Health Effects of Mycotoxins: A Toxicological Overview

Journal of Toxicology CLINICAL TOXICOLOGY Vol. 42, No. 2, pp. 217 234, 2004 REVIEW Health Effects of Mycotoxins: A Toxicological Overview Frederick Fung, M.D., M.S., 1,* and Richard F. Clark, M.D. 2 1 Sharp Rees-Stealy Medical Group, San Diego University of California, San Diego, California, USA 2 Division of Medical Toxicology, University of California, San Diego, California, USA ABSTRACT Diseases caused by fungi are spread by direct implantation or inhalation of spores. Fungi can cause adverse human health effects to many organ systems. In addition to infection and allergy, fungi can produce mycotoxins and organic chemicals that are responsible for various toxicologic effects. We reviewed the published literature on important mycotoxins and systemic effects of mycotoxins. Scientific literature revealed a linkage between ingesting mycotoxin contaminated food and illness, especially hepatic, gastrointestinal, and carcinogenic diseases. Issues related to mycotoxin exposure, specific diseases, and management are discussed. Although there is agreement that diet is the main source of mycotoxin exposure, specific health effects and risk assessment from indoor nonagricultural exposure are limited by the paucity of scientific evidence currently available. Further research on the health effects of inhaling mycotoxins in indoor settings is needed. Key Words: Mycotoxins; Toxicological; Gastrointestinal; Fungi; Mycotoxicosis; Indoor; Agricultural. INTRODUCTION Mold, mildew, and mushrooms are scientifically known as fungi. They are neither plants nor animals. Fungi are saprophytic decomposers and are unable to generate food from sunlight or other energy sources (1). Diseases caused by fungi are spread by direct implantation or inhalation of spores. Similar to all infectious agents, a spectrum of medical conditions can result from fungal exposure. This may range from a superficial skin disease such as tinea to invasive internal organ pathology such as pulmonary aspergillosis. *Correspondence: Frederick Fung, M.D., M.S., Sharp Rees-Stealy Medical Group, San Diego University of California, 2001 Fourth Ave., San Diego, CA 92101, USA; Fax: (619) 699-1514; E-mail: fred.fung@sharp.com. 217 DOI: 10.1081/CLT-120030947 Copyright D 2004 by Marcel Dekker, Inc. 0731-3810 (Print); 1097-9875 (Online) www.dekker.com

218 Fung and Clark From an allergy/immunology perspective, fungus has been implicated as allergens capable of inducing immune and allergic reactions. Depending on genetic susceptibility, individuals can develop allergic diseases such as rhinitis, sinusitis, asthma, and hypersensitivity pneumonitis (2). Mycotoxins are naturally occurring environmental contaminants. They are widely found in foodstuffs. Surveys have shown elevated mycotoxins in foods and feeds from countries throughout the world (3). Toxic effects due to mycotoxins have been linked to ingestion of contaminated food. Over the last few years, some reports have associated mycotoxin-producing fungal species found indoors with poorly defined symptoms in potentially exposed individuals. These fungal species include Aspergillus, Penicillium, Fusarium, Stachybotrys, Alternaria, and others. In general, mycotoxicosis refers to toxic effects or poisoning due to mycotoxins (4). Mycotoxicosis has long been studied and published in food and veterinary toxicology literature. This review will focus mainly on human toxicity due to mycotoxins. However, due to a lack of specific human effects and some reporting of nonspecific effects of putative indoor mycotoxin exposure, results of animal studies are discussed to illustrate toxicologic aspects of mycotoxicosis. The goal of this review is to summarize the current understanding of human mycotoxicosis. Because much of the recent controversy regarding the potential for human disease from mycotoxins centers on inhalation of the spores and toxins, this route of exposure will be discussed in detail in the last section of this review. Toxic effects due to mold or fungal products have been recognized for centuries. Eating moldy grain has long been suspected to have caused illnesses ranging from gastrointestinal upset to hallucination and convulsions. St. Anthony s Fire or sacred fire documented since the Middle Ages is caused by the fungus Claviceps purpurea. The chemical ingredients responsible for ergot poisoning or ergotism were not isolated until the nineteenth century. Mycotoxins and their impact have been under constant research, especially in agricultural and veterinarian settings. The primary objectives of these research efforts are to minimize animal and human exposures to naturally occurring mycotoxins that are invariably present in animal feeds and foodstuffs (5). Since these mycotoxins are natural contaminants, they cannot be totally eliminated without damaging the food. On an international basis, the World Health Organization (WHO) constantly monitors food safety programs and evaluates mycotoxin levels in various food sources. These important mycotoxins include aflatoxins, trichothecenes, ergots, ochratoxins, and patulin. TOXICOLOGY OF IMPORTANT MYCOTOXINS In general, mycotoxins are complex organic compounds that are not volatile at normal ambient temperatures. Except in a few famine stricken areas, most outbreaks due to mycotoxins in the last century were related to agricultural and veterinarian practices (6). There have been several case series of human mycotoxicosis secondary to eating mycotoxin-contaminated grain (7). Concerns of illness due to inhaling mycotoxins did not rise to national notoriety until 1997 when the Centers for Disease Control reported an unusual cluster of infants suffering idiopathic pulmonary hemorrhage (IPH) who were also reported to have been exposed to mold (8). Claviceps species and their ergot alkaloid mycotoxins have for unknown reasons never been associated with indoor air-related complaints, even though they may represent the most common form of human mycotoxicosis. Excellent reviews have been published on the toxicology of ergot alkaloids (9,10); therefore, these compounds are not discussed in detail in this paper. The following mycotoxins are discussed in terms of chemical characteristics, toxicology, detection, and standards, if available. Potential toxicological effects of important mycotoxins are discussed under specific organ toxicity. AFLATOXINS There are several aflatoxins (AF) and their metabolites (such as AFB1, AFB2, AFG1, AFG2, AFM1, AFM2) that are capable of producing human disease. Aflatoxins are named after their fluorescence as blue or green under UV light and other analytical characteristics. Aflatoxin metabolites found in mammalian milk are named AFM, where M denotes Figure 1. Structures of mycotoxins: Aflatoxins B and M.

