IN VITRO ASSESSMENT OF BINDING CAPACITY OF COMBINED ADSORBENT (BENTONITE WITH YEAST CELL WALL EXTRACTS) AND AFLATOXIN B1

Th e contamination of animal feed with mycotoxins is a worldwide problem in the animal husbandry, but it also represents a serious threat for the whole food chain. Th e health of both animals and humans is potentially endangered. From this point of view afl atoxins are a class of mycotoxins especially well known. Th erefore, new strategies to combat these natural contaminants are constantly being developed. Th e most applied method to protect animals against afl atoxicosis is the utilization of feed additives aimed to adsorb afl atoxins. In order to estimate adsorbing potential of feed additive “MycoStop DUPLO”, designed for the prevention and/or alleviation of adverse eff ects of afl atoxin B1 in animal nutrition, in vitro trial was conducted. As a result of the experiment, conducted at pH 5 during 120 minutes of incubation at 37°C, the optimal formulation of the adsorbent was revealed. Th is product, in low concentration and in the presence of high amounts of toxin, met the stringent European regulation requirements for minimum 90% afl atoxin binding effi ciency (90.1% achieved with 0.02% adsorbent and 4 mg/L toxin concentration). In higher adsorbent (0.2%), and lower toxin (0.2 mg/L) conditions, adsorption was 99.6%. Such outcome indicated the validity of in vitro experimental approach which can serve as a reliable fast tool for triage of adsorbents and preselect them for in vivo tests.


INTRODUCTION
Mycotoxins contaminate food chain through food and feed crops, mainly cereals, which become infested prior to and during harvest, or during (improper) storage. Th ey are produced as secondary metabolites of diff erent types of fungus under the favourable environmental conditions, when temperature and moisture are appropriate. Climate changes during the last decade in particular contributed to the escalation of this problem (Nešić et al., 2014;Nešić, 2018;Jakšić et al., 2017Jakšić et al., , 2018Jakšić et al., , 2019. Afl atoxins are strong (Class I, IARC, 2002) carcinogens in mammalian species, difuranocoumarin derivatives produced by diff erent species of Aspergillus (Aspergillus fl avus, Aspergillus parasiticus, Aspergillus nomius and Aspergillus pseudotamarii). Several types of afl atoxin (14 or more) are found in nature, and B 1 , B 2 , G 1 , G2 and M 1 are of major importance. Afl atoxins B 1 , B 2 , G 1 and G2 are direct secondary metabolites of fungi, whereas afl atoxin M1 is produced by metabolizing afl atoxin B 1 (AFB 1 ), which is usually a major product of toxigenic strains (WHO, 2018). Th e presence of mycotoxin in feed results in huge economic losses for animal breeders caused by decreased performance and production, increased susceptibility to diseases and other adverse eff ects (Rawal et al., 2010).
Since the mycotoxins have an important impact, there is a continuous effort to develop various ways to alleviate and/or prevent their harmfulness. Th e approach by using diff erent feed additives, which either adsorb mycotoxins on their surface or foment enzyme degradation of mycotoxins proved to be particularly eff ective (EFSA, 2009;Nedeljkovic-Trailovic et al., 2015). Th e output depends mostly on the chemical structure of the adsorbent, as well as on the type of present mycotoxin. Mineral adsorbents (e.g. hydrated sodium calcium aluminosilicate, sodium bentonit, dietary clay and zeolites) and active charcol are among the most used for this purpose. Th ese are the substances that are not resorbable from the gut and that physically bind target chemicals and consequently block their resorption (Nešić et al., 2014). Th e feasibility of utilizing organic adsorbents has also been examined, particularly those isolated from the yeast cell wall that possess signifi cant adsorption capacity (Devegowda et al., 2004;Nešić et al., 2008). Recently, some new types of additives which contain microorganisms have been developed. Th ey have the ability to enzymatically modify the mycotoxin structure (Fuchs et al. 2002;Nešić et al. 2011Nešić et al. , 2012. Th e ability of diff erent adsorbents to ameliorate afl atoxin B 1 toxicity was tested in in vitro and in vivo conditions and the fi ndings mostly correlated (Vekiru et al., 2015). Supplementation of diets with selected adsorbents, especially of the bentonite type, seems to almost fully protect animals against afl atoxicosis, so the EFSA Scientifi c Report gives an actual and comprehensive overview on this topic (EFSA, 2009). Bentonites are composed predominantly of smectite. However, a wide variety of other minerals may occur as impurities. Th e dioctahedral smectite mineral montmorillonite is present in most bentonites. Depending on the dominant exchangeable cations, bentonite may be referred to as calcium or sodium bentonite. Sodium bentonite swells and expands to a greater degree than its calcium equivalent. Calcium bentonite may be converted to sodium bentonite, then termed sodium activated bentonite. Th e type of the cation on the surface of the aluminium sheet (Ca or Na) may aff ect the binding capacity of the montmorillonite (EFSA, 2011).
Besides its excellent nutritional value, yeasts and yeast cell wall can also be used as adsorbents for mycotoxins. Th e adsorption of mycotoxins can be enhanced by using yeast cell walls instead of whole cells. Th e cell walls harbouring polysaccharides (glucan, mannan), proteins and lipids exhibit numerous diff erent and easily accessible adsorption centers including diff erent adsorption mechanisms, e.g. hydrogen bonding, ionic, or hydrophobic interaction (Huwig et al., 2001). Regarding polysaccharides, including β-D-glucan and α-mannan, it has been proposed that their antigenotoxic action mechanism is related to their action as antioxidant agents (Pereyra et al., 2012). Th e ability of β-D-glucan to partially prevent DNA damage induced by AFB 1 in mouse hepatocytes was determined in a trial (Madrigal-Bujaidar, 2015). Th e data suggested the formation of a supramolecular complex between AFB 1 and β-Dglucan. Mannan oligosaccharide is a potent immunomodulator which alleviates the damages of AFB 1 (Sun et al., 2019).
Th e aim of the presented in vitro trial was to estimate adsorbing potential of "MycoStop DUPLO". It is a feed additive which combines bentonite and yeast components and is intended for prevention and/or alleviation of adverse eff ects of afl atoxin B 1 in animal nutrition.

