovine (Bos taurus) ABCG2 transporters were not available for the transporter assay, an 16 alternative model had to be found. It was determined that commercial preparations of human, 17 rat, or mouse ABCG2 were available and so genetic alignments were conducted to determine  4 None of the five compounds showed any cytotoxicity by MTT assay (Fig. 1)

hour treatment of HepG2 cells
12 To extend the investigation, HepG2 cells were also exposed to compounds for a period of 13 72hr. In this model of sub-chronic exposure, doses were decreased to be more relevant to an 14 in vivo situation. Under these conditions no compound showed any cytotoxicity by MTT 15 assay (Fig. 2). There was no significant difference between the control solvent-only and the 16 treatments with the compounds. 17 18 19 Combination treatment of HepG2 cells 20 To test for possible synergism between the individual compounds, which would have been 21 ingested simultaneously, a combination treatment was performed. Each chemical was cross-22 tested in a dual exposure (equal concentrations) with each of the other chemicals and as 23 overall mixture containing all the test compounds. HepG2 cells were exposed to the mixtures 24 for 72-hr at concentrations up to 20 µM. Again, no toxicity in terms of cell death was 25 reported with any treatment (Fig. 3).  6 It is clear from the results of the ABCG2 transporter assay that none of these glucosinolate 7 derivatives inhibit ABCG2 at concentrations up to 1 mM. This is the first study to investigate 8 the effects of these compounds on the ABCG2 transporter. As high levels of ABCG2 is 9 closely correlated with multi-drug resistance in cancer cells, numerous studies have 10 investigated the structural requirements for inhibition of this protein. The protein structure of 11 the ABCG2-FTC complex has recently been published (19). This shows that FTC binds into 12 the active site (competitive inhibition) and prevents conformational changes required for the 13 transportation activity (19). It is presumed that this is the primary mode of inhibition due to 14 the fact that the most potent ABCG2 inhibitors contain several key structural similarities 15 which resemble the FTC molecule (20). To date, there is little evidence that the presence of a 16 nitrile or epithionitrile species alone is predictive of ABCG2 inhibition. However, the current 17 study did not include an evaluation of the nitrile/epithionitrile metabolites of the 18 glucosinolate indole-3-carbinol (21) which have a structural backbone more closely 19 resembling known ABCG2 inhibitors. The evaluation of these compounds against ABCG2 20 activity may be an avenue for further study.

22
Conclusions 23 The results of this study indicate that direct liver cell toxicity or the inhibition of ABCG2 24 transporters in the liver by nitrile or epithionitrile derivatives of progoitrin and three other 25 glucosinolates was not the likely cause of the cattle deaths, photosensitivity or liver disease in 1 the poisoning outbreak in New Zealand in 2014. We have been unable to show any evidence 2 of in vitro toxicity of CHB, CHEB, 3-B, 4-P, or 5-H. This suggests that toxic mechanism is 3 something that we are unable to replicate in vitro or alternative metabolites are responsible 4 for the toxicity.