The Compromised Intestinal Barrier Induced by Mycotoxins

Mycotoxins are fungal metabolites that occur in human foods and animal feeds, potentially threatening human and animal health. The intestine is considered as the first barrier against these external contaminants, and it consists of interconnected physical, chemical, immunological, and microbial barriers. In this context, based on in vitro, ex vivo, and in vivo models, we summarize the literature for compromised intestinal barrier issues caused by various mycotoxins, and we reviewed events related to disrupted intestinal integrity (physical barrier), thinned mucus layer (chemical barrier), imbalanced inflammatory factors (immunological barrier), and dysfunctional bacterial homeostasis (microbial barrier). We also provide important information on deoxynivalenol, a leading mycotoxin implicated in intestinal dysfunction, and other adverse intestinal effects induced by other mycotoxins, including aflatoxins and ochratoxin A. In addition, intestinal perturbations caused by mycotoxins may also contribute to the development of mycotoxicosis, including human chronic intestinal inflammatory diseases. Therefore, we provide a clear understanding of compromised intestinal barrier induced by mycotoxins, with a view to potentially develop innovative strategies to prevent and treat mycotoxicosis. In addition, because of increased combinatorial interactions between mycotoxins, we explore the interactive effects of multiple mycotoxins in this review.


Introduction
Mycotoxins are the non-enzymatic poisonous metabolites produced by fungi such as Aspergillus, Penicillium, and Fusarium genera [1,2]. In recent years, approximately 500 mycotoxins derived from these fungal species and others have been identified [3]. Of these, considerable attention has been given to several common mycotoxins, which affect both human and animal health, as well as economic growth. These include aflatoxins (AFs), ochratoxin A (OTA), deoxynivalenol (DON), fumonisins (FBs), zearalenone (ZEN), patulin (PAT), nivalenol (NIV), and citrinin (CTN) [4,5]. In terms of carcinogenicity, mycotoxins have been classified into five groups by the International Agency for Research on Cancer  [5,6] PAT hepatotoxicity Group 3 PMTDI = 0.4 µg/kg bw/d [1,5,9] NIV immunotoxicity, hematotoxicity, myelotoxicity Group 3 TDI = 1.2 µg/kg bw/d [2,5,10] T-2, HT-2 toxic on the skin and mucous membranes Group 3 PMTDI (T-2+HT-2) = 0.06 µg/kg bw/d TDI (T-2+HT-2) = 0.1 µg/kg bw/d [5]  As depicted (Figure 1), the term chemical barrier refers to the mucus layer of antimicrobialrelated proteins (e.g., mucins and AMPs) secreted by intestinal epithelial cells [21,28]. This barrier prevents luminal bacteria from coming into direct contact with the intestinal epithelium [29]. Mucins are the main constituents of the mucus' thick matrix and are produced and secreted by goblet cells [30,31]. Mucin 2 (MUC2) is the most abundant protein in the intestine and plays an important role in mucus layer integrity and function [30]. In mice lacking MUC2, luminal bacteria come into direct contact with epithelial cells, thereby inducing inflammationrelated diseases such as spontaneous colitis and ulcerative colitis (UC) [32]. AMPs are primarily As depicted (Figure 1), the term chemical barrier refers to the mucus layer of antimicrobial-related proteins (e.g., mucins and AMPs) secreted by intestinal epithelial cells [21,28]. This barrier prevents luminal bacteria from coming into direct contact with the intestinal epithelium [29]. Mucins are the main constituents of the mucus' thick matrix and are produced and secreted by goblet cells [30,31]. Mucin 2 (MUC2) is the most abundant protein in the intestine and plays an important role in mucus layer integrity and function [30]. In mice lacking MUC2, luminal bacteria come into direct contact with epithelial cells, thereby inducing inflammation-related diseases such as spontaneous colitis and ulcerative colitis (UC) [32]. AMPs are primarily secreted by Paneth cells, which are only located in the small intestine, while low AMP levels are secreted by enterocytes [28].
The immunological barrier comprises immune cells in LP (e.g., dendritic cells, resident macrophages, B cells, and T cells) and the secreted immune mediators (e.g., secretory immunoglobulin A (sIgA) and cytokines) [23,33,34]. SIgA is generated by B cells in the LP, transposed across the epithelium, and secreted into the mucus layer [35]. Cytokines are the important regulators in inflammatory and immune responses, and they are secreted by immune cells and epithelial intestinal cells [18].
More than 10 14 commensal bacteria, known as the microbiota, comprising hundreds of different species, colonize the mammalian intestinal tract to form the intestinal microbial barrier. This microbial ecosystem plays a key role in maintaining intestinal health, as microbial dysbiosis leads to intestinal inflammation [36][37][38]. To comprehensively understand mycotoxin-mediated damage to the intestine, we systematically analyzed dysfunction mechanisms related to these physical, chemical, immunological, and microbial barriers.

