JFS Special Issue: 75 Years of Advancing Food Science, and Preparing for the Next 75

Advances in Mycotoxin Research: Public Health Perspectives Hyun Jung Lee and Dojin Ryu

Aflatoxins, ochratoxins, fumonisins, deoxynivalenol, and zearalenone are of significant public health concern as they can cause serious adverse effects in different organs including the liver, kidney, and immune system in humans. These toxic secondary metabolites are produced by filamentous fungi mainly in the genus Aspergillus, Penicillium, and Fusarium. It is challenging to control the formation of mycotoxins due to the worldwide occurrence of these fungi in food and the environment. In addition to raw agricultural commodities, mycotoxins tend to remain in finished food products as they may not be destroyed by conventional processing techniques. Hence, much of our concern is directed to chronic health effects through long-term exposure to one or multiple mycotoxins from contaminated foods. Ideally risk assessment requires a comprehensive data, including toxicological and epidemiological studies as well as surveillance and exposure assessment. Setting of regulatory limits for mycotoxins is considered necessary to protect human health from mycotoxin exposure. Although advances in analytical techniques provide basic yet critical tool in regulation as well as all aspects of scientific research, it has been acknowledged that different forms of mycotoxins such as analogs and conjugated mycotoxins may constitute a significant source of dietary exposure. Further studies should be warranted to correlate mycotoxin exposure and human health possibly via identification and validation of suitable biomarkers.

Abstract:

Keywords: food safety, mycotoxins, public health

Introduction Mycotoxins are secondary metabolites, or simply chemicals, produced by certain filamentous fungi that may cause adverse effect in animals and humans. Approximately 400 mycotoxins are known to date with vast structural diversity, although it is well established that not all fungi produce toxic metabolites, and not all secondary metabolites are toxic (Hussein and Brasel 2001). Only a limited number of mycotoxins occur frequently at significant concentrations in food. Among all mycotoxins, aflatoxins, ochratoxins, fumonisins, trichothecenes (particularly deoxynivalenol), and zearalenone are considered most important in food safety and public health due to their occurrence and toxicity. Mycotoxins are of public health concern mainly due to their worldwide occurrence and prevalence that may lead to adverse effects as a result of chronic exposure even when they contaminate foods at low levels. The important mycotoxin-producing fungi are Aspergillus, Penicillium, and Fusarium spp. (Table 1). It should be noted that a toxigenic species may produce one or more mycotoxins; whereas multiple species can produce one of the mycotoxins. Consequently, co-contamination of mycotoxins poses another layer of significant concern in public health, for example synergistic effect. In a recent survey, aflatoxins, ochratoxins, fumonisins, deoxynivalenol, and zearalenone were found in 33%, 45%, 59%, 64%, and 28% of the samples, respectively, from MS 20151535 Submitted 9/10/2015, Accepted 10/20/2015. Authors are with School of Food Science, Univ. of Idaho, 875 Perimeter Drive MS 2312, Moscow, ID 83844, U.S.A. Direct inquiries to author Ryu (E–mail: [email protected]).

T: Toxicology & Chemical Food Safety

This article was originally published online on November 13, 2015. Subsequently, an error in the presentation of Figure 1 and typographical errors were corrected and the article was published on November 25, 2015.

T2970

Journal of Food Science r Vol. 80, Nr. 12, 2015

7049 samples sourced in North and South America, Europe, and Asia from 2009 to 2011 (Rodrigues and Naehrer 2012). It should be noted that 81% of all samples were positive for at least one mycotoxin. Failure to prevent fungal growth and toxin production in the field or during storage will inevitably lead to a health risk to the consumer, together with often significant economic losses. Once ingested, mycotoxins are generally absorbed from the gastrointestinal tract which may be excreted following metabolisms in varying length of time. Mycotoxins can affect single or multiple target organs of animals and human with varying degree of cytotoxic, mutagenic, teratogenic, carcinogenic, and/or immunosuppressive potency. Hence, many countries have adopted regulations to limit mycotoxins in foods to protect consumers from health risks associated with these toxins.

Background of Major Mycotoxins Aflatoxins After the identification of aflatoxins in 1960, aflatoxins are most studied mycotoxins and of greatest concern in public health worldwide. Aflatoxins are a group of mycotoxins produced mainly by Aspergillus flavus and A. parasiticus (Frisvad and others 2006). A. flavus is known as the most common aflatoxigenic species in agricultural commodities, including corn, producing only aflatoxin B series, whereas A. parasiticus has been found mostly in peanuts producing both B and G series (Peraica and others 1999). Among all analogs of aflatoxins, aflatoxin B1 (AFB1 ) is most toxic and occurs most frequently with highest concentrations. Aflatoxin M1 (AFM1 ) is the hydroxylated metabolite of AFB1 following ingestion, found in the milk and urine of humans and animals (Qian and others 1994; Thirumala-Devi and others 2002). R  C 2015 Institute of Food Technologists

doi: 10.1111/1750-3841.13156 Further reproduction without permission is prohibited

Advances in Mycotoxin Research . . . Table 1–Summary of the major mycotoxins, associated fungi and their physiological effects. Mycotoxins

Fungi

Aflatoxins

Aspergillus flavus, A. parasiticus

Fumonisins

Fusarium verticillioides (syn., moniliforme), F. proliferatum A. ochraceus, A. carbonarius, Penicillium verrucosum

Ochratoxins Deoxynivalenol Zearalenone

F. sporotrichioides, F. graminearum, F. culmorum, P. poae, F. roseum, F. tricinctum, F. acuminatum F. granimearum, F. culmorum, F. crookwellense

Toxicity

Reference

Carcinogenic, acute hepatotoxic, immunology suppression Carcinogenic, hepatotoxic

