Fish farming accounted in 2018 for 52% of the total world fish used for food, excluding China, with an estimated production of 82 million metric tons (FAO, 2020). These values have steadily increased over the last decades. However, the present world challenges, such as the decreasing and availability of wild fishery stocks (Tacon and Metian, 2008) and the ongoing climate change, have changed the production paradigm. Therefore, the aquaculture industry is switching to a more sustainable model. The first step to achieve this is the generalized use of plant-based aquafeeds and fish feed additives (Tacon et al., 2011; Daniel, 2018). While this seems the right strategy, it raises new concerns on feed quality, such as pesticide use, market competition, deforestation, and mycotoxins presence.

Mycotoxin contamination and proliferation in aquafeeds represent increasing fish farming concerns (Nácher-Mestre et al., 2015, 2018). Recent numbers estimate contamination by molds and fungi in 25% of the crops (Pandya and Arade, 2016). Some mold species produce mycotoxins (including aflatoxins) as secondary metabolites (Zhang et al., 2014). Liposoluble aflatoxins have low molecular weight circa ~0.7 kD (Zhang et al., 2014) and readily cross the skin, and the gastrointestinal epithelium (Ditta et al., 2019). Aflatoxin B1 (AFB1) is the most potent carcinogen and genotoxic among aflatoxins (Flores-flores et al., 2015; Ostry et al., 2017).

Previous studies focused on the dietary effects of AFB1 in farmed fish, including the gilthead seabream (Sparus aurata), showed primarily species-specific growth impairment (Santacroce et al., 2012; Gonçalves et al., 2018a,b; Barany et al., 2020). Despite the growth inhibition, there is a lack of apparent pathological signs, and the adverse impact of AFB1 usually goes unnoticed in farmed fish (Anater et al., 2020). Especially interesting is the absence of AFB1 in fish fillets or liver in the gilthead seabream or salmonids after chronic dietary feeding (Nácher-Mestre et al., 2018; Barany et al., 2020), which contrasts with the findings in other species (Huang et al., 2014; Bedoya-Serna et al., 2018) and makes this issue more disturbing. We suggest that the putative absence of AFB1 residue in edible flesh could be due to high AFB1-tolerance implicit in intrinsic-species metabolism (Huang et al., 2011; Deng et al., 2018a,b) or that derived AFB1 bioactive metabolites are responsible for its effects. Such could be the case for AFB1 Exo-8,9-epoxide metabolite (Guengerich et al., 1998), for which commercial detection tests are yet undeveloped.

Anti-nutritional factors present in feeds or the environment (e.g., natural substances or toxicants) impact digestion, the intermediary metabolism, and cause inflammation (Kiron et al., 2020). Many types of anti-nutrient compounds appear in aquafeeds, especially in plant protein-rich feeds. Fermentation, heat treatments, antioxidants inclusion, and clay sequestrants are some methods used for anti-nutrient removal (Estensoro et al., 2016; Chakraborty et al., 2019). The effectiveness of the methods is dependent on the anti-nutritional compound nature. Specifically, several sequestrants or additives have been tested with aflatoxins to counteract their adverse effects (Rodrigues et al., 2019; de Freitas Souza et al., 2020).

Recent studies have pointed to the intestinal epithelium as a potential target of the actions of mycotoxins in humans, specifically in epithelial integrity (Akbari et al., 2017). Intestinal epithelia display two distinctive features: polarity across membranes and tightness among epithelial cells. The asymmetric distribution of proteins that regulate ion movements across the apical and basolateral membranes generates polarity. In contrast, tight junctions (TJs), which are multiprotein junctional complexes, prevent cell-to-cell leakage (Shen et al., 2011) and determine epithelial integrity and selectivity. The TJs permeability dynamic is mediated by claudins and occludins, among other elements (Mukendi et al., 2016). Both integral membrane proteins are the major structural and functional proteins within TJs to form the intercellular seal and permeability control (Claude, 1978; van Itallie and Anderson, 2006, 2013; Zhang and Guo, 2009).

According to their function, claudins are classified as barrier claudins that reduce permeability or pore claudins that promote selective ion permeability (van Itallie and Anderson, 2006; Hou et al., 2013). They can also interact with intracellular signaling molecules, linking extracellular transport, and intracellular signaling events (Collins et al., 2017). More than 63 genes encoding claudins exist in teleost fish (Kolosov et al., 2013; Lu et al., 2013). Of those, 30 claudins show some expression level in the gastrointestinal tract (GIT), but there is no evidence of their function (Tipsmark et al., 2010). The role of occludins is linked to ion selectivity and permeability, based on morphological, physiological, and mRNA expression results on studies performed in frog urinary bladder and in weaning piglet intestine (Claude, 1978; Zhang and Guo, 2009). Overall, information on occludins distribution and physiology is scarce at best.