Physico-chemical characterization tests
Th e physico-chemical properties of adsorbents were examined as moisture content, acidity and swell index. Moisture content in adsorbents was determined by drying in oven (Memmert UNB 500, Germany) at 105°C to constant mass. Th e acidity of the adsorbent samples was measured in 1:10 adsorbent: water suspension (De Mil et al., 2015). Th e suspensions were shaken for 2 h and were left to sediment for next 2h under closed lid. Th e pH of the supernatant was measured using pH meter (Consort, Turnhout, Belgium). For determination of the swell index, the ASTMD5890, 2011 method was used.

In vitro experiment design
Th e assessment of afl atoxin B 1 adsorption capacity was accomplished in accordance with the Regulation (EU) No 1060/2013 (European Commission, 2013), an approved method for the evaluation of bentonites authorized as feed additives against AFB 1 . Th e test was carried out in a buff er solution at pH 5.0, at 37°C, for 120 minutes, with a concentration of 4 mg/L for AFB 1 and 0.02% (w/v) for the adsorbent (phase I). Th e best performing adsorbents from the phase I were examined in the second phase of the experiment. In this phase, II binding capacity was investigated using a standard solution of 0.2 mg/L AFB 1 and the adsorbent in the concentration of 0.2% (w/v; Prapapanpong et al., 2019). All the tests were done in triplicate.
Aft er incubation (shaking for 2 h at 37°C), samples were fi ltered (syringe fi lters 0.22 μm; LLG-Labware, Meckenheim, Germany) and the solution was analyzed by an HPLC Dionex UltiMate 3000 Series system equipped with a FLD 3100 detector (Th ermo Scientifi c, Germering, Germany) at 30ºC, and λ ex 365 nm, λ em 435 nm. Th e HPLC column was Supelcosil TM LC-18-DB, 250 x 4.6 mm (particle size 5 μm; Merck, Darmstadt, Germany), fi tted with a guard column. Th e mobile phase was acetonitrile: water (50:50, v/v) fi ltered through 0.22 μm membrane fi lter, at a fl ow rate of 1.2 mL/min. Th e system was controlled by Chromeleon® 7 soft ware (Th ermo Scientifi c, Germering, Germany). Th e peak areas at afl atoxin B 1 retention times were compared to the corresponding calibration curves. Calculation of AFB 1 adsorption rates (%) was performed according to the following equation: BC AFB1 = (1-C I /C 0 ) x 100% BC AFB1 = binding capacity; C I = concentration of free AFB 1 aft er the incubation period; C 0 = initial fortifi ed concentration of the AFB 1 .

