Ochratoxin A (OTA), a mycotoxin produced by toxigenic species of Aspergillus
fungi, is a common contaminant in cereals, fruits, and nuts intended for human and animal consumption. For both these fungal species, moisture and temperature are key factors for growth and toxin production. Thus, OTA can be produced during plant growth, harvest, and storage [1
]. The European Commission provides maximum guidance levels for OTA in animal feeds. For broiler chickens, the OTA concentration in complete feed and in individual cereal ingredients used for feed formulation should not exceed 0.1 and 0.25 mg/kg, respectively [3
]. Nevertheless, the presence of OTA at a chronic concentration, whilst not necessarily posing a health issue, could be disadvantageous to bird performance and as such contribute to a negative economic output in a broiler operation. In this context, the use of appropriate mitigation programs is often recommended. Toxin adsorbents are frequently supplemented to animal diets to prevent the deleterious effect of mycotoxins, among them specific yeast cell wall extracts have shown promising results [4
Several experiments were carried out to study the effects of OTA on production parameters in animal husbandry. Studies have shown that administration of diets contaminated with OTA reduces feed intake and body weight gain in broiler chickens in a dose-dependent manner and negatively affects weight gain in swine [7
]. OTA has been shown to exert toxicity by inhibition of mitochondrial function, increased oxidative stress, and inhibition of protein synthesis [8
]. At a molecular level, OTA pathology includes damaged membrane lipids, DNA mutations, and nitrosylated proteins [8
]. In both mammalian and avian animals, sub-acute exposure to OTA has nephrotoxic, immunosuppressive, and teratogenic effects [9
]. Carcinogenic effects of OTA have been demonstrated in mice and rats; while in humans, renal and testicular cancer have been associated with OTA exposure [9
To date, absorption of OTA has been studied mainly in mice and rats. The results for rats indicate that the major sites of OTA absorption are the duodenum and proximal jejunum, but the toxin is also absorbed from the stomach [10
]. Studies of OTA absorption show that the toxin can be quickly detected in peripheral blood, with the maximum concentration in blood following oral administration of OTA being reached after 20 min in chickens, 1 h in rabbits, and 10 h in pigs [11
]. Absorbed OTA binds to blood proteins, such as albumin, and distributes throughout the animal. The half-life of OTA in blood after oral administration depends greatly on animal species, e.g., it is 6.7 h in quail, 39 h in mice, 120 h in rats, and 510 h in Rhesus monkeys (Macaca mulatta
]. OTA is mainly detoxified into ochratoxin α (OTα) by animal tissues and by intestinal anaerobic bacteria. Other metabolites of OTA are formed in the liver by cytochrome P450 and these metabolites differ from one animal species to another [13
]. OTA and OTα are excreted in faeces, whereas other OTA metabolites, together with OTA and OTα, are found in urine [14
Mycotoxin adsorbents or sequestrants are used in animal husbandry due to the common occurrence of mycotoxins in feedstuffs. The purpose of the mycotoxin adsorbents is to bind to the toxin in the digestive tract of the animal, after which the adsorbent-toxin complex is transported intact through the gastro-intestinal tract of the animal and excreted. In poultry, adsorbents such as yeast cell walls, esterified glucomannan, and aluminosilicate have been tested [15
]. For the toxin adsorbent to be functional, it is essential that the digestive processes of the animal do not alter its properties. In order to analyse the effects of digestion, toxin, and adsorbent have been subjected to digestion treatments—such as low pH—in in vitro conditions, and the results compared with a control without digestion treatment [16
Yeast cell wall extract (YCWE) has been studied with emphasis on its toxin adsorbent properties [18
]. The YCWE has been shown efficacious in sequestering mycotoxins in in vitro and decreasing the adverse effects of mycotoxins in vivo conditions [15
]. In the current study, the adherence of YCWE to OTA were examined both in vitro and in silico. The aim was to obtain accurate information on the interaction between OTA and YCWE during digestion treatments and on the effect of low pH on OTA-YCWE complex formation. Unlike previous studies, the effectiveness of YCWE in decreasing the OTA bioavailability was subsequently assessed in vivo in broiler chickens, at a dose that did not exceed the recommendation for commercial broiler feeds presented in the European Union guidance [3
]. In accordance with the EFSA dossier preparation instructions for Substances for the Reduction of Mycotoxin Contamination (SMRC) [19
], the efficacy of YCWE in reducing bioavailability of OTA was determined with analysis of OTA deposited in broiler liver tissue.