Experimental Models Used to Assess the Intestinal Barrier
Several in vitro, ex vivo, and in vivo approaches have been used to recapitulate animal intestinal environments to assess and characterize the intestinal barrier (Table 2). Table 2. The characteristics of in vitro, ex vivo, and in vivo intestine models.

Models Types Advantages Limitations References
In vitro-2D  intestinal model   Caco-2 cells, IPEC-1 cells,  IPEC-J2 cells, IPI-2I cells and  PSI-1 cells, co-culture of  different cell lines) well-established and relatively cheap only containing a single cell type without villus and crypt domain [39][40][41][42][43][44] In vitro-3D intestinal model enteroids, also known as organoids or mini-guts partially recapitulate the anatomy of native epithelium, have the ability to passage at an almost unlimited scale the effects of substances on the luminal side are poorly investigated, considerable cost, do not contain the immune and stromal cells [41,[45][46][47][48][49] Ex vivo applied in humans, rodents, swine, poultry and horse a more accurate model to mimic the physiology in vivo fail to achieve long-term culture, careful and laborious preparation [45,50,51] In vivo commonly used models include mouse, rat, chicken, turkey, fish, pig, sheep and bovine provide the information based on the whole animals, thus they could corroborate the toxicity in humans effectively the use of live animals should follow 3R (replacement, reduction and refinement) principle

In Vitro Intestine Model
In vitro models, representing the intestinal barrier, commonly refer to intestinal epithelial cells cultured in transwell chambers with a semipermeable membrane filter. They are typically referred to as two-dimensional (2D) cultures. To date, several intestinal epithelial cell lines from different animal species and humans have been generated. Among these, the porcine intestinal epithelial cell lines, IPEC-1 and IPEC-J2, and the well-established human Caco-2 cell line are the most frequently used intestinal models for studying barrier functions [39,44]. After seeding onto transwell chambers for 1-3 weeks, these cell lines spontaneously differentiate to form polarized monolayers, representing in vitro intestinal physical and chemical barriers [52][53][54]. In addition to these models, IPI-2I and PSI-1 cells from the adult boar ileum and adult pig small intestine, respectively, have also been used as intestinal barrier models [40,55]. However, in considering the complexity interaction between different cell types, monocultures may only partially represent the intestinal barrier. Therefore, multi-culture systems are often preferred [56]. The co-culturing of polarized IPEC-J2 cells with porcine peripheral blood mononuclear cells (PBMCs), and the co-culture of Caco-2 cells and intestinal HT29-MTX cells, has been successfully applied to evaluate intestinal barrier modulation by mycotoxins [42,43,57]. While these aforementioned cell lines are commercially available, freshly isolated intestinal primary cells from various animal species are considered as more accurate models in mimicking in vivo physiology. However, they fail to achieve long-term culture. Therefore, different cell lines are chosen based on specific research purposes.
In addition to the traditional approach of 2D intestinal epithelial cell monolayers, 3D-intestinal models (enteroids), also known as "organoids" or "mini-guts", which are derived from individual intestinal stem cells, have been developed [41,47]. Currently, based on the enteroids technique of human and mouse models, 3D culture systems have been established for various species including pig, chicken, cow, sheep, and horse [58]. Furthermore, enteroids can be cultured in a multitude of tissues including liver, stomach, colon, etc. [59]. These systems are advantageous for the following reasons: (i) enteroids harbor most, if not all, intestinal cell types [48]; (ii) it is almost impossible to accumulate mutations; (iii) enteroids can be passaged to an almost unlimited scale [45,48,49]. Although enteroids are regarded as a new exploratory intestinal model, they have several limitations: (i) when compared with the well-established and relatively cheap 2D approaches, enteroid systems are considerably costly [41,46]; (ii) the effects of molecules on the enteroid luminal side are poorly investigated [49]. To circumvent this latter hurdle, a combinatorial 2D culture and 3D enteroid approach was developed, where the enteroids were dissociated into single-cell suspensions and then cultured in the transwell chambers [60]. When compared with conventional 2D cultures, 3D enteroids partially recapitulate Toxins 2020, 12, 0619 5 of 42 the anatomy of native epithelium. If researchers require better models to simulate the intestine, and the corresponding technology is available, then 3D enteroid approaches are the best choice.

