Identification of a novel human deoxynivalenol metabolite enhancing proliferation of intestinal and urinary bladder cells

The mycotoxin deoxynivalenol (DON) is an abundant contaminant of cereal based food and a severe issue for global food safety. We report the discovery of DON-3-sulfate as a novel human metabolite and potential new biomarker of DON exposure. The conjugate was detectable in 70% of urine samples obtained from pregnant women in Croatia. For the measurement of urinary metabolites, a highly sensitive and selective LC-MS/MS method was developed and validated. The method was also used to investigate samples from a duplicate diet survey for studying the toxicokinetics of DON-3-sulfate. To get a preliminary insight into the biological relevance of the newly discovered DON-sulfates, in vitroexperiments were performed. In contrast to DON, sulfate conjugates lacked potency to suppress protein translation. However, surprisingly we found that DON-sulfates enhanced proliferation of human HT-29 colon carcinoma cells, primary human colon epithelial cells (HCEC-1CT) and, to some extent, also T24 bladder cancer cells. A proliferative stimulus, especially in tumorigenic cells raises concern on the potential impact of DON-sulfates on consumer health. Thus, a further characterization of their toxicological relevance should be of high priority.

The trichothecene deoxynivalenol (DON, vomitoxin) is a frequent contaminant of grains and cereal products world-wide. Since DON constitutes a major issue for food and feed safety, different international expert bodies, including those of the FAO/WHO and EFSA, extensively evaluated its occurrence, exposure, metabolism, and toxicity [1][2][3] . As a result, regulatory limits were introduced in many countries to manage the concentration of DON in food and feed 4 and a provisional maximum tolerable daily intake (PMTDI) for DON and its acetylated metabolites of 1 μ g/kg body weight was established 1 . Exposure to DON was clearly associated with the consumption of cereals 5 . Recent surveys, applying innovative LC-MS/MS based biomarker approaches, revealed that significant parts of several European populations exceeded the PMTDI in various years [6][7][8][9][10] . In humans, DON has been associated with gastroenteritis, whereas in animal models acute DON intoxication causes emesis and chronic low-dose exposure elicits anorexia, growth retardation, immunotoxicity as well as impaired reproduction 11 . Although chronic exposure is evident globally, the effects of low-dose DON exposure on humans are still unknown.
The primary mode of DON action is the efficient inhibition of protein synthesis by binding to eukaryotic ribosomes 12 . Thereby, the synthesis of macromolecules as well as cell signaling, differentiation, and proliferation are impaired. However, DON also activates intracellular protein kinases which mediate selective gene expression and apoptosis 11 . DON has been reported to inhibit several intestinal transporters in the human epithelial intestinal cell line HT-29-D4 while in Caco-2 cells it was found to induce IL-8 secretion 13 . In the human Jurkat T-cell line the induction of oxidative stress was recently confirmed by studying the nuclear translocation of the transcription factor NRF2 and its binding protein KEAP1 as well as by changes in cell levels of reduced glutathione 14 .
It is known since a long time that DON is extensively metabolized to glucuronide conjugates (DON-GlcA) as the predominant products of phase II metabolism in animals 15 . However, the first assay to measure DON and its glucuronides indirectly using enzymatic hydrolysis in human urine was developed by Meky et al. 16 only a decade ago. During the last years, the structures of these conjugates in human urine have been identified with DON-15-GlcA as the major metabolite and minor contributions of DON-3-GlcA and DON-7-GlcA 6,7,17 . The overall 24 h urinary excretion rate of total DON (i.e. the sum of DON and its glucuronides) was estimated to be on average 72% in a moderately exposed UK population 18 . In the cited study β -glucuronidase from E. coli (Type IX-A), which is typically free of sulfatase activity, was employed. This estimate was confirmed in other studies either utilizing direct quantification of glucuronides by LC-MS/MS 19 or enzymatic hydrolysis and GC-MS instrumentation 20 . Also the bacterial detoxification product deepoxy-DON (DOM-1) was found in lower numbers and concentrations in some studies after enzymatic hydrolysis 21,22 or via a direct approach 10,23 . To the best of our knowledge, a DON-sulfate conjugate has not been reported as a human metabolite before. However, literature reports of a tentatively identified DON-sulfate conjugate in sheep urine based on an indirect approach using enzymatic de-conjugation with sulfatase 15 and samples obtained from chicken tissues 24 were published. Furthermore, Schwartz-Zimmermann et al. 25 demonstrated DON-3-sulfate as the major DON metabolite in different poultry species and the formation of DOM-3-sulfate. Very recently, DON-3-sulfate and DON-15-sulfate were also unambiguously identified as plant metabolites formed in DON treated wheat 26 utilizing chemically synthetized reference standards 27 for structure confirmation and absolute quantitation.
Based on the formation of DON-sulfates as phase II metabolites in animals, we tested the hypothesis that DON may be converted into a sulfate conjugate in humans as well. Hence, we developed a highly sensitive LC-MS/MS method for the direct quantification of DON and its urinary metabolites including DON-sulfates and applied it to two sets of urine samples which have been well-characterized before. We present experimental evidence for the existence of DON-3-sulfate in human urine, which has not been described as a human metabolite of the major trichothecene DON before. Furthermore, we performed a preliminary toxicological characterization of the DON-sulfates which unraveled potential implications on cellular growth.