Functional feed additives, such as toxin adsorbents, must remain active throughout the digestive tract of the animal and as such evaluating their integrity and susceptibility to the digestive environmental condition is necessary. Therefore, the YCWE product tested in this study was subjected to chemical, enzymatic, and physical digestion treatments simulating in vitro broiler chicken digestion. Tests of the toxin sequestration properties of YCWE after each digestion step revealed no significant changes in the level of radioactive OTA bound to YCWE, indicating that YCWE functionality and activity are preserved in the gastro-intestinal tract of broilers. Measuring changes in radioactivity is a highly sensitive method for assessing alterations in the concentration of radiolabelled compound (tritium-labelled OTA was selected as a signal molecule when assaying in vitro product efficacy). The method allowed analysis of the experimental samples directly after one centrifugation step, thus avoiding more laborious extraction procedures, isolation, and detection methods.
Analysis of the adsorption of OTA to YCWE during simulated consecutive digestion steps revealed a substantial increase in adherence at low pH. At pH 2.5, more than 80% of OTA was found to have bound to YCWE, whereas the enzymatic digestion treatments tested (pepsin, pancreatin) had no effect on binding performance. However, when the pH was re-adjusted to 6.5, OTA release from YCWE was observed. In broiler chickens, it has been shown that through reverse peristalsis digesta can also occasionally move from distal to proximal digestive tract. Thus, it is possible that digesta moves back from neutral duodenum to acidic gizzard [20
]. Although our model did not specifically mimic such reverse peristalsis, we showed that when changing pH from neutral to acidic, and again to neutral, OTA binding was reversible, which is to say it becomes bound and is then released again. Thus, our model enables also predicting such a situation. This observation shows similarities with results obtained by Oh et al. [17
], where low pH in combination with YCWE was found to decrease the toxicity of OTA in cultured macrophages. Oh et al. [17
] suggest that the mechanism of decreased toxicity involved conformational changes in OTA at pH below 5.0, which in turn affected hydrogen bonding and van der Waals forces between OTA and YCWE. The increased adherence of OTA to YCWE at pH 2.5 and the detachment of OTA at pH 6.5 observed in the present study may support the mechanism suggested by Oh et al. [17
]. However, it appears from our in silico studies that the changes in OTA conformation tended to happen independently of pH, due to several degrees of liberty from the presence of seven rotatable bonds around the amide bond. The observed changes in affinity could then be attributed rather to protonation level of the molecule than OTA conformational changes. The OTA molecule has a carboxylic acid group with pKa1 at 4.4 [21
] and, according to our result, has a strong acidic pKa of 3.17. This functional group increasingly occurs in protonated form at pH lower than the pKa1 and protonation affects the molecule, making it electrochemically uncharged and decreasing its polarity. When pH is increased up to pH 6.5, OTA has a theoretical charge of −1.03 involving a deprotonation of the carboxyl group. If changes in the molecular structure due to deprotonation are minimal compared to protonated form (Figure S1
), the glucan conformational structure can be dramatically affected, as demonstrated by the molecular dynamics simulation using two protonation states of OTA (Figure 5
and Figure 6
, Table 1
), as well as the stability of interaction that involves in its majority, the carboxylic group of OTA in the formation of hydrogen bonds. All these factors may have an effect on OTA adherence to YCWE and demonstrated that the changes in β-(1,3)-d
-glucan organization could transform the docking occurring at pH 2.5 into a surface interaction at pH 6.5. Moreover, the deprotonated form of OTA was more prone at interacting with the solvent environment, with an increased amount of H-bond formed, that directly competed with the capacity of the molecule at binding to the YCWE.