Health Effects of Mycotoxins 219 milk or mammalian metabolites. There are two broad categories of aflatoxins according to their structures. Aflatoxins B 1,2 (AFB 1, AFB 2 ) and aflatoxins M 1,2 (AFM 1, AFM 2 ) are within the difurocoumarocyclopentenone series (Fig. 1). Aflatoxins G 1,2 (AFG 1, AFG 2 ) are of the difurocoumarolactone series. These toxins are produced by certain strains of the fungi Aspergillus flavus and A. parasiticus. Aflatoxins were discovered during an epidemic of disease that wiped out more than 100,000 turkeys in the 1960s. The disease was traced to turkey feed made of moldy Brazilian peanuts. Eventually it was discovered that all crops and foodstuff, including corn, rice, wheat, barely, and nuts contain these naturally occurring mycotoxins. It has become a matter of how hard one looks in food including meat, milk and eggs to discover mycotoxins. Aflatoxins are commonly found in crops in the field prior to harvest. Post harvest contamination can also occur if crop drying is delayed, and during storage when moisture is present, allowing for the mold growth. Aflatoxins are detected in storebought corn (maize), peanuts, cottonseed, almonds, figs, and a variety of other foods and feeds. Milk, milk products (non-fat dry milk, cheese, and yogurt), eggs, and meat products are contaminated due to animal consumption of aflatoxin-contaminated feed (11). Aflatoxins found in tobaccos have been shown to be toxic to animals after inhalation and ingestion (12). Exposure to aflatoxins is typically by ingestion of contaminated foodstuff. Dermal exposure results in slow and insignificant absorption (13). Inhalational exposure in humans has not been studied because of a lack of relevancy in food toxicology. After absorption, 65% of AFB 1 is cleared from the blood in 90 min and the plasma half-life is short. The exact in vivo half-life of aflatoxins is unknown. The half-life of aflatoxins tested in human liver homogenates is approximately 13 min (11,14,15). In vitro metabolism studies have shown the following metabolic reactions for AFB 1 : reduction produces aflatoxicol (AFL); hydroxylation produces AFM 1 ; hydration produces AFB 2a ; and epoxidation produces AFB 1-2,3-epoxide. The epoxide is the most reactive metabolite and is thought to be responsible for both the acute and chronic toxicity of AFB (15). AFB 1 epoxide occurs in endo and exo forms. The exo-epoxide is highly electrophilic and reacts with the DNA guanine moiety to form covalent bonds at the N-7 guanine residue leading to depurination and carcinogenesis (16). Depurination is a process where the purine base of a DNA molecule is lost, potentially leading to a somatic mutation and carcinogenesis. It was estimated that ingestion of 2 6 mg/kg/day of AF over a month produced hepatitis in India, where some fatalities were reported (17). However, a suicide attempt with 1.5 mg/kg of pure aflatoxin resulted only in nausea, headache, and rash (18). Aflatoxins can be detected in the laboratory using thin layer chromatography (TLC), high-performance liquid chromatography (HPLC), gas chromatography mass spectrometry (GC-MS) or liquid chromatography electrospray ionization tandem-mass spectrometry (LC-ESI-MS-MS) (19,20). The detection limits vary from picogram (pg) per mg (dust) or pg per ml (liquid) to ug/kg (peanuts). Immunological methods provide a rapid screening for aflatoxins. Commercial enzyme-linked immunoassay (ELISA) test kits have a detection limit of about 2 ug/kg. Despite a rapid result, it may not be sufficiently reliable as a quantitative method of detection (21). Currently, maximum concentrations of aflatoxins allowed in food are set by U.S. Food and Drug Administration (FDA). Food for human consumption is allowed 20 ppb of AF and milk 0.5 ppb. High levels up to 300 ppb are allowed in feed for cattle, hogs, and poultry. Currently, there are no standards for work place or environmental aflatoxin exposures (22). European countries have considered setting permitted levels for mycotoxins in foodstuffs (23,24). OCHRATOXINS Ochratoxin A (OTA) is a pentaketide-derived dihydroisocoumarin moiety linked via the 12-carboxy group by a peptide bond to L-phenylalanine. There are several OTA analogues, ochratoxins B, C, and alkyl esters of ochratoxins that have similar structure but are less toxic. OTA was the first mycotoxic compound isolated from Aspergillus ochraceous, and later it was found in other Aspergillus and Penicillium species such as Penicillium verucosum. OTA is a main contaminant of cereals (corn, barley, wheat) and to some extent beans (coffee, soy, and cocoa). The levels of contamination are typically less than 200 ug/kg (25).