Pier and others (1976), Cullen and others (1993), IARC (1993) IARC (1993), Marasas (1995)

Cacinogenic, nephrotoxic, hepatotoxic, teratogenic Gastrointestinal haemorrhaging, immuno-depressants

Mayura and others (1984), Lea and others (1989), Gekle and Silbernagl (1993), IARC (1993), JECFA (2001) Arnold and others (1986), Forsell and others (1987), Hughes and others (1999)

Estrogenic activity

Mirocha and others (1971)

Table 2–European regulations for major mycotoxins in food (EC 1991; EC 1996; EC 1998; EC 1999; EC 2000; EC 2001a; EC 2001b; EC 2004; EC 2006). Mycotoxins

Commodity

Maximum limit (µg/kg)

Aflatoxins Groundnuts (peanuts), dried fruits and processed products thereof Groundnuts subjected to sorting or other physical treating Nuts, dried fruits, and corn subjected to sorting or other physical treating Cereals and processed cereal products Milk Cereal based baby foods for infants and young children Infant formulae and infant milk Fumonisins (Sum of FB1 and FB2 )

Unprocessed corn Corn grits, meal, and flour Corn-based breakfast cereals and corn-based snacks Processed corn-based baby foods for infants and young children

AFB1 2 8 5 2 − 0.1 −

Total AFsa 4 15 10 4 − − −

AFM1 − − − − 0.05 − 0.025

4,000 1,000 800 200

Ochratoxin A Unprocessed cereals; Roasted coffee beans, ground roasted coffee All product derived from unprocessed cereals intended for direct consumption Dried vine fruits (raisins, currents, sultanas); Instant coffee Wine; Grape juice (not concentrated) Cereal based baby foods for infants and young children

5 3 10 2 0.5

Deoxynivalenol Durum wheat, oats, and corn Unprocessed cereals other than durum wheat, oats, and corn Cereal flours used as raw material in food products Cereal products as consumed and other cereal based products as retail stage Cereal based baby foods for infants and young children

1,750 1,250 750 500 200

Zearalenone Unprocessed cereals other than corn; Corn intended for direct human consumption, corn-based snacks and corn-based breakfast cereals Unprocessed corn Cereal flours other than corn flour Corn flour All product derived from unprocessed cereals intended for direct consumption (excluding processed corn-based foods) Cereal based baby foods (including processed corn-based foods) for infants and young children

350 75 200 50 20

Total AFs means sum of AFB1 , AFB2 , AFG1 , and AFG2 .

AFB1 is not only carcinogenic but also teratogenic, hepatotoxic, mutagenic, and immunosuppressive, and the main target organ is the liver (Eaton and Gallagher 1994). Based on human epidemiological studies, AFB1 is one of the most potent hepatocarcinogen among naturally occurring toxicants and known to cause increased hepatocellular carcinoma in populations exposed to AFB1 via contaminated foods (Peers and Linsell 1973; Groopman and others 1988). Hence, AFB1 is classified by the International Agency of Research on Cancer (IARC) in Group 1, as carcinogenic to humans, whereas AFM1 is listed in Group 2B, as possibly carcinogenic to humans (IARC 1993). As shown in Table 2, the European Commission (EC) set maximum limits for AFB1 and total aflatoxins at 2 and 4 μg/kg, respectively in nuts, dried fruits, and cereals (EC 1998, 2001a), and have regulated a maximum limit of 0.025 μg/kg for AFM1 for infant formula and

infant milk (EC 2006). The Codex Alimentarius Commission has established a maximum limit of 15 μg/kg total aflatoxins for raw shelled nuts destined for further processing (Codex 2001). The U.S. Food and Drug Administration (FDA) has set action levels for aflatoxins; total aflatoxins of 20 μg/kg in foods including nuts and AFM1 of 0.5 μg/kg in milk (FDA 2011).

Ochratoxins Although ochratoxin A (OTA) is known to be most toxic and most prevalent among the forms of ochratoxin in food, this class of mycotoxins is produced by a number of species in 2 distinctively different genera of Aspergillus and Penicillium under varying environmental conditions (Frisvad and others 2006). In turn, ochratoxins have been found in an exceptionally wide variety of agricultural commodities and their products including all Vol. 80, Nr. 12, 2015 r Journal of Food Science T2971

T: Toxicology & Chemical Food Safety

a

100

Advances in Mycotoxin Research . . . Table 3–The U.S. Food and Drug Administration action levels for major mycotoxins in food (FDA 2011). Mycotoxins

Commodity

FDA action level (µg/kg)

Aflatoxins Foods, peanuts and peanut products, brazil and pistachio nuts Milk Fumonisins Degermed dry milled corn products (corn grits, meal, and flour) Cleaned corn intended for popcorn Dry milled corn bran Cleaned corn intended for masa production

Total AFsa 20

AFM1 −

−

0.5 Total fumonisinsb 2,000-4,000 3,000 4,000 4,000

Trichothecenes (Deoxynivalenol) Processed wheat-based products

1,000

a Total b

AFs means sum of AFB1 , AFB2 , AFG1 , and AFG2 . Total fumonisins means sum of FB1 , FB2 , and FB3 .