In the broiler trial, no acute OTA toxicity occurred in the birds, since all measured growth parameters, including mortality, were similar between the treatment groups. The dose of OTA fed to the broiler chickens was 0.1 mg/kg of feed, which is the maximum recommended level in commercial broiler feed in the European Union. In earlier studies, decreased body weight gain has been reported for broilers fed OTA at a dose of 4 mg/kg [22
] or 0.5–0.25 mg/kg [23
]. However, literature data on body weight changes are not conclusive, as no change in body weight gain was observed in a study by Politis et al. [25
] in which 0.5 mg/kg of OTA was fed to broilers for 42 days. Therefore, the finding that the low dose of OTA applied in the present study did not change broiler performance parameters was expected and is in line with the previous research. However, these findings apply to a single mycotoxin and it is hypothesised that when several toxins are present at the same time, even if their levels are within the statutory or guidance limits, there may be synergistic forms of toxicity, as demonstrated by the impact of OTA in conjunction with the presence of penicillic acid (Penicillium
spp. mycotoxin) in a macrophage culture assay [26
] and negative effects on animal performance. In such cases, toxin binders may significantly improve the performance of production animals.
Analysis of the liver of broilers fed an OTA-containing diet for 14 days showed clear accumulation of OTA in liver tissues. It is thus evident that OTA is deposited in broiler liver tissues even when the dose of toxin is present in feed below the guidance level of the European Commission. However, the concentration of OTA in liver samples (0.072 ± 0.003 µg/kg) was low relative to the maximum recommended level of 2–10 µg/kg in products intended for human consumption in the European Union and Canada [27
]. Several-fold higher OTA concentrations than were found in the present study have been detected in broiler products and chicken eggs in Pakistan, where 41% of broiler products have been found to be contaminated with OTA and the OTA concentration in broiler liver is as high as 3.56 µg/kg [28
]. As the concentration of OTA in broiler chicken tissues is a result of toxin concentration in feed [29
] it is clear that OTA concentrations exceeding 0.1 mg/kg are present in broiler feeds in some geographical regions.
When the broiler chickens in the present study were fed both YCWE and OTA, the OTA deposits in the liver decreased significantly. This decrease in OTA deposition was dependent on the dose of YCWE. The optimal YCWE dose tested was 4.0 kg/T, because the concentration of OTA detected in broiler liver was found the lowest with this dose. Lowered OTA deposition in the liver has also been reported in previous studies in which broiler chickens were fed OTA-contaminated feed together with toxin-deactivating products such as Trichosporon mycotoxinivorans
] plant extracts [31
], and yeast products [32
]. However, in all those studies the concentration of OTA in broiler feed was at least five-fold higher than in the present study. Thus, a novel finding in the present study is that OTA deposition can be reduced using YCWE at chronic level of exposure of OTA. A low level can still be potentially damaging, e.g., in a study by Pozzo et al. [33
], an OTA level of 0.1 mg/kg feed did not induce clinical signs but was shown to induce mild histological pathology in immune organs.
In the present study, a clear correlation was observed between results from in vitro and in vivo studies. The relative decrease in free OTA in the adsorption models used herein was comparable to the reduction in OTA concentration found in broiler livers. Dose dependence was observed both in vitro and in vivo, as an increase in the YCWE dosage decreased the measured OTA concentration. Previous studies have revealed a specific dose dependency in the fraction of OTA deposited in the liver, e.g., in a study by Hanif et al. [29
] using mycotoxin deactivators, the effect of the deactivators was greater when the OTA dosage in broiler feed was increased. In the present study, there was an indication of a similar effect with the increase in YCWE inclusion from 2.0 to 4.0 kg/T but no difference was observed with a further increase of the level inclusion up to 8.0 kg/T. At 4.0 kg/T, the highest measured OTA concentration in the feed, OTA deposition in the liver was the lowest observed in the study, apart from the control. Thus, addition of YCWE to animal feed has the potential to decrease OTA deposition in the liver.
Overall, this work provides a methodological and practical insight to the demonstration of the binding properties of YCWE, showing an ability to chemically interact with OTA, dependent on the digestive physiological conditions and pH. This was demonstrated in vitro and by both biochemical and computational mechanistic/dynamic experiments. As reported here, changes in pH can induce conformational changes not only of the OTA molecule but also of the β-d-glucans, the bioactive component of YCWE, affecting the geometry of the binding site and consequently the affinity of interaction. A complementary in vivo study demonstrated that YCWE, when fed to broiler chickens at different inclusion levels, could tightly correlate these in vitro/in silico observations and mitigate the impact of OTA by significantly reducing accumulation in liver tissue. Collectively, these results showed that YCWE could sequester OTA and prevent its deposition in liver, which in turn can protect the animal from a mycotoxin challenge when present at low and chronic levels in the diet.
