Mycotoxins are toxic secondary metabolites produced by filamentous fungi, including species within the Aspergillus, Penicillium and Fusarium genera. These microorganisms are known to infect cereals at multiple stages of production, ranging from field to storage; and result in a persistent risk of mycotoxin contamination throughout the entire production chain (Daou et al. 2021; Xu et al. 2023). Considering that cereals are major components of regular livestock diets, animals become vulnerable to the ingestion of contaminated grains. This exposure may lead to reduced performance, such as lower feed intake, milk yield, and growth efficiency and/or fertility (Vila-Donat et al. 2018; Xu et al. 2022a; Kihal et al. 2022).

Among the most commonly detected toxins in feed are Aflatoxins (AFs), Ochratoxin A (OTA), Trichothecenes, Zearalenone (ZEN) and Fumonisins (FBs) (Awuchi et al. 2022). Considering FBs, Fumonisin B₁ (FB₁) is the predominant compound encountered, especially in maize and its by-products (Beccaccioli et al. 2021; Pamphile and Azevedo 2002). Therefore, the occurrence of FB₁ in animal feed has been documented worldwide, with reports from Europe, the Americas, Asia and Africa showing contamination levels ranging from 24.5% to 100% (Gao et al. 2023).

The presence of FB₁ in products intended for animal consumption is concerning, as it has been linked with equine leukoencephalomalacia and porcine pulmonary edema. Additionally, its ingestion can also result in the reduction of feed intake, lower egg production, hepatic necrosis, and thymic cortical atrophy in poultry (Awuchi et al. 2021; Chen et al. 2021; Schrenk 2022; Yang et al. 2020). Effects on ruminants are briefly described, with previous studies showing lower milk yield after FB₁ ingestion by dairy cattle; whereas adult beef cattle seem to be more resistant, exhibiting only mild hepatic necrosis (Diaz et al. 2000; Osweiler et al. 1993; Smith 2018).

Therefore, economic pressures, particularly in the livestock industry, are motivating producers to seek effective solutions to prevent the adverse effects of mycotoxins on animal health and productivity. However, if grain contamination has already occurred, preventative measures are no longer feasible. Consequently, the implementation of mitigation strategies intended for the removal, degradation or inactivation of toxins becomes crucial. Therefore, integrated approaches that combine multiple strategies are essential to guarantee safe mycotoxin levels in animal feed (Shi et al. 2018; Xu et al. 2022a).

Described methods include chemical, biological and physical treatments of grain. Chemical agents, such as ammonia, hydrogen peroxide, and organic acids, are highly effective in reducing mycotoxin levels. Nevertheless, the use of such reagents can make raw materials inedible for animals and contributes to environmental pollution. Biological strategies involve the use of enzyme-producing microorganisms to degrade or bio-transform toxins, which presents challenges for in-field application due to environmental interactions. Lastly, physical methods include sorting and separation, floating and segregation by density, irradiation, ultrasound treatment, dehulling, milling and adsorption (Luo et al. 2018; Peng et al. 2018).

Adsorption involves the use of an adsorbing agent (AA), which forms mycotoxin–adsorbent complexes through direct binding with the toxins. This process reduces mycotoxin bioaccessibility and, therefore, intestinal absorption; as the formed complexes are excreted via feces. As a consequence, this also leads to a decrease in mycotoxin uptake and in its spread to target organs (Boudergue et al. 2009; Xu et al. 2022a). In contrast with other mitigation strategies, which aim to diminish mycotoxin levels during feed production, AAs are used as feed additives. Additionally, the main advantages of such binders are their cost, safety and ease of administration through inclusion in feed (Peng et al. 2018).

Among employed AAs are organic (yeast cell wall, yeast cell wall beta-D-glucan fraction, oat and alfalfa fibers) and inorganic (bentonites, montmorillonites, zeolite and activated carbon) compounds (Kihal et al. 2022). In general, inorganic binders are highly efficient in adsorbing AFs, due to the high polarity and small molecule size of these toxins. However, such products are relatively inefficient in adsorbing other mycotoxins, especially Fusarium toxins, such as Deoxynivalenol (DON) and ZEN. In this context, the use of organic binders is recommended (Vila-Donat et al. 2018).