Ex Vivo Intestine Model
Ex vivo studies involve the isolation of living functional tissue or organs from an organism and culturing in an Ussing chamber [61]. This model has been applied to various species, including humans [50], rodents [62], swine [51], poultry [51], and horses [63]. A main advantage of this intestinal explant culture systems is that the model maintains the histological architecture and complex intestinal barrier integrity under in vivo conditions [64]. Similarly, researchers can isolate specific segments of the gut, from the duodenum to colon, that could answer their research questions [61]. Small bowel explant cultures have been successfully used for celiac disease studies [65], and colon explants investigations have been powerful in studying the effects of intestinal microorganisms [66,67]. However, major drawbacks exist for intestinal explants, thereby restricting their application. These include short-term preservation during culture and careful and laborious preparation [39,61]. While these limitations are technical in nature, ex vivo intestinal models are widely used to evaluate gastrointestinal toxicity induced by mycotoxins [68][69][70][71][72].

In Vivo Intestine Model
In vivo studies, based on information from whole-animal systems, can effectively emulate the toxicity mechanisms in humans. The role of live animals for education and research purposes has been recognized by the World Organisation for Animal Health (OIE) [73]. In general, several animal species, such as mice, rats, chickens, turkey, fish, pigs, sheep, and cows, have been used to evaluate the toxicology of mycotoxins. When choosing a particular animal model for toxicological research, it must be borne in mind that animal species often show different susceptibilities to mycotoxins [1,74]. In much of the contemporary research on mycotoxicology, DON has been the most frequently analyzed mycotoxin. Researchers have observed that traditional in vitro systems, based on cell models, cannot comprehensively characterize pharmacokinetics, bioavailability, and in vivo metabolism, thus necessitating the requirements for in vivo animal models [75]. However, there are the drawbacks to such experimentation (i.e., time-consuming, expensive, and individual differences cannot be ignored). Additionally, from an ethical and animal welfare perspective, the use of live animals should follow the 3R (replacement, reduction, and refinement) principle. Therefore, toxicological assessments cannot be completely dependent on animal tests. Critically, experimental models should be selected based on the scientific area, e.g., in vivo models focus on the overall changes in an organism, while in vitro model investigations are ideal if particular pathways are affected by mycotoxins. Figure 2 shows information on an intestinal barrier compromised by mycotoxins, including disrupted intestinal integrity (physical), thinned mucus layer (chemical), imbalanced inflammatory factors (immunological), and dysfunctional bacterial homeostasis (microbial). In the following sections we discuss the effects of common mycotoxins such as AFs, OTA, DON, ZEN, FB1, PAT, and CTN.