Mycotoxins, including OTA, are highly prevalent in animal feed. A three-year survey by Rodrigues et al. [34
] found that up to 93% of feed samples were contaminated with OTA, while other mycotoxins, such as aflatoxin and fumonisin, were similarly common. As a consequence, correct harvest and storage methods are essential in controlling mycotoxin contamination of feedstuffs [35
]. However, as this study indicates, including mycotoxin adsorbents or deactivators provides further assurance of good animal health and diminished toxin presence in human foodstuffs.
5. Materials and Methods
5.1. In Vitro Assessment of OTA Sequestration by YCWE
Yeast cell wall extract (YCWE) (Mycosorb A+®, Alltech Inc., Nicholasville, KY, USA) used in the study was added to 50 mM Na-phosphate buffer (pH 6.5) in silyated test vials (Supelco, Sigma-Aldrich, Saint Louis, MO, USA) in a final reaction volume of 10 mL. Buffer alone was used as a control. A mixture of tritium ([3H])-labelled OTA (Moravek Biochemicals, Inc., Brea, CA, USA) and non-radiolabeled OTA (AppliChem GmbH, Darmstad, Germany) was prepared in Na-phosphate buffer to give a final OTA concentration of 20 ng/mL and [3H] activity of 0.08 µCi/mL. The OTA solution was added to test treatment vials (see treatment condition below) for a final YCWE concentration of 5 and 15 mg/mL. Reaction vials were mixed and incubated at +37 °C for 2 h on a rotary shaker (GFL, Burgwedel, Germany). After incubation, the vials were centrifuged for 10 min at 3000× g. The [3H] activity of unbound [3H]-OTA was analysed by taking a 50 µL sample of the supernatant, adding it to 3.5 mL of scintillation cocktail (OptiPhase SuperMix, Perkin Elmer, Waltham, MA, USA) and measuring the [3H] activity using a scintillation counter (Microbeta 1450, Perkin Elmer, Waltham, MA, USA).
5.2. In Vitro Digestive Simulation Conditions and OTA Sequestration Activity of YCWE
In the first in vitro model experiment, YCWE at two final concentrations (5 and 15 mg/mL) was subjected to three successive environmental conditions before the resultant adsorption efficacy was calculated based on the measure of the amount of free OTA remaining in the supernatant: (1) a 50 mM Na-phosphate buffer environment maintained at pH 6.5, considered as the control treatment; (2) a simulated gastric environment initiated by adjusting the pH to 2.5 with 1 M HCl and reacting with YCWE for 3 h at +37 °C, which represented the acid alone gastric digestion treatment; (3) a simulated enzymatic digestion treatment in an initial conditioning step through the addition of pepsin (Sigma-Aldrich, Saint Louis, MO, USA) at 7.5 mg/mL under the same environmental conditions as with (2), followed by a second step of pH adjustment to 6.5 with 1 M NaOH and addition of 2.5 mg/mL of pancreatin (Sigma-Aldrich, Saint Louis, MO, USA), representing the concomitant digestive enzymatic treatment. YCWE was incubated during the two successive enzymatic digestion steps at +37 °C for 3 h on a rotary shaker. After incubation in the pH was adjusted to 6.5 and the free OTA was measured.
In the second in vitro model experiment, the adsorption efficacy of YCWE (15 mg/mL) was evaluated based on the amount of free OTA remaining in the supernatant after each stage of digestive simulation: (1) starting stage, Na phosphate buffer maintained, pH 6.5; (2) first intermediate stage, after the acid gastric digestion treatment at pH 2.5; (3) second intermediate stage, after acid gastric digestion and neutralisation at pH 6.5; (4) third intermediate stage, after the one-step enzymatic (pepsin) treatment at pH 2.5; (5) final stage, after the two-step enzymatic (pepsin followed by pancreatin) treatment at pH 6.5.