Considering the FB₁ molecule size and its structural conformation, adsorption may be a challenging issue, especially with inorganic products. To address this limitation, adsorption can be increased through the incorporation of biological components, such as algae extracts, to inorganic binders. This modification enhances the interlayer spaces of clays, increasing surface area and resulting in more binding sites for mycotoxins (Cai et al. 2024; Oguz et al. 2022; Rasheed et al. 2020; Wang et al. 2023; Xu et al. 2022b).

Algal extracts are appealing options for AAs, as they are well-known for their biosorption/bioremediation of heavy metals (Cheng et al. 2019; Lin et al. 2020). These organisms’ cell walls contain a wide range of proteins and polysaccharides, such as β-D-glucans, that are potential binding sites for mycotoxins. Additionally, other algal compounds could also be responsible for adsorption capacity, such as chlorophyll and chlorophyllin; which may also form complexes with mycotoxins, reducing their bioaccessibility (Simonich et al. 2007).

Furthermore, the inclusion of algae as feed additives may not only help prevent the adverse effects of mycotoxins but also promotes improved animal health (Perali et al. 2020). Algae display high bioactive compound content, which can act as prebiotics and therefore enhance animal immunity. In addition, production performance (i.e. weight gain, feed intake, and feed conversion rate) may also be increased, since algae are rich in proteins, vitamins, and minerals (Fraga-Corral et al. 2023; Makkar et al. 2016; Yadavalli et al. 2023).

Estimating the adsorbent efficiency of different products is indispensable, since it depends on the type of adsorbent, its physico-chemical properties, and the target mycotoxins. In vivo testing of mycotoxin binding efficacy of a large number of adsorbents is challenging, due to the complexity and cost; making in vitro analyses powerful tools for assessing and ranking the efficacy of various AAs. Nevertheless, experimental conditions for in vitro tests reported in the literature may vary widely, making it difficult to establish reliable protocols (Boudergue et al. 2009; Faucet-Marquis, 2014). Considering the information cited above, the aim of this study is to develop an in vitro protocol for FB1 adsorption to evaluate and compare the binding efficacy of products composed of inorganic clay and algae extracts, in order to identify the most effective formulation for FB₁ adsorption in livestock feed.

LC-MS/MS for mycotoxin quantification was performed using the Agilent 1290 Infinity LC-System (Agilent Technologies, Santa Clara, USA) coupled to a 6460 triple quadrupole (Agilent Technologies, Santa Clara, USA) mass spectrometer, at the Food Toxicology Laboratory, Food Engineering School, State University of Campinas (UNICAMP).

The Zorbax SB-C8 column (3.0 mm x 100 mm; 1.8 µm) was used for chromatographic separation at 25 ºC. The analyses were performed in isocratic mode, using 30% of 0.1% (v/v) formic acid in water (mobile phase A) and 70% of 0.1% (v/v) formic acid in methanol (mobile phase B), with a flow rate of 0.35 mL/min and injection volume of 3 µL.

Mass spectrometry conditions were: Electrospray Source Ionization in the positive mode (ESI+) with gas temperature at 300°C, gas flow at 10 L/min, nebulizer at 35 psi, sheath gas flow at 10 L/min, sheath gas temperature at 350°C, capillary voltage at 3.0 kV and nozzle voltage at 0.5 kV. Ultrapure nitrogen was used in the nebulizer and as collision gas.

For FB₁, m/z 722 (precursor), m/z 353 (quantifier transition) and m/z 334 (qualifier transition) were monitored. The fragmentor applied was 185 V and collision energies were 40 eV for both quantifiers and qualifier transitions. Data acquisition and analyses were performed using MassHunter software workstation version 7.00 (Agilent Technologies, Santa Clara, USA) in selected reaction monitoring (SMR) mode, using a dwell time of 100 ms per channel.