220 Fung and Clark OTA is easily absorbed through the gastrointestinal tract mainly in the duodenum and jejunum based on animal studies (26). There are no studies on skin or inhalational absorption of OTA. When absorbed, OTA has a high binding affinity for plasma protein. OTA was found in decreasing order of concentrations in kidney, liver, fat, and muscle tissues (25). Excretion is mainly via renal elimination (27). The elimination halflife in an animal model has been reported between 23.6 to 28.7 h (28) to as long as 35 days in monkeys (29). The LD 50 of OTA ranges from 0.5 mg/kg for dogs to over 50 mg/kg for mice (29). The toxicity of OTA involves several mechanisms. OTA inhibits protein synthesis by competing with the phenylalanine aminoacylation reaction catalyzed by Phe-tRNA synthase (30). This results in inhibition of protein as well as DNA and RNA synthesis. OTA also disrupts hepatic microsomal calcium homeostasis by impairing the endoplasmic reticulum membrane via lipid peroxidation (31). OTA is the major ochratoxin component and is the most toxic among the analogues. However, it has been estimated that an infant could eat up to 10 kg of food contaminated with 20 ppb without significant adverse health effects (32). OTA can be analyzed using TLC, HPLC, and ELISA. The limits of detection on TLC are in the ug/kg range. HPLC has a detection range in ng/kg (33). Immunoassays such as ELISA can detect OTA in pg ranges. However, the possibility of cross-reactions cannot be fully ruled out. Other techniques should be used to confirm the levels of OTA (34). The tolerance levels for OTA have been suggested as 1 ug/kg for infant foods and 5 ug/kg for cereals (23). Both the Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO) Joint Expert Committee recommends a provisional tolerable weekly intake of 100 microgram/kg body weight of OTA (35). Currently, there are no standards for workplace or environmental OTA exposures. epoxy ring commonly known as the 12,13-epoxytrichothecene. Trichothecenes are classified into four groups. Group A includes T-2 toxin and diacetoxyscirpenol (DAS). Group B includes 4-deoxynivalenol (DON) and nivalenol (NIV). Many Fusarium species produce group A and B trichothecenes. Baccharis megapotamica (plant species) produces the Group C trichothecene baccharin that is the least common. Group D mycotoxins include roridins produced by Mycothecium roridum, verrucarin produced by M. verrucaria, and satratoxins produced by Stachybotrys atra (36). It is important to point out that the more common and potent trichothecenes are produced by Fusarium species. Baccharin is not a fungal trichothecene and is not discussed here. Trichothecenes are found in crops, food, and animal feed contaminated with Fusarium species. Although T-2 is highly toxic and more well known than other trichothecenes, DON, NIV, DAS are more frequent contaminants of food and animal feed than T-2 toxin. Levels in the range of 0.5 40 mg/kg of T-2, DON, and NIV are detected in foodstuffs such as corn, peanuts, rice, and wheat. In commercial foods such as corn flour, popcorn, potato, wheat flour, breakfast cereals, and baby food, trichothecene levels are much lower, in the range of 0.03 0.5 mg/kg. Only group D trichothecenes, particularly satratoxins found in hay and straw, are produced by Stachybotrys atra. Other trichothecenes are produced mainly by Fusarium species (36). Group A and B trichothecenes are rapidly absorbed from the GI tract. Although there are no human data on absorption through inhalation and skin contact, in vitro TRICHOTHECENES Trichothecenes are a large group of sesquiterpenoid chemicals characterized by a tetracyclic 12-13-

Health Effects of Mycotoxins 221 and animal studies show that trichothecenes are poorly absorbed through intact skin (37). Trichothecenes undergo de-epoxidation and glucuronidation resulting in less toxic metabolites. A dog model showed that trichothecenes are distributed to the liver, kidneys, intestines, spleen, and other organs. The elimination half-life is estimated at 1.6±0.5 h after IV injection of the toxin (38). Another model on swine and cattle showed a half-life of 13 and 17 min, respectively (39). Trichothecenes do not require metabolic activation to exert their toxicity. The presence of a reactive electrophilic, 12-13 epoxide, accounts for a rapid onset of trichothecene toxicity. The mechanism of toxicity again involves inhibition of protein and DNA synthesis (40). They also produce general cytotoxicity by inhibiting the mitochondrial electron transport system (41). The 12-13-epoxide of the trichothecenes is essential for the toxicologic activity. The de-epoxidation of T-2 in mammalian systems results in loss of toxicity (42). The dose of trichothecene needed to cause symptoms in humans is unknown. There is great variability in the toxicity of these compounds in animal studies. In general, the relative order of decreasing toxicity is T-2, DAS, DON, and NIV (43). The LD50 of trichothecenes ranges from 0.5 to 300 mg/kg, depending on the route of administration and animal model (36). LD50 ranges from 13 ug to 140 ug are found when trichothecenes are applied directly to rat brain (44). A recent review outlines the dietary and non-dietary exposure to trichothecenes (45). TLC methods are nonspecific and insensitive for detecting trichothecenes. GCMS and LCMS have been used for the determination and identifications of trichothecenes in trace concentrations. HPLC is also a reliable and sensitive method for detecting trichothecenes (20). Supercritical fluid chromatography combined with MS can be used for testing biological specimens down to pg levels (46). Immunoassays such as RIA and ELISA are useful as screening tests for T-2 and DON in cereals (47,48). Setting tolerance levels for trichothecenes appears to be difficult due to the mixtures of toxins with different toxicity. The European Community is considering an action level of 500 ppb for cereals for DON and 750 ppb for flour consumed by people (European Mycotoxin Awareness Network at www.lfra.co.uk.eman accessed April 7, 2003). The FAO/WHO Joint Expert Committee recommends a provisional maximum tolerable daily intake for DON at 1 mg/kg body weight and 60 mg/kg of body weight for T-2 toxin (35). Currently, there are no standards for workplace or environmental trichothecene exposures. CITRININ Citrinin was initially found and isolated from Penicillium citrinum. It was later found in other Penicillium species. It typically contaminates rice, wheat, barley, corn, rye, and other foodstuffs. Despite its co-contaminating cereals with OTA, citrinin has been implicated as a nephrotoxin. Its LD50 in animal models has been reported as about 50 mg/kg by oral route. Pathologically, citrinin damage includes hepatic