major cereal grains, fruits and fruit juices, dried fruits, nuts, spices, coffee, cocoa, wine, beer, and meats (Aragu´as and others 2005; Gonzalez and others 2006; Duarte and others 2011). A. ochraceus occurs in stored cereal grains and coffee beans (Van der Merwe and others 1965; Taniwaki and others 2003). Another important OTA producer, A. carbonarius, occurs in grapes as well as in coffee beans (Horie 1995; Joosten and others 2001). P. verrucosum is also a major OTA-producer particularly in cereal grains (Lund and Frisvad 2003). OTA is a potent renal carcinogen based on experimental animal studies and has been classified as a possible human carcinogen, Group 2B (IARC 1993; NTP 1989). In addition, OTA is known to be teratogenic, embryotoxic, genotoxic, neurotoxic, and immunosuppressive in various animal models (Mayura and others 1984; Lea and others 1989; Gekle and Silbernagl 1993). However, its impact on human health is largely unclear. Although OTA has long been linked to renal diseases in humans including Balkan endemic nephropathy (BEN) and chronic interstitial nephropathy (CIN), Grollman and others (2007) suggested aristolochic acid, from the weeds in genus Aristolochia commonly found in the field, as the most likely causative agent of BEN. Nonetheless, the Joint FAO/WHO Expert Committee on Food Additives (JECFA) evaluated toxicity of OTA and established a provisional tolerable weekly intake (PTWI) of 100 ng/kg of body weight (approximately 14 ng/kg bw/day; JECFA 2001). The EU Scientific Committee for Food (SCF) also proposed the tolerable daily intake (TDI) of 5 ng/kg of body weight per day (SCF 1998). Recently Canada proposed similar regulatory guidelines (Canada Health 2009). At present, there are no guidance or action levels for OTA in foods in the United States. It would be of interest for public health perspectives to note that oats and oat-based products may be contaminated with OTA with high incidence and levels. According to recent surveys conducted in the United States, OTA was found in 201 out of 489 (42%) retail breakfast cereal samples in the range of 0.1 to 9.3 ng/g (Nguyen and Ryu 2014a; Lee and Ryu 2015). More significantly, 70% (142/203) of oat-based breakfast cereals were contaminated with OTA, the highest incidence among all product groups tested. It should also be noted that oat-based infant cereals were highly contaminated with OTA (59% or 30/51) in the range of 0.6 to 22.1 ng/g (unpublished data). Such high incidence and levels of OTA contamination are of particular concern in public health in considering the EU’s maximum limit of OTA is 0.5 ng/g for infant foods.

considered plant pathogens, as other Fusarium species, as they can infect the host plant to cause diseases, for example ear rot in corn. It should be noted that fumonisins are produced in corn during the preharvest and postharvest periods, even without visible growth or symptoms on the plant as the Fusarim spp. are endophytes (Bacon and others 2008). Fumonisins are found most frequently in corn worldwide as the producers can grow over a wide range of temperatures and at relatively high water activities (aw > 0.9; Marin and others 2013). Among all analogs in this group, fumonisin B1 (FB1 ) is known to be most toxic and occurs most frequently (Shephard and others 2007). FB1 has been found to be a potent cancer promoter (Gelderblom and others 1988). Based on studies with experimental animals, the kidney and liver are considered as the major target organs with carcinogenicity of FB1 found to be the prevalent endpoint (Gelderblom and others 1991). Hence, FB1 is listed as a Group 2B carcinogen (IARC 1993) and the EC has recommended guidance levels for FB1 and total fumonisins in foods (Table 2; EC 2006). JECFA established a provisional maximum tolerable daily intake (PMTDI) for fumonisins (FB1 , FB2 , and FB3 , alone or in combination) of 2 μg/kg of body weight per day (JECFA 2011a). Recommended maximum levels for total fumonisins were also adopted by FDA for corn and processed corn-based products intended for human consumption as shown in Table 3 (FDA 2011). Nonetheless, the effect of fumonisins on human health is unclear although several epidemiological studies reported its association with esophageal cancer and liver cancer in Africa and China (Sydenham and others 1990; Ueno and others 1997). More recently, FB1 has been suggested as a risk factor in the development of neural tube defect (Gelineau-van Waes and others 2009; Voss and others 2014).

T: Toxicology & Chemical Food Safety

Deoxynivalenol Deoxynivalenol (DON, also known as vomitoxin) is one of the trichothecenes, a class of more than 200 structurally related hematotoxic and immunosuppressive mycotoxins produced by Fusarium spp. (Frisvad and others 2006). DON has received particular attention due to its toxicity and widespread occurrence. F. graminearum and F. culmorum are known as consistent producers of DON in all major cereal grains including wheat, barley, oats, and corn. These fungi are important plant pathogens, causing diseases in the host plant such as Fusarium head blight, that grow in the crop in the field under temperate climates (Eriksen 1998). DON causes vomiting, feed refusal, and teratogenicity in Fumonisins animals (Forsyth and others 1977; Khera and others 1986), Fumonisins are produced mainly by Fusarium verticillioides and and gastroenteritis with vomiting in humans (Pestka and F. proliferatum (Frisvad and others 2006). These organisms are also Smolinski 2005). Moreover, DON has been shown to possess T2972 Journal of Food Science r Vol. 80, Nr. 12, 2015

Advances in Mycotoxin Research . . .

Zearalenone Zearalenone (ZEN) is produced mainly by F. graminearum and F. culmorum (Frisvad and others 2006). Although the major sources of ZEN are wheat, rye, and oats in European countries, corn and wheat are more frequently contaminated with ZEN in the United States and Canada. These fungal species can grow under cool and moist conditions when these crops are in bloom as well when the harvested grain is in storage (Panel 2011). Public health concern from ZEN is attributed to its estrogenic activity, binding competitively to estrogen receptors in various animal models, although its affinity is much lower than that of physiological concentration of 17β-estradiol (Kuiper-Goodman and others 1987; EFSA 2004). ZEN is classified under Group 3 (IARC 1993), although carcinogenic, hepatotoxic, genotoxic, and immunosuppressive properties of this endocrine disruptor have been documented (Becci and others 1982; El-Makawy and others 2001; Kostro and others 2011). Established PMTDI for ZEN is 0.5 μg/kg bw/day (JECFA 2010) whereas its maximum limits in foods in the EU range from 20 to 350 μg/kg (Table 2).