The linearity of FB₁ was obtained based on calibration curves in the matrices (Figure S1; citrate buffer 0.1 M with 10% methanol at pH 3, 5, and 7). FB₁ standard was prepared at seven concentrations (0.5, 1, 2.5, 5, 7.5, 10, and 12.5 µg/mL) in triplicates, and subsequently analyzed under the conditions previously described (Sun et al. 2018). All linearities (R²) obtained were ≥ 0.99 at pHs 3, 5, and 7 (Figure S1). Recovery was evaluated in triplicates on the same day as the analysis, with results ranging from 95% to 110% for citrate buffer at pH 3, 5, and 7 (EC 2006). The limits of detection (LODs) and the limits of quantification (LOQs) for each matrix are provided in Table S1; these were calculated through the minimum concentration detected with a 3x and a 10x signal-noise ratio, respectively.

Furthermore, precision was assessed at three FB₁ concentration levels (2.5 µg/mL, 5 µg/mL and 7.5 µg/mL) by analyzing five replicates of each level in citrate buffer pH 5, without the presence of adsorbent. Intra-day precision was assessed by analyzing all replicates on the same day under the same conditions. Inter-day precision was determined by repeating the procedure on three different days. The results were expressed as the coefficient of variation (CV%) for each level (Table S2). All CV% values were within the acceptable limits of ≤ 20%, in accordance with SANTE/11312/2021 (EC 2021) (Table S2).

The selectivity of the method was assessed by analyzing blank and fortified samples (2.5 µg/mL of FB₁) in citrate buffer (pH 3, 5, and 7). The fortified samples exhibited well-defined peaks in the qualifier transition (m/z 334) with consistent retention times ranging from 1.928 to 1.955 minutes (Figure S2). No interfering peaks were observed in the blank samples at the corresponding retention times, demonstrating the selectivity of the method for FB₁ analysis (Figure S2).

pH

Three different pH levels (3, 5 and 7) were evaluated, to determine the optimal level to proceed with the trials. For this purpose, three FB₁ working solutions (5 µg/mL) were prepared in citrate buffer with 10% methanol at pH 3, 5 and 7, separately. Next, 5 mL of each solution was pipetted into 50 mL polypropylene tubes containing either: 0.1% (w/v) of A2, according to manufacturer instructions; 0.1% (w/v) of activated charcoal (AC; AC positive control); or no adsorbent (control without adsorbent); in triplicates (Faucet-Marquis et al. 2014)

Samples were then incubated (30 minutes; 37ºC; 8 x g) and centrifuged (15 minutes; 8586 x g). Subsequently, 1 mL of supernatant from each sample was retrieved, filtered by syringe filter (0.22 µm) and analyzed with HPLC-MS/MS (item 2.3.). After chromatographic analyses, the adsorption percentage of each test was calculated based on Eq. 1:

where C_ads is the adsorbed mycotoxin yield (µg/mL) and C₀ is the concentration of FB₁ in the controls without adsorbent (µg/mL) (Joannis-Cassan et al. 2011).

Four FB₁ concentrations (1 µg/mL, 2.5 µg/mL, 5 µg/mL, and 10 µg/mL) were prepared in 5mL of pH 5 citrate buffer (10% methanol) as solvent. Considering the evaluated product should demonstrate a high affinity and capacity to adsorb mycotoxins at its low inclusion rate in livestock diets, 0.1% (w/v) of product A2 was employed.

The tests were performed in triplicates, using 0.1% (w/v) of AC as positive control (AC positive control) and mycotoxin working solution without adsorbent as control without adsorbent. Next, the tubes were incubated (30 minutes; 37 ºC; at 8 x g) and centrifuged (15 minutes; at 8586 x g). After, approximately 1 mL of the supernatant was withdrawn, filtered by syringe filter (0.22 µm), and taken for HPLC-MS/MS analysis using the conditions described above (item 2.3.); and adsorption percentage was calculated based on Formula 1.

Considering the previously determined optimal pH (item 2.4.1.), 5 mL of increasing amounts of FB₁ (0.5, 1, 2.5, 5, and 7.5 µg/mL) were incubated in 50 mL polypropylene tubes with 0.1% (w/v) of adsorbent (A2), 0.1% (w/v) of AC (AC positive control) or without adsorbent (control without adsorbent). All tests were performed in triplicates and incubation conditions were 37 ºC, with rotation 8 x g for 30 min. Next, all samples were centrifuged (15 minutes; 8586 x g) and 1 mL of their supernatants were withdrawn separately, filtered by syringe filter (0.22 µm), and taken for HPLC-MS/MS analysis; as formerly described (item 2.3.).