Results
Identification of DON-3-sulfate as novel human metabolite and potential biomarker. As illustrated in Fig. 1, DON-3-sulfate was detected in human urine and identified based on comparison with authentic reference standards which have been chemically synthetized and confirmed by NMR before 27 . The retention time as well as the intensity ratio of the selected reaction monitoring (SRM) transitions and the MS/MS spectra identified the detected metabolite as DON-3-sulfate. Whereas glucuronide formation in humans mainly occurs at C-15, sulfates are bound predominantly to the C-3 carbon. Interestingly, no DON-15-sulfate was identified in any of the investigated samples in this study. This means that the unknown human sulfotransferases 28 , mediating conjugation of DON, seem to follow a different stereoselectivity than the involved UDP-glucuronosyltransferases 29 . This is to the best of our knowledge the first report of a DON-sulfate metabolite in any human sample material. Since DON-3-sulfate was only determined in artificially DON-treated wheat but not in any naturally contaminated food sample intended for human consumption and the transfer via chicken meat or eggs 24,25 seems highly unlikely, we propose that the identified conjugate is an endogenous human metabolite produced in the intestine or liver. Natural occurrence and excretion rate of DON-3-sulfate in human urine. To investigate the natural occurrence of DON-sulfates, first morning urine samples obtained from Croatian women (n = 40) were analyzed by the newly developed LC-MS/MS based method. DON-3-sulfate was quantified in 28 out of the 40 urine samples (70%). The maximum concentration was 58 μ g/L while the average concentration was 4.5 μ g/L (0.012 μ M), when for samples below the limit of detection (LOD) the half LOD was deployed for average calculation. As mentioned above no DON-15-sulfate was detected in any sample.
Besides the investigation of the natural occurrence of DON-sulfates in human urine, the method was also utilized to re-investigate urine samples from an eight-day duplicate diet survey 19 . This study has been designed initially to unravel the toxicokinetics of DON in vivo especially focusing on the formation of glucuronide conjugates. The DON-3-sulfate metabolite was determined in this set of samples frequently as well and its urinary 24 h excretion rate was estimated to be approximately 4% of the DON quantity ingested through the contaminated food ( Table 1). The fast elimination of the sulfate conjugate was verified by its absence in the urine sample obtained on day seven, the first day after the consumption of DON contaminated food was stopped.