222 Fung and Clark fatty infiltration and necrosis. It decomposes at 175 C and degrades during heating processes. Epidemiologically, citrinin has been associated with yellow rice syndrome, but there has been no systematic investigation to the actual mycotoxin or agent responsible for this ill-defined illness (49,50). Citrinin appears to be destroyed by food processing. Currently, there are no regulations or guidelines for this mycotoxin. No reports or studies on the toxicity of this mycotoxin by inhalational exposure are available. FUMONISINS PATULIN Patulin is a polyketide lactone, produced by Penicillium, Aspergillus, and other mold species that grow on fruits such as apples, pears, and grapes. The LD50 of patulin ranges from 15 to 25 mg/kg and varies with animal species and route of exposure. Toxicity includes congestion and edema of pulmonary, hepatic, and intestinal blood vessels and tissues. Sarcomas were observed when large doses of patulin were injected into animals. As a result, there have been concerns over the possibility of carcinogenicity to children and adults who drink large amounts of fruit juice, especially apple juice, for many years (51). WHO Codex Alimentarius Commission recommends a limit of 50 ug/kg of patulin in apple juice and cider (52). Patulin is degraded by sulfur dioxide or sulfide, a common food preservative for dry fruits and juices. No reports or studies on the toxicity of this mycotoxin by inhalational exposure are available. Fumonisins are a group of mycotoxins produced by Fusarium species. They are common contaminants of corn and maize. Although they are not as potent as aflatoxins, their concentrations frequently reach hundreds of parts per million up to 300 mg/kg of maize. The chemical structure has a long hydroxylated hydrocarbon chain with methyl and amino groups. Fumonisin B1 is the most important and potent mycotoxin in this group. It has been implicated in sporadic animal diseases (53), especially equine leucoencephalomalacia (ELEM). Toxic mechanisms of these compounds likely involve the inhibition of the sphingolipid synthesis resulting in the disruption of sphingomyelin (54). Autopsy of animals with ELEM shows cerebral edema and liquefaction necrosis (55,56). Other organs can also be affected. Centrilobular necrosis and hepatic fibrosis are observed in severe cases. In humans, fumonisins have been associated with an increase risk of esophageal carcinoma in areas (57) where consumption of fumonisin-contaminated corn and maize are prevalent. Fumonisins are relatively heat stable and are detectable in tortilla flour. FAO/WHO Committee on Food Additives recommend a maximum tolerable intake at 2 ug/kg body weight per day consumption on the basis of a NOEL and safety factor of 100 (35). There are no reports of human toxicosis due to fumonisins. No reports or studies on the toxicity of this mycotoxin by inhalational exposure are available.

Health Effects of Mycotoxins 223 MYCOTOXIN-INDUCED ORGAN TOXICITY It has been well documented that ingestion of mycotoxin contaminated foodstuffs can produce acute and chronic effects. These effects cannot be attributed to fungal growth within the host or allergic reaction to fungal proteins. Mycotoxin-related diseases postulated to affect humans are summarized in Table 1. Although mycotoxicosis secondary to ingestion is well documented, we will comment on ingestion-related mycotoxicosis and animal studies to put inhalational exposure in perspective. Pulmonary Toxicity Possible relationships between mold and respiratory disease have been recognized for years, especially in the industrial setting. Early Russian literature describes toxicoses associated with inhaling dust heavily contaminated with fungal spores (Aspergillus fumigatus and Stachybotrys atra). Pulmonary mycotoxicoses was used to describe a patient who had inhaled massive amounts of fungi and developed respiratory symptoms (58). Massive amounts of fungi was only used as a qualitative description of the exposure in this report, and no quantification was performed. Clinically, symptoms and signs in this case presented as significant lung injury in an agricultural setting. This individual presented with cough, respiratory distress, fever, fatigue, interstitial or alveolar infiltrates, and leukocytosis. The condition was selflimited and the patient recovered without residual deficits. This condition is different from hypersensitivity pneumonitis in that it is transient and occurs after massive exposure to fungal materials that contain a mixture of components. Neutrophils, but not lymphocytes, are found on bronchoalveolar lavage (BAL), and fungal precipitin testing is negative. In 1997, the CDC reported a cluster of 10 infants from the Cleveland area diagnosed with idiopathic pulmonary hemorrhage (IPH) that presented with respiratory distress requiring intensive care treatment (8). The preliminary analysis of this case control study indicated that IPH was associated with mold, especially Stachybotrys exposure. Further analyses of the raw data revealed that the case and control infants were different in terms of sex, race, breast-feeding, birth weight, and exposure to secondhand smoke. The odds ratio (OR) calculation included one extreme outlier in the case group. If this outlier was excluded, the OR drops from 9.8 to 1.5. Case house sampling was biased in that they were sampled twice. Furthermore, the investigators used aggressive, nonstandardized methods to generate artificial aerosols for sampling (vacuuming carpets and pounding on furnace ducts and furniture). This resulted in a potentially higher exposure assessment of cases. Additionally, water damage in case homes was not defined. No other family members were clinically ill, and this would be inconsistent with a toxic mechanism as described in prior mycotoxin literature. There is also no consistent case definition of lung disease making case control analysis inconclusive. From a toxicological perspective, toxigenicity of mold was not demonstrated and no airborne mycotoxins were identified. As a result of further reviews by internal and external panelists, the CDC recanted the original conclusions and stated that the association between Stachybotrys exposure and pulmonary hemorrhage has not been proved (59). No further epidemiologic studies by CDC regarding mold and IPH have been reported since that time. The mechanism of mycotoxin-induced lung injury is still unclear. Although trichothecenes produced by Stachybotrys are potent protein synthesis inhibitors, pulmonary hemorrhage cannot be fully explained by this mechanism. Another hypothesis consisted of trichothecenes inhibiting collagen synthesis in rapidly growing young lungs, leading to capillary fragility. With further exposure to irritants such as tobacco smoke, it could have resulted in hemorrhage. However, this hypothesis is not supported by animal