Investigation of Ruminant Xenobiotic Metabolism in a Modified Rumen Simulation System ( RUSITEC )

The approval of plant protection agents requires xenobiotic metabolism and residue studies in rats, farm animals and plants (e.g., EU regulation 1107/2009) performed according to OECD guidelines. The intestinal metabolism of ruminants can produce specific residues, which must be investigated in detail. This is usually done by performing additional in vivo studies. The aim of the present work is to assess whether a modified in vitro rumen simulation system (RUSITEC) can generate this information. Rumen constituents from sheep were incubated over 8 days. The pH (6.70 ± 0.07), the redox potential (301 mV ± 30 mV), the microbial composition, and β-glucosidase activity were monitored. After an equilibration period of four or five days the fermenters were probed with 14C-labelled triazole derivatives, i.e., common metabolites of azole fungicides. Only triazole-alanine was cleaved to 1,2,4-triazole, while triazole-acetic acid and triazole-lactic acid remained stable for up to 96 hours. The two glucosides octyl-β-D-glucopyranoside and polydatin, which are common residues found in plants, were both rapidly cleaved in the in vitro rumen system. The data shows that the modified RUSITEC system is stable, viable and maintains metabolic capacity over a longer period of time (at least 8 days). This makes many animal experiments obsolete and represents a significant contribution to the 3Rs (refine, reduce, replace). The modification of the RUSITEC system enables safe use of unlabeled but also of radiolabeled test compounds. This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 International license (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium, provided the original work is appropriately cited.


Introduction
In the development of plant protection products extensive tests are required by the approving agencies such as e.g.European food and safety agency (EFSA) in Europe and United States Environmental Protection Agency (US-EPA) in USA, to prove the safety to the environment, animals and humans.Xenobiotic metabolism studies in rats, farm animals and plants are important to identify and quantify the active ingredient and its related degradation products, called metabolites.The active ingredient or metabolites may enter the food chain;therefore, a detailed assessment of the quality and quantity of the residues is important.The results are compared with the rat metabolism studies to determine the appropriate coverage of the relevant metabolites by toxicological studies in rats (EFSA, 2016).If not, additional toxicological studies are required.All xenobiotic metabolism studies are performed according to OECD TG 417 (rat, OECD, 2010), 503 and 505 (farm animals, OECD, 2007a,b), 501 and 502 (plants, OECD, 2007c,d).Based on the outcome of these studies, specific aspects of xenobiotic metabolism in ruminants may arise: 1) are the observed metabolites ruminant specific and formed directly in the rumen?2) are ruminants able to cleave plant specific metabolites like glycosides to the respective aglycon?These questions are commonly investigated by performing additional in vivo studies in ruminantsaccording to OECD 503 (OECD, 2007a), which are treated with radiolabeled test item, housed in metabolic cages for at least 5 days and sacrificed at the end of the study.To avoid these animal experiments in the future and contribute significantly to the 3R principle, an alternative method for replacing this type of studies was developed by adapting and improving an existing method called the rumen simulation technique RUSITEC.Ruminant metabolism differs from that of other animals and man due to their specific forestomach system (anaerobic condition) and contribution of microorganisms to e.g.cellulose degradation as reported by Czerkawski and Breckenridge (1977).A rumen simulation system (RUSITEC) is an in vitro system containing rumen, solid feed and buffer incubated under anaerobic conditions at 39°C.In the past this method was mainly used to measure methane production (Becker, 2012) or the microbial environment (Oeztuerk et al., 2005).In the present study the original RUSITEC was further refined to make it safer and simplified for routine analyses for xenobiotic metabolism, including handling of radiolabeled compounds.The modifications were performed with the aim to establish a further in vitro test system to the frame work of alternate testing, thus avoiding extensive animal experiments with of plant protection agents (Clippinger et al., 2016).For answering the first question (are the observed metabolites ruminant specific and formed directly in the rumen?) the metabolism of triazole derivative metabolites (TDM) were investigated.Azoles are developed and produced by different companies and are an important group of fungicides for farmers.These azoles are forming common metabolites so called triazole derivative metabolites (TDM) (Stroeher-Kolberg et al., 2014;FAO, 2007;TDMDG, 2016).TDMs and their xenobiotic and toxicological properties are in the focus of the agencies, so therefore additional safety studies were required (EFSA,2015, TDMDG, 2016).For answering the second question (are ruminants able to cleave plant specific metabolites like glycosides to the respective aglycon?) the cleavability of formed glucoside conjugates was investigated.Glycosides are important plant constituents (Jenner et al., 2005, Bennett andWallsgrove, 1994) and rumen contains microorganisms with β-glucosidase activity that can potentially cleave the glycosides to the respective aglycon.As a model compound to analyze the stability and enzymatic activity of the in vitro system radioactively labelled octyl-β-D-glucopyranoside was incubated.As a further non-labelled compound the metabolism of polydatin, the glucoside-derivate of resveratrol, was analyzed.