This step was performed to determine the ideal intermediate concentration of FB₁, to evaluate the adsorption capacity of different algae-based products. For this purpose, the adsorption percentage of each concentration was calculated based on Formula 1 (item 2.4.1). Moreover, an isotherm curve was also elaborated by plotting the variations in the residual FB₁ concentration after adsorption in µg/mL (C_eq) against the quantity of adsorbed FB₁ per gram of adsorbent (Q_eq), calculated using Formula 2:

where C₀ is the concentration of FB1 in the blank controls (µg/mL), m is the mass of adsorbent in grams and V is the final volume of the solution in liters.

Additionally, the data obtained for the isotherm curve was fitted into two different models, Freundlich and Hill (Joannis-Cassan et al. 2011). In this regard, the first assumes the sorption occurs on a heterogeneous surface and is described by Formula 3:

where K_F is a constant representing the capacity of the adsorbent to bind to the mycotoxin (milligram ^{1 −} (^{1 ∕ n}_F) × liter ^{1 ∕ n}_F per gram) and n_F is a constant concerning the affinity of the adsorbent (Joannis-Cassan et al. 2011).

As for the Hill model, it is usually employed to describe the adsorption of diverging compounds onto a heterogeneous substrate (Joannis-Cassan et al. 2011), using Formula 4:

where Q_Hmax is the maximal mycotoxin adsorption corresponding to the site saturation (mg/g), K_D symbolizes the Hill constant (mg/L) and n_H corresponds to the cooperativity coefficient of the interaction among product and toxin (Joannis-Cassan et al. 2011).

In vitro adsorption tests are generally used as a screening method for the efficacy assessment of mycotoxin binders. Some of the main advantages of performing such assays are their lower cost and quicker results, when compared to in vivo analyses. However, most of the studies on in vitro adsorption trials do not present standardized and repeatable conditions, especially regarding the pH levels and mycotoxin concentration (Faucet-Marquis et al. 2014; Kihal et al. 2022). This is particularly important, as these assays represent the gold standard used by most companies developing adsorbents for livestock feed, since they mimic the animal gastrointestinal system. During digestion, mycotoxins ingested with feed can be bound by the adsorbent, reducing their absorption and subsequent biotransformation in the liver. Consequently, lower toxin levels reach systemic circulation, decreasing their transfer through the food chain into animal-derived products such as milk, eggs or meat (Avantaggiato et al. 2006; Kihal et al. 2023). Thus, validating a robust adsorption protocol is essential to ensure both animal and food safety.

The adsorption capacity of mineral products, such as bentonites, is directly linked to their physico-chemical characteristics, especially their interlayer spaces’ conformation/charge and cation exchange capacity. The latter is highly influenced by the pKa of the adsorbent (point zero charge) and the medium’s pH level. In this regard, when the pH level of the medium is lower than the pKa of the adsorbent, it results in a loss of charge and lower cation exchange capacity. Conversely, when the medium’s pH level is higher than the bentonite’s pKa, it becomes more electronegative and adsorption capacity is increased (Du et al. 2021; Kihal et al. 2022).

Considering that FB₁ is a highly polar compound, minor pH alterations can result in structural modifications as well as protonation/deprotonation of carbonyl groups. In acidic conditions, the FB₁ molecule is protonated, while in neutral-alkaline pH it exists as an anion (Momany and Dombrink-Kurtzman 2001; Šegvić and Pepeljnjak 2001). The results of the present study, show that adsorption was higher at pH 3, which is not consistent with the physico-chemical characteristics of FB₁ and bentonites. At this pH range, the cation exchange capacity and electronegativity of the AA should be low; which considerably reduces the binding performance when only bentonite is utilized (Barrientos-Velázquez et al. 2016).

