2.1. Screening of Adsorbent Materials as Multi-Mycotoxin Detoxifying Agents
Preliminary work was performed to study the effect of a cocktail of mycotoxins on adsorption and to select promising multi-mycotoxin-adsorbing materials. Experimental conditions of our previous adsorption studies [
18,
19] were used as a standard. Hence, twelve Na/Ca smectites (comprising seven di-octahedral smectites and five tri-octahedral smectites) and six lignocellulose-based materials were tested using a fixed amount of material (1 mg mL
−1) and of mycotoxins (1 µg mL
−1). Mycotoxins tested as a pool were AFB
1, ZEA, OTA, FB
1, and DON, and were dissolved in buffers at pH 3 and 7 to simulate gastric and intestinal pH values (
Table 1). T-2 and HT-2 were not assayed in the preliminary study.
As expected, excluding the bentonite labelled 2 (a Na di-octahedral smectite), most smectites showed more than 90% adsorption of AFB
1 at both pH values, regardless of their structure (di- or tri-octahedral). The in vitro efficacy of bentonite clays, especially di-octahedral smectites (i.e., montmorillonite), as AFB
1 binders has been widely reported in the literature. So far, only montmorillonite has been authorized as a technological feed additive for aflatoxin decontamination of feeds for ruminants, poultry and pigs [
5]. However, all smectites do not adsorbing aflatoxins equally and can differ depending on their intrinsic properties [
6]. Recently, it has been proven that the efficacy of di-octahedral smectites in sequestering aflatoxins depends on their physico-chemical and mineralogical properties, including geological origin [
6]. Tri-octahedral smectites have been poorly evaluated as mycotoxin binders, including for aflatoxins [
12,
13]. This may be due to the relative abundance of di-octahedral smectites compared to tri-octahedral varieties. In accordance with the findings of our screening study, a report by Vila-Donat et al. [
12] proved the ability of tri-octahedral smectites in sequestering AFB
1 to be even better than that of bentonites with di-octahedral structures.
Although in our experimental conditions, di- or tri-octahedral smectites did not seem to differ in sequestering AFB
1, they differed substantially in binding mycotoxins other than AFB
1, which could be due to their origin and physiochemical properties, i.e., cation exchange capacity, pore volume and expandability [
1,
29]. Four out of twelve smectites efficiently adsorbed ZEA (>70%), regardless of pH. Interestingly, all these bentonites belonged to the group of tri-octahedral smectites, which is in line with the study by Vila-Donat et al. [
13] that found good ZEA adsorption values for these minerals. In addition to tri-octahedral smectites, only heat-activated bentonites and organophilic bentonites can bind ZEA, to some extent [
30,
31,
32,
33].
All bentonites (except the bentonite labelled 2) showed high efficacy (>80%) in removing FB
1 from medium at pH 3. When recorded at pH 7, FB
1 adsorption values dropped to less than 20%. Three tri-octahedral smectites bound about 26% of the toxin amount in solution. The effect of medium pH on FB
1 adsorption has been observed in several adsorbents, including minerals and organic materials [
3,
18,
19,
21,
34].
OTA was adsorbed by most bentonites but to a different extent depending on pH of the medium, being significantly high at gastric pH and negligible at intestinal pH. Eight out of twelve bentonites adsorbed more than 70% of OTA at pH 3. The highest adsorption values (≥80%) were recorded for three bentonites, all belonging to the tri-octahedral smectite group. At neutral medium pH, OTA adsorption was less than 25%. The effect of pH on OTA adsorption has been widely reported in the literature for different types of adsorbents, including clays [
18,
19,
35,
36].
In accordance with previous reports [
37,
38,
39], DON was poorly adsorbed by all bentonites, including the tri-octahedral smectites. Vila-Donat et al. [
13] also found negligible DON adsorption by tri-octahedral smectites.
In conclusion, the screening of smectites for simultaneous removal of mycotoxins allowed the selection of sodium smectites with tri-octahedral structures (materials labelled 9–12 in
Table 1) as promising multi-toxin adsorbents. With respect to di-octahedral smectites, the smectites with tri-octahedral structures contain lower Al
2O
3 and higher MgO contents, which leads to different layer charges and physicochemical properties [
4]. It is well known that the properties of smectites change not only with the magnitude of the layer charge but also with its distribution throughout the layer, with the exchangeable cations and with their hydration status [
6]. Thus, for octahedral charged smectites (such as the di-octahedral smectite montmorillonite), the negative charge is delocalized over surface oxygens so that only weak hydrogen bonds can form with interlayer water. For tetrahedral charged smectites (such as the tri-octahedral smectites, beidellite and saponite), the charge is more localized and stronger hydrogen bonds can form between surface oxygens and interlayer water [
4]. These different distributions of interlayer charge, together with different hydration statuses, are responsible for the different physicochemical properties of smectites and for their adsorbing features, including mycotoxins uptake.