Animals
The three sheep used in this study were housed on a farm of the "LandwirtschaftlicheUntersuchungsund Forschungsanstalt (LUFA)" in Speyer, Germany.The animals were fed hay and concentrate, housed individually with contact to each other and bedded on straw.Animals were checked on daily basis on any signs of discomfort by the trained staff (>30 years experience) and inspected regularly by a veterinarian.The sheep were fitted with permanent rumen cannulas (approved by the ethics committeeat NiedersächsischesLandesamtfürVerbraucherschutz und Lebensmittelsicherheit (LAVES) Reference Number. 33.4-42502-04-13A373).The animals are used to the procedure due to regular sampling for tests on animal feed, so the sampling (performed within 5 minutes) of the rumen fluid did not stress the animals significantly.Beside the liquid rumen, solid rumen constituents, which accumulate in the cannulas were collected into pre-warmed isolated flasks (neoLab-Weithals-Isoliergefäß; neoLabMiggeLaborbedarf-Vertriebs GmbH, D-69123 Heidelberg).The liquid rumen was withdrawn by aspiration and approx.400 mL from each sheep combined into another isolated flask (neoLab-Edelstahl-Isolierflasche 1,0 l; neoLabMiggeLaborbedarf-Vertriebs GmbH, D-69123 Heidelberg).Prior to being filled, all flasks were rinsed with nitrogen to enable anaerobic conditions.The transport to the laboratory took usually maximal~30 min.However, it is known by internal studies (datanot shown), that the rumen fluid is stable also over a longer period (~3 h).

Modified RUSITEC system
The classical RUSITEC has volumes between 500 and 1000 mL (Lengowski et al., 2016;Becker, 2012;Oetztuerk et al., 2005;Czerkawski and Breckenridge, 1977).For the purpose described herein, where the metabolism of radioactively labelled compounds was to be analyzed, the system was modified to a smaller volume of 250 mL to reduce the total required amount of rumen contents.In parallel the connections within the system were made safer to correspond to the radiosafety requirements.The fermenters were provided by Institute of Physiology, University of Veterinarian Medicine Hannover, Foundation, Germany (see Fig. 1a, 1b).To this end the tubing connecting the buffer reservoir with the fermenter was changed to medical supplies and luerlock connections (ZVD-Messsystem; Sarstedt AG & Co. KG, D-51588 Nümbrecht).This enabled cheap sterile and single use tubing.The original plastic tube for the buffer was replaced by a 25 cm long steel tube soldered to a luer-lock connection.This enabled an easy and safe daily change of the nylon feed bags without the risk of radioactive spillage.(see Fig 1c).Six fermenters were placed in a water bath kept at 39.5° C by means of a thermostat (ED; JULABO GmbH, D-77690 Seelbach) and monitored and documented with a temperature data logger (EBI 300; ebro Electronic GmbH, D-85055 Ingolstadt) and an external sensor (TPC 300; ebro Electronic GmbH, D 85055 Ingolstadt) (see Figure 1a).The simulation of stomach movement was achieved with a motor so that the buffer was pumped into the fermenter at a speed of 0.

