1. Introduction

Mycotoxins are secondary fungal metabolites (i.e., metabolites not essential to the normal growth, development, and reproduction of fungus) that cause pathological changes in animal species (1). They are low molecular weight compounds with diverse chemical structures and biological properties (Figure 1). A large number of fungal metabolites have been identified as mycotoxins, contaminating agricultural commodities worldwide. Fusarium, Pencillium, and Aspergillus are the three most important mycotoxin-producing fungal genera with respect to animal and human health (2). Fungal invasion and toxin production in crops can occur preharvest, at harvest, and during processing, transportation, and storage. Temperature and moisture can greatly affect the growth rate of fungi as well as the types and amounts of mycotoxins produced (1).

Figure 1 Structures of major mycotoxins.

Mycotoxins have attracted worldwide attention because of the significant economic losses associated with their impact on human health, animal productivity, and both domestic and international trade (3). It has been estimated that the impact of mycotoxins on the feed and livestock industries causes annual losses of $5 billion in the United States and Canada (3).

Because cereal grains and nuts are often used as ingredients in commercial pet food, companion animals such as cats, dogs, birds, rabbits, and guinea pigs are often exposed to the effects of mycotoxins. Cereal byproducts, furthermore, may be diverted to animal feed even though they may contain concentrated levels of mycotoxins compared to raw cereals (4, 5).

There have been several reviews of the physiological effects of mycotoxins in pet species in recent years (6-8). The current review, in contrast, addresses the worldwide prevalence of mycotoxin contamination in pet food and technological advances in preventing mycotoxicoses in companion animals. Mycotoxin outbreaks in companion animals, mycotoxin surveys of commercial pet food, challenges faced by the pet food industry, and some practical preventive strategies for mycotoxicoses will be discussed.

2. Mycotoxins in Pet Food

Mycotoxins commonly found in pet food include aflatoxins, ochratoxins, and the Fusarium mycotoxins. They differ in their toxic effects and their prevalence across regions, entering feed-manufacturing processes through contaminated raw materials (Table 1). These mycotoxins may have chronic effects on animal health. When highly contaminated ingredients are accidentally used in feed production, animals can develop acute mycotoxicosis (Table 2). Aflatoxins, ochratoxins, and tremorgenic myocotoxins have also resulted in several food-poisoning cases in pet animals due to the consumption of spoiled foods.

2.1. Aflatoxins. Aflatoxins are a group of mycotoxins produced by Aspergillus spp., mainly A. parasiticus and A. flavus (2). They are common fungal contaminants of nuts but are also found in many other feedstuffs (Table 1). Aflatoxins B₁, B₂, G₁, and G₂ are four naturally occurring forms of aflatoxins, with aflatoxin B₁ being the most potent, prevalent, and carcinogenic (6, 9).

The chemical structure of aflatoxin incorporates dihydrofuran and tetrahydrofuran moieties coupled to a substituted coumarin (10) (Figure 1). Aflatoxins may be converted to more reactive, electrophilic epoxides by phase I metabolism occurring primarily in the liver (1). The unstable epoxides may bind covalently to intracellular macromolecules such as DNA, RNA, and protein, resulting in adduct formation, cellular damage, and hepatic necrosis (11).

Aflatoxins are hepatotoxic and carcinogenic. Dogs exposed to 0.5-1 mg of aflatoxin/kg of body weight (bw) typically die within days, exhibiting vomiting, depression, polydipsia, polyuria, and hepatitis (7). Anorexia, lethargy, jaundice, disseminated intravascular coagulation, and death have also been described in dogs exposed to 0.05-0.3 mg of aflatoxin/kg of feed over 6-8 weeks. Newberne and Wogan (12) experimentally induced malignant tumors in rats with an 80-week exposure to 15 g of aflatoxin B₁/kg of feed. The sensitivity of companion animals to aflatoxin B₁, expressed as oral LD₅₀, varies considerably, with rabbits (0.3 mg/kg of bw), mink (0.5-0.6 mg/kg of bw), cats (0.55 mg/kg of bw), dogs (1.0 mg/kg of bw), and guinea pigs (1.4-2.0 mg/kg of bw) being more sensitive than mice (9.0 mg/kg of bw) or hamsters (10.2 mg/kg of bw) (10).