The results obtained from the current study suggest that the adsorption of FB₁ by the algae-based product complex occurs mainly due to the presence of algae in the interlayers of bentonites. Algae mycotoxin adsorption mechanisms are not yet well understood. However, it is known that algal cell walls contain polymers, such as β-D-glucans, xylans, galactanes and mannans. These components are likely key contributors to the adsorption capacity of algae, as they can bind to mycotoxins through Van der Walls interactions or hydrogen bonds (Cheng et al. 2019; Fraga-Corral et al. 2023; Jouany et al. 2005; Yiannikouris et al. 2006). Such interactions are not based on charge exchange but rather depend on the structural stability of the polysaccharide molecules. In the case of β-D-glucans and other glucose-based polymers, this stability is pH-dependent. Under acidic conditions (pH 3), the structural rigidity of these polysaccharides increases, which improves their ability to bind to mycotoxins. Conversely, at neutral pH levels (5–7), the molecules are less stable and can suffer conformational changes, which might hinder mycotoxin adsorption (Faucet-Marquis et al. 2014; Yiannikouris et al. 2004; Yiannikouris 2006).

Therefore, the low FB₁ adsorption observed at pH 5–7 likely reflects the reduced activity of the algae component, due to pH-sensitive structural changes. Simultaneously, the bentonite may not have attained sufficient surface electronegativity to effectively contribute to adsorption at this pH interval. Altogether, this suggests that pH 5–7 likely represents an intermediate zone, where neither components perform optimally, leading to the observed decrease in overall adsorption efficiency. Moreover, our results also support the evidence that the modification of inorganic mineral products (i.e. bentonites) by adding organic extracts may increase their adsorption capacity; since previous reports indicate FB₁ adsorption by clays alone is moderate to low (Elliott et al. 2020).

In this respect, Rasheed et al. (2020), through the addition of orange peel extracts to pure bentonite, obtained efficient in vitro adsorption of Aflatoxin B1 (AFB₁), FB₁, and OTA. In their study, FB₁ adsorption exceeded 80% in buffered solutions at pH 2.5 and pH 7, as well as in simulated gastric fluids (acidic conditions). However, when the reaction was performed in simulated intestinal fluids (near-neutral conditions), the performances was considerably reduced to under 20% (Rasheed et al. 2020). These observations are consistent with our results, in which higher adsorption efficiency was also observed at acidic pH 3. Structurally, the organic components were found within the interlayers of the clay, like the algae-based products, generating additional binding sites for mycotoxins. The authors suggested that FB₁ might have formed peripheral interactive forces, especially Van der Waals interactions with the organic fraction of the product (Rasheed et al. 2020).

Additionally, Oguz et al. (2022) evaluated the adsorption efficacy of clays, plant extracts, and glucomannans, individually and in combination, using in vitro methods. For FB₁, the authors reported moderate binding levels for clays and glucomannans alone. However, when plant extracts and glucomannans were added to a mixture of clays, there was a 20% increase in the adsorption capacity of the product, particularly under acidic conditions (pH 3), where FB₁ adsorption reached 54.61%. (Oguz et al. 2022). These findings corroborate the results of the current study, in which FB₁ adsorption by algae-modified clay was greatest at pH 3.

As for in vivo studies, Tsiouris et al. (2021), by adding yeast cell walls and silymarin to clays, observed the reduction of adverse effects related to AFB₁ and OTA in broiler chicks. Treated animals also showed healthier intestinal conditions after ingesting the formulated product, which could hinder the development of gut pathogens, such as Escherichia coli, Clostridium perfringens, Salmonella spp. and Campylobacter spp. (Tsiouris et al. 2021). In addition, El-Nekeety et al. (2017) also formulated an organically modified clay and, as a result, the product was able to efficiently adsorb FB₁ and ZEN, reducing their toxic effects in rats.

An isotherm curve was constructed to characterize FB₁ adsorption by one algae-based formulation, to select an optimal concentration for testing different products (Boudergue et al. 2009; Joannis-Cassan et al. 2011). The shape of the curve resembled those observed in other studies involving FBs and organically modified clays (Baglieri et al. 2013). According to Baglieri et al. (2013) and Lemke et al. (1998), such a profile indicates the binding of the mycotoxin to specific sites. Furthermore, the data revealed that binding site saturation was not achieved at the tested concentrations; however, our preliminary results suggest that an FB₁ concentration of 10 µg/mL combined with 0.1% (w/v) of adsorbent may saturate the binding sites of the product.