Feed
To mimic regular feed intake 2.5 g hay (DeinBestes -Alpen-Wiesenheu; dm-drogeriemarkt GmbH & Co. KG, D-76185 Karlsruhe) and 1 g pellets (sheep fodder S2; Raiffeisen-KraftfutterwerkMittelweser Heide GmbH, D-27316 Hoya) were filled into nylon bags with 50 µm mesh (In-Situ-Bags 5 × 10 cm R510; ANKOM Technology, 14502 Macedon, NY, USA) closed and stored in Nalgene bottles at room temperate until used.Prior to filling, hay was ground for 7 s at 7 000 U/min (Grindomix GM 200; RETSCH GmbH, D-42781 Haan) into particles of different sizes.The pellets were ground in a mortar.This particle distribution was intended to mimic rumination and thus the content of the rumen in an intact animal.

Liquid and solid rumen content
The liquid rumen was filtered through a viscose cloth (Tork Long-Lasting Cleaning Cloth 90477; SCA GmbH, D-85737 Ismaning).Then the pH and the redox potential of the filtrate was determined (see below) and 250 mL filled into each of four fermenters.Two nylon bags with feed were put into the fermenters.To one of these 7 g of solid rumen contents were added to inoculate with bacteria associated with solids.The fermenters were placed into the water bath and connected to the pump and buffer reservoir and the motor was imitating peristaltic set at three strokes/min.The bag containing the solid rumen constituents was changed to a fresh feeder bag after 24 h, the second one after 48 h, such that each feed bag was in the fermenter for 48 h and each day one was changed.During the entire experimental period pH and redox potential were determined in samples collected at the same time and stored at -20° C.

Incubations
To equilibrate the system a pre-incubation of four or five days was performed during which pH, redox potential, particles and colony formation, protozoa and glucosidase activity were monitored (details see below).During this phase a stable population of micro-organisms adapted to the in vitro conditions is achieved (Lengowski et al.,2016;Becker, 2012;Martínez et al., 2011).The amount of compound to be added was calculated under consideration of the loss occurring by the circulating buffer according to the following formula:

Determination of pH and the redox potential
These were determined daily directly in the cultivated rumen fluid when the fermenter was opened to change feed bags.The pH was determined with a portable pH-meter (SG2 SevenGo; Mettler Toledo, CH-8603 Schwerzenbach) and a pH-electrode (InLab Expert Pro ISM; Mettler Toledo, CH-8603 Schwerzenbach), the redox potential with a redox-electrode (LE 510; Mettler Toledo, CH-8603 Schwerzenbach) after equilibration for 1 min.

Determination of particles
Samples were analyzed every 24h with an automatic coulter counter (Moxi Z; ORFLO Technologies, 83340 Ketchum, ID, USA) in thin layer cassettes type S (MXC002; ORFLO Technologies, 83340 Ketchum, ID, USA) in the small particle mode.This setting measures particles of 2-10 µm.Rumen contains particles between 0.4 and 10 µm and some protozoas and flagellates are between 4 and 14 µm large (Breves andLeonhard-Marek, 2010, Hungate, 1966).This method should therefore register any changes in the microbial composition of the rumen, taking into account that no distinction is possible between live cells and other particles.Protozoa were counted every 24 h in a Neubauer chamber directly in 10 µL rumen and digitally recorded (AxioCamMRc; Carl Zeiss Microscopy GmbH; D-07745 Jena) and analyzed manually with the software (AxiVision V. 4.9.1.0;Carl Zeiss Microscopy GmbH, D-07745 Jena).

Analyses of metabolites
Metabolites of the triazoles and glycosides were separated by HPLC and quantified by flow through scintillation or UV detection as follows: The HPLC system for analyses of TA, TAA and TLA was composed of a Dionex Ultimate 3000 pump, a Dionex UCI-100 interface, a Dionex ASI-100 autosampler, a Dionex STH 585 oven, steered by GynkotekChromeleon V6.80 SR5 software.

Statistics
Means of the fermenters and SD were calculated with Excel 2013 (Version 15.0.4763.1003;Microsoft Corporation, 98052 Redmond, WA, USA).The automatic determination of the numbers and the size of particle, outliers were identified by Grubbs according to the statistical program Minitab (Version 17.2.1;Minitab Inc., 16801 State College, PA, USA, significance levels α=0,05).In accordance with the relevant OECD-regulations for studies on metabolism a single HPLC determination was run.