models. Although some experimental animal studies with Stachybotrys have shown hemorrhagic changes in lungs, these exposures were intratracheal or intranasal instillation of large quantities of spores or toxins (60,61). It is unclear if the hemorrhagic exudates were the toxic effects of the mycotoxin or compensatory mechanisms of inflammation. These animal models typically lack proper controls and statistical power. Direct instillation of mold spores or mycotoxins through the trachea or nose has questionable physiological and clinical relevance. Spores larger than 10 microns in aerodynamic diameter are not likely to reach the lower respiratory tract by inhalation exposure. More importantly in the above studies, histopathologic changes were not noted in other target organs of mycotoxins (spleen and intestines). Without pathophysiologic changes in other organs and tissues, isolated pulmonary hemorrhage is not consistent with toxicity due to trichothecenes. Other studies and reports involving the respiratory system have not provided pathological and laboratory findings supportive of pulmonary diseases (62 69). Current epidemiologic and toxicologic evidence suggests a transient inflammatory or irritant effect secondary to mold spores or

224 Fung and Clark Table 1. Organ toxicity secondary to mycotoxins. Organ toxicity Mycotoxins Proposed mechanism Comments Pulmonary hemorrhage Trichothecenes Protein and collagen synthesis inhibition High levels of airborne toxin needed Encephalopathy Aflatoxins Cytotoxicity Consumption of toxins Ergot alkaloids Vasoconstriction CNS depression Microbial volatile organic chemicals (complex alcohols and aldehydes) Decrease activity of CNS neurons similar to alcohols and aldehydes Sufficient/exceed concentrations to induce mucous membrane irritation Hematologic/immunologic Trichothecenes Protein and enzyme synthesis inhibition High levels of airborne toxin needed Suppression Aflatoxins or consumed Cancer Liver Aflatoxins Electrophilic binding of DNA/RNA Esophageal Fumonisins nucleophilic sites Consumption of mycotoxin-contaminated food Nephropathy Ochratoxins Direct cytotoxicity Consumption of food contaminated with mycotoxins Teratogenicity Ergots Binding of nucleophilic sites Consumption of mycotoxin-contaminated food Trichothecenes Aflatoxins Gastrointestinal toxicity Most mycotoxins Direct cytotoxicity Consumption of mycotoxin

Health Effects of Mycotoxins 225 mycotoxins. However, pulmonary hemorrhage or other severe lung diseases have not been causally linked to mycotoxin exposure in indoor nonagricultural settings. Neurologic Toxicity Only a few cross-sectional surveys report that mold exposure is associated with neurologic symptoms (70). Much of the concerns are based on ergotism, ingestion of contaminated food, and perhaps other toxic mushroom poisoning. Ergot poisoning can cause convulsions and hallucinatory effects. Other fungal components such as volatile organic chemicals (VOCs) at sufficient concentrations and durations may produce neurologic effects (71 73). 3-Nitropropionic acid produced by Arthrinium species has been implicated to cause moldy sugar cane poisoning, also known as Kodua, with symptoms including dystonia, convulsion, carpopedal spasm, and coma (74,75). Cyclopiazonic acid produced by Penicillium and Aspergillus species has been linked to Kodua poisoning (76). Clinically, patients manifested somnolence, tremors, and giddiness. Although not reported in humans, verruculogen and penitrem A (Penicillium and Aspergillus species) are known tremorgenic mycotoxins and may be responsible for tremors, ataxia, and convulsions in animals (77). A major limitation of these animal studies is the route of administration being intraperitonel injection, thus having questionable clinical relevance in humans (78). Recent reports of trichothecene or Stachybotrys exposure do not describe specific neurologic findings. Even in potential pediatric cases of exposure in which toxigenic molds were found, no neurological symptoms or signs were reported (79). At this time, there is no substantive evidence of neurologic injury caused by indoor mold or mycotoxin exposure (80). Hematologic and Immunologic Toxicity The hypothesis linking mycotoxin (gliotoxin from Penicillium) and AIDS in the early 1980s has been proved to be false (81). However, there are hypotheses claiming immunosuppressive effects of mold spores without any controlled studies. Concerns over hematologic and immunologic effects of mycotoxins are based on the early observation of pancytopenia after ingesting trichothecene-contaminated grain (82). Although some animal studies suggest that mycotoxins affect the immune system (83,84), implying a susceptibility to bacterial infection because of mycotoxicosis has not been well documented because of a wide variety of confounding factors such as hygiene and nutritional status of animals (85). Although aflatoxins are the most extensively studied mycotoxins, they do not produce significant immunosuppression. The mechanism of toxicity for aflatoxins is similar to other mycotoxins (i.e., protein synthesis inhibition). In general, aflatoxin is metabolized into epoxide intermediates that bind DNA and RNA. These metabolites impair DNAdependent RNA polymerase, therefore inhibiting RNA and protein synthesis. Theoretically, this could impair the proliferation and differentiation of immune cells, immunoglobulin, and cytokines. Trichothecenes inhibit protein synthesis by blocking RNA and DNA synthesis through peptidyltranferase inhibition. The result is similar to radiation exposure (radiomimetic) in that rapidly dividing cells are impaired first (86). Animal studies on immune modulating effects of mycotoxins are not representative of human exposure. These animal studies used purified high-dose toxins, administered orally or by intraperitoneal injection, making clinical application impossible. Observational studies (87) on Stachybotrys-exposed humans have not detected the radiomimetic effects in alimentary toxic aleukia (ATA) after trichothecene ingestion. Studies on the effects of CD4/CD8 ratio, natural killer (NK) cells, and other cytometric parameters were not clinically and statistically significant (88). Many confounding factors such as secondhand smoke, presence of pets and dust mite antigen, endotoxins, VOCs, and diet were not considered in studies suggesting immunomodulation by mycotoxins. Currently, there is no evidence of significant immunological suppression or modulation resulting from inhaled mycotoxin in the indoor setting (89). Cancer Aflatoxins have been most extensively studied in animals. Extrapolation of animal studies to humans should be interpreted with caution (90). Trout, salmon, and rats are prone to develop hepatocellular carcinoma upon aflatoxin (AF) exposure. The adult mouse is most resistant to AFB 1 hepatotoxicity (90). Epidemiological studies done in Asia and Africa have demonstrated that chronic low-level AFB exposure in foodstuff increases the risk of developing hepatocellular carcinoma (91). The foodstuffs were mostly contaminated with Aspergillus flavus and Aspergillus parasiticus that are known to produce AFB. AFB and AFM have been detected in blood, urine, and feces of individuals who consumed heavily contaminated food. However, co-infection with hepatitis B virus is an