Effects of Mycotoxins on the Physical Barrier
Numerous studies have shown that mycotoxins disrupt the intestinal physical barrier (Tables 3  and 4). In general, intestinal epithelial cells are self-renewed every 3-5 days [76], except for Paneth cells, which undergo renewal every 18-23 days. Thus, rapidly proliferating and regenerating intestinal epithelial cells maintain the intestinal physical barrier. Besides these cells, TJ proteins also play roles in the intestinal physical barrier. A previous study has reviewed the effects on intestinal permeability induced by mycotoxins [34]; therefore, in the present study, we will pay more attention to the latest research. Figure 2 shows information on an intestinal barrier compromised by mycotoxins, including disrupted intestinal integrity (physical), thinned mucus layer (chemical), imbalanced inflammatory factors (immunological), and dysfunctional bacterial homeostasis (microbial). In the following sections we discuss the effects of common mycotoxins such as AFs, OTA, DON, ZEN, FB1, PAT, and CTN.        Several mycotoxins modulate the intestinal physical barrier by affecting in vitro intestinal cell proliferation. AFB1 and AFM1 significantly inhibited cell growth, causing cell cycle arrest in Caco-2 cells [77][78][79][80]. In addition, AFM1 and OTA inhibited the cell viability of the co-culture of Caco-2/HT29-MTX [43]. Equally, OTA promoted apoptosis in IPEC-J2 and Caco-2 cells by inducing mitochondrial reactive oxygen species (ROS) and arresting the cell cycle [81][82][83]. Apart from mycotoxin effects in human intestinal cells, DON inhibited intestinal epithelial cell viability and induced apoptosis in rat and pig cell models [84][85][86][87][88]. ZEN-induced cell death was confirmed in IPEC-J2, IPEC-1, and human colon carcinoma cells (HCT116 cells) [89][90][91][92]. FB1, PAT, and CTN induced cell death and apoptosis in the human colon proliferating intestinal cell line (HT-29), Caco-2 cell, and HCT116 cell [93][94][95][96]98,126,127]. These studies showed that mycotoxins caused intestinal epithelial cell death, resulting in damage to the intestinal physical barrier.
Apart from cell proliferation, TJ-mediated intestinal permeability is also modulated by mycotoxins. AFM1 and OTA exposure to both apical and basolateral surfaces increased intestinal permeability, reduced TJ proteins protein expression levels, and affected the TJ protein distribution pattern [54]. These TJ protein localization effects were confirmed by transmission electron micrographs [57]. OTA treatment resulted in intestinal barrier dysfunction, which was confirmed by increased cell permeability and microvilli disruption and TJ proteins in various cell culture systems [82,105]. These observations could be explained by ROS/Ca 2+ -mediated myosin light chain kinase (MLCK) activation [128]. Exposure to DON decreased transepithelial electrical resistance (TEER) values and the abundance of TJs in a concentration-and time-dependent manner [106,108,109]. It has been reported that ZEN has no effect on TEER values of IPEC-1 cell, while its metabolites (α-ZOL and β-ZOL) led to time-dependent decreased TEER values [118]. FB1 and its metabolite, hydrolyzed FB1 (HFB1), damaged the intestinal integrity in different intestinal cell systems [42]. Previous studies have shown that PAT and T-2 toxin increased differentiated Caco-2 cells permeability, via the destruction of TJs, and was accompanied by MLC2 phosphorylation [94,95,122,124]. These results indicated that TJ expression levels are associated with TEER values and potentially represent a change in intestinal permeability.

Effects of Mycotoxins on the Ex Vivo Physical Barrier
An increasing number of ex vivo studies have determined the effects of mycotoxins on the intestinal physical barrier. Recently, DON was reported to inhibit growth and reduce ZO-1 and claudin-1 expression levels in enteroids isolated from jejunal crypts in porcine and mice [109][110][111]. When jejunal explants from weaning piglets were exposed to DON, decreased jejunal scores and MAPK activation were recorded [72]. FB exposure to an ex vivo rat large intestine induced lipid peroxidation, which could alter cell membrane permeability [129]. Apical villi necrosis was found in pig jejunal explant exposure to PAT [123]. DON appears to be the most studied mycotoxin in ex vivo studies. Therefore, it could be speculated that an ex vivo model is also highly applicable for other mycotoxins.

Effects of Mycotoxins on the In Vivo Physical Barrier
AFB1 and AFM1 have the ability to evaluate caspase-3 and caspase-9 mRNA expression and reduce the Bcl-2/Bax ratio in mice intestinal tissues, suggesting these mycotoxins induce apoptotic events [100]. When broiler chicks were exposed to AFB1, researchers observed increased lactulose:rhamnose (L:R) ratio and diamine oxidase (DAO) in plasma, and reduced clauin-1 and occludin mRNA expression levels in the mid-jejunum [101,102]. DON-mediated impairment of barrier function in rats and grass carp was associated with the depressed crypt depth ratio and TJ amount in the jejunum, which was associated with Wnt/β-catenin and MLCK signaling pathways [108,110,111,113]. In addition, DON-treated weaned piglets exhibited lower disaccharidase (maltase, sucrase, and lactase) activity, suggesting detrimental effects on gut health [112]. Similarly, the effects caused by ZEN in the post-weaning gilts intestinal physical barrier were partially elucidated by ZEN-induced oxidative stress mechanisms [120]. Similar results were shown in rats challenged with ZEN; TJ proteins in the jejunum exhibited down-regulated mRNA expression [119]. The chronic ingestion of FB1 decreased villi height and occludin expression in the piglet intestine [121]. In addition, T-2 toxin treatment disrupted intestinal histology in turkeys and mice, including the duodenum, jejunum, ileum, and colon [124,125].
The adverse effects on intestinal morphology in various in vivo models have been recorded for different mycotoxins such as AFs [99,103,104], OTA [107], and DON [72,[114][115][116][117]. In addition, a variety of clinical chemistry analytes such as urine, feces, and blood may also reflect defects in intestinal barrier function in vivo (e.g., fatty acid-binding protein and C-reactive protein levels in serum, and fecal hemoglobin and fecal calprotectin) [130,131]. However, these parameters have not been widely used to evaluate the effect of mycotoxin effects. Therefore, they may provide new insights in future studies.
These observations show that when the intestine is affected by short-term, low-concentration mycotoxins, self-regulating cellular abilities maintain the intestinal physical barrier integrity. However, when such damage is sustained and exceeds the self-regulating capabilities, intestinal epithelial cells and TJ proteins become disrupted, leading to barrier compromise. Similarly, long-term exposure of animals to low mycotoxin doses could also adversely affect intestine health.