Additionally, two different models were applied for isotherm curve fitting: the Freundlich model and the Hill model. The Freundlich model is applicable to non-linear multilayer adsorption, showing the exponential distribution of the active binding sites and their energies. In contrast, the Hill model describes cooperative interactions in biological systems and is mainly related to the binding of different species. Therefore, the Freundlich model appears to better fit the conditions and experimental data of the current study when compared to the Hill model (Al-Ghouti and Da’ana 2020; Rajahmundry et al. 2021).

After determining the FB₁ concentration to be 2.5 µg/mL from the isotherm curve, five different algae-based products were tested. The results showed adsorption percentages from 48% (A2) to 34% (A5). Moreover, green algae-based products exhibited significantly higher adsorption compared to those derived from red algae. This contrast may be attributed to the distinct compositions of the red and green algae cell walls (Romera et al. 2008). Red algae cell walls contain sulfated polysaccharides composed of galactanes serving as potential adsorption sites, whereas green algae have glycoprotein-rich walls with diverse functional groups (amino, carboxyl, sulfate, hydroxyl) acting as binding domains (Romera et al. 2008). At present, there are no studies comparing mycotoxin adsorption by multiple products containing different types of algae. However, both algal extracts and live algae have been widely used for heavy metal biosorption, with results showing that green algae generally display higher adsorption capacity than red algae, consistent with our findings (Boukarma et al. 2024; Romera et al. 2006).

Considering the use of only algae or algae-based products for mycotoxin adsorption, Perali et al. (2020) tested the binding efficacy of Lithothamnium calcareum, for the adsorption of AFB₁ in broiler chicks. As a result, the authors showed not only a reduction of adverse effects related to mycotoxin ingestion, but also an improved body weight, weight gain, and feed intake in the animals treated with L. calcareum (Perali et al. 2020). Conversely, another study involving DON and algae-modified clay, showed that the AA was not effective in adsorbing the toxin nor in avoiding the adverse effects related to DON in nursery pigs; possibly due to a low affinity of the product to this mycotoxin (Frobose et al. 2016).

Altogether, when preventive measures against mycotoxin contamination in animal feed are ineffective, adsorbents are recommended to reduce toxin uptake and prevent productivity losses (Boudergue et al. 2009; Jouany, 2007; Vila-Donat et al. 2018). FB₁, one of the most commonly found toxins in feed, has a complex and elongated structure, which makes its adsorption challenging. Therefore, it is essential to test and evaluate novel AAs to mitigate its adverse effects and enhance livestock performance (Gao et al. 2023; Oguz et al. 2022; Shi et al. 2018; Xu et al. 2022a). In this context, incorporating algae into livestock diets may simultaneously contribute to mycotoxin adsorption and improve animal health and immunity (Fraga-Corral et al. 2023; Yadavalli et al. 2023).

In vitro adsorption tests are useful tools for evaluating the adsorption capacity of multiple agents on a small scale, with lower cost and analysis time when compared to in vivo assays. However, adsorption mechanisms are complex, especially when dealing with organic-based products, with conditions such as pH, saturation of the adsorbent, and type of mycotoxin influencing the product performance being important. In this regard, the evaluation of multiple test parameters in the present study resulted in an adequate in vitro FB1 adsorption protocol to screen different algae-based products. The results also showed different adsorption capacities of the formulations evaluated, showing that green algae were more efficient than red algae-based products. Such findings may indicate that the different compositions of these organisms’ cell walls can influence their performance in FB₁ adsorption. In summary, the evaluated products, particularly those containing green algae, show promising potential for reducing FB₁ exposure in livestock. However, further research is needed to assess their efficacy across different feed matrices and under simulated digestion conditions, as well as through in vivo studies on FB₁ bioavailability. This is especially relevant given that the inclusion of algae in livestock diets has already been linked to improvements in animal health.