Our recent work in seabream (Sparus aurata) showed that chronic ingestion of AFTB1 impairs growth and metabolism (Barany et al., 2020) as well as the physiological responses to additional stressors, such as crowding densities (Barany et al., 2021). In this line, we hypothesized that the gastrointestinal system is at the base of the previously reported growth impairment caused by AFB1 in fish since it establishes the first contact with food and its components. Therefore, this study aimed to characterize the direct effects of AFB1 on intestinal physiology by 1) comparing electrophysiological properties in vivo vs. in vitro expositions (i.e., short vs. long-term); 2) the morphological modifications caused in vivo after long-term exposure; and 3) the mRNA expression alterations in situ on putative proteins that keep cell-to-cell adhesion along with the gastrointestinal architecture.

Dietary AFB1 did not significantly change plasma osmolality or Na⁺ levels (Table 2).

The present study shows that AFB1 has an impact on intestinal physiology in the seabream. An observation in line with our previous work in this species showing adverse effects of AFB1 on seabream growth, disruption of metabolism, and tissue integrity, for example, liver (Barany et al., 2020). Here, our experimental approach was twofold; on the one hand, we used tissues from healthy fish to assess the immediate effects of AFB1 in vitro in Ussing chambers. On the other hand, we assessed the intestinal response to dietary AFB1 using fish obtained from the same experimental procedure employed by Barany et al. (2020).

An in vitro experiment using Ussing chambers was devised to test the putative direct effects of AFB1 on the anterior intestine of fish without previous contact with the toxin. Although the advantage of this approach is that intestinal explants maintain the original function and architecture of the tissue, they have a limited life span of ~4 h (Fuentes et al., 2006). The concentrations used in vitro were in the low micromolar range (e.g., 8 and 16 μM of apical AFB1). These values were derived from calculations of the dietary leaching of AFB1 used in feeding experiments (Barany et al., 2020), considering feed AFB1 content (1 or 2 mg/kg of feed), feeding rates, intestinal fluid volume, and putative residence time of food (Gilannejad et al., 2019). Short-term exposure to apical AFB1 at these levels did not affect the primary markers of tissue integrity TER or intestinal permeability. Our results contrast with previous studies in human colon carcinoma Caco-2 cell cultures (Romero et al., 2016), where AFB1 exposure in vitro significantly reduced TER but at more prolonged periods. Therefore, this apparent disparity in effect might be related to the timing of exposure or apical AFB1 doses employed since the highest concentrations in our tests were about 6-fold lower (16 vs. 100 μM) than those used in Romero et al. (2016).

Furthermore, we measured apical to basolateral fluxes to determine putative changes in the paracellular pathway. The results with FITC (4 kD) showed no significant effects in permeability. This apparent lack of direct action on paracellular permeability might be related to unchanged permeability for larger organic solutes and/or the intrinsic biology of different intestinal regions. It has been previously shown how the permeability might differ from smaller to larger organic solutes and within other intestinal regions in mice (Tanaka et al., 2015). Although we did not find statistical differences in permeability, it appears that high concentrations of apical AFB1 (16 μM) may lower the paracellular fluxes for larger molecules according to the TER increase observed in the in vivo feeding experiments. Similarly, an inverse relationship between gut permeability and TER has been reported in parasitized fish intestines (Sitjà-Bobadilla et al., 2019) or intestinal mycotoxicosis (Gao et al., 2018, 2020).

In contrast, long-term feeding experiments showed that dietary AFB1 modified intestinal selectivity severely by increasing TER but was without effect on Isc. TER is the gold standard to assess barrier integrity and function (Wijtten et al., 2011). Interestingly, in vivo dietary AFB1 significantly enhanced TER in the anterior but not in the posterior intestine, showing up anterior-to-posterior differences in barrier function previously reported in seabream (Carvalho et al., 2012; Gregorio et al., 2012). To our knowledge, no other fish studies have been performed in this context. Although, the increase in tissue resistance in response to dietary AFB1 is somehow surprising considering that in Caco-2 cells, AFB1 in the culture medium decreased TER after 48 h of exposure (Gao et al., 2018). However, Galarza-Seeber et al. (2016), using gut leakage of FITC in an in vivo chicken model, showed that the integrity of the gut barrier function was unaffected by dietary AFB1. We suggest that these apparent disparities across studies in TER may be linked to intestinal inflammation due to a rebound effect caused by a secondary response in the long-term exposure to AFB1. Therefore, eliciting a putative biphasic inflammatory response characterized by an initial transient TER decrease that precedes a subsequently TER increase if the aflatoxicosis persists. Intestinal inflammation decreases the tissue capacity for absorption and might even revert to secretion, thus causing diarrheic disorders (Schulzke et al., 2009; Verkman et al., 2013) that ultimately impair growth performance (Barany et al., 2020). Rebound effects are recurrent physiological responses in order to adjust the organism to face a new situation or as a detrimental side effect (Fenwick et al., 1999; Montserrat et al., 2007).