In both experiments, a mixture of OTA (final concentration 20 ng/mL, [3H] activity 0.08 µCi/mL) was added to the digested YCWE and non-digested control. The activity of unbound [3H] OTA was measured in five replicates following the above described assessment of OTA sequestration by YCWE. The pH adjustments between digestive phases were monitored, and the indicated values represent the actual pH measurements.
5.3. In Silico Assessment of the Sequestration Properties Investigated by Means of Molecular Mechanics and Dynamics
Molecular mechanics investigations were carried out using several open-source programs: (1) Constructs of the β-(1,3)-d
-glucan chain branched with a side chain of β-(1,6)-d
-glucan previously generated on a Silicon Graphics computer with the Accelrys package (formerly Accelrys, now Biova Dassault Systemes, San Diego, CA, USA) in CFF91 force-field with steepest descent minimization [36
]. This structural assembly was considered in the present study as the receptor, with three distinct conformations being investigated, a β-(1,3)-d
-glucan chain alone, and two different β-(1,6)-d
-glucan branched β-(1,3)-d
-glucan chain conformations, as modelled in previous work; (2) OTA three-dimensional structure was downloaded from Chemspider compound repository under the permalink record 390954 (Royal Society of Chemistry, Burlington, VT, USA, http://www.chemspider.com/Chemical-Structure.390954.html
) with OTA considered as the ligand in our study; (3) Molecular docking experiments were run on AutoDock Vina (The Scripps Research Institute, La Jolla, CA, USA, [37
]). Before starting the docking process, all molecules were added with their polar hydrogens and were evaluated for their rotatable bonds. Seven degrees of liberty were found on the particular OTA molecule. The receptor conformation was constrained to the previously minimized conformation [38
]. A grid box with a spacing of 60.0 Å in every x, y, z direction was parameterized and centered on a site of interaction at the proximity of the β-(1,3)- and β-(1,6)-d
-glucans branching but also including in the grid box dimension adjacent binding sites over the entire single helical chain length consisting of 36 glucopyrannose +/− 3 side chain residues. PDB extension files were converted into pdbqt files before performing docking using Autodock Vina (v1.5.6, The Scripps Research Institute, La Jolla, CA, USA), an empirical scoring function that calculates affinity. The docking experiment was performed under C++ generic programming in a Microsoft Windows®
(10 Pro, 1809, Microsoft Corporation, Redmond, WA, USA) environment and produced a maximum number of nine binding modes with a maximum energy range of 3 kcal/mol, and for which an affinity (kcal/mol) and distances from the best modes were measured.
Molecular dynamics were investigated using several open-source programs: (1) doGlycans (v1.0, Open Source program under GNU General Public License, v3.0, Free Software Foundation Inc, Boston, MA, USA), a tool to prepare carbohydrate structures under an Optimized Potentials for Liquid Simulations for All Atoms (OPLS-AA) modified force-field and related topology files for simulation [39
] was used under Python 3 script and the Conda management system (Anaconda Inc., Austin, TX, USA) applied to the β-d
-glucans receptor; (2) TopolGen (v1.1, Open Source Program under GNU General Public License, v3.0, Free Software Foundation Inc., Boston, MA, USA), a Perl script, and LigParGen (Open Source Program,http://zarbi.chem.yale.edu/ligpargen/
) were used to produced GROMACS-formatted topology from the above generated PDB files docking experiment, compatible with an OPLS-AA force-field for OTA [40
]; (3) GROMACS 2019.1 (GROMACS, Open Source Program under GNU General Public License, v3.0, Free Software Foundation Inc., Boston, MA, USA, www.gromacs.org
]) molecular dynamics simulation package was used to simulate receptor-ligand complex over 10 ns in a neutralized solvated system (water) at 37 °C and to study the initiation of the interaction energy at different charge states corresponding to a pH simulation performed at pH 3.0 and 6.5 to match our in vitro evaluations [44
]. OTA was restrained using a force constant of 1000 kJ mol−1
while the constrained and unconstrained structure of β-d
-glucans were evaluated. Interaction energy was evaluated according to ionic coulomb energy potential and Lennard-Jones neutral molecule interaction potential energy expressed in kJ/mol.