Effect of Food Processing Once harvested, the majority of agricultural crops are subjected to one or more forms of food processing before consumption. It is important to understand the effects of food processing on public health because the physical and/or chemical properties as well as the concentrations of mycotoxins in foods may be altered during the processing, which may lead to altered toxicities in animals and humans. Therefore, it is important to understand the effect of processing, on both chemical and toxicological changes, in setting regulatory limits as summarized in Table 2 and 3. Although regulatory limits are largely determined based on the toxicity data and estimated human consumption or exposure, unrealistic or excessively stringent limits would result in higher price of the crops in trade and ultimately cause an economic burden to consumers. The effect of specific food processes on mycotoxins will not be discussed here because review articles dedicated to this subject are available elsewhere (Ryu and others 2008; Kabak 2009). In general, the rate of reduction is not expected to be significant as mycotoxins are known to be stable under most conventional food processing conditions other than a few exceptions including nixtamalization, roasting, and extrusion.

Conjugated Mycotoxins One of the most recognizable advances in mycotoxin research is in better understanding of different forms of mycotoxins present in various commodities and food products. Until recently, mycotoxin analysis has assumed nearly complete recovery of the amount of naturally occurring toxins present in food. However, mycotoxins may present as conjugated forms in plants; either as extractable forms (also called masked mycotoxins) or as unextractable forms

(also known as bound mycotoxins). In addition, mycotoxins, in free or bound forms, may undergo chemical transformations during food processing or bind with food matrix, both of which might result in underestimation of mycotoxin bioavailabilities and toxicitities. Hence, the importance of understanding different forms of mycotoxins in various foods and commodities should not be overlooked. Because Lu and others (1997) observed reduced toxicity of FB1 fructose adducts formed during heat processing, a series of studies have been conducted to elucidate the chemical and toxicological fate of fumonisins during food processing. With a particular interest in thermal processing, reaction of FB1 in the presence of reducing sugars was characterized and the major reaction products were identified as N-(carboxymethyl)-FB1 and N-(1-deoxyd-fructos-1-yl)-FB1 (Howard and others 1998). In addition to reacting with reducing sugars in the Maillard reaction, fumonisins may also bind to other constituents in foods, for example proteins. For example, Kim and others (2003) found an average of 2.6 times more FB1 in a protein-bound (or hidden) form than free FB1 in retail corn flakes; these forms which should be considered when estimating human exposure. The formation of such “bound” form of FB1 and potentially other mycotoxins during thermal process prompted concerns in food safety because (i) these reaction products may not be detected by conventional analytical techniques such as high-performance liquid chromatography (HPLC) due to the chemical transformation and (ii) may remain toxic or release the parent compound (mycotoxin) during digestion. Bound fumonisins had been widely believed to occur only during thermal processing, and it was suggested that a covalent bond formed between the tricarboxylic moiety of fumonisins and hydroxyl group of starch or the amino groups in protein (Park and others 2004). However, Dall’Asta and others (2008) found bound fumonisins in unprocessed corn, and suggested that noncovalent interactions may partly be responsible for the formation of bound fumonisins.

Methodology Mycotoxins are present in food in very low concentrations, that is parts per million (ppm or mg/kg) to parts per billion (ppb or μg/kg). However, mycotoxins may cause toxic effects at such low levels, thus regulatory limits are set accordingly for each mycotoxin in various commodities or foods to protect public health. Therefore, development of reliable methods for detection and quantification of mycotoxins is essential in research, surveillance, and regulation. As mycotoxins tend to remain not only in raw agricultural commodities but in their products manufactured in the downstream processes, the development and application of reliable method is a very challenging task in considering the diverse physical and chemical nature of the raw and processed products as well as the properties of mycotoxins of interest. Significant advances have been made in the last few decades in the development of methods for detection and qualification of mycotoxins. By in large, these methods are based on chromatographic separations such as thin-layer chromatography, gas chromatography, HPLC, and liquid chromatography coupled with various types of mass spectrometry (LC/MS). Because most of these chromatographic methods require expertise and considerable investment in instrumentation, enzyme-linked immunosorbent assay (ELISA) has become a method of choice in routine analyses in the field, for example in grain elevators, as it offers sensitive and rapid analyses with simplicity and low cost. Notable advances in these conventional methods are discussed in this Vol. 80, Nr. 12, 2015 r Journal of Food Science T2973

T: Toxicology & Chemical Food Safety

immunosuppressive properties in animal and human lymphocytes (Bondy and Pestka 1991; Meky and others 2001). The IARC placed DON in Group 3 (not classifiable as to its carcinogenicity to humans; IARC 1993). The EC has recommended maximum limits for DON and its derivatives in foods including cereals as showed in Table 2 (EC 2006). The FDA also has established an advisory level of 1000 μg/kg for DON for finished wheat products (FDA 2011). SCF established a recommended tolerable daily intake (TDI) for DON as 1 μg/kg bw/day whereas PMTDI was set at the same level for DON and its acetylated derivatives by JECFA (SCF 2002; JECFA 2011b).