The GIT is the first contact point with feed components. Therefore, dietary aflatoxins are promptly absorbed into the bloodstream from the GIT (Battacone et al., 2012; Rodrigues et al., 2019). In tetrapods, the small intestine (homologous to the anterior fish intestine) has a greater surface area due to the presence of villi (Kiela and Ghishan, 2016), while in fish, it has also been reported a greater number of villi and globet cells within this same region (Verdile et al., 2020). Note that the villi are conformed mostly by absorptive cells (Kiela and Ghishan, 2016). Thus, in our feeding experiments, we assumed the anterior intestine was exposed to much higher doses of AFB1 released from the feed than the posterior since it is in this region where the AFB1 is firstly absorbed (Rodrigues et al., 2019).

In the intestine, TJs separate the paracellular space from the intestinal lumen acting as a divider of apical and basolateral domains of plasma membranes (van Itallie and Anderson, 2013), thus regulating the paracellular movement of ions, water, and small molecules. Claudins and occludins are essential elements of a functional tight junction and determine the TER of the intestinal epithelium. Therefore, the modification in barrier properties reflected by TER is tightly linked with changes in these key structural proteins of the TJ (Curry et al., 2020). To understand the underlying molecular network of altered tissue resistance in response to dietary AFB1 observed in feeding experiments in the seabream, we performed a qPCR array that included several claudins and occludins. Bearing in mind that in most studies performed to date, claudin gene mRNA expression usually reflects protein levels and can be considered a proxy for their function (Alexandre et al., 2005; Tanaka et al., 2015). In our study, dietary AFB1 had a wide range of effects on claudin and occludin effects in expression (measured by qPCR) in the seabream intestine. Specifically, it seems that the anterior intestine is predictably more responsive than the posterior intestine to AFB1 exposure, in line with the observed histopathological effects (e.g., inflammation). This observation is unsurprising since AFB1 has a low molecular weight and is assumed to be rapidly absorbed first by the proximal intestine (Rodrigues et al., 2019).

Previous studies in grass carp showed that mycotoxins downregulated barrier-forming TJ proteins expression, such as zonula occludens 1, zonula occludens 2b, ocln, cldnc, cldnf, cldn7a, cldn7b, and cldn11, whereas cldn12 and cldn15a were upregulated (Huang et al., 2018). Based on several overexpression studies, certain claudins, including cldn1, cldn3, cldn4, and cldn15, increased TER, whereas cldn2 had opposite effects (Singh and Harris, 2004; Tamura et al., 2008; Milatz et al., 2010). Additionally, certain claudins’ functions have also been linked to intestinal crypt stem cell survival, self-renewal, and epithelial differentiation (Tamura et al., 2008; Xing et al., 2020). Remarkably, differential overexpression among claudins possibly modified ionic paracellular selectivity, for example, Cl⁻ and Na⁺ (Alexandre et al., 2005), and organic solute fluxes (Tanaka et al., 2015).

In our study, dietary AFB1 in the seabream significantly downregulated cldn3 in the anterior intestine and cldnb in the posterior intestine, while cldn7b, cldn15, and cldn34 were upregulated in the anterior intestine. Thus, some of these changes in expression probably have physiological implications that may trigger a switch in tissue selectivity reflected by the in vivo TER increase and/or the apparent decrease in the intestinal tissue permeability properties for larger organic solutes (~4 kD) as previously shown by Curry et al. (2020). Similarly, previous studies have provided a relationship between TER decreased and permeability enhancement (Villarroel et al., 2009).

Few studies have investigated the role of occludin-proteins/genes in fish to the best of our knowledge. Overall, their distribution is tissue-specific, and they play a role in hydromineral balance through the regulation of epithelial tightness. Such is the case in fish, where occludins play an essential role in regulating paracellular solute movements (Chasiotis and Kelly, 2008; Tipsmark and Madsen, 2012) and show the unambiguous distribution to apical sections of osmoregulatory tissues, such as the gill, kidney, and GIT. Previous studies have correlated low protein expression with leaky epithelia, whereas high protein expression appears in tight epithelia (Chasiotis and Kelly, 2008). Occludin downregulation parallels decrease TER in Caco-2 epithelial monolayers (Musch et al., 2006; Zeissig et al., 2007).