RESULTS
Th e results of physico-chemical characterization tests showed diff erent properties of adsorbents regarding moisture content, swell index and acidity (Table 1). Test results for the adsorption of afl atoxin B 1 (4 mg/l) by diff erent adsorbents 1 -5 (0.02% w/v) at pH 5 aft er 120 minutes (phase I) were from 9.1 ± 1.9 % to 90.1 ± 0.2 %. In the second phase (phase II) of the trial, which was performed with high adsorbent (0.2% w/v) and low toxin concentration (0.2 mg/L), the best performance was confi rmed for the sample number 4 as binding capacity was 99.6 ± 0.03 % ( Table 2). 27.0 ± 0.5 / Th e characterization study showed that all samples diff ered in physicochemical properties, such as moisture content and swelling index. Besides that, their ability to adsorb afl atoxin B 1 also varied greatly.

DISCUSSION
Th e variable properties of adsorbents are the result of their diff erent composition in terms of the ratio of organic and inorganic components. Sample number 1 mostly diff ered, as it contained zeolite for mineral component instead of montmorillonite. However, more samples need to be investigated for correlation assessment between physico-chemical properties of adsorbents and the amounts of adsorbed toxin.
According to Regulation (EU) No 1060/2013 (European Commission, 2013), AFB 1 good binding capacity (BC AFB1 ) should be above 90%, which was achieved with the product number 4. Th e adsorbent was tested under "intensifi ed conditions" (low binder concentration and high toxin concentration), as described by Vekiru et al. (2015) and such concept was carried out to get closer to the limit of the product's adsorption capacities. In the Phase II, an extremely high percentage of binding for the same sample confi rmed the previous good outcome.
Based on the obtained results, the material labelled with number 4 was categorized as good and its composition proved to be the best of all the examined samples. Bentonite, which is the main component of this product, is well known for its ability to bind afl atoxins (EFSA, 2009(EFSA, , 2011, while natural yeast extracts, a cell wall derivatives of Saccharomyces cerevisiae, show considerable binding ability with several commonly occurring mycotoxins (Devegowda and Murthy, 2005) and are benefi cial in minimizing their adverse eff ects in animals (Nešić et al., 2008). Such multilevel mechanism of action, achieved through the complex composition of this adsorbent, indicate various usage potentials and also enable further effi ciency testing of this feed additive.
Th e poorest performing of the zeolite sample was sample (No 1), which is in accordance with the results of Th ieu and Pettersson (2008) who reported that bentonite has a better ability to adsorb AFB1 than zeolite. According to Marroquín-Cardona et al. (2009), this could be due to the smaller size of the zeolite pores (4 -7 Å in case of natural clinoptilolite) in comparison with AFB1 size (10.4 -12.8 Å), which limits the adsorption to the external surface only.
As reported in the case of charcoal, in vitro success is not always a suffi cient criterion to choose an adsorbent for practical use, so in vivo trials should verify its usefulness (Diaz et al., 2002;. Even among good binders, there were diff erences in in vivo effi cacy, indicating that in vitro testing alone is not always adequate for complete evaluation of the additive (Vekiru et al., 2015). Nevertheless, the advantage of in vitro test is the possibility of rapid screening efficiency of a large number of adsorbents. In this way, the reduction of mycotoxin toxicity is also indirectly confi rmed. Th us, in vitro experimental approach can serve as a reliable fast tool for triage of adsorbents. Vekiru et al. (2015) showed in their trial that it is a helpful to preselect an AFB 1 adsorbent and to predict the in vivo AFB 1 detoxifying performance.

CONCLUSION
As a result of the presented experiment, the optimal formulation of the adsorbent No 4 "MycoStop DUPLO" was found, which at low concentration and in the presence of high amounts of toxin met the stringent European regulation requirements for the minimum 90% afl atoxin binding effi ciency. Based on good in vitro afl atoxin B 1 adsorption results, it seems pertinent to extend in vivo studies of the selected adsorbent. As, according to literature data, it combines bentonite and yeast polysaccharides, it is reasonable to perform assays with other mycotoxins in the future and expect promising results.
Although there is a regulation on the in vitro testing of bentonite in the EU, many national regulations worldwide do not cover estimation of binding capacity of adsorbents used as additives in animal feed. Also, there is no unique methodology for analyzing adsorbents and variously designed experiments could be found in the literature. Th erefore, the information on the adsorption capacity is obtained in diff erent ways and is not always comparable. It would be necessary to standardize this procedure and establish regulations to cover this signifi cant area.