Advances in Mycotoxin Research . . . Table 4–Summary of mycotoxin detection methods. Type of detection technique Conventional techniques

Selectivity

Sensitivity

High

High

Very high

Very high

Enzyme-linked immunosorbent assay (ELISA)

Good

Good

Immunochemical techniques (Flow-through assay and lateral flow assay)

Good

Fair

Electrochemical immunoassay

Good

Good

Relatively low cost Decreased reproducibility by Feasibility for miniaturization and repeated use multiplexing Ease of application

Fluorescence polarization immunoassay (FPIA)

Good

Good

Rapid analysis Ease of application Portable

Matrix interference Not a high throughput method

–a

–

More stable over high temperatures and array of organic solvents Greater solubility Ease of modification by simple recombinant DNA

–

Sensor technologies

Low

Low

Rapid analysis Relatively low cost Nondestructive analysis

Dried blood spot (DBS)

Low

Low

Simple sample treatment Long-term storage (over years)

High matrix dependence Lack or limitation of appropriate calibration materials Nonlinear calibration curves Response to a broad range of compounds Require improved method validation

High-performance liquid chromatography (HPLC)

Liquid chromatography coupled with mass spectrometry (LC/MS)

Emerging techniques

Nanobody

a

Advantage

Disadvantage

Low detection limit High cost Possible quantification of multiple Require clean-up of sample and mycotoxin in single analysis derivation to improve sensitivity Use of organic solvent Time consuming Require training/expertise Very high cost Very low detection limit May require clean-up of sample Possible to quantify multiple extract to improve sensitivity mycotoxins in a single analysis Use of organic solvent Identify and characterize Time consuming mycotoxin metabolites or Require training/expertise degradation products Rapid analysis Cross-reactivity with other Not requiring clean-up or compounds structurally similar concentration step to the antigen Ease of application Matrix interference Relatively low cost Possibility of under- or Portable overestimation Not requiring complex Limited storage condition and equipment and special training shelf-life of the kits Limited use of organic solvent Limited detection range Only qualitative or Extended shelf-life semiquantitative analysis Not requiring clean-up or Matrix interference concentration steps Possible overestimation Rapid analysis Ease of application Relatively low cost Not requiring complex equipment and special training Portable

Not applicable.

T: Toxicology & Chemical Food Safety

section followed by a summary of new or emerging technologies considered from food safety and regulatory standpoints. The EC and FDA have established detailed science-based sampling plans as (Table 4). well as the methods of analyses for various mycotoxins including Sampling aflatoxins, OTA, and Fusarium-toxins (FDA 2004; EC 2006). For Prior to the discussion of analytical methods, importance of example, aflatoxins are very heterogeneously distributed in a lot, sampling should be noted because most mycotoxins are not evenly particularly for food products with a large particle size such as distributed in crops due to sporadic infestation of toxigenic fungi dried figs or groundnuts. In order to ensure samples are reprein most agricultural commodities. Therefore, identifying a method sentative, the quantity of the aggregate sample from the batches for collecting representative samples is an important factor to be with larger particle size should be proportionally larger than that

T2974 Journal of Food Science r Vol. 80, Nr. 12, 2015

Advances in Mycotoxin Research . . .

Conventional techniques The term “conventional techniques” often refers to the ELISA and the chromatographic separation coupled to suitable detection systems. Most analytical methods involve solvent extraction and clean-up steps except ELISA. The solvent extraction methods used to separate target mycotoxins from the sample matrix is dependent on chemical structure of mycotoxins, the type of sample matrix, and following clean-up method. Organic solvents, such as methanol and acetonitrile, are used frequently for extraction as mixtures in varying ratios with or without water. In addition, pH may also need to be considered in extraction as exampled in higher recovery of OTA under acidic conditions (Bazin and others 2013). Consequently, the extraction methods need to be validated according to the target mycotoxins and the sample matrix. Clean-up or purification of initial extracts usually follows to increase analytical sensitivity by reducing background noise. If necessary, concentration of the extract may also be achieved by reducing the volume of extract after clean-up. A variety of clean-up techniques using liquid–liquid partitioning, column chromatography, solid phase extraction (SPE), and immunoaffinity columns (IAC) are available. Among these, IAC based on the use of monoor poly-clonal antibodies has been preferred due to its specificity to the target mycotoxin and greater recovery (Valenta and others 2003; Chan and others 2004). Meanwhile, using stir bar sorptive extraction (SBSE) method was suggested for determination OTA (Nguyen and Ryu 2014b). This equilibrium technique is based on extraction of the analyte from an aqueous matrix onto a sorbent phase coated on a magnetic stir bar. In contrast to extraction with an adsorbent, where the analytes interact with active sites on a surface, in the SBSE method the solutes migrate into the sorbent phase, which are then recovered using organic solvent mixtures (David and Sandra 2007). Although the analysis time is longer than other routine methods, SBSE method is simpler to perform and avoids the clean-up step. Advances made in detection and quantification of mycotoxins is most visible in HPLC and LC/MS offering high sensitivity, accuracy, and efficiency. The limit of detection in modern chromatographic methods may reach sub-ppb levels following appropriate sample preparation and purification steps (Brera and others 2011). In particular, LC/MS has become a universal approach for mycotoxin analysis and confirmation (Z¨ollner and Mayer-Helm 2006). It provides higher sensitivity and selectivity than HPLC coupled with UV or fluorescence. It is also possible to identify and characterize mycotoxin metabolites or degradation products. Thus, an increasing number of researchers have used LC/MS for toxicokinetic and metabolism studies as well as for identification and quantification of reaction products formed during food processing (Warth and others 2012; Han and others 2013; Bittner and others 2015). Moreover, it can provide a platform to quantify multiple mycotoxins in a single analysis (Al-Taher and others 2013). This is particularly desirable because mycotoxins may occur in various combinations produced by a single or several fungal species in food matrices (Njumbe Ediage and others 2011). ELISA has found widespread use for rapid screening and monitoring of mycotoxin contamination in raw material and final products. Commercial ELISA kits are available for the most of the