226 Fung and Clark important synergistic factor affecting carcinogenisity of AF mycotoxins (92). The carcinogenesis of AF involves several mechanisms. The formation of AF epoxide binds at the critical nucleophilic sites of DNA and RNA leading to mutation and subsequent cancer formation. AF can activate proto-oncogenes (c-mys, c-ha-ras, etc.) and inhibit tumor suppression genes (16,54,93). Acute large exposure to AF can cause severe liver damage with high mortality and morbidity (94). This type of exposure has been proposed to induce hepatoencephalopathy similar to Reyes syndrome (95). However, clinical evidence of this is limited. In addition, there is only weak evidence regarding carcinogenicity of inhaled AF. As noted above, animal studies have shown that AF can produce mucosal and tracheobronchial damage. A mortality study on peanut workers showed increased rates of overall and lung cancer mortality. However, smoking was not controlled and the workers did not have an increased rate of hepatocellular carcinoma, making the results inconsistent with other studies and animal models (96). In a study of agriculture workers exposed to aflatoxins, a two- to three-fold increase in risk of hepatocellular carcinoma and biliary cancer was found. However, the exposure assessment was historical and not measured (97). A case report of alveolar cell carcinoma was reported as a result of inhaling aflatoxin in a laboratory worker (98). Associations with other mycotoxins and cancer have been proposed. Ochratoxins produced by Aspergillus ochraceous and some Penicillium species are associated with renal tumors and Balkan endemic nephropathy (29). Patulin produced by some Penicillium species typically contaminates apple juice and is an animal carcinogen (99). Zearalenone produced by Fusarium sp. (F-2 toxin) possesses estrogenic activity that has the potential to stimulate estrogen sensitive tumors such as breast and cervical neoplasia (100 102). However, animal studies are equivocal and human studies are lacking. Fumonisins found in corn produced by Fusarium species have been postulated to be associated with esophageal cancer (103). However, other contaminants were also present in corn, confounding factors were not controlled, and the data was equivocal (34,103). Trichothecenes have not been associated with human cancer based on World Health Organization reports (36). Interestingly, some mycotoxins are antineoplastics. For example, ergosterol peroxide from Paecilomyces was found to be more potent than cisplatin against gastric tumors (104). T-2 toxins have been shown to inhibit several cancer cell lines such as human myeloid leukemia K562, human cervical carcinoma, and HeLa cells (105). In summary, epidemiological evidence suggests an association between chronic consumption of AFcontaminated foodstuff and hepatocellular carcinoma. Laboratory animal toxicological evidence also suggests that ingestion of large doses of mycotoxins can produce tumors. However, there is no evidence linking inhaling mycotoxins with human malignancy in a nonagricultural or nonindustrial setting. It cannot be assumed that people exposed to indoor mold spores have also been exposed to mycotoxins. Lifetime cancer surveillance is not even recommended in the industrial setting, where airborne mold exposure can be intense. Cancer surveillance for individuals exposed to indoor mold is not supported by scientific evidence and such practice is therefore not warranted or medically indicated (106). Renal Toxicity Ochratoxins (found in cereals, coffee, bread, and meat) produced by Penicillium and Aspergillus are associated with Balkan endemic nephropathy (99). There is a case report suggesting acute renal failure was due to inhaling airborne ochratoxins. It was postulated that the patient was exposed to ochratoxins that were later isolated from wheat (107). Infants with IPH related to Stachybotrys exposure suffered electrolyte imbalance, non-ionic gap acidosis, and proteinuria, but this may be related to shock syndrome rather than mycotoxin effects (63,65). It should be pointed out that no airborne trichothecenes have been isolated or correlated with renal toxicity. Teratogenicity and Effects on Pregnancy Many mycotoxins are animal teratogens. Animal studies have also demonstrated mycotoxic effects on fertility and embryo implantation (108,109). Ergots have been shown to induce abortion due to their oxytocic or uterine contraction effects (110). Zearalenone may cause infertility and fetal malformation due to its estrogenic effects (101,102). Trichothecenes have been shown to cause limb and tail abnormalities in animal studies (111). Aflatoxins are teratogenic to most animal models studied (36). Animals fed with Stachybotrys mycotoxins were noted to have low fertility and a decrease in litter size in addition to increased frequency of fetal death or resorption (102). However, no maternal effects were documented and no mycotoxins were measured in either fetus or mother (102). It is also

Health Effects of Mycotoxins 227 important to point out that these exposures were all by oral administration. Extrapolation of animal data with these major limitations cannot be applied in human inhalational exposure. There are no human epidemiological studies on fetal and maternal effects associated with consumption of mycotoxin in foodstuff. In summary, there is only weak evidence on the teratogenic effects from mycotoxin ingestion. There is no evidence that inhaling mycotoxins in indoor nonagricultural settings has teratogenic and adverse affects on pregnancy. Gastrointestinal Toxicity Mold-contaminated food products and potentially mycotoxins are known to cause nausea, vomiting, abdominal pain, and diarrhea when ingested (74,112). The mechanism of toxicity is related to the direct toxic effects on GI mucosal surfaces. Mushroom toxicity exerts similar toxic effects (113,114). ISSUES RELATED TO MYCOTOXIN EXPOSURE AND DISEASE Exposure Assessment In the discipline of clinical toxicology, issues related to exposure and disease are important. These issues include exposure assessment, toxin identification, dose-response relationship, mechanism of toxicity, and toxic syndrome or disease resulting from the exposure. Some degree of accuracy and consistency on exposure assessment and toxicity are needed for risk assessment and