Many food-poisoning outbreaks in dogs have been associated with aflatoxin contamination because aflatoxins have a high toxicity in canine species (Table 2). Anorexia and depression are the two most visible symptoms of aflatoxicosis in dogs, although sudden deaths are also often reported. Dogs are typically exposed to aflatoxins through accidental inclusion of contaminated corn in the feed-manufacturing processes (13-19) and, in some cases, through improper preparation of domestically prepared dog food (20-22). In the former cases, aflatoxin outbreaks may persist for up to several months before being diagnosed, thereby affecting a large number of animals (16, 17, 19). Early diagnosis by veterinarians and prompt recall procedures are crucial in controlling such outbreaks.

Aflatoxins have been detected in commercial pet food worldwide, notably in North and South America (Table 3). Wild bird feed was found to be the most contaminated among different types of pet foods in several surveys, possibly due to the use of corn, nuts, and seeds as significant ingredients (23-25). Up to one-fourth of the wild bird feed samples were contaminated with more than 100 g of aflatoxin B₁/kg in two surveys. This presents a potential health threat to the birds (24, 25).

Commercial dog and cat foods, in contrast, generally have lower aflatoxin content. Although the percentage of aflatoxin-positive samples varies by survey, almost all of the positive samples contained <20 g of aflatoxin B₁/kg (23, 24, 26-28). Such a low level of aflatoxin exposure appears to be insufficient to cause noticeable symptoms in companion animals, but the chronic hepatotoxic and carcinogenetic effects should not be overlooked.

A number of aflatoxin-producing fungi have also been detected in commercial pet food (29, 30). Aspergillus flavus, for example, was found in 14 of 60 pet food samples purchased in Portugal (30). Scudamore et al. (23), in another study, identified significant growth of Aspergillus spp. in commercial pet food with 20-25% moisture content after a 4-week incubation. Although the presence of toxigenic fungi does not necessarily result in mycotoxin production, preventing fungal growth in pet food can certainly minimize the risk of mycotoxicoses. Bueno et al. (29) proposed that a moisture content below 11.50% could suppress fungal growth in pet food.

2.2. Ochratoxins. Ochratoxins are produced by numerous Aspergillus and Penicillium spp., including A. ochraceus and P. verrucosum (10). Ochratoxins are found principally in cereal grains but also in animal byproducts because of their tight binding with plasma proteins and long half-life in animal tissues (Table 1). There are four ochratoxin homologues: A, B, C, and D. Ochratoxin A is the most prevalent, whereas ochratoxins A and C are the most toxic (1).

Ochratoxins are phenylalanine-containing dihydroisocoumarins (Figure 1). Although their mechanism of action remains unclear, it has been suggested that ochratoxins act by disrupting phenylalanine metabolism (31). Ochratoxin A also possesses a chlorinated phenol moiety, a procarcinogenic group that causes electrophilic attack on DNA (28). Ochratoxins accumulate primarily in the kidneys as a result of high blood flow and tubular re-uptake (32). Ochratoxin A has also been found to interfere with signal transduction pathways at nanomolecular concentration in renal cells, leading to specific changes in function and phenotype but not to necrosis (33).

Ochratoxin A is a potent nephrotoxin in pet species. A 2-week exposure to 0.3 mg of ochratoxin A/kg of bw is lethal to young male beagle dogs, which exhibit severe kidney damage, anorexia, vomiting, weight loss, tenesmus, intestinal hemorrhage, tonsillitis, dehydration, and prostration (34). A 1-year exposure to 0.8 mg of ochratoxin A/kg of feed caused mild nephropathy in pigs (35). In addition to acute and chronic nephrotoxic effects, ochratoxin A has also been shown to be immunotoxic, teratogenic, and nephrocarcinogenic in animals (8, 32, 36, 37).

Although ochratoxin A is widely found in pet food (Table 3) and has a long half-life in vivo, the significance of the effect of ochratoxin A on the health of pet animals remains unclear. Several studies in Europe revealed that most pet food did not have a detectable amount of ochratoxin A or contained <5 g of ochratoxin A/kg (23, 30, 38, 39), and the chronic effect of such a low dietary exposure has not been investigated in companion animals. Pühringer et al. (39) investigated the ochratoxin A content of 101 feline kidneys in Austria, and 16 of the kidney samples contained 0.31-5.18 g of ochratoxin A/kg. This finding suggested a high dietary ochratoxin exposure in the cats but did not correlate with the pathological findings in the kidney samples. Occasional cases of acute ochratoxicosis in companion species have been reported in the literature (40-42; Table 2).

2.3. Fusarium Mycotoxins. The Fusarium mycotoxins are a chemically and biologically diverse group of compounds (Figure 1). The trichothecenes, zearalenone, and fumonisins are considered to be particularly important in animal health worldwide (43). The trichothecenes, a family of over 100 tetracyclic sesquiterpenoid compounds including deoxynivalenol, diacetoxyscirpenol, and T-2 toxin, can be produced by F. graminearum, F. sporotrichioides, and Stachybotrys chartarum (44). Zearalenone is produced primarily by F. graminearum, F. avenaceum, and F. nivale. Fumonisin and fusaric acid are produced by a variety of Fusarium species.

Many cereal ingredients of pet food are known to contain the Fusarium mycotoxins. Corn, for example, is commonly contaminated with fumonisins (45-49). Corn, wheat, and barley produced in North America are often contaminated with deoxynivalenol, the most common trichothecene (50, 51).

The Fusarium mycotoxins have different modes of action. Trichothecenes can cause vomiting, feed refusal, gastrointestinal irritation, and immunosuppression (1). Hughes et al. (52) reported anorexia and vomiting in dogs and cats exposed to 4.5 and 7.7 mg of deoxynivalenol/kg of feed, respectively. Zearalenone, on the other hand, is an estrogenic mycotoxin (53). A 7-day dietary exposure to 200 g of zearalenone/kg of bw/day was found to produce pathological changes in the canine reproductive system (54, 55). Fumonisins inhibit sphingolipid synthesis and metabolism and damage various organs in animals (1). Equine leukoencephalomalacia can be caused by <5 mg of fumonisin B₁/kg of feed and is characterized by feed refusal, depression, ataxia, blindness, and hysteria (1). Fumonisin B₁ was also found to cause liver and kidney cancer in rodents (56). Fusaric acid is a potent inhibitor of dopamine--hydroxylase, suppressing norepinephrine synthesis in the brain of dogs, cats, rabbits, and rats (57-59). Reduced appetite, vomiting, suppressed weight gain, and hypotension were observed after the feeding of young beagle dogs 50 mg of fusaric acid/kg of bw for 30 days and 6 months (60, 61).

Despite their marked prevalence in the cereal ingredients of pet foods, less research has been devoted to the Fusarium mycotoxins compared to aflatoxins and ochratoxins. A number of surveys, however, have shown that they are present in significant amounts in commercial pet food (Table 3). Zearale-none and fumonisin B₁, for instance, were found in 84 and 100% of samples, with the highest levels being 299.5 and 1410 g/kg of feed respectively (62, 63). Although acute Fusarium mycotoxicoses have not been reported to date in pet species, the chronic toxicity to pet animals should be of concern.

2.4. Tremorgenic Mycotoxins. Tremorgenic mycotoxins are produced by Penicillium, Aspergillus, and Claviceps (6). Unlike aflatoxins, ochratoxin A, and the trichothecenes, tremorgenic mycotoxins are rarely found in feed ingredients but are usually produced during food spoilage. Dogs are exposed to tremorgenic mycotoxins when eating moldy garbage, such as dairy products, nuts, grains, bread, and spaghetti (64). More than 20 mycotoxins have been classified as tremorgens, with penitrem A and roquefortine being the most important ones for companion animals (6, 65, 66).

Most tremorgenic mycotoxins are neurotoxic indole alkaloids (Figure 1). Penitrem A, for instance, causes slight salivation, vomiting, ataxia, and tremors in dogs, possibly by inhibiting inhibitory neurotransmitters, such as glycine and -aminobutyric acid in the central nervous system (6, 64, 67, 68). The mechanism of action of roquefortine, however, has not been studied in detail.

Oral exposure to 0.175 mg of penitrem A/kg of bw was sufficient to induce muscle tremors in dogs (67). The oral LD₅₀ values for penitrem A and roquefortine are 1.1 and 20 mg/kg of bw in mice, respectively (69, 70). There have been at least 10 reports of tremorgenic mycotoxicoses in dogs due to garbage consumption since 1979, mostly in North America (64, 67, 68, 71-78).

3. Government Regulation of Mycotoxins in Pet Food

Regulation of mycotoxin content of animal feed worldwide mainly focuses on farm animals, with less attention to companion species. In most countries, pet food is regulated by a maximum mycotoxin contamination for all feedstuffs rather than pet-specific legislation. Both the United States and Canada, for example, have a 20 g/kg legal limit for aflatoxins B₁, B₂, G₁, and G₂ for all animal feed (79, 80; Table 4). The U.S. recommendations on deoxynivalenol and fumonisins also apply to all feedstuffs (79, 81). The Canadian recommendations on deoxynivalenol, ochratoxin A, and zearalenone, in contrast, apply only to cattle, poultry, and swine feedstuffs (81). The European Union has imposed a 20 g/kg legal limit for aflatoxin B₁ in animal feed (82) and issued guidance values for deoxynivalenol and fumonisin (83).

Mycotoxin regulations vary among different countries. Whereas at least 50 countries regulate the aflatoxin content of feedstuffs, only 9, with 6 in Europe, have similar regulations for ochratoxin A (81). Switzerland and Brazil, for instance, have legal limits of 10 and 50 g/kg, respectively, for aflatoxin B₁ in all animal feeds (84, 85). Government regulations of mycotoxin contamination appear to reflect analytical detection limits and regional prevalence as well as trade relationships among different countries, but do not necessarily represent the safe limit for mycotoxin exposure in pet animals.

4. Challenges in the Pet Food Industry

The pet food industry has the same risks regarding mycotoxin contamination of raw materials as the feed industry for agricultural species. The most common challenges in the pet food industry include grain processing, sampling error, analytical methods, synergistic interactions, storage conditions, and conjugated mycotoxins.

4.1. Grain Processing. Cereal byproducts such as bran and feed grade flour are often included in the formulation of pet food, although they may contain concentrated levels of mycotoxins compared to grits and flour for human consumption. Brera et al. (4, 5) demonstrated that aflatoxins, zearalenone, and fuminosin B₁ were concentrated by at least 3-fold in bran, germ, and feed grain compared to the raw corn kernels during an industrial milling process, whereas the mycotoxin level in the grain fractions for human consumption was reduced. Fungal growth and mycotoxin contamination are more severe in the outer part of grains, contributing to the concentrating effect in the bran and germ fractions. The inclusion of bran and germ in turn causes a higher mycotoxin level in animal flour. The grain fractions intended for animals may, therefore, contain higher mycotoxin levels than raw grains.

4.2.1. Challenges in Sampling. The objective of sampling raw materials is to obtain samples that truly represent the grains that are arriving in a shipment or that are in storage. Characteristics to be measured include proximate composition, moisture content, mold and insect damage, and the presence of mycotoxins.

Unlike nutritional characteristics, mycotoxins are not evenly distributed in grains. Areas of concentrated contamination are commonly found in storage bins, arising from the contamination of certain areas in the crop, the mixture of contaminated grains with uncontaminated grains, and the formation of high-moisture areas. Cucullu et al. (86) reported that 18% of cottonseed kernels were contaminated with aflatoxin B₁, with individual concentrations ranging from 150 to 5 750 000 g/kg. Johnson et al. (87), in another study, reported aflatoxin B₁ contamination in 72 samples of corn from two bins ranging from 0 to 376 g/kg with an average of 21 g/kg in the first bin and from 0 to 332 g/kg with an average of 15 g/kg in the second bin. It has also been reported that uncontaminated kernels surround aflatoxin-contaminated kernels. Sixteen of 140 individually analyzed corn kernels were contaminated with aflatoxin B₁, ranging from 260 to 38000 g/kg, and 12 of the 140 kernels were contaminated with zearalenone, ranging from 9 to 1700 mg/kg (88).

Sampling operation is often the biggest source of error in mycotoxin analysis due to uneven distribution of mycotoxins in grains (89). Miraglia et al. (90) purposed that an ideal sampling protocol should include (1) objective-orientated analysis to determine why, where, and when to obtain samples and (2) a sampling method to take into account the heterogeneous nature of mycotoxin contamination.

The U.S. Department of Agriculture has a test procedure consisting of a 5 kg sample, a 250 g subsample, and a single TLC analysis (91). The total associated variance of this procedure is 630.7, including 521.4 (82.7%) due to the sampling, 59.2 (9.4%) due to subsampling, and 50.1 (7.9%) due to the TLC analysis (92). Increasing the sample size to 20 kg, using a vertical cutter mixer to take a 100 g subsample instead of the USDA/Agricultural Marketing Service 250 g subsample, and replacing TLC with HPLC can reportedly reduce the total variance to 176. Distribution of variance will then be 130.4 (74.1%) due to sampling, 25.5 (14.5%) due to subsampling, and 20.1 (11.4%) due to analysis (92).

Feed mills usually have a maximal tolerance for a given mycotoxin in a shipment to determine acceptance. If the sample is not taken appropriately, it is possible that a shipment with higher concentrations of mycotoxins than the maximum permitted will be accepted. A shipment that truly has lower concentrations of mycotoxins may also be rejected. Reducing the variability will, therefore, reduce the risk for both the seller and the buyer of rejecting an acceptable shipment or accepting an unacceptable shipment, respectively (89, 91).

4.2.2. Sample Preparation for Analysis. Variation in subsampling is less than that seen in sampling. Subsamples should be prepared by (1) coarse grinding of 1 kg of the sample to a particle size that can pass a standard 14 mesh screen; (2) mixing to homogenize material; (3) subdividing to take a portion for further fine grinding that can pass a standard 20 mesh screen; and (4) taking a subsample for analysis that can be 25-100 g (93). The use of a vertical cutter mixer can reduce the variance of subsampling (91).

4.3. Analytical Methods. There are different methods to analyze mycotoxins. To analyze aflatoxins, TLC is uncomplicated and requires less expensive equipment than other analytical methods. The lower limit of detection of aflatoxin both in corn and in cottonseed meals is 1 g/kg with TLC (87). The accuracy of TLC, however, is not as high as more expensive methods such as HPLC. The limit of detection of HPLC for aflatoxins is as low as 0.3 g/kg (94). To analyze zearalenone, TLC is less sensitive with a detection limit of 50 g/kg. GC has a detection and quantification limit of 20 g/kg, and GC-MS has a limit of 0.5 g/kg (95).

The use of TLC and GC to analyze trichothecenes is limited because interfering substances can cochromatograph with toxins, thereby confounding identification (96). The use of HPLC is considerably more difficult because these mycotoxins do not absorb or fluoresce in UV light (96). A more appropriate method for trichothecene analysis is the combination of GC-MS that eliminates most of the complication related to those substances that interfere with the detection of trichothecenes (96). Simultaneous determination of deoxynivalenol, 3-acetyldeoxynivalenol, nivalenol, fusarenon-X, T-2 toxin, neosolaniol, diacetoxyscirpenol, and zearalenone using GC-MS has been described (97). This provides a convenient tool for the detection of trichothecenes and zearalenone in cereals.

Various TLC, LC, and HPLC protocols have also been developed to detect and quantify fumonisins and ochratoxin A. Preis and Vargas (98), for instance, described a TLC protocol for fumonisin B₁ with a detection limit of 0.1 mg/kg. In another study, Visconti et al. (99) lowered the detection limit to 0.05 mg/kg using LC. Martins et al. (100) developed a HPLC method for ochratoxin A with a detection limit of 0.2 mg/kg.

Further development of purification protocols, chromatographic methods, and detection techniques may further reduce the detection limit of different mycotoxins. Ventura et al. (101), for instance, described a detection method to detect as little as 0.1 g/L of aflatoxins and ochratoxin A simultaneously using ultraperformance liquid chromatography-tandem MS and a solid-phase extraction. The present trend in chromatographic mycotoxin analysis is to use immunoaffinity columns (IAC) as a sample purification technique to improve sensitivity, although there are concerns that IAC may affect the accuracy of analysis (102).

Commercial enzyme-linked immunosorbent assay (ELISA) kits are simple, cheap, rapid, and widely adopted as in-house screening methods for individual mycotoxins (103). Direct competitive ELISA is most often used in mycotoxin analysis (104). ELISA may produce false-positive results, however, due to antigen-antibody cross-reactions, so further confirmation using HPLC or GLC is sometimes necessary (103). Recent development of molecular imprint materials and recombinant antibodies for specific mycotoxins promises novel applications in ELISA and sample purification (105, 106).

4.4. Conjugated Mycotoxins. The possibility of conjugated mycotoxins in plants has been suggested. Savart (107) was the first to synthesize in vitro eight fatty acid esters and two glucoside conjugates of deoxynivalenol. In another study, detection of deoxynivalenol-3-glucoside was reported in both naturally and artificially contaminated wheat and corn, suggesting that plants can transform deoxynivalenol into conjugated forms (108). Detection of conjugated mycotoxins is difficult because these are more polar compounds than the precursor toxins, and the extraction of conjugated mycotoxins with conventional solvents is more difficult due to losses in the sample purification process (108).

The natural occurrence of conjugated zearalenone, such as zearalenone-4--D-glucopyranoside, has been reported in both corn and wheat (109). In the same study, the authors found that the conjugated zearalenone was present in 42% of contaminated wheat samples. It was concluded that metabolic processes in the wheat cell produce the glucopyranoside conjugate of zearalenone.

Fumonisins have been found to react with the food matrix in such a way that the amino group essential for detection by derivatization is blocked (110). Shier et al. (111) showed that fumonisins may bind to protein in roasted corn meal dough and escape HPLC detection. Seefelder et al. (112) suggested that fumonisins could conjugate to proteins as well as to polysaccharides via the free tricarballylic acid side chains.

Conjugated forms of ochratoxin A have been discovered in fungal culture (113). There is no information, however, on the natural occurrence of conjugated ochratoxin A in grains or flour.

4.5. Storage Conditions. Conditions of grain storage are very important to prevent mycotoxin production and mold spoilage. The most important physical factors involved in the production of mycotoxins are temperature, water activity or moisture content, and atmospheric composition (114). In stored cereals, a moisture content of >15% is required to maintain fungal viability (115). There is an upper limit of moisture content for the stable storage of grains, which may be close to the moisture content of grains at harvest. When crops are harvested with moisture content in excess of the safe level, grains must be dried immediately (115). There are two types of transfer of heat and/or moisture in stored grains: (1) forced cold drying and (2) forced hot drying and aeration.

Transfer of water vapor and heat occurs mainly between the grains and the air that circulates around the grains. The air and grains, therefore, tend toward a hydric and thermal equilibrium (116). Natural transfer occurring in storage in the absence of forced air is uncontrolled and unpredictable. Differences in moisture content result in the transfer of moisture from humid to dry grains when dry grains are mixed with wet grains. More than 1 week is required to reach the moisture equilibrium. When grains of different moisture levels are stored together without mixing, the time to accomplish moisture equilibrium is longer (116).

The formation of wet spots results from moisture transfer when grains of different temperatures are contained in the same bin. It is common to find areas of cold grain and areas of warm grain in the same bin, and air moves from warm grains to cold grains. When warm air passes through cold grains, air moisture condenses and the moisture content of the cold grains increases (117).

The so-called "cold wall" effect introduces temperature gradients to stored grains. This effect is caused by diurnal and seasonal variations of grain temperature. One side of a bin may be exposed to sunlight with the other side in the shade, resulting in a temperature difference. During transport by boat, there can be differences between the sides of the ship that are in contact with cold water and in the internal part of silos that are warmer. Heat and moisture are transferred from the hottest to the coldest grains, and dry grains are moistened. When entering air currents have a higher dew point than the grains through which they are circulating, condensation of water occurs on the surface of grains and the water is absorbed (116). Good storage can prevent mold growth and mycotoxin production. It is clear that storing grains with moisture content below 15% or water activity below 0.7 prevents the formation of mycotoxins in stored grains (118). Good management practices for stored grains, therefore, reduce the risk of mycotoxin production.

4.6. Synergistic Interactions among Mycotoxins. It is known that synergistic interactions among mycotoxins occur in some domestic species. A toxicological synergism between deoxynivalenol and fusaric acid has been demonstrated in 8 kg piglets, where deoxynivalenol toxicity was augmented when fusaric acid was added in the diet (119). In another study, it was found that chickens fed combinations of deoxynivalenol and T-2 toxin from hatching to 3 weeks had significantly reduced body weight gain (120). This variable was not reduced, however, when either deoxynivalenol or T-2 toxin was fed singly, thereby suggesting a synergistic interaction.

Aflatoxins, ochratoxins, and other mycotoxins have also been demonstrated to interact synergistically. The feeding of diets containing both aflatoxins and ochratoxin A to chickens from hatching to 3 weeks of age resulted in significantly greater relative weight of gizzard and kidneys as well as less weight gain compared to either mycotoxin fed singly (121). Citrinin and penicillic acid were also found to potentiate the nephrotoxic and carcinogenic effects of ochratoxin A, resecptively (122, 123).

5. Preventative Strategies

Mycotoxins are, in general, chemically and thermally stable compounds. Once a mycotoxin-contaminated ingredient is screened and enters the milling process, mycotoxins are likely to be retained in the finished product, and further removal of mycotoxins is practically impossible. Most preventative strategies, therefore, focus on removing contamination early in processing or preventing mycotoxicoses in vivo without compromising nutritional quality.

5.1. Processing Techniques. There are three major processing techniques to reduce mycotoxin content of cereals: sieving, washing, and pearling. Cracked, damaged, and improperly developed kernels are good substances for fungal growth. Most mycotoxins are found at high concentration in this fraction as well as in dust and debris formed during grain handling. The separation of these more contaminated fractions from kernels may markedly reduce the mycotoxin content of cereal grains. Trenholm et al. (124) investigated the efficacy of sieving on corn, barley, and wheat contaminated with 5-23 mg of deoxynivalenol/kg and 0.5-1.21 mg of zearalenone/kg. After being sieved with a series of screens, the retained fractions contained 67-83% less mycotoxins with a 34-69% material loss of the original kernel.

Washing methods, on the other hand, take advantage of the fact that mycotoxins are primarily found on the outer surface of grains. Treholm et al. (125) were the first to report a washing technique to decontaminate grains containing deoxynivalenol and zearalenone. After a first wash with 1 M sodium carbonate solution and two extra washes with distilled water, barley and corn showed 72-74 and 80-87% reductions in deoxynivalenol and zearalenone contents, respectively. In another study, a 1-day hot water (80 C) treatment with 1 M sodium carbonate reduced the deoxynivalenol content of barley from 18.4 to 1.4 mg/kg (126). A major disadvantage of washing methods, however, is that the wetted cereal grains need drying, causing an extra cost to feed manufacturers.

House et al. (127) developed an abrasive pearling procedure that removed 66% of deoxynivalenol content with a 15% material loss of barley in 15 s without wetting the cereal. Siwela et al. (128) reported that the aflatoxin content of corn was reduced by 93.4% after a physical dehulling procedure.

Ozonation has also been proposed as a detoxifying method for mycotoxin-contaminated grains. Procotor et al. (129) evaluated the effectiveness of ozonation in breaking down aflatoxins in peanut kernels and flour. Up to 60% of aflatoxin B₁ was degraded after a 15-min gaseous ozonation treatment at room temperature. An aqueous treatment of 25 mg/mL ozone was found to destroy the epoxy group of trichothecenes, although the toxicity of residual compounds required further investigation (130).

Acid-based mold inhibitors are commonly used by the animal industry to prevent mycotoxin formation in animal feed. Weak organic acids including benzoic, acetic, sorbic, and propionic acids are known to inhibit fungal growth by acidifying the cytoplasmic content of fungal cells (131, 132). In addition to applying these processing techniques, it is also important to purchase high-quality feedstuffs and to handle the grain carefully to prevent kernel damage, thereby minimizing mycotoxin formation.

5.2. Nutrient Supplementation. A number of nutrient supplements have been proposed to act therapeutically to minimize mycotoxin-induced tissue damage and altered behavior. These nutrients include large neutral amino acids, antioxidants, and polyunsaturated fatty acids (PUFA).

Supplemental amino acids primarily alleviate the effects of mycotoxins on the central nervous system. Fusaric acid, T-2 toxin, and deoxynivalenol elevate blood and brain tryptophan concentrations, increase brain serotonin concentrations, and thereby result in feed refusal and vomiting (133, 134). It is possible, therefore, to supplement feedstuffs with large neutral amino acids, such as leucine, isoleucine, valine, tyrosine, and phenylalanine, to compete with tryptophan for active transporter carriers across the blood-brain barrier, thereby preventing the mycotoxin-induced brain uptake of tryptophan and subsequent overt toxicity (135).

Because a number of mycotoxins including aflatoxin B₁, fumonisin B₁, deoxynivalenol, T-2 toxin, and ochratoxin A are known to damage cell membranes by lipid peroxidation, the therapeutic properties of antioxidant nutrients against mycotoxins have been investigated (136). Selenium and vitamins A, C, and E have been demonstrated to reduce the toxic effects of mycotoxins in vitro and in vivo (137-141).

The effectiveness of antioxidants in the prevention of mycotoxin-induced toxicity has also been investigated. A few reports have also suggested that antioxidants can prevent aflatoxin-induced carcinogenicity (142-144). Pretreatments with vitamins A, C, and E in mice and rats were found to reduce DNA adduct formation in the kidney and liver caused by ochratoxin A and zearalenone (145). Antioxidants appear to act as superoxide anion scavengers and thereby protect cell membranes and DNA from mycotoxin-induced damage (136).