PyMOL (The PyMOL Molecular Graphics System, v2.2.3 Schrödinger, LLC., New York, NY, USA) was used for visualization of the 3D molecular structures for all stages of molecular mechanics and molecular dynamics simulations. The chemical properties of OTA were calculated using the Chemicalize Chemoinformatic platform (ChemAxon, Cambridge, MA, USA, http://www.chemaxon.com
) for the calculation of pKa of the molecules and chemical properties at different pH values.
5.4. In Vivo Dietary Treatments
The trial was conducted with a basal diet (Table 3
) that was used as a negative control and also to generate the mycotoxin challenge dietary treatments. For these latter treatments, pure OTA (Sigma-Aldrich, Saint Louis, MO, USA) was dissolved in ethanol and mixed into a small amount of broiler feed. This mixture was step-wise mixed into larger batch of feed and eventually to final OTA challenge diet. The different doses of YCWE were added to small batches of broiler feed which were then mixed into final dietary treatments.
5.5. Animal Trial
The feeding trial was conducted in the research facility of Alimetrics Ltd. in Southern Finland, in accordance with EU Directive 2010/63/EU. Following the standard operating procedures of Alimetrics Ltd., ethical approval or animal trial permit was not required as the substance under investigation is an approved feed ingredient in the EU and the level of Ochratoxin A included in the diets were below EU guidance levels (European Commission, 2006, [3
]). 240 newly hatched male Ross 508 broiler chicks were randomly allocated to five feeding treatments divided between 40 pens (8 pens per treatment, 6 birds per pen). Temperature and lighting programs followed the Aviagen recommendations for Ross broilers (Aviagen Group, Huntsville, AL, USA). For the first 7 days, all birds were fed a basal diet (Table 4
). From day 8 onwards, the birds on the control diet continued a basal diet while the other treatment groups received a targeted 0.090 mg/kg OTA-contaminated diet added to three varying amounts of YCWE (2.0, 4.0, 8.0 kg/T). The birds were individually weighed on days 1 and 21. Feed conversion ratio (FCR) was corrected for mortality by including the weight of deceased birds in the calculation. FCR was calculated over the period 1–21 days.
5.6. Mycotoxin Analysis of Dietary Treatments
Confirmation of the levels of OTA in final feed via analysis (Table 4
). Basal diets were also analysed for their content in aflatoxin B1 and B2, deoxynivalenol, nivalenol, zearalenone, fumonisins B1 and B2, T-2 toxin, and HT-2 toxin by means of LC-MSMS (Nutricontrol, Veghel, The Netherlands). Considering the measurement uncertainty announced by Nutricontrol, we can conclude that the content of OTA in the diets did not significantly differ from each other.
5.7. OTA Analysis of Liver Tissues
Livers were collected from four broilers per pen. The livers were combined in pairs, resulting in two analytical samples per pen. The tissues were ground and stored at −20 °C until analysis. OTA analysis was performed by extracting 5.0 g of ground liver with 25 mL of 1% NaHCO3: methanol (30:70, v/v). After centrifugation, 10.0 mL of the supernatant was mixed with an equal volume of PBS and applied onto an OchraTest WB immunoaffinity column (Vicam, Milford, MA, USA). The effluent was filtered through a 0.45 µm syringe filter and analysed by high pressure liquid chromatography equipped with a fluorescent detector (HPLC Prominence, Shimadzu, Kyoto, Japan). The HPLC separation was performed by means of a Phenomenex Luna C18(2) 3 µm, 150 × 4.60 mm column equipped with a Gemini C18, 4 × 3 mm SecurityGuard pre-column (Phenomenex Inc., Torrance, CA, USA) at a flow rate of 0.8 mL/min. Injection volume was set to 40 µL and column oven temperature 30 °C. Fluorescence detection was carried out at an excitation wavelength of 333 nm and emission wavelength of 460 nm. The limit of detection measured from standard deviation of background signal was 0.4 ng/kg and the limit of quantification was 2 ng/kg.
5.8. Data Analysis
Statistically significant differences between mean values of the parameters tested were analyzed with ANOVA using Tukey’s honestly significant difference (HSD) post-hoc test in the SPSS statistical software package (IBM, version 22, Armonk, NY, USA).