public health policymaking. To begin with, the terminology of units of measurement of fungi can be confusing. Industrial hygienists may use the unit designations of spores/m 3 or CFU (colony forming unit)/m 3 in reports of mold presence. These units measure different things. Spores/ m 3 is the unit of measurement for total mold spores (both viable and nonviable). CFU/m 3 measures viable fungal spores only, or the potential infectivity of the fungal species. Since mycotoxins may still be present in dead (nonviable) fungal spores, the total (viable plus nonviable) mold spore count measured in spores/m 3 is a better indicator of potential mycotoxin exposure, assuming that the mold spores contain mycotoxins. Although protocols for fungal sampling and collection are available, there are significant limitations with current methodologies. Fungal capturing devices may have a variation of up to 1,000 fold between specimens obtained from the same source (115). Single samples from either the suspected or control area therefore cannot provide scientifically meaningful conclusions due to lack of statistical and practical significance. This is especially important if only one or a limited number of samples are obtained in selective areas without proper methodological and statistical considerations prior to fungal sampling. Exposure to mold and mycotoxin has not been compared between agricultural and occupational settings. It has been shown that mold exposure in agricultural environments can reach 10 10 spores/m 3 (116,117). Aggressive or destructive testing for mold spores overestimates realistic exposure levels. Isolation of toxigenic mold while ignoring other potential indoor contaminants (chemicals, endotoxins, bacteria, viruses, animal and insect products) may invalidate study conclusions. Microorganisms other than fungi can also produce toxic materials. For example, fungi as well as building materials and human activities are sources of volatile organic chemicals. Since mycotoxins are of low volatility, isolation or detection of mycotoxins in dust and building materials does not imply inhalation mycotoxin exposures. In addition to factors important for mycotoxin transport, an association has not been established between biological effects and surface area mold counts (118). The conditions for mycotoxin production also vary widely. Toxin production is, in general, dependent on nutrients available to the fungus, moisture level, ph, temperature, substrates, and presence or absence of specific gases or essential metals. Therefore, the presence of potentially toxigenic fungi does not imply the presence of mycotoxins, nor has the finding of mycotoxins in the amounts typically measured ever been shown to cause mycotoxicosis. In addition, the finding of mycotoxins does not prove that a particular fungal species was or is present. Fungal species identification is not a simple process, and may require the expertise of specialized medical mycologists (115,119). Lastly, mycotoxin detection is difficult and has many limitations. There are no standardized protocols by which to collect mycotoxins from the field (120). It is important to emphasize that mycotoxins may or may not be present in fungal spores. Even if present, mycotoxin concentration per spore is only in trace amounts. As a result, a large quantity of spores is necessary to be processed in order to achieve analytical detectable limits, and the present technology of analyzing for many mycotoxins does not have validated reference ranges or detection limits. Therefore, the concentrations reported may not have scientific validity or clinical relevancy. Except for a few well-studied mycotoxins such as aflatoxins, gold standard mycotoxins may not be readily available as

228 Fung and Clark references when analyzing mycotoxins collected from the field. The analytical sensitivity, specificity, and detection range have not been adequately characterized. In summary, the fundamental issues of mycotoxin exposure in terms of collection, sampling, and analysis have not been standardized or defined. Currently, mycotoxin exposure assessment is only experimental. Without further epidemiological and clinical research, application of these methods to human risk assessment is unjustified and potentially very misleading. Dose-Response Relationship Mold spores consist not just of toxin but also sugar, water, and other materials like pollens. Mold spores frequently do not contain mycotoxins, thus for example, not all Stachybotrys spores contain trichothecenes (121). Furthermore, not all spores can reach lung tissues because of their size. However, for the sake of assessing risk, let s assume all mold spores 1) contain trichothecenes, 2) reach the lung tissues, and 3) are fully absorbed. The minimal levels of Stachybotrys spores that must be present in the air for health effects to occur in both acute (one time) and chronic (24 h for weeks) exposures have been calculated (106). For a one-time exposure, the concentration is estimated to be 15.310 6 or 15.3 million spores in one cubic meter of air. For chronic exposure, the concentration is anticipated to be 6810 3 or 68 thousand spores in one cubic meter of air. A dose-response relationship can be obtained using an inhalational risk assessment model for fungal exposure in indoor environments. This model considers fungal particle size and maximum potential dose of mycotoxins (122). Under laboratory conditions, one spore of Stachybotrys atra contains 3 fentigrams (fg) of trichothecene (123). Using this model, a 1200 h exposure of 1,000 Stachybotrys atra spores would give a potential exposure of 10,800 fg or 0.01 nanogram of trichothecenes (122). Acute toxicity (LD50) for T- 2 toxin per inhalation is reportedly 20 mg/m 3 (124). To reach a 20 mg/m 3 concentration, the number of spores required to reach an airborne mycotoxin concentration that could potentially induce acute toxicity would be 2010 12 fg/m 3 divided by 3 fg/spore or 6.610 12 spores/m 3. To put this exposure into perspective regarding airborne particles, in an agricultural setting, total mold concentrations exceeding 110 6 spores/m 3 are considered irritating and the environment as dusty. If the spore concentration exceeds 110 9 spores/m 3 the environment would be so cloudy that a person could not see more than a few yards (116,117). The calculated level of mold exposure and dose of mold toxin (trichothecene, T-2) required to induce toxicity is consistent with the U.S. Army Medical Research Institute of Infectious Diseases data on the Comparative Lethality of Toxins. The data show that mycotoxin T-2, with an LD50 of 1.2 mg/kg is about one million times less toxic than bacterial toxins such as Botulinum toxin (125). Mycotoxicosis Case Definition Patients with mycotoxicosis due to consumption of mycotoxin typically