In addition to tri-octahedral smectites, lignocellulose-based materials were found quite effective for simultaneously adsorbing the target mycotoxins (
Table 1), although they appeared less effective than clays. In particular, AFB
1 adsorption was in the range 47–81% at pH 3, and 60–87% at pH 7. Regardless of pH, a significant ZEA adsorption was also recorded, comparable to that obtained by smectites. ZEA adsorption values were in the range 37–77% at pH 3 and 61–71% at pH 7. FB
1 and OTA adsorptions were again influenced by pH, being high at pH 3 (19–39% for FB
1 and 35–73% for OTA), and negligible at pH 7 (less than 8%). The pH of the medium influences the extent of FB
1 and OTA adsorption, since it affects the degree of ionization of their molecules, and the surface charge of the lignocellulose-based materials, which are characterized by the presence of phenolic and carboxylic groups. According to previous studies using agricultural by-products as mycotoxin binders [
18,
19], lignocelluloses did not adsorb DON.
As indicated by the supplier of the lignocelluloses, the materials were high-purity lignin products prepared by a patented technology used industrially in conjunction with the manufacture of high-quality papers. The materials were provided in the form of fine light brown powders (with <200 µm particle size) and differed depending on the location of production facilities (India, USA and Canada), and plant feedstock (annual fibers such as wheat straw, sugarcane bagasse, and wood). They were chemically characterized by the supplier as mixtures of natural polyphenols, composed of medium-to small polymer aggregates, particularly rich in OH phenolics, but also containing aliphatic and carboxylic groups. The chemical composition of the lignocelluloses did not differ substantially, thus they acted similarly in sequestering mycotoxins.
The ability of lignocelluloses in simultaneously adsorbing AFB
1, ZEA, OTA and FB
1 is in accordance with the findings of our previous reports showing that certain fibre-rich agricultural by-products, tested at medium-low dosages (1–10 mg mL
−1), were able to sequester different mycotoxins [
18,
19]. All these studies confirm that food plants that are available in large quantities, or certain waste products from agricultural and industrial operations (such as the lignocelluloses assayed herein), particularly rich in undegradable fibres (lignin, cellulose), are potentially low-cost products for mycotoxin adsorption.
Under experimental conditions of the screening study, two materials (one from each group of organic and inorganic binders) were selected as promising MMDAs, since they simultaneously adsorbed more than 70% of three different mycotoxins. These materials were the tri-octahedral smectite labelled 9 (which exhibited mycotoxin adsorption values comparable to the other tri-octahedral smectites) and the lignocellulose-based material labelled 6. The latter was one of the best adsorbents in its group. Because multi-toxin adsorption by the selected smectite was higher than the lignocellulose, especially for FB
1 uptake, the materials were combined in mixtures containing smectite as the main component. Several combinations of these materials were attempted to prepare mixtures with satisfying multi-mycotoxin adsorption features. These composites were assayed for mycotoxin adsorption as single components (data not shown). The product obtained by mixing the smectite and the lignocellulose in the weight ratio 70:30 was found to be the best combination. As expected, the mixing of the smectite with the lignocellulose yielded a composite with mycotoxin adsorption values slightly lower than the mineral clay. However, the lignocellulose possessed unique biological properties (antioxidant activity), which justifies its use (see
Section 2.2). Choosing such a component with mycotoxin-adsorbtion properties also had the effect of limiting the decrease in effectiveness for the best performing adsorbent (the smectite) due to its dilution in the mixture.