major mycotoxins in various food matrices. Without clean-up or concentration steps, ELISA has emerged as a rapid test with its high sensitivity, low cost, and ease of application that may also be used in field conditions (Zheng and others 2005). There are 2 type of ELISA analysis: direct and indirect. The benefits of direct ELISA are very quick analysis times and reduced cross-reactivity by using only one antibody. Although the difficulty of labeling and signal amplification on the primary antibody in the direct ELISA analysis brings less sensitivity, indirect ELISA uses the labeled secondary antibody to be more sensitive by enhancing signal amplification. The ELISA is based on the reaction between antigens and antibodies, but antibodies often show cross-reactivity with compounds similar to the mycotoxins (Lee and others 2014). The presence of components structurally related to mycotoxins can interfere with antigen–antibody binding and may lead to erroneous measurements. In addition, “matrix effect” or “matrix interference” may cause under- or overestimation of mycotoxins concentrations. As food matrixes contain various naturally occurring antigens that are a mixture of macromolecules with several epitopes, HPLC analysis is often employed for confirmation of mycotoxins. Hence, it is important to apply validated ELISA methods for targeted mycotoxin in a specific commodity.

Emerging techniques Research and development of rapid screening methods for the detection and quantification of mycotoxins have increased considerably during the past decades. Together with the elevated interests and wide-spread use of conventional ELISA methods, efforts have been directed to overcome inherent limitation of the technology, that is storage conditions and shelf-life of the kits, limited use of organic solvents that may affect the recovery, cross-reactivity, etc. Such effort resulted in development of different platforms based on the familiar antigen–antibody interaction, especially direct ELISA analysis, such as the flow-through assay and the lateral flow assay (Paepens and others 2004). These assays utilize antibody coated membranes to allow reaction with the target antigens following application of liquid sample extracts. Hence, it may provide a platform to detect multiple mycotoxins simultaneously by coating the membranes with different antibodies for the target molecules including AFB1 , OTA, FB1 , DON, ZEN, and T-2 toxin (Schneider and others 2004). These immunochemical techniques do not require any clean-up or concentration steps, whereupon the analysis time is quietly short. Extended shelf-life and ease of use may be considered as major advantages of this platform while it may only provide qualitative or semiquantitative results. Electrochemical immunoassay is based on the combination of antigen–antibody interaction and an electrode that can convert the binding reaction to an electrochemical signal. Catalytically inactive materials, such as gold and some carbon materials can be used as such converters. The advantages of electrochemical immunoassay may include low cost, feasibility for miniaturization and multiplexing, and ease of application. However, repeated use may cause decreased reproducibility over time due to the reduced binding affinity of the antibodies immobilized onto the surface of the electrode as they need to be regenerated by a solvent mixture after each use (Urusov and others 2010). Fluorescence polarization immunoassay (FPIA) shares the same principle as competitive ELISA—that is the mycotoxin or antigen in a sample competes with fluorescein-labeled antigen for binding with the antibody. As small molecules (free mycotoxins in sample) rotate more rapidly and give lower polarization than larger molecules (fluoresceinlabeled mycotoxin) at a given temperature, the polarization value Vol. 80, Nr. 12, 2015 r Journal of Food Science T2975

T: Toxicology & Chemical Food Safety

of batches with a smaller particle size. The total sample size ranges from 4.5 kg (10 lb) to 34 kg (75 lb), depending on the expected heterogeneity of the sample. In general, processed products show less heterogeneous distribution than raw agricultural commodities; therefore, processed products require simpler sampling provisions.

Advances in Mycotoxin Research . . . is inversely proportional to mycotoxin concentration (Maragos and others 2002). FPIA has 2 differences from ELISA: (i) no enzymatic reaction for detection and (ii) no separation between bound and free compounds, result in quick analysis time. FPIA may provide easy-to-use and high throughput platform, although background interference needs to be addressed with care. More recent advances involve development of novel mycotoxin binding molecules and recombinant antibodies, for example polymers, aptamers, and nanobodies, to replace traditional antibodies produced in animals for improved stability and higher expression yield. Nanobody in particular is known to be more stable over high temperatures and array of organic solvents as it is constructed based on the variable domain of heavy chain antibodies. Because this engineered antibody can be constructed in smaller molecular weight than traditional antibodies chemically labeled with enzymes, it can also provide greater solubility and ease of modification. Several researchers have already developed and validated nanobodies for major mycotoxins (Fan and others 2013; He and others 2014). It may be noted that Liu and others (2015) developed a rapid and sensitive direct competitive fluorescence enzyme immunoassay for OTA based on a nanobody-alkaline phosphatase fusion protein with a detection limit of 0.04 ng/mL and a quantification range of 0.06 to 0.43 ng/mL. In addition, a number of “sensors” have been developed for rapid and low-cost determination of mycotoxins in various food matrices. These sensor technologies are based on the bio- and/or chemical reactions of the mycotoxin or other metabolites produced by fungi. For instance, electronic noses adsorb volatile organic compounds of low molecular weight, which are released by many fungi as products of secondary metabolism, and measure the concentration of these compounds with a variety of transduction systems based on electrical-, optical-, or mass-transduction, such as with metal oxide sensors or surface acoustic wave sensors (Olsson and others 2002; Logrieco and others 2005). Similarly, electronic tongue instruments have also been developed for liquid samples (S¨oderstr¨om and others 2003). Optical analysis methods, such as Fourier Transform mid-infrared spectroscopy (Kos and others 2003) and near-infrared transmittance spectroscopy (Pet˚ tersson and Aberg 2003), may also provide fast and nondestructive detection of mycotoxins. The major restrictions for most of these analyses are the high matrix dependence and the lack or limitation of appropriate calibration materials. These techniques respond to a broad range of compounds the sensor array generates patterns of responses that can be distinguished for different samples. It should be noted that biomonitoring is important in estimating human exposure to mycotoxins and for quantitative risk assessments. For example, the dried blood spot (DBS) technique in which the blood samples are blotted and dried on filter paper on site, can easily be stored for long time (over years) and analyzed in the lab by HPLC or LC/MS (T¨or¨ok and others 2002). Recently, Cramer and others (2015) detected OTA and its degradation product 2’R-OTA in blood from coffee drinkers using the DBS method.