Effects of Mycotoxins on the Chemical Barrier
The contribution of mycotoxins to intestinal chemical barrier disruption (e.g., mucins and AMPs) has been extensively conducted across in vitro, ex vivo, and in vivo studies (Table 5). To date, hundreds of AMPs have been identified including LEAP-2A, LEAP-2B, hepcidin, and β-defensin1.

Effects of Mycotoxins on the In Vitro Chemical Barrier
In the co-culture of Caco-2/HT29-MTX cells, AFM1, OTA, and ZEN down-regulated the protein expression of intestinal mucins MUC2, MUC5AC, and MUC5B [43,57,135,136]. Similar results were also observed for DON where it inhibited MUC1, MUC2, and MUC3 mRNA levels in human goblet cells (HT29-16E cells) [134]. T-2 toxin exposure led to a thinned MUC2 layer in HT-29 cells and reduced MUC2 protein expression in Caco-2 cells [124]. Up-regulated porcine β-defensin 1 and β-defensin 2 mRNA expression was also observed following exposure to DON, ZEN, FB1, and NIV, whereas no significant increases in β-defensin 1 and β-defensin 2 secretion occurred in IPEC-J2 cells [137]. In addition, in co-culture studies with Caco-2/HT29-MTX cells, MUC5AC and MUC5B mRNA expression was decreased by DON and NIV exposure, while their protein levels were increased [136]. These differences in mRNA and protein levels may be partly explained by post-transcriptional or post-translational regulatory mechanisms, or protein degradation pathways. Equally, discrepancies in quantification techniques between studies may have also led to these differences, as protein quantification sensitivity is not as high as transcript measurements [144].

Effects of Mycotoxins on Ex Vivo and In Vivo Chemical Barriers
Few studies have investigated mycotoxin effects on the intestinal chemical barrier. One study showed that, in jejunal explants from pigs, DON time-dependently inhibited mucin mRNA expression levels [134]. In vivo exposure to a mycotoxin-contaminated diet also regulated mucin and other protein productions. Both AMPs and MUC2 mRNA expression levels were inhibited in the intestine of broiler chickens, mice, and juvenile grass carp upon DON exposure [110,111,[138][139][140]. Furthermore, reduced goblet cells in the small intestine were observed in OTA and DON treatment [107,112,115,121,140,141]. In contrast, no changes in goblet cells and crude mucin production were observed in laying hens and gilts exposed to AFB1, DON, or ZEN [133,142]. Furthermore, goblet cell hyperplasia (increased goblet cell number) was observed in AF/FB1-challenged broilers and mice exposed to combined DON and ZEN [104,121,132,145]. In general, goblet cell hyperplasia led to increased mucins secretion. T-2 toxin exposure resulted in increased MUC2 mRNA levels in the jejunum of chickens [143]. These observations suggest that the continuous hyper-secretion of mucins is likely to deplete goblet cell number, ultimately disrupting the mucus layer [146].
A variety of responses (increase, no change, and/or decrease) were observed in mycotoxin-induced mucins and goblet cells. Different experimental models using different species could account for these inconsistencies. Precise mechanisms remain unclear, however, and more studies are required to resolve these issues.

Effects of Mycotoxins on the Immunological Barrier
Mycotoxin effects on the intestinal immunological barrier in in vitro and in vivo models are summarized (Table 6).