The composite with the smectite/cellulose weight ratio of 70:30 was tested at 1 mg mL
−1 of dosage with a multi-mycotoxin working solution containing 1 µg mL
−1 of AFB
1, FB
1, ZEA, OTA, DON, T-2, and HT-2; and at pH 3, 5 and 7. With respect to the preliminary study for the screening of MMDAs, T-2 and HT-2 were included in the adsorption assay. The latter was performed at pH 5, being the physiological pH value of the duodenal compartment of monogastric animals, where main absorption of mycotoxins takes place. As shown in
Table 2, the mixture retained the efficacy of its components in adsorbing simultaneously AFB
1, FB
1, ZEA and OTA, and showed a moderate efficacy (23–45%) in binding T-2. DON and HT-2 were poorly sequestered. AFB
1, ZEA and T-2 adsorptions were unaffected by pH in the range 3–7, while FB
1 and OTA adsorptions were higher at acid pH values. Interestingly, at pH 5 the new composite was still able to adsorb a high amount of FB
1 (91%) and OTA (32%).
2.2. Evaluation of Cytotoxicity and Antioxidant Properties of Selected Multi-Mycotoxin Detoxifying Agents
The aim of this study was to assess whether certain materials acting as mycotoxin adsorbents have biological effects (antioxidant properties) and can be used to fortify mycotoxin-detoxifying agents in preventing or counteracting toxic effects. Because the intestine system represents the specific and primary target of mycotoxin-contaminated feeds, the assessment of protective effect of mycotoxin-adsorbent materials was carried out on a cell line of this target organ (Caco-2TC7). The materials selected as MMDAs, i.e., the tri-octahedral smectite and the lignocellulose (items 9 and 6 in
Table 1, respectively), were subjected to a gastro-intestinal digestion process and, subsequently, their chyme samples were analyzed for biological activity. Ascorbic acid was tested as a standard natural compound, as it is widely used in feed formulation as a preservative.
Prior to determining the antioxidant activity, a cytotoxic test was performed to assess the safety of the agents and to select the highest no-cytotoxic dilution of chyme samples. Cytotoxic assays were performed on the intestinal cell line Caco-2TC7 using the MTT test. A 1:10 dilution of the chyme samples, corresponding to 1 mg mL−1 of equivalent concentration of digested materials (tri-octahedral smectite and lignocellulose-based material) did not affect the viability of intestinal cells after 24 h of exposure. These results suggest that these materials did not exert toxic effects in vitro. On the contrary, in the same experimental conditions, the chyme sample of ascorbic acid at 1 mg mL−1 induced an 80% decrease in resorufin fluorescence, indicating cytotoxicity.
The antioxidant activity of chyme samples was measured using the Cellular Antioxidant Activity (CAA) assay and was expressed as median effective dose (EC50), which is the concentration of the material (mg mL−1) that produces a 50% reduction of induced ROS. The lignocellulose-based material showed good antioxidant activity, although lower than ascorbic acid, with EC50 values at 0.248 ± 0.090 mg mL−1 and 0.008 ± 0.001 mg mL−1 of the digested material, respectively. The smectite did not show anti-oxidant properties up to 1 mg mL−1 of the digested product.
As reported by the supplier, the selected lignocellulose-based material consisted of phenylpropanoic polymer aggregates with a high grade of purity (about 95–97%) and high polyphenol content, as determined by the Folin-Ciocalteu assay and expressed as Gallic acid equivalent (mean value at 35 GAE/100 g of lignocellulose). These values are in accordance with data reported in the literature for analogous materials [
40,
41,
42]. Because the lignocellulose contained a high level of polyphenols, it was supposed that they can be released during the digestion process and can determine biological properties of the ingested/digested material, including the antioxidant capacity. The lignocellulose was first digested, and the polyphenol content of the centrifuged chyme samples (relevant to 1 mg mL
−1 of digested material) was measured by Folin-Ciocalteau assay and expressed as catechin equivalent. Interestingly, these chyme samples showed a high total content of polyphenols (210.9 ± 9.7 mg mL
−1 of catechin equivalent), suggesting high bioaccessibility. The biological properties of polyphenols depend on their bioaccesibility, the process of releasing polyphenols from the food matrix in the GI through enzymatic hydrolysis, which may be at least partially absorbed. High bioaccessibility for total phenol content is of great importance since it is linked to the ability of these active compounds to counteract oxidative stress, pathogens, and infections.