T: Toxicology & Chemical Food Safety

Bioassays A variety of methods mentioned above can measure the amount of mycotoxin in food and identify the mycotoxin itself or its metabolites in clinical samples. However, these methods do not provide insight on the toxicity of mycotoxins and the mechanisms by which the act in vivo. Therefore, it is essential from a public health perspective to develop suitable bioassay methods to assess T2976 Journal of Food Science r Vol. 80, Nr. 12, 2015

biological activities or toxicity of mycotoxins. Toxicity studies in humans are not ethical due to the need to expose subjects to potentially highly toxic fungal toxins. Hence, bioassays frequently involve the use of cell lines from vertebrates (in vitro) or various live experimental animals (in vivo) as shown in Table 5. Several bioassay methods have been based on yeast and bacteria strains as well (Madhyastha and others 1994; Engler and others 1999). In considering variations of any biological system, either in vitro or in vivo, reproducibility must be considered in selecting a model in addition to the sensitivity of the cell-line or whole animal to the targeted mycotoxin. In vitro cell systems have their strength as a screening tool for toxicity and as a way to observe mechanism pathways. In general, a pathway of toxicity is a cellular mechanism that is expected to result in an adverse effect at a cellular level and which may lead to an adverse health effect for the organism. Methods to evaluate toxicity or mechanism in the body, utilizing derived or primary cell lines from animals, have been well established. Although continuous cell lines lost their original functions during immortalization or transformation, primary cell lines does not have this issue. As a result, primary cell lines often better able to demonstrate specific organ related effects which might be important to animals but they are more difficult to maintain. In addition, application of more than one cell lines of different origin for screening purposes has been recommended as different cell lines have been observed to have diverse sensitivities (Kamp and others 2005; McKean and others 2006). Cytotoxicity tests using continuous or primary cell lines have advantages: more convenient, cheap, rapid, and easily repeatable. However, the results or end points from the cell-based bioassays are difficult to extrapolate to the whole animals, particularly to humans. This is partly due to the fact that some of continuous cell lines have lost part of their original biochemical functions found in the tissue from which they were derived. Chick embryo test may provide a useful tool in screening toxicity of mycotoxins. This method uses fertilized eggs at specific stage of embryo developments. It has been found that 10-day-old chick embryos are most sensitive to toxic substances. A small window is cut in the shell directly above and sample solution is injected directly below the caudal region of the embryo. Then the procedure is carried out under a dissecting microscope following desired incubation period. In determination of cytotoxicity, cell cultureMTT [3-(4,5-dmethyl-2-thiazolyl)-2,5-diphenyltetrazolium bromide] test has been commonly used. Cytotoxicity can be assessed by treating the cells with MTT which then reduced to formazan (a purple product) by the mitochondrial enzyme succinate dehydrogenase in living cells. Cell lines from different animals or human can be used in this colorimetric test. Elucidating metabolic pathways and toxicity in animal studies is helpful to understand mycotoxin metabolism and toxicity. Similar to other xenobiotics, methods to study toxicity in experimental animals have been standardized and are highly effective in estimating toxicity that may occur in humans. According to the type of toxicity, different species of animals may be needed because each species may respond differently to the same mycotoxin. Among these animal models, rodents are the most commonly used for testing toxicity due to their availability, low cost, and relatively short lifespan. Nonetheless, the data obtained from animal study may not be directly applicable to humans although experimental animals may provide more relevant data than those from in vitro systems. Moreover, it is difficult to set the end points of animal studies than can extrapolate or relevant to humans.

Advances in Mycotoxin Research . . . Table 5–Summary of typical bioassay models associated with major mycotoxins. Mycotoxins

End point/target organ(s)

Aflatoxins

Carcinogenicity

Liver

Ochratoxins

Carcinogenicity, Cytotoxicity, Oxidative stress

Liver

Kidney

Fumonisins

Carcinogenicity

Liver Kidney

Deoxynivalenol

Immunosuppressive properties

Zearalenone

Oestrogen-responsive proliferation Carcinogenicity

Animal model Male F344 rats

LLC-PK1 (pig kidney); HK-2 (human kidney)

Male F344 rats; Male SD rats; Male Wistar rats

–

Male F344 rats; Male SD rats; Male Wistar rats

Female B6C3F1 mice

LLC-PK1 (pig kidney); BHK-21 (baby hamster kidney) –

Male F344 rats

Female B6C3F1 mice

MCF-7 (human breast cancer) –

Pig Female B6C3F1 mice; Female F344 rats

References IARC (1976); NTP (1980); Reddy and others (2006) Abdel-Wahhab and others (2005, 2008); Gautier and others (2001); Meki and Hussein ¨ ¸ elik and others (2001); Ozc (2004); Soy¨oz and others (2004) NTP (1989); Gautier and others (2001); Meki and Hussein (2001); Petrik and others ¨ ¸ elik et al. (2004); (2003); Ozc Soy¨oz and others (2004); Abdel-Wahhab and others (2005); Costa and others (2007); Abdel-Wahhab and others (2008) Howard and others (2001); Voss and others (2002) Abeywickrama and others (1998); He and others (2001); Howard and others (2001), Norred and others (1993); Riley and others (2001); Voss and others (2002) Forsell and others (1987); Pestka and others (1987) Welshons and others (1990); JECFA (2000) NTP (1982)

There is nothing of typical bioassay model.