suffer severe constitutional symptoms and present with objective findings of illness. In contrast, indoor mold exposure patients typically have ill-defined symptoms with minimal objective findings. Most studies published in peer review literature on the health effects of mold exposure have relied on subjective and retrospective questionnaires. A recently published paper reviewed all human studies and reports associating indoor mold exposure and health effects from an epidemiological perspective. These authors concluded that there is insufficient evidence to support mycotoxicosis due to indoor mold exposure (126). Only a few studies have included physical examinations and diagnostic testing. It is unclear in these studies whether these objective findings were related to fungal exposure. The validity of self-reported and retrospective questionnaires on symptoms may be fraught with flaws such as recall bias, systematic bias (127), and potential secondary gain in disputed situations. While most cross-sectional studies reported more wheezing, cough and cold symptoms, conclusions were made despite no supportive objective evidence between mold exposure and symptoms. Currently, there are several symptom-driven syndromes such as chronic fatigue syndrome (CFS) and multiple chemical sensitivity (MCS) that also suffer similar case definition weaknesses. Case definition is important even when the causation is unknown. Case definition clarifies the scope of the problem and patients can be managed appropriately. Other than allergic reactions to mold, and mycosis or mold infections, many patients who experience a constellation of nonspecific symptoms have been labeled as mycotoxicoses despite a lack of objective findings, exposure assessment, and toxicity evaluation. These patients are clinically different from those who have consumed mycotoxin-contaminated foodstuffs. Management Management of potentially toxic mold exposure is mainly supportive. There are no antidotes for mycotoxins. However, it is important to listen to a patient s

Health Effects of Mycotoxins 229 complaints and obtain a detailed history in terms of exposure to mold and other potentially toxic substances. Except in unusual circumstances such as outbreaks of ingestion of mycotoxin-contaminated food, physical examination and laboratory testing are usually nonspecific and not diagnostic. A good differential diagnosis is critical because patients experience various symptoms that have other possible etiologies. When there is no evidence of significant objective findings of disease after a thorough history-taking, exposure assessment, physical examination, and appropriate diagnostic testing, patients should be informed that it is impossible to conclude with any reasonable medical certainty that their symptoms are related to any fungal exposure. Furthermore, on the basis of current knowledge, it is very unlikely that fungal exposures have caused any toxicity in this patient, nor will any toxicity occur in the future based on this exposure. CONCLUSION Fungi can cause adverse human health effects to many organ systems. In addition to infection and allergy, fungi can produce mycotoxins and organic chemicals that are responsible for various toxicologic effects under specific circumstances. Review of scientific literature on mycotoxin-related human diseases clearly reveals a linkage between ingesting mycotoxin-contaminated food and illness, especially hepatic, gastrointestinal, and carcinogenic diseases. Currently, there is no supportive evidence to imply that inhaling mold or mycotoxins in indoor environments is responsible for any serious health effects other than transient irritation and allergies in immunocompetent individuals. Several academic and professional organizations have published position papers stating similar conclusions (106,128,129). Further research on the health effects of inhaling mycotoxins should focus on the following issues. Objective biomarkers of disease should be identified and correlated with mycotoxin exposure. Appropriate animal models adhering to inhalation toxicology principles are needed to better understand mycotoxin absorption, metabolism, and disease relationship in indoor environment. Sound epidemiologic methodology is urgently needed to associate individuals exposed to known concentrations of mycotoxin with well-defined end points or disease entities. Other indoor aerosols and mold components such as endotoxins, volatile chemicals, allergens, glucans, and ergosterols should be studied individually and in combination to elucidate specific diseases resulting from fungal exposure. Finally, exposure assessment techniques, laboratory testing, analyses on mycotoxins, and other aerosols should be standardized. Diagnostic tools for evaluating mycotoxin exposure patients should have appropriate sensitivity, specificity, and predictive values so as to be clinically useful for establishing exposure-disease relationships. Public policy and resource allocation can then be set based on sound science. REFERENCES 1. Kendrick B. Chapter 1, kingdoms and classification. In: The Fifth Kingdom. 2nd ed. Waterloo, Ontario, Canada: Mycologue, 1992:1 6. 2. Solomon WR, Platts-Mills TAE. Chapter 19, aerobiology of inhalant allergens. In: Middleton E, Reed CE, Ellis EF, Adkinson NF, Yunginger JW, Busse WW, eds. Allergy-Principles and Practice. 4th ed. St. Louis: Mosby, 1993:469 528. 3. Jelinek CF, Pohland AE, Wood GE. Review of mycotoxin contamination worldwide occurrence of mycotoxins in foods and feeds an update. J Assoc Anal Chem 1989; 72(2):223 230. 4. Kendrick B. Chapter 21, mycotoxins in food and feed. In: The Fifth Kingdom. 2nd ed. Waterloo, Ontario, Canada: Mycologue, 1992:316 331. 5. FDA: FDA/CFSAN Food Compliance Program. Mycotoxins in Imported Foods. Center for Food Safety Applied Nutrition Food Compliance Program. Issued January 20, 1999. 6. Ueno Y. The toxicology of mycotoxins. CRC Crit Rev Toxicol 1985; 14(2):99 132. 7. Hadidane R, Roger-Regnault C, Bouattour H, Ellouze F, Bacha H, Creppy EE, Dirheimer G. Correlation between elementary mycotoxin contamination and specific diseases. Human Toxicol 1985; 4(5):491 510. 8. MMWR. Pulmonary hemorrhage/hemosiderosis among infants Cleveland, Ohio, 1993 1996. Morb Mort Wkly Rep 1997; 46:33 35. 9. Zavaleta EG, Fernandez BB, Grove MK, Kaye MD. Saint Anthony s fire (ergotamine induced leg ischemia) a case report and review of the literature. Angiology 2001; 52(5):349 356. 10. Williams SR. Chapter 102, Ergotamines. In: Dart RC, ed. Medical Toxicology. 3rd ed. Baltomore: Lippincott, Williams & Wilkins, 2004. 11. Dennings DW. Aflatoxin and human disease. Adv Drug React Ac Pois Rev 1987; 6(4):175 209. 12. Forgacs J, Carll WT. Mycotoxicoses: toxic fungi in tobaccos. Science 1966; 152(729):1634 1635. 13. Riley RT, Kemppainen BW, Norred WP.