2.3. Effect of pH on Mycotoxin Adsorption and Desorption Study
The removal of mycotoxins from aqueous mediums through an adsorption process is, in most cases, dependent on pH value. The pH of the medium can affect the surface charge of adsorbents as well as the degree of ionization of toxins, and subsequently it can lead to a shift in reaction kinetics and equilibrium characteristics of the adsorption process. As reported by Greco et al. [
18], it is important to underline that a good adsorbent should have high adsorption efficacy in a pH range similar to the one encountered along the gastrointestinal tract of monogastric animals (pH 1.5–7.5). Therefore, it should be able to bind mycotoxins at low pH (simulating the range of pH typical of the gastric compartment) and retain the bound toxins during the transit of the food bolus through the intestinal compartments.
In this study, the effect of pH on toxin adsorption by the smectite–lignocellulose composite was investigated using a 5 mg mL
−1 of dosage and a series of multi-mycotoxin solutions (1 µg mL
−1) with pH values varying in the range 3–9. As shown in
Figure 1, mycotoxin adsorptions differed according to the degree of ionization of toxins. AFB
1, ZEA and T-2, which are mainly non-ionizable molecules, were adsorbed to the same extent regardless of pH, whereas FB
1 and OTA were adsorbed mainly at acid pH. DON adsorption was negligible (<20%), while HT-2 adsorption was not recorded.
AFB
1 is a planar, hydrophobic, and non-ionizable molecule, therefore a change of medium pH is not expected to affect its adsorption [
43]. Indeed, AFB
1 was completely sequestered (100%) from the mediums at all the assayed pH values. High ZEA adsorption values were also recorded (86–93%). ZEA is a weak acid with a pKa of 7.62 [
19], it should be in the protonated (non-ionic) form in the pH range between 3 and 7, and negatively charged at pH 8–9. T-2 adsorption was also stable (69–76%) along the pH range of the study. T-2 toxin has a tetracyclic sesquiterpenoid 12,13-epoxytrichothene ring system. It is a weak basic (essentially neutral) compound with a pKa of 13.2. Therefore, hydrophobic interactions may be responsible for ZEA and T-2 adsorptions by the product in pH ranges typical of the GI tract. FB
1 adsorption values ranged from 99 to 8% and were significantly higher when pH was ≤5. Due to the presence of carboxylic, hydroxyl, and amino functional groups in its structure, FB
1 was the most polar mycotoxin among those tested. As discussed by Avantaggiato et al. [
19], the anionic and cationic form of the toxin strongly depends on the pH, in particular the anionic form should be prevalent at pH values ≥5, and the cationic form at pH values <6. The findings of our study suggest that FB
1 adsorption by the new binder occured mainly when pH was between 3 and 5, and electrostatic interactions or hydrogen bonds involving the carboxylic functional groups may take place. The trend of OTA adsorption is comparable to FB
1. OTA adsorption values were 74–69% at pH 3–5, about 30% at pH 6–7, and ≤17% at pH ≥ 8. OTA is an ionizable molecule that should be in the anionic form in buffer solutions near pH 7, and in the uncharged form in acid solutions (pH < 4) [
19]. Taking into account our findings, OTA adsorption seems to occur mainly when in the uncharged form and decreases when pH increased up to 7, where the anionic form is predominant. Hydrophobic interactions may drive the adsorption of OTA onto the product.
To investigate whether a pH change in the medium can produce a toxin release from the adsorbent material, a desorption study was performed as described by Greco et al. [
18]. Values of mycotoxin adsorption at pH 3 and desorption at pH 7 were calculated for each toxin and expressed in percent (
Table 3). Mycotoxin adsorptions were >80% for all toxins, and a change of pH did not produce a release of AFB
1, ZEA and T-2. For these mycotoxins, desorption values were lower than 11%. Despite the low FB
1 and OTA adsorptions recorded at pH 7 (
Figure 1), the elevation of pH from 3 to 7 did not result in a total release of these toxins. Desorption values were 31% for FB
1 and 38% for OTA.
These findings suggest that mycotoxins can be adsorbed by the new binder in the gastric compartment of monogastric animals (at low pH) and retained during transit through the small intestine. This should prevent mycotoxin absorption at intestinal level, preserving intestinal integrity [
44,
45,
46].
2.4. Effect of Adsorbent Dosage and Toxin Concentration on Adsorption of AFB1, ZEA, OTA, FB1 and T-2
The effect of adsorbent dosage on simultaneous adsorption of AFB
1, ZEA, OTA, FB
1 and T-2 was investigated using equilibrium isotherms. Adsorption experiments were performed in triplicate, at pH 5, testing a fixed amount of toxins (1 µg mL
−1 each) in a multi-mycotoxin solution, with different adsorbent dosages (0.005–10 mg mL
−1). pH 5 was chosen as it represents the critical pH value above which FB
1 and OTA adsorptions significantly dropped (
Figure 1). From a biological point of view, pH 5 is the physiological pH of the duodenum compartment, where mycotoxin absorption mainly occurs. Mycotoxin adsorption data are listed in
Table 4.