Here we identified three different occludins, for example, ocln1, ocln2, and ocln3 in the GIT system of seabream, without an apparent pattern of anterior–posterior expression differentiation. On analyzing the dietary effects of AFB1, we found a significant increase of relative mRNA expression of ocln1 and ocln3 in the anterior intestine, likely associated with the TER enhancement in response to dietary AFB1 as previously reported in cells in vitro with androgens (Kaitu’u-Lino et al., 2020). Note that occludin 1b in the human distal intestine might also play an essential role in intestinal leakage across TJs (Muresan et al., 2000). In addition, corticosteroids (e.g., cortisol) and other hormones can also directly modulate epithelial permeability characteristics reflected by changes in TER (Chasiotis et al., 2010; Trubitt et al., 2015). We believe this is not the case in our study since we found no differences in plasma cortisol in response to dietary AFB1 (Barany et al., 2021). Thus, and besides endogenous hormones, anti-nutritional factors, such as AFB1, seem to alter the intestinal barrier function properties throughout diet alone. Similar suppressive effects on occludins relative expression have also been reported in the intestine of Atlantic salmon (Moldal et al., 2018). Specifically, and among other effects, deoxynivalenol was suggested to diminish the expression of occludins in the distal intestine.

Mycotoxins also alter the functions and histomorphology of the intestine (Liew and Mohd-Redzwan, 2018; Alassane-Kpembi et al., 2019). In the present study, dietary AFB1 induced desquamation with exudate in the intestinal lumen of the anterior region, while goblet cells (GC) decreased in the epithelial lining of the intestinal villi. Previous studies on AFB1 dietary effects on fishes also observed sloughing of surface epithelial in the intestine of Catla catla (Andleeb et al., 2015) and described hypoplasia of GC villi sloughing in carps exposed to AFB1 (Khan et al., 2019). In our study, dietary AFB1 caused in the seabream edema in the villi, necrosis of epithelial cells, and mononuclear cells infiltration in the submucosal layer of the villi. Also, subepithelial spaces were observed due to the separation of mucosa and submucosa layers (Akinrinmade et al., 2016). Thus, we suggest some of these gut-architectural affections might be caused by the specific toxic binding of AFB1 to constituents of TJs directly exposed to the toxicant (Sonoda et al., 1999), plus that observed claudins dysregulation may also interfere with TJs signaling to regulate local cell turnover (Tanaka et al., 2015; Xing et al., 2020).

In summary, our results showed that dietary AFB1 modifies fundamental mechanisms of intestinal physiology in the seabream. AFB1 impacts intestinal histomorphology and alters barrier function. Interestingly, no short-term effects were observed, indicating that chronic exposure, and molecular alterations, are needed to reveal its impact. Indeed, we reported that long-term dietary AFB1 modifies barrier function as shown by TER measurements and scrambles the expression of key components of the tight junction, claudins, and occludins. Based on these results, we conclude that dietary AFB1 in the gastrointestinal system is the base of the previously reported growth impairment of AFB1 in the seabream (Barany et al., 2020) that later led to altered physiological stress responses to crowding densities (Barany et al., 2021). Further studies regarding the specific physiological involvement of the described claudins and occludins in the current study are warranted to fully understand the likely complex combinatorial interactions linked to intestinal permeability regulation.

All data generated or analyzed during this study are available as Supplementary Material for this article.

The animal study was reviewed and approved by Junta de Andalucía (reference number 28-04-15-241) and by the Ethics and Animal Welfare Committee from the Spanish Government (RD53/2013). In addition, all animal manipulations were carried out in compliance with the European (Directive 2010/63/EU) and Portuguese legislation for the use of laboratory animals. All animal protocols were performed under Group-C licenses from the Direção-Geral de Alimentação e Veterinária, Ministério da Agricultura, do Mar, Ambiente e Ordenamento do Território, Portugal.

AB, JM, and JF contributed to the conception and design of the research. AB performed osmolality, enzyme activity, electrophysiology, molecular analysis, data curation, statistical analysis, and drafted manuscript. AB and SFG performed Na⁺ and permeability analysis. MO performed the histopathological examination. GM-R supervised molecular analyses and performed phylogenies. AB, MO, SFG, GM-R, JM, and JF edited and revised the manuscript. All authors contributed to the article and approved the submitted version.

This work was funded by the Spanish Ministry of Economy and Business-MINECO (AGL2016-76069-C2-1-R) awarded to JM. The authors (AB and JM) belong to the Fish Welfare and Stress Network (AGL2016-81808-REDT), supported by the Agencia Estatal de Investigación (MINECO, Spanish Government). AB was supported by the University of Cadiz Ph.D. scholarship (PIF UCA/REC02VIT/2014). CCMar is supported by national funds from the Portuguese Foundation for Science and Technology (FCT) through project UIDB/04326/2020.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.