Effects of Mycotoxins on the In Vitro Immunological Barrier
OTA down-regulated the gene expression of cyclooxygenase-2 (COX-2) and lipoxygenase-5 (5-LOX) genes in Caco-2 cells, which have been regarded as inflammatory mediators [105]. The mRNA expressions of pro-inflammatory cytokines including interleukin-1 beta (IL-1β), IL-6, IL-8, COX-2, and tumor necrosis factor-alpha (TNF-α) were increased in the DON-exposed porcine intestinal epithelial cells, with DON-mediated inflammation partially dependent on ROS production and MAPK pathway activation [106,[151][152][153][154]. Another study also reported that ZEN exerted no effects, or only marginal ones, on cytokine gene expression in IPEC-1 cells [156]. However, other studies reported an increasing tendency towards inflammatory cytokine gene expression and synthesis in ZEN-exposed IPEC-1 cells [91,118,157]. In addition, the ZEN metabolites α-ZOL and β-ZOL decreased pro-and anti-inflammatory cytokine expressions [118]. The exposure of FB1, in a concentration-dependent manner, to IPEC-1 cells depressed IL-8 mRNA and protein expression, while no changes in IL-8 secretion were recorded in HT-29 cells [93,158]. These studies show that mycotoxins, especially ZEN, exert different inflammatory cytokine expression profiles, suggesting underlying mechanisms are complex and require further in-depth study. Additionally, there is no evidence on the impact of AFs on intestinal inflammation. Therefore, more studies are required to fill this important knowledge gap.

Effects of Mycotoxins on the In Vivo Immunological Barrier
AFB1 and AFM1 down-regulated pro-inflammatory cytokine production in the small intestine of mice and pigs [100,147]. AFB1 also affected sIgA and polymeric immunoglobulin receptor (pIgR) expression, thereby affecting sIgA transport in epithelial cells into mucus layers. Transcriptome analysis revealed that seven processes or pathways related to the immune system were grouped in Litopenaeus vannamei as the result of AFB1 challenge [148]. OTA exposure resulted in up-regulated IL-1β and TNF-α mRNA expression in the small intestine of broiler chickens and pekin ducklings [107,150]. Similar results were also observed for DON and FB1 exposure, where cytokine production was significantly up-regulated in the intestine of piglets and juvenile grass carp [121,138,155]. In other studies, in the intestines of pregnant dams, weaned pups, and pigs treated with ZEN and FB1, pro-inflammatory cytokine mRNA expression was significantly decreased [119,158].
CD4 + cells are involved in cytokine production, and CD8 + cells play a crucial role in protecting host target cells [160,161]. Mycotoxin-contaminated diets led to decreased CD4 + and CD8 + cell percentages in chickens [162]. In addition, CD4 + cell numbers were decreased in the jejunum of broilers treated with AFB1 [149]. In addition, CTN administration altered immune cell (CD4 + , CD8 + , and CD19 + ) populations in the small intestine of mice [159]. Similarly, cell morphology effects have been observed in these models; in AF-challenged broiler chicks, the lamina propria lymphoid follicle diameter (LLFD) and lamina propria lymphoid follicle number (LLFN) in the jejunum were increased [104]. In addition, DON exposure increased the lymphocyte number in the intestine of pigs fed DON-contaminated diets [141,142], indicating mycotoxin-mediated inflammatory responses.
These data reflect the limited research on disrupted intestinal chemical barriers induced by mycotoxins in ex vivo models. While these studies have enlightened the literature on AFs, OTA, DON, and ZEN, the effects of other mycotoxins such as FB1 and PAT require greater attention.

Effects of Mycotoxins on the Microbial Barrier
Several studies have shown the impairment to the intestinal microbial barrier is caused by mycotoxins (Table 7).

Effects of Mycotoxins on the In Vitro Microbial Barrier
The simulator of the human intestinal microbial ecosystem (SHIME) was used to investigate OTA effects, and it showed that gut microbiota diversity was altered, with the loss of beneficial species Lactobacillus reuteri [167]. In another study, Escherichia coli (E. coli) were translocated across IPEC-J2 cell monolayers when induced by DON [152]. This finding was consistent with previous studies, demonstrating that bacterial translocations across IPEC-1 and Caco-2 cell monolayers are induced by DON, NIV, and PAT [166,169]. Importantly, bacterial translocation appears to be concentration-and species-dependent [179], with incubation times and cell line types potentially affecting the results [166].

Effects of Mycotoxins on the In Vivo Microbial Barrier
In broiler chicks, the dietary supplementation of AFs resulted in markedly increased ileal bacterial counts of E. coli, Klebsiella, Salmonella, and total negative bacteria [104,132,165]. In addition, AFB1 reduced intestinal bacterial flora diversity in rats, shrimp, and mice [163,164,180]. In AFB1-treated shrimp, the application of high-throughput sequencing analysis showed that Bacteroidetes relative abundance decreased, and the abundance of Proteobacteria and Firmicutes increased [180]. These results were consistent with a previous study demonstrating that OTA exposure significantly increased the relative abundance of Lactobacillus, but decreased overall gut microbiota diversity in rats [168].
Cecal species' richness and evenness were decreased upon DON exposure in broiler chickens and mice, Bacteroidetes and Firmicutes abundance increased, while Proteobacteria decreased, suggesting an overall impaired gut microbiota community [170,171]. A fecal microbiota analysis explored DON-induced gut microbiota changes in nude mice and pigs and showed that DON treatment generated a higher abundance of Clostridiales, Lachnospiraceae, and fecal aerobic mesophilic bacteria [140,172]. In addition, DON exposure increased Bacteroides and Prevotella genera levels but decreased E. coli levels in a model of human microbiota-associated rats [174].
The biolog EcoPlate method showed that ZEN exposure significantly reduced E. coli, Clostridium perfringens, and Enterobacteriaceae levels in gilts [177]. Additionally, ZEN increased Desulfovibrio and decreased Lactobacillus in mouse colon tissue [176]. After exposure to FBs, the digestive microbiota balance was impaired in pigs [178]. In contrast, AFB1 and DON exerted no effects on gut microbiota diversity and relative abundance in broilers and rats [102,173,175].
In general, microbiota communities in the colon, cecum, and feces are representative of the intestinal bacterial flora. However, microbiota differences in different intestinal sections are still unclear. Similarity, studies on mycotoxin (i.e., ZEN and FB1) mediated alterations in the intestinal microbial barrier are limited. Therefore, more research is required in this area.

Contribution of a Leaky Gut to Intestinal Inflammatory Disease
Currently, links between mycotoxin exposure and some human carcinogenic and teratogenic diseases, including Reye' syndrome, cirrhosis, hepatitis, and esophagus cancer, have been demonstrated [181]. However, potential links between mycotoxins and human chronic intestinal inflammatory diseases remain unclear. Among the chronic intestinal inflammatory diseases, inflammatory bowel diseases (IBDs) including ulcerative colitis (UC), Crohn's disease (CD), and celiac disease (CeD) are well studied. The dysfunction of intestinal barrier and increased intestinal permeability, also known as "leaky gut", are believed to play a prominent role in the etiology of these diseases [182,183].
Intestinal epithelial cell death and abnormal TJ expression cause increased transcellular and paracellular transport. These perturbations may cause increased access of different molecular weight xenobiotics and bacterial translocation, ultimately activating local and systemic immune responses [23,184]. Among the abnormalities of TJs, claudin switching is the most notable, which refers to the balance between different members of the claudin family [185,186]. In particular, the down-regulation of several tight claudins, such as claudins 3, 4, 5, 7, and 8, and the increased expression of leaky claudins, such as claudin 2 and 15, are often associated with an inflamed gut, including UC, CD, CeD, and irritable bowel syndrome [185,[187][188][189][190]. In addition, reduced TJ expression could accelerate mucosal inflammation (ulceration and colitis), as observed in claudin 2 and claudin 7 knockout mice, and claudin 2 overexpression investigations in mice [191][192][193].
In addition to intestinal physical and immunological barriers, chemical barrier disruption also leads to inflammatory disease. Perturbations in intestinal mucus layers, including mucus composition and production, contribute to human chronic intestinal inflammatory disease. In addition, significant modifications to gut microbiota number and composition were observed in IBD patients and were characterized by decreased anaerobic bacteria and increased Enterobacteria numbers [181]. The intestinal microflora may also affect mucus production and TJ expression, potentially leading to inflammatory responses [204,205]. As such, the intestinal barrier is an interconnected system. Therefore, disruption to one or more components may result in human intestinal disease [20,206]. Several factors are involved in the induction and persistence of the chronic intestinal inflammatory diseases: (i) abnormal TJs composition, (ii) altered production of mucus layers, (iii) changes in intestinal immunity, and (iv) modification of intestinal microflora or increased bacterial translocation (Figure 2). Given that healthy relatives of CD patients manifested increased intestinal permeability without disease [207,208], we conclude that a compromised intestinal barrier is not the sole factor causing inflammatory diseases, but intestinal dysfunction could exacerbate inflammation and enhance its severity [21,183].

Interactive Effects of Mycotoxins on Intestinal Barrier
Most fungi produce one or more mycotoxins. Therefore, the co-occurrence of these molecules in disease should receive more attention. A previous study demonstrated that approximately 28 mycotoxins were quantified by a liquid chromatography tandem mass spectrometry (LC-MS/MS) multi-toxin method, in samples collected from Burkina Faso and Mozambique [209]. A recent study also showed a greater co-exposure phenomenon, where the analyzed maize samples were contaminated by 5 to 41 mycotoxins in Malawi [210]. In addition, mycotoxin co-occurrence was also identified in oil, dairy milk, and infant formulas [211][212][213]. After ingestion of mycotoxin-contaminated food, these molecules may be absorbed into the blood. To corroborate this, UHPLC-MS/MS analysis identified 26 mycotoxin biomarkers, including AFB1, OTA, FB1, ZEN, and DON, in plasma and urine samples from 260 rural residents in China [214]. Equally, similar data have been observed in different populations across the globe [215][216][217][218].
The interactive effects of multi-mycotoxins are classified as synergistic, additive, and antagonistic, and they represent the effects of a components' mixture are higher, equal, and lower than the sum of the effects induced by individual components. For the intestinal physical barrier, AFM1 cytotoxicity against Caco-2 cells was enhanced in the presence of OTA, ZEA, and α-ZOL, with additive and synergistic effects demonstrated by most combinations [219]. The additive and synergistic effects on reduced cell viability were demonstrated by combining DON and ZEN in a bi-culture Caco-2, THP-1, and HepaRG cell system [220]. The effects on increased intestinal permeability and decreased TJ expression in differentiated Caco-2 cells exposed to a combination of AFM1 and OTA were significantly stronger than the individual mycotoxins, performing additive and synergistic effect [54]. In addition, non-toxic OTA aggravated DON-induced TJ disruption in IPEC-J2 cells, suggesting a synergistic effect was at play [106].
From a chemical barrier perspective, an AFM1 and OTA combinatorial approach produced a synergistic effect by depressing mucin expression in Caco-2/HT29-MTX co-cultures [43]. In contrast, antagonistic interaction of combined DON and ZEN was observed for MUC5AC expression levels [135]. Non-additive (including synergistic and antagonistic) effects were also shown for the modulation of mucins and β-defensins expression, induced by combinations of DON, FB1, ZEN, and NIV [136,137]. Furthermore, different interactions were observed in the combination of DON and FB on intestinal immunological changes [121]. Additionally, the antagonist effect was observed in the combined DON and ZEN on gut microbiota levels [177].

Conclusions
Mycotoxin contamination of human food and animal feed results in economic loss and health detriments. In addition to the carcinogenic, mutagenic, and genotoxic effects mediated by these molecules, mycotoxin-induced impairment of the intestinal barrier has also gained considerable attention in recent years. To assess intestinal toxicity, in vitro epithelial barrier investigations frequently use human (Caco-2 and HT29 cells) and porcine epithelial cells (IPEC-1and IPEC-J2) as well as enteroids. Ex vivo models are represented by explants from mice and pigs. The predominant in vivo models use rodents (mice and rats) and pigs, followed by chickens and fish. Based on these varied models, our review revealed that the mycotoxins, i.e., AFs, OTA, ZEN, DON and FB1, decreased TJ protein expression levels, depressed mucins and AMPs secretion, activated cytokine production, and altered gut microflora composition. These elements represent the intestinal physical, chemical, immunological, and microbial barrier targeted by mycotoxins.
It is entirely plausible that mycotoxins are implicated in human chronic intestinal inflammatory disease. In this review, we showed evidence linking a compromised intestinal barrier to inflammatory disease, but further studies are required to fully confirm this hypothesis. Therefore, underlying mechanisms remain an issue. More studies are required to investigate how to control and decrease mycotoxin-contaminated foods and animal feeds, to protect human and animal health, and to reduce economic loss. CXCL10 C-X-C Motif Chemokine