Figure 2 shows the plots of adsorption data, expressed as percentage and plotted as a function of the adsorbent dosage. Mycotoxin adsorption was significantly affected by the dosage of the binder, and the percentage of mycotoxins removed from the multi-mycotoxins solution increased with increasing dosages.
Experimental values of mycotoxin adsorption were in the following ranges: 9–100% AFB1, 56–100% FB1, 9–95% ZEA, 6–85% OTA, 5–70% T-2. Some DON adsorption (22%) was recorded at the higher dosages (5 and 10 mg mL−1). HT-2 adsorption did not occur even at these dosages.
Isotherm adsorption plots were well fitted by the Langmuir model (R
2 > 0.990). This model allowed the calculation of two adsorption parameters, i.e., the Ads
max and the C
50, which are the theoretical estimated maximum adsorption and the theoretical dosage of adsorbent, providing a 50% reduction of toxin. Ads
max and C
50 calculated for AFB
1, FB
1, ZEA, OTA and T-2 are listed in
Table 5 and expressed as percentage. C
50 could not be calculated for DON and HT-2 since their adsorption was negligible. For all toxins, Ads
max and C
50 values calculated by the Langmuir model were in accordance with the experimental values listed in
Table 4.
These results suggest a high efficacy of the composite in simultaneously adsorbing five out of the seven target mycotoxins. A dosage of 2 mg mL
−1 should assure efficient (≥50%) and simultaneous adsorption of AFB
1, FB
1, ZEA, OTA and T-2 from a multi-mycotoxin solution containing 1 µg mL
−1 of each toxin. This dosage is lower than the C
50 values calculated for some multi-mycotoxin biosorbents tested in our previous studies [
18,
19], and in accordance with the findings of the two recent works using tri-octahedral bentonites as binders [
12,
13]. In the latter studies, some tri-octahedral bentonites when tested at 2 mg mL
−1 adsorbed more than 90% of ZEA and FB
1 [
13], while low dosages (from 0.2 up to 2 mg mL
−1) adsorbed 96–100% of AFB
1 and 15–75% of OTA [
12].
Equilibrium adsorption isotherms were also used to describe the effect of toxin concentration on adsorption of AFB1, ZEA, OTA, FB1 and T-2 by the new adsorbing agent.
Adsorption is a complex process of transferring specific contaminants (such as mycotoxins) from a fluid phase to a solid phase. In order to successfully simulate and optimize the adsorption of a chosen adsorbing agent, adsorption equilibrium isotherms must be studied. The results of these studies are used to assess the affinity or capacity of an adsorbent and select a suitable adsorbent and adsorbent dose. Indeed, information obtained by these studies can also enable the estimation of the economic feasibility of an adsorbent’s commercial application for mycotoxins.
Experimental adsorption data of AFB
1, ZEA, OTA, FB
1 and T-2 isotherms were fitted by the Freundlich, Langmuir, and Sips adsorption models to calculate the parameters involved in the adsorption process, i.e., Ads
max and K
L. These parameters represent the predicted maximum adsorption capacity and the affinity of the adsorption process, respectively [
19]. Most mycotoxin adsorption isotherms, excluding those obtained for T-2, were characterized by a typical L-shape (
Figure 3), and were well fitted by the Langmuir model as it provided the lowest standard error statistics and the highest R
2 values (R
2 ≥ 0.98) compared to the other models. Adsorption parameters calculated by the Langmuir model for AFB
1, ZEA, OTA, and FB
1, and by the experimental adsorption values for T-2, are listed in
Table 6. Following the assumptions of the Langmuir model, we can assume that the adsorption of AFB
1, ZEA, OTA, and FB
1 took place at definite localized sites of the adsorbing agent, which were equivalent [
47].