Vitality and stability of the RUSITEC system
The pH values determined every 24 h up to 192 h were stable between pH 6.52 and 6.83 with a mean of pH 6.70 (± 0.07 SD; n=66).The redox potentials measured at the same time were also stable between -240 and -345 mV, with a mean of -301 mV (± 30 mV SD; n=66).The colony forming assay to determine bacterial population showed an increase in colonies from 1.1 × 10E7 colonies / mL at the start of fermentation to 4.1 (± 0.8 SD) × 10E7 after 120 h to 4.8 (± 1.4 SD) × 10E7 after 192 h.The shape, color and structure of the colonies were unchanged as a function of time.
The concentration of particles in the liquid phase of the fermenter decreased significantly in the first 24 h from 178.6 × 10E5 particles /mL to 36.3 × 10E5 during the equilibration phase and then further decreased to a stable value of 9.2 × 10E5 particles within the following timeframe up to 192h (see Fig. 2 a).The size distribution was also conspicuously different at 0 h from the distribution at later time points (see Fig. 3).Generally small particles were more abundant than larger particles.At the start of fermentation particles of 2-3.5 µm constituted 90 %, after 24 h they constituted 70 % of all particles (see Fig. 2b).
The number of protozoa determined by counting in the Neubauer chamber decreased significantly over time from 63 × 10E4 / mL at time 0 to 1× 10E4 at 120 h.The decrease was most significant after 24h (see Fig. 3).
β-Glucosidase activity was expressed as units, where 1 unit cleaves 1 µmol 4-nitrophenol-β-D-glucoside in 1 min at pH 7.0.Glucosidase activity showed a decrease corresponding to the decrease in particles after 24 h incubation but remained constant at 4.8 units/L until the end of the experiments (see Fig. 4).The resveratrol glucoside polydatin was incubated at two different concentrations.When only 15 mg were added to the fermenter, polydatin was cleaved to resveratrol by β-glucosidase in 5 min, whereas after adding 142 mg polydatin a time dependentcleavage over 1 h was observable (see Fig 7).The metabolite resveratrol seemed stable over 6 h since no further peak with an absorption at 306 nm was observed.After 24 h no resveratrol was measurable.Here, as in incubations with the labelled compounds, no quantitative measurements were possible due to the dilution of the rumen fluid with buffer over time.

Discussion
The modified RUSITEC system was developed to facilitate incubations with radioactive compounds.The key features are a smaller volume, safe luer-lock connections and single use cheap sterile tubing.Because the conditions differ from those of the original RUSITEC system developed by Czerkawskiand Breckenridge (1977), vital parameters such as pH, redox potential and microbial populations were analyzed.The pH was as described in the literature and remained stable at pH 6.70, which is in the range of 5.0 to 7.5 recommended by Czerkawski (1986).The redox potentials equally did not vary during the experiment and remained stable between -240 and-345 mV.Descriptions of the redox potentials in rumen of ruminants are controversial.Breves and Leonhard-Marek (2010) determined values between -250 and -300 mV, while Barry et al. (1977) determined potentials in sheep of -150 to -260 mV and Marounek et al. (1982) found values between −145 and −190 mV in goats.Marden et al. (2005) describe these discrepancies in detail and explain them byusing different reference electrodes with and without corrections.A study using the same electrode as here obtained values in the same range as those measured here in our modified RUSITEC system (-170 to -300 mV; Becker, 2012).Therefore, the anaerobic conditions necessary for rumen metabolism were stable during the experimental procedures.
The determination of bacterial colonies showed an increase in colonies in the first 24h followed by a stable population over time.This is in accordance with studies with a comparable approach of measurement (Hungate, 1966;Kistner, 1960).But there are also values in the literature, which show higher numbers of 10E10 to 10E11 cells/mL fresh rumen.The reason for the much lower numbers determined after plating in culture like in this approach, are the lack of growth of many of the rumen bacteria on standardized culture plates (Hungate, 1966).As with other entero-bacteria the culture conditions of rumen bacteria are not yet developed to enable stable growth in vitro (Breves and Leonhard-Marek, 2010).Molecular biology allows analysis of these bacteria (Wang et al., 2015) and a recent publication (Lengowski et al., 2016) has shown the dynamics of several rumen micro-organisms in the RUSITEC system.These authors showed a decrease in most species in the first 24 h of cultivation, leading to a stable population at later incubation times.The loss of protozoa was most drastic with a factor of 10-50.This was also observed in our modified RUSITEC system, where protozoa decreased from 6.3 × 10E5 protozoa to 1 × 10E4 protozoa per mL rumen.The physiological range is wide between 10E5 to 10E6 protozoa per mL rumen (Hungate, 1966).Others have described also the loss of protozoa in the RUSITEC system (Lengowski et al., 2016;Becker, 2012;Martínez et al., 2010;Ziemer et al., 2000).The reason discussed for these observations are loss due to buffer circulation, their long generation time of two to three days (Fuchigami et al., 1989), or adherence of protozoa to feed particles in the nylon bags (Czerkawski and Breckenridge, 1979) thus not accessible to the liquid withdrawn.There is, however, agreement that loss of protozoa does not significantly influence metabolism in the rumen (Breves and Leonhard-Marek, 2010).
The number of particles determined by coulter counting also diminished drastically in the first 24 h.Here again the flow of buffer surely dilutes particles.In addition, the automatic system does not differentiate between inert particles and microorganisms.After 24 h a stable value was reached.The size distribution also stabilized with particles between 2 and 3.5 µm being most abundant.According to Hungate(1966) most of them are 0.4-1.0µm wide and 1-3 µm long.Larger bacteria and smaller protozoa such as flagellates which are 4 to 14 µm in size (Breves and Leonhard-Marek, 2010) will therefore constitute most of the counted microorganisms.Our data cannot fully be compared with others, nevertheless the obtained data can be used to check changes in the fermentation conditions and used as internal quality check.The size distribution with the smallest particles comprising 70 % of all particulate matter suggests that these were indeed live cells which attained a stable population during our experiments.
The β-glucosidase activity profile closely follows that of the particles in the rumen and reaches a stable value after 48 h of 4.8 (± 1.0 SD) units/l.Values in the literature are much higher but cannot be directly compared, since assay conditions are different as well as unit definitions.Agarwal et al. (2002) and Wang et al. (2015) defined 1 unit as the cleavage of 1.0 µmol substrate at pH 4.5 and 70° C. In addition, these authors lysed cells in the rumen by vortexing and a freeze-thaw cycle, thereby liberating microbial β-glucosidase.

ALTEX preprint published March 12, 2018 doi:10.14573/altex.1712221
Metabolism studies of triazole derivatives metabolites like triazole-alanine (TA), triazole-acetic-acid (TAA) and triazole-lactic-acid (TLA) (Stroeher-Kolberg et al., 2014;FAO 2007) in the modified RUSITEC system are important contributions to the investigations requested by EFSA for plant metabolites of azole-fungicides (EFSA, 2013).Our studies showed that triazole-alanine (TA) is metabolized to 1,2,4-triazole and that after 96 h half of the added TA is cleaved.The amounts incubated are however larger compared to the amounts expected to be ingested by ruminants in vivo.Analyses of TA in fodder resulted in values of 0.1 mg TA/kg wet weight (Stroeher-Kolberg, 2014).Ruminants would therefore eat less than 1 mg/ kg fodder.Sheep which eat 1-2 kg feed per day would ingest maximum of 2 mg TA into 5 l of rumen (Orpin and Letcher, 1984).In our in vitroexperiment, this would amount to 0.1 mg TA per fermenter of 250 mL, which would then probably be very rapidly metabolized to triazole.The other two test compounds triazole-acetic acid (TAA) and triazole-lactic acid (TLA) were completely stable and were not degraded (see Fig 5b and c).The selective metabolism, namely cleavage of TA and the stability of TAA was proven in vivo by OECD 505 study in cows (TDMDG, 2016).This shows clearly the predictivity of this method.No data on ruminant metabolism of TLA are currently available, neither OECD 501 nor OECD 505 studies.But, since data are requested by authorities, the current data of this publication have been submitted to the British Chemicals Regulation Directorate (CRD) (TDMDG, 2016).If finally approved, a feeding study with 30 cows according to OECDTG 505 could be avoided and significant number of animal can be replaced.
Glycosides are very common in plants (Jenner et al., 2005;Bennett and Wallsgrove, 1994) and can be formed by intrinsic plant constituents but also from exogenous compounds such as plant protection active ingredients or their metabolites.During the approval process of plant protection products, the fate of such glycosides, ingested by ruminants is requested.The metabolism of octyl-β-D-glucopyranoside as model compound and polydatin the glucoside of the plant constituent resveratrol wasanalyzed.Since β-glucosidase activity in the modified RUSITEC system was stable at a constant level after equilibration an effective cleavage of both compounds was expected.This was indeed the case: octyl-β-D-glucoside was cleaved after 5 min to a metabolite with lower polarity than the glucoside (see Fig. 6).Unequivocal identification of this metabolite as octanol was not possible, it is therefore denoted as octanol(-derivative) in the figure.After 30min the parent compound had disappeared completely and only octanol or its derivative remained.Here also the amount of substrate added to the fermenter was high to enable detection after the dilution steps caused by exchange of buffer.At physiological concentrations of a glucoside a very rapid cleavage to the aglycon can be expected.
Also, polydatin was rapidly cleaved and resveratrol was detectable after 5 min incubation of 142 mg polydatin.After 1h only resveratrol was detectable (see Fig 7).When more physiological concentrations of 15 mg polydatin were fermented only resveratrol was measurable after 5 min, showing a rapid metabolism of this glucoside under physiological conditions.Resveratrol itself seemed stable up to 6h since no further metabolite detectable at 306 nm was visible, but after 24 h no resveratrol was detectable leading to the assumption that resveratrol is indeed further metabolized in the RUSITEC system.The loss of resveratrol is higher than expected by dilution of the liquid phase.Since no radioactive polydatin was available, any further metabolism could not be investigated.Nevertheless, these data show that the system has a very high metabolic capacity and seems to well mimic the in vivo situation of ruminants.

Conclusions
The regulatory agencies require often further animal studies in ruminants to answer the fate and therefore further xenobiotic metabolismof formed metabolites of applied active ingredients to conduct a data-based risk assessment for pesticides (EFSA, 2015).The modified RUSITEC system presented herein is applicable to replace these additional requested xenobiotic metabolism studies in livestock animals for the registration process of plant protection products or veterinary drugs.The system maintains the particular properties of rumen and after a pre-incubation time is stable up to 192 h.The modification described herein has not changed the metabolic capacity described over many years for the conventional 500-1000 mL RUSITEC system.The simplification and safer handling of the RUSITEC makes it possible to use the system in routine applications and the handling of radiolabeled and unlabeled compounds.Since the selectivity of the metabolism was proven by in vivo studies, BASF will replace in vivo animal studies on ruminant metabolism studies beyond OECD 503 by performing these RUSITEC studies.The method is already included in the common method portfolio of BASF leading to a significant contribution to the 3R strategy.Further studies comparing data with in vivo observations will show the usefulness of the system and will in future prevent many animal studies and provide alternatives to animal experiments in the sense of the 3Rs.

Fig. 1 :
Fig.1: The RUSITEC system A, Schematic of the RUSITEC system; B, Picture of the RUSITEC system; C, Modified RUSITEC system

Fig. 2 :
Fig. 2: Particle density and distribution in the fermenter over time A, Particle density in the fermenter over time (x ̅ ± SD); B, Particle distribution according to size in the fermenter over time

Figure
Figure 6: Incubation of 14C-octyl-β-Dglucopyranoside in the modified RUSITEC system Incubation for 5 min (A), after 10 min (B), after 15 min (C) and 30 min (D).For HPLC conditions see Materials and Methods