Omega-3 PUFA have been suggested to reduce IgA-induced nephropathy. Shi and Pestka (146) demonstrated that dietary supplementation with 3% (w/w) eicosapentaenoic acid (EPA) suppressed the development of IgA nephropathy in mice consuming 20 mg of deoxynivalenol/kg of feed. In another study, docosahexaenoic acid (DHA) and EPA, but not -linolenic acid, attenuated the proinflammatory response and renal damage induced by 10 mg of deoxynivalenol/kg of feed (147). It is believed that DHA and EPA reduce deoxynivalenol-induced interleukin-6 production and therefore alleviate the inflammatory damage to the kidney (148, 149).

Despite the experimental findings of the protective effects of nutrient supplementation against mycotoxins, there are limited clinical feeding trials to verify their efficiency in treating mycotoxicosis in domestic and pet animals. The feasibility of using these dietary supplements to treat mycotoxicosis in pet animals, therefore, remains to be confirmed.

5.3. Mycotoxin-Sequestering Agents. Mycotoxin-sequestering agents, such as activated charcoal, silicate minerals, and cholestyramine, can prevent intestinal mycotoxin absorption in animals (150). Hydrated sodium calcium aluminosilicate (HSCAS) has been shown to significantly reduce the adverse effects of feeding 7.5 mg of aflatoxin B₁/kg of feed on broiler chickens with an inclusion level of 0.5% (w/w) (161). Further experiments using 3.5 mg of aflatoxins/kg of feed, combined with 2.0 mg of ochratoxin A/kg of feed and 8.0 mg of T-2 toxin/kg of feed, however, showed that the 0.5% HSCAS inclusion could diminish many of the adverse effects of dietary aflatoxins but had no effect on ochratoxin A or T-2 toxicity in the chickens (152).

Most mycotoxin-sequestering agents either only adsorb specific mycotoxins, require a high inclusion rate in animal feed, cause other health complications, or are too expensive for industrial applications. Naturally occurring glucomannan-containing polymers extracted from yeast cell wall, in contrast, have a high adsorptive capacity for binding a combination of different mycotoxins and can prevent mycotoxicoses with a low inclusion rate, thereby showing great promise as a practical solution for the pet food industry.

The efficacy of glucomannan-containing polymers as mycotoxin-sequestering agents in feeds has been investigated in several studies with domestic animals. Raju and Devegowda (153) reported that the feeding of 1 g of a polymeric glucomannan mycotoxin absorbent (GMA)/kg of feed significantly improved feed intake, body weight gain, and biochemical and hematological parameters of broiler chickens during in a 35-day exposure to aflatoxin B₁ (0.3 mg/kg of feed), ochratoxin A (2 mg/kg of feed), and T-2 toxin (3 mg/kg of feed). In another study, Swamy et al. (164) demonstrated that 0.2 g of GMA/kg of feed could prevent the changes in neurochemistry and serum antibody concentrations caused by a dietary combination of deoxynivalenol, 15-acetyldeoxynivalenol, fusaric acid, and zearalenone (5.5, 0.5, 26.8, and 0.4 mg/kg of feed, respectively) in immature swine. Similar benefits have been seen in a multiple fusarial challenge to broiler chicken (155), laying hens (156), turkeys (157), and horses (158).

5.4. Microbial Deactivation of Mycotoxins. The application of microorganisms capable of detoxifying mycotoxins into nontoxic metabolites in animal feed has been proposed to prevent mycotoxicosis (159). These microorganisms act in the intestinal tract of animals prior to mycotoxin absorption. A strain of bacteria (BBSH 797), for instance, was isolated from bovine rumen fluid and found to deactivate trichothecenes by reducing epoxide rings (160, 161). Trichosporon mycotoxinivorans (Trichosporon MTV, 115), a novel yeast strain, was found to be capable of degrading ochratoxin A and zearalenone (159). The detoxifying effectiveness of these strains was demonstrated in yeast and gut tissue cultures (159, 162, 163).

6. Conclusions

It has been shown that mycotoxins have various pathological effects on pet species. The worldwide prevalence of mycotoxins in pet food has also been well-documented to cause mortality and economic losses due to acute mycotoxicosis in extreme cases. Specific mycotoxin regulations on pet food should, therefore, be put in force to address the problem. Further research to develop more accurate sampling techniques, more efficient decontaminating processes, and more effective dietary supplements would protect companion animals from the harmful effects of mycotoxins until mycotoxins can be completely eliminated from the raw materials used in pet foods.

Acknowledgment

We thank Josef Böhm, John Hilton, John Leslie, Richard Manderville, Peter Scott, and Michael Spandern for advice and assistance in the preparation of the manuscript.