LC-MS/MS method development and validation. The MS/MS parameters of DON-sulfates as well
as the other analytes ( Table 2) included in the method were optimized in both, the positive and the negative ESI mode. All analytes investigated in this study yielded higher absolute signals and better signal to noise ratios in the negative ionization mode. To differentiate between the two isomers the fragment ion at m/z 345 (− 30 amu) was used. This corresponds to [M-CH 2 O-H] − with a loss of CH 2 O from the -CH 2 OH group attached to the carbon at the C-6 position of the DON-3-sulfate as described before 26 .
The eluents were optimized in order to maximize the retention, recovery and signal to noise ratio of all analytes, however, DON-sulfates were regarded as the most relevant targets. One important objective was to chromatographically baseline separate the DON-sulfate and DON-glucuronide isomers. This task was successfully accomplished by careful optimization of the mobile and stationary phases. Acidified methanolic eluents and the same stationary phase with biphenyl chemistry have been reported recently to exhibit excellent separation of DON and its polar conjugates 25 . Since higher concentrations of acetic acid resulted in decreased signal intensities only a low concentration (0.05%) was chosen for the final method.
The proposed method was validated thoroughly to estimate the linear range, matrix effects, intra-and interday precision, selectivity, as well as the LOD and limit of quantification (LOQ) values. Detailed results are presented in Supplementary Table 1. The method proved to be linear over three orders of magnitude when measuring reference standards in pure solvent. It has been reported before that DON and its polar conjugates are prone to severe matrix effects in biological samples 23,30,31 . Interestingly, DON-sulfates have been described being susceptible to signal enhancement rather than ion suppression during electrospray ionization in samples derived from animal material 25 and wheat samples 26 . This behavior was confirmed in human urine in this work albeit in a less pronounced manner with acceptable and very stable apparent recoveries ranging from 107-111% and 114-117% for DON-3-sulfate and DON-15-sulfate, respectively. Also the intra-and interday precision with relative standard deviations of 6-15% and 5-12%, respectively can be regarded as acceptable when taking the fast and effective sample preparation and the challenging biological matrix into account. The obtained LODs (DON-3-sulfate: 0.45 μg/L; DON-15-sulfate: 0.35 μg/L; see Supplementary Table 1) were judged to be applicable to quantify even low DON exposures. The retention times were stable with a maximum shift of less than 1.2% for DON-sulfates which is typically regarded as acceptable for LC separations. Overall, the results clearly indicated that the chosen 'dilute and shoot' approach was feasible and did not require any further sample clean-up or enrichment step.

Effect of DON and its sulfates on the translation efficiency in mammalian cells.
Since the primary mode of DON and trichothecene action is the inhibition of protein biosynthesis by eukaryotic ribosomes, we tested Interestingly, the two sulfate conjugates demonstrated a marked proliferative stimulus on HT-29 colon carcinoma cells in a concentration range between 0.1 and 25 μ M (Fig. 3a). This effect was confirmed in HCEC-1CT and T24 cells albeit less pronounced (Fig. 3b,d) while it was not significant in Caco-2 cells (Fig. 3c).

Discussion
This is to the best of our knowledge the first report of a DON-sulfate metabolite in any human sample. Based on the chromatographic retention behavior and the MS/MS spectra displayed in Fig. 1, the isomer occurring in human urine was identified as DON-3-sulfate. In principle, also the formation of DON-7-sulfate might be possible. However, the unreactivity of the C7 position to chemical sulfation has been demonstrated before 27   is unlikely that a potentially occurring DON-7-sulfate, for which no reference standard is available yet, co-elutes with DON-3-sulfate under the tailored chromatographic conditions and shows the same MS/MS spectrum. DON-3-sulfate was found to be present in 70% of the investigated samples obtained from Croatian women with a high maximum concentration of 58 μ g/L, corresponding to 0.15 μ M. In addition, it was detected frequently in a set of samples from an in vivo toxicokinetics study utilizing urine samples obtained from a male Austrian volunteer. Thereby, the urinary 24 h excretion rate was estimated to be approximately 4% of the DON quantity ingested through the consumption of contaminated food (Table 2). This likely indicates that sulfation is a minor metabolic pathway compared to glucuronidation 19,32 , although the fraction of DON-3-sulfate excreted in the bile was not estimated. However, the contribution in human urine is higher than the 2% reported for sheep 15 . Besides, it might be possible that sulfates but not glucuronides are transferred through the cell membrane by specific transporters. In the investigated population sulfation was more relevant than de-epoxidation as no DOM-1 was detected in any sample. DOM-1 was first demonstrated in the urine of French farmers, representing on average < 5% of the total urinary DON in individuals with detectable DOM-1 levels 21 . Since then it was demonstrated in a limited number of studies mainly in its glucuronide form 10,22,23 .
The sulfotransferases responsible for mammalian xenobiotic metabolism are cytosolic enzymes forming a gene superfamily. Differences in substrate specificity between the different sulfotransferases can be relevant for tissue-specific toxicological effects 28 . Ten distinct human sulfotransferase forms are known, however, currently it is unknown which gene product is mediating the conjugation with DON. This information would also be of relevance since the distribution of sulfotransferases may strongly differ between tissues. As one example hP-PST (human phenol sulfotransferases) exhibit high expression levels in the liver while it is detected typically in lower levels in other tissues.
Meky et al. 16 reported that rat urine incubated with sulfatase resulted in no change of DON related chromatographic peaks. Hence, in the past most bio-monitoring studies focusing on the indirect quantification of DON employed β -glucuronidase from E. coli, which is essentially free of sulfatase activity 7,18,33 . Based on the identification of DON-3-sulfate in this study the use of β -glucuronidase/sulfatase from from Helix pomatia is recommended for future studies as already described by some groups [34][35][36] .
While DON-3-sulfate does not inhibit in vitro protein synthesis at concentrations up to 100 μ M (Fig. 2) DON-15-sulfate was found to be a moderate inhibitor of mammalian ribosomes with an IC 50 of about 47 μ M. The inhibition observed in this experiment is slightly lower than that observed on wheat ribosomes, where the IC 50 of DON-15-sulfate was about 66 μ M 26 . When compared to DON, these figures demonstrate that DON-sulfates can be regarded as detoxification products with respect to their effect on protein translation.
In order to enable a preliminary characterization of the biological activity of the DON-sulfates in comparison to the parent compound DON, additional cytotoxicity experiments were performed. Four cell lines derived from the intestinal tract (HT-29, HCEC-1CT, Caco-2) and from the urinary bladder (T24) were selected to give a comprehensive overview. In agreement with the effect of the three compounds on mammalian ribosomes (Fig. 2), DON was cytotoxic in all the tested cell types while DON-3-sulfate did not exert any toxic effect and DON-15-sulfate showed only a limited effect in T24 cells. Intriguingly, when incubated in the nanomolar range DON triggered a proliferative stimulus in the cells of cancerous origin used in the present study.
In addition, the two sulfate metabolites demonstrated a distinct proliferative stimulus on human colorectal adenocarcinoma HT-29 cells over a concentration range between 0.1 and 25 μ M. This effect was present, even if . This is of particular interest since this concentration range seems to be coherent with the concentration of the urinary DON metabolites that can be found also in the bladder in vivo as suggested by the urinary concentrations reported in this paper. Taking into account that several recent bio-monitoring studies reported on individuals exceeding the proposed PMTDI established for DON [6][7][8][9][10]37 and the frequent occurrence of DON-3-sulfate in the urine of exposed individuals in the study at hand, this highlights the urgent need for further studies and a deeper toxicological characterization of DON-sulfates. It should also be considered that we were able to demonstrate the capacity of wheat plants to form both, DON-3-sulfate and DON-15-sulfate conjugates in a previous study 26 . In a wheat suspension culture additionally 15-acetyl-DON-3-sulfate was reported very recently 38 . Hence, it seems plausible that this new class of masked/modified mycotoxins might enter the body via contaminated food in addition to the proposed endogenous production of DON-3-sulfate in the human body.
A proliferative effect of DON on tumor cells at very low concentrations has been reported for the parent compound DON in recent experiments as well 39,40 . However, according to literature and confirmed by our data this effect disappears once the cytotoxicity of DON outstrips the growth stimulus at a concentration of 1 μ M. The sulforhodamine B (SRB) assay applied in this work measures the cellular protein content and is a standard assay of the National Cancer Institute for in vitro anticancer-drug screening. It provides a sensitive measure of cytotoxicity induced by drugs or xenobiotics and is frequently used to quantify clonogenicity 41 .
DON has been considered to be non-carcinogenic (Group 3) by the International Agency for Research on Cancer for a long time 42 . Even though the impact of DON and its metabolites on the growth of different cells types remains to be clarified with respect to the mechanisms sustaining it and future risk assessment, this is the first report on DON metabolites which potentially promote the cellular growth at concentrations occurring in vivo due to widespread chronic exposure.
In summary, we demonstrated for the first time that DON-3-sulfate is a human metabolite of the abundant food contaminant DON. Using a newly developed, highly sensitive and selective LC-MS/MS method this new potential biomarker was quantified in the majority of tested urine samples. To evaluate the potential consequences of this unexpected finding for consumers of mycotoxin contaminated food, preliminary toxicological testing was performed. Interestingly, and maybe of high importance for public health and future DON risk assessment, it was found that the DON-sulfates can trigger cellular proliferation in vitro in a concentrations range that seems to be relevant in vivo as suggested by the obtained urinary concentrations of DON-3-sulfate. Urine samples. The samples used in this study originated from two different experiments. To generally investigate the occurrence of DON-sulfates in a population exposed to high levels of DON, first morning urine samples obtained from Croatian women (n = 40) were utilized. Volunteers were all healthy, non-smoking pregnant women in their final trimester of gestation who resided in the eastern area of Croatia (from and around the city of Osijek; age: 26-33 years old). These samples have previously been tested on multiple mycotoxin biomarkers using an advanced LC-MS/MS method 6 as well as on ochratoxin A and ochratoxin alpha using HPLC-FLD 44 . They partly exhibited high concentrations of DON (max. 275 μ g/L), DON-3-GlcA (max. 298 μ g/L), and DON-15-GlcA (max. 1238 μ g/L) and are thus ideally suited to screen for novel metabolic products of DON. Samples were taken in February 2011 and stored at − 20 °C until analysis. Informed consent was obtained from all participants. The study was approved by the Ethics Committee of the Faculty of Food Technology, University Josip Juraj Strossmayer Osijek and the measurements were carried out in accordance with the approved guidelines.
The second set of samples originated from an in vivo case study which investigated human DON and zearalenone (ZEN) metabolism in detail through the analysis of urine samples obtained from one Austrian volunteer following a naturally contaminated diet containing 138 μ g DON and 10 μ g ZEN over a period of four days 19 . Sulfate conjugates were not included in the original study due to a lack of an authentic reference standard at that time. The study was conducted on a 27 year old, healthy male volunteer whose diet consisted of cereals with wheat bran for breakfast, maize porridge (including maize flour) for lunch and bread, beer and pop-corn in the evening as described in detail before 19 . For calculating average DON-sulfate excretion rates 24 h urine samples were used in the study at hand. Samples were taken in July 2011 and stored at − 20 °C until analysis. This study was approved by the ethics commission of the government of Lower Austria. Measurements were carried out in accordance with the approved guidelines after informed consent was obtained.
Sample preparation. The time-and cost-effective sample preparation procedure chosen was based on a protocol for the simultaneous quantification of multiple mycotoxins and metabolites 45 . In brief, samples were allowed to reach room temperature, centrifuged for 3 min at 10.000 rpm, and diluted 1:10 with a neat dilution solvent (ACN/H 2 O: 10/90).

LC-MS/MS instrumentation.
Method development, validation, and sample analysis was carried out using a Thermo TSQ Vantage LC-MS/MS triple quadrupole system (Thermo, San Jose, CA, USA) coupled to an Accela 1250 LC system. Data acquisition was performed using the Xcalibur software (version 3.0) whereas the evaluation of data was done using LCquan (version 2.9). The mass spectrometer was equipped with a heated electrospray (hESI) interface which was operated in negative ionization mode. Nitrogen was used as drying and argon as collision gas. The parameters of the ion source are reported in Supplementary Table 2.
Analytes were separated on a Kinetex Biphenyl column (3.0 × 150 mm, Phenomenex, Torrance, CA, US) with 2.6 μ m particle size and a SecurityGuard ULTRA pre-column (Phenomenex). Gradient elution at 40 °C was performed within 17 min. Eluent A (H 2 O/MeOH; 9/1) and eluent B (MeOH) both contained 0.05% acetic acid and the flow rate was set to 400 μ L/min. After an initial time period of 1.0 min at 100% A, the percentage of B was linearly raised to 16% until minute 10.0. Then, eluent B was raised to 95% until minute 12.0 followed by a hold-time of 2.0 min and subsequent 3.0 min column re-equilibration at 100% A. A volume of 10 μ L of the diluted samples corresponding to 1 μ L undiluted urine was injected. ESI-MS/MS was performed in selected reaction monitoring (SRM) mode for all analytes investigated in this study. At least two individual transitions were monitored for each analyte. Analyte dependent MS/MS parameters were optimized via direct infusion of reference standards. Quantification of all analytes for which reference standards were available was done by external calibration curves (1/x weighted) as described in the validation section below and all results were corrected for the apparent recovery of the respective analyte.
Two QC samples were included in each batch of 20 samples within an LC-MS/MS measurement sequence. One was the same pooled blank urine used during validation while the other was the blank urine spiked with working standard solution. The results of the spiked QC sample required to be within 15% of the assigned values. In case of non-accordance the whole sequence was rejected for the affected analyte.

Method validation.
In-house validation of the developed method was carried out to determine the parameters linear range, precision, recovery, selectivity, and sensitivity. Intra-and interday precision as well as the apparent recovery of analytes were evaluated by measurements of a pooled blank urine sample spiked with the working solution at three concentration levels: Low (3 μ g/L), middle (30 μ g/L) and high (300 μ g/L). When taking the urine dilution into account this corresponds to 0.3, 3, and 30 μ g/L, covering a wide range of concentrations. Spiking experiments were performed in triplicate and on three different days. Intra-(n = 9) and interday (n = 27) precision were expressed as the relative standard deviation of the obtained recoveries for each metabolite. The selectivity of the chosen product ions was evaluated throughout method development and validation and was Scientific RepoRts | 6:33854 | DOI: 10.1038/srep33854 continued during the application of the method to experimental samples. LOD and LOQ values were calculated from chromatograms of spiked blank urine samples based on a signal to noise ratio of 3:1 and 6:1, respectively. Calibration curves (1/x weighted) were constructed from peak areas of the reference standards in solvent plotted against their concentrations. Each calibration was carried out at seven concentration levels covering three orders of magnitude. The calibration range was 0.1-100 μ g/L for all analytes.

Effect of DON and its sulfates on the translation efficiency of mammalian cells. The TnT ® T7
Coupled Reticulocyte Lysate System (Promega, Madison WI, USA) was used for in vitro translation experiments. The assays were performed as described before 46 with minor modifications. Six independent experiments were performed. All data were tested on normality by the Shapiro Wilk test. Different doses of DON and DON-sulfates were tested on significant differences to the water control by One-Way ANOVA. Significant differences between selected DON and DON-sulfate concentrations were tested by Student's t-test.

Effect of DON and its sulfates on cell viability (sulforhodamine B assay).
In order to provide an initial characterization of the biological effects of the DON-sulfates at cellular level SRB experiments were performed as described previously 41,49 . HT-29, HCEC-1CT, Caco-2 and T24 cells were seeded in 96-wells plates and incubated with different concentrations of the parent compound or the metabolites for 24 h. At the end of the incubation cells were rinsed twice with PBS (100 μ L) and fixed for 30 min at 4 °C with 50 μ L of 50% trichloroacetic acid (TCA) per well. To remove TCA cells were repetitively rinsed with water and 100 μ L of SRB reagent (SRB 0.4% in 1% acetic acid) were added to each well. After 1 h of incubation, stained cells were rinsed with acetic acid (1%) and water to remove the unbound SRB and, subsequently, the protein-bound SRB was solubilized with 100 μL Tris (10 mM). Single wavelength absorbance (570 nm) was read on a Cytation 3 Imaging Multi Mode Reader (BioTek, Bad Friedrichshall, Germany). Results are presented as mean of at least three independent experiments performed in quadruplicate ± SE and analyzed applying the Kruskal-Wallis-ANOVA test with OriginPro software (version 9.1). Cellular metabolism. HT-29 cells were seeded in 24-well plates (50.000 cells per well). After 48 h cells were incubated in triplicate with 10 μ M of either DON, DON-3-sulfate, or DON-15-sulfate for 0, 3, 24, and 48 h. After the respective time the supernatant and the cell lysate were analyzed separately to evaluate if deconjugation occurred under cell culture conditions. The supernatant was diluted 1:1 (v/v) with MeOH, centrifuged (18.000 rpm, 5 min) and 10 μ L of the diluted supernatant were injected into the LC-MS/MS system. The HT-29 cells were washed with ice cold PBS and detached from the culture plates with 100 μ L trypsin. After addition of 100 μ L ice cold MeOH the cells in suspension were disrupted by shock freezing in liquid nitrogen twice. The cell lysate was centrifuged (18.000 rpm, 10 min) and 100 μ L of the supernatant were transferred to an autosampler vial with micro insert and subsequently analyzed by LC-MS/MS.

Ethics statement.
All experiments involving human urine samples were approved by the responsible ethics commission. The study involving Croatian samples was approved by the Ethics Committee of the Faculty of Food Technology, University Josip Juraj Strossmayer Osijek whereas the Austrian study was permitted by the ethics commission of the government of Lower Austria.