The experimental values of AFB
1 adsorption registered at pH 3 and 7 were in the ranges of 66–12% and 87–22%, respectively (
Figure 3), and were obtained using an extremely low amount of product, i.e., 0.05 mg mL
−1. The values of Ads
max and K
L differed depending on the pH of the medium (
Table 6) and were higher at pH 7. Ads
max was 67.9 ± 2.9 µg mg
−1 (218 ± 9 mmol Kg
−1) at pH 7 and 24.6 ± 0.9 µg mg
−1 (79 ± 3 mmol Kg
−1) at pH 3. Different AFB
1 adsorption capacities at pH 7 and 3 were also observed for the tri-octahedral smectites examined by Vila-Donat et al. [
12]. K
L constant, related to the adsorbent affinity, was increased by decreasing the pH. At pH 3, the K
L was 2-fold higher compared to that calculated at pH 7, being 1.3 ± 0.3 L mg
−1 ((41 ± 9) × 10
4 L mol
−1) and 0.6 ± 0.1 L mg
−1 ((19 ± 3) × 10
4 L mol
−1), respectively. Experimental values of adsorption isotherms displayed in
Figure 3 suggest that more than 50% of adsorption occurred when AFB
1 and the product were in a weight ratio ≤80 µg toxin/mg of product at pH 7, and ≤20 µg toxin/mg of product at pH 3. From a practical point of view, these results suggest that more than 50% of the toxin could be adsorbed by 1 mg of the additive supplemented per 1 g of a feed (1 kg ton
−1) containing up to 20 µg g
−1 of AFB
1, regardless of medium pH. This concentration is 1000 times higher than the EU-MLs (20 µg kg
−1) and can pose serious risks in terms of livestock health and human toxicity, due to the carry-over of metabolites in food products from animals fed with aflatoxin-contaminated feed. AFB
1 adsorption by the composite is expected to be due to smectite being the most prevalent component in the mixture. This mixture shows Ads
max values for AFB
1 higher than those recorded by the study that focused on the use of tri-octahedral bentonites as mycotoxin adsorbents [
12]. In the latter study, Ads
max values for AFB
1 did not exceed 10 and 31 µg mg
−1 at pH 3 and 7, respectively. However, these values are lower than those obtained by testing di-octahedral bentonites with sedimentary origin [
6].
In accordance with the preliminary study, FB
1 adsorption occurred mainly at acid pH (
Figure 3). The experimental values of FB
1 adsorption were in the ranges of 98–28% and 62–21% at pH 3 and 7, respectively, and were obtained using a dosage of 0.05 mg mL
−1 at pH 3 and 1 mg mL
−1 at pH 7. At these pH values, Ads
max was 91 ± 6 µg mg
−1 (126 ± 8 mmol Kg
−1) and 0.60 ± 0.02 µg mg
−1 (0.80 ± 0.03 mmol Kg
−1), respectively (
Table 6). The Langmuir K
L was not pH-dependent, being 1.8 ± 0.2 L mg
−1 ((12 ± 3) × 10
5 L mol
−1) and 1.6 ± 0.4 L mg
−1 ((13 ± 1) × 10
5 L mol
−1) at pH 3 and 7, respectively. Taking into account the experimental values of FB
1 adsorption (
Figure 3), it can be supposed that 1 kg ton
−1 of the product can adsorb more than 50% of the toxin in an acidic system containing up to 160 µg g
−1 of FB
1, a toxin concentration that exceeds the higher EU-guidance level (60 µg g
−1) two-fold. At neutral pH, such a dosage of the product can reduce by 50% a toxin concentration lower than 0.2 µg g
−1. Previous FB
1 adsorption studies using tri-octahedral bentonites as binders reported lower adsorption capacities, with Ads
max not exceeding 50 µg mg
−1 [
13].
The experimental values of OTA adsorption were in the ranges of 47–21% (pH 3) and 11–5% (pH 7) and were recorded by testing the product at 0.05 and 1 mg mL
−1, respectively (
Figure 3). The respective values of Ads
max and K
L were 11.1 ± 0.4 µg mg
−1 (28 ± 1 mmol Kg
−1) and 1.5 ± 0.1 L mg
−1 ((63 ± 5) × 10
4 L mol
−1) at pH 3; and 0.30 ± 0.03 µg mg
−1 (0.8 ± 0.1 mmol Kg
−1) and 0.4 ± 0.1 L mg
−1 ((15 ± 2) × 10
4 L mol
−1) at pH 7 (
Table 6). The experimental values of OTA adsorption isotherms and the Langmuir adsorption constants varyied depending on the pH, and allowed us to calculate the maximum toxin concentration that can be reduced by 50% by 1 kg ton
−1 of the product. At acid pH this value was 1 µg g
−1, which is four times higher than the higher EU-guidance level (0.25 µg g
−1). These results show the good efficacy of the product in adsorbing OTA, and are in accordance with the findings of the preliminary study performed by Vila-Donat et al. [
12].
At pH 3 and 7, the experimental values of ZEA adsorption were in the ranges of 73–45% and 77–48%, respectively, and were obtained by testing the adsorbent at 0.5 mg mL
−1. Values of Ads
max were 4.6 ± 0.1 µg mg
−1 (14.3 ± 0.4 mmol Kg
−1) at pH 3 and 7.2 ± 0.5 µg mg
−1 (23 ± 2 mmol Kg
−1) at pH 7. These values are comparable to those recorded by Vila-Donat et al. for tri-octahedral smectites [
12]. The values of ZEA adsorption affinity, K
L, were similar when calculated at pH 3 and 7, being 0.9 ± 0.1 L mg
−1 ((28 ± 2) × 10
4 L/mol) and 0.5 ± 0.1 L mg
−1 ((17 ± 2) × 10
4 L mol
−1), respectively. Finally, taking into account the results of adsorption isotherms, it can be calculated that 1 kg ton
−1 of the product can bind more than 50% of ZEA with a concentration up to 4 µg g
−1, regardless of medium pH. This toxin concentration is eight times higher than the higher EU-guidance level (0.5 µg g
−1).
T-2 toxin adsorption was slightly affected by pH and was well described by linear adsorption isotherms (
Figure 3). A linear isotherm (also known as the Henry isotherm) assumes that the concentration of adsorbate on the adsorbent surface remains constant with the adsorbate concentration, and it is usually applied to the adsorption of hydrophobic analytes in water systems. It provides no information on the monolayer adsorption capacity, in contrast to the Langmuir model. Experimental adsorption values of T-2 adsorption were in the range 66–51% at pH 3 and 62–50% at pH 7. These values were obtained testing the product at 5 mg mL
−1 dosage. At both pH values, the maximum adsorption capacity determined experimentally was 1.0 µg mg
−1 (2.1 mmol Kg
−1). Isotherms in
Figure 3 plotting the percentage of adsorption as a function of the weight ratio between the toxin (µg) and the product (mg) suggest that 1 kg ton
−1 of the product can reduce by 50% a fixed toxin concentration up to 2 µg g
−1. To date, this is the first study focused on the efficacy of a new additive containing a tri-octahedral smectite as the main component for sequestering T-2 toxin.
The new multi-mycotoxin adsorbing agent, prepared by mixing a selected tri-octahedral smectite with a biosorbent obtained from vegetal biomasses, showed maximum adsorption capacities (Ads
max) for target mycotoxins higher than those reported in the literature for di-octahedral smectites with hydrothermal origin, tri-octahedral smectites or biosorbents [
6,
12,
13,
18,
19].
The overall experimental data of this in vitro study seem to support the hypothesis that the new additive can be a valuable dietary approach to remove mycotoxins from contaminated feed. The work sheds light on the additive’s mode of action and may help to interpret the results of the in vivo study recently published by our group [
48]. In this preliminary study with poultry, the additive was assessed for its efficacy in counteracting the deleterious effects of an aflatoxin-contaminated diet (0.02 mg kg
−1) supplemented with the additive at a dose of 5 kg ton
−1 in the feed, and administered for 10 days. As described in the report published by Longobardi et al. [
48], the additive reverted the nephrotoxicity induced by AFB
1. Poultry is very sensitive to AFB
1 intake, and oxidative stress caused by AFB
1 plays a crucial role in chickens’ kidney damage by generating lipid peroxidation accompanied by a concomitant increase in the antioxidant enzymes involved in ROS metabolism (NADPH oxidase isoform 4 (NOX4) and its regulatory subunit p47-phox). The inclusion of the additive into the contaminated diet down-regulated both the transcription and the expression of NOX4 in chicken kidneys, together with its p47-phox subunit.
Taking into account the results of the present study, it can be suggested that the additive sequestered AFB1 at the GI level, thus reducing dietary exposure to AFB1 and counteracting oxidative stress in poultry. The antioxidant properties and the high polyphenol content in the lignin included in the final formula of the additive could have helped ameliorate the toxic effects of AFB1.