Biomarkers As mycotoxins are occurring naturally and frequently contaminate foods, adverse effects may be observed after ingestion of the contaminated foods. Because any biological assay may not provide direct evidence of toxicity at an individual level, in a strict sense, adverse effects on human health can only be estimated by observing responses in the human body followed by exposure to the mycotoxin(s). Such biomonitoring may be possible with suitable biomarkers. Because biomarkers are used to quantify exposure to a toxin and measure the extent of any toxicity, they must be related to the biochemical mechanism pathway, react at realistic doses, be specific and sensitive, quantitative and easily measurable (Timbrell 1998). Biomarkers of exposure to mycotoxins can be the substance itself, their metabolite(s), or compounds obtained by reaction with biological molecules. Hence, the identification of suitable biomarker(s) is important and essential step for quantitative risk analysis. In this section, biomarkers available to date are described briefly in the context of biochemical mechanisms of mycotoxins of concern.

Aflatoxin B1 (AFB1 ) Absorbed AFB1 in the gastrointestinal track is transported to the liver and metabolized by cytochrome P450 enzymes to AFB1 -8, 9-epoxide. This important metabolite reacts with DNA, specifically guanine residues, to generate the adduct, AFB1 -N7 -Gua (Groopman and others 1993; Walton and others 2001), or with serum albumin to generate lysine adducts (Sabbioni and others 1990). Approximately 0.2% of dietary AFB1 was found to be excreted as AFB1 -N7 -Gua form in human urine, whereas 1% to 2% of dietary AFB1 existed as covalently bonded to serum albumin. In addition, AFM1 may also be a useful biomarker as 1% to 2% of dietary AFB1 was observed to be excreted as AFM1 form in

human urine. Zhu and others (1987) reported a strong correlation between dietary intake of AFB1 and urinary excretion of AFM1 , a metabolite of AFB1 , in 252 urine samples from residents of China. Gan and others (1988) reported good correlations between dietary intake of AFB1 and level of AFB1 -lysine adduct in serum, and between AFB1 -lysine adduct level in serum and AFM1 excretion in urine. Groopman and others (1992b) also observed a correlation between dietary exposure to AFB1 and amount of AFM1 or of AFB1 -N7 -Gua adduct excreted in the urine. Another study in Gambia analyzed 20 urine samples of people who had hepatitis B virus (HBV) showed the correlation between the levels of dietary exposure of AFB1 and AFB1 -N7 -Gua adduct in urine (Groopman and others 1992a). Although the presence of the AFB1 -lysine adduct in serum is considered an accurate biomarker of AFB1 dietary exposure over a period of months based on the long half-life of albumin in serum and higher yield than urinary biomarkers of AFB1 (Sabbioni and others 1987; Leong and others 2012), urinary biomarkers of AFB1 better reflect short-term dietary exposure to AFB1 (around 1 day before) because of their short half-life (Qian and others 1994).

Ochratoxin A OTA is mainly absorbed in the gastrointestinal tract and binds with serum albumin in the blood (Kumagai and Aibara 1982; Roth and others 1988). As the binding between OTA and serum albumin is rather strong and stable, the half-life (t1/2 ) of OTA has been found to be about 800 h (35 d) in humans (Schlatter and others 1996). Although its metabolic pathways are variable depending on the route and dose of administration, kidney is known to be the major target organ as toxicokinetics of OTA is mainly influenced by enterohepatic circulation and reabsorption by the kidney (Gekle and Silbernagl 1994; Ringot and others Vol. 80, Nr. 12, 2015 r Journal of Food Science T2977

T: Toxicology & Chemical Food Safety

a

Liver

Cell culture (origin) HepG2 (human hepatoma) –a

Advances in Mycotoxin Research . . . 2006). Urine and fecal excretion are major excretory pathways for plasma clearance of the OTA. Profiles for OTA biomarkers in humans have been obtained using blood samples. However, the amount of OTA contributed by foods OTA was not positively correlated with the OTA concentration in the blood (Duarte and others 2011). In a U.K. study, no significant correlation between the OTA concentration in plasma and the amount of OTA exposure was observed (Gilbert and others 2001). Meanwhile, urinary biomarkers are capturing more attention because they are noninvasive and simple procedures used to obtain them. The level of OTA excretion in urine is relatively low comparison with the level of OTA in blood as OTA is reabsorbed in kidney (Dahlmann and others 1998). According to Pfohl-Leszkowicz and Manderville (2007), OTA was found in urine several days after OTA consumption. Daytime OTA excretion is now a valuable tool in humans (Duarte and others 2011). Several researches also reported that the presence of OTA metabolites such as OTB in human urine (Jonsyn-Ellis 2001). Still, the correlation of the OTA level in urine with the OTA intake remains a complex issue.

Fumonisin B1 (FB1 ) FB1 is poorly absorbed by the intestine with low bioavailability, although the amounts absorbed are eliminated rapidly from circulation with estimated half-life of 18 to 26 min for rats (Marasas and others 2000). Most of eliminated FB1 (35% to 84% of dose) was recovered from feces within 96 h (mostly first 24 h) via bile (Norred and others 1993; Chamberlain and others 1996). Absorbed FB1 does not undergo major metabolic changes whereas most mycotoxins are metabolized in the liver, kidney, or other organs (Norred and others 1993; Prelusky and others 1996; Dilkin and others 2010). Consequently, identification of biomarkers for FB1 exposure has not followed the metabolite profile approach used for other mycotoxins but relied on the measurement of FB1 concentration in bio-fluids as the parent form was found in urine

in humans (Gong and others 2008). Xu and others (2010) also reported that 1% to 2% of ingested FB1 was found in urine collected from human. Nevertheless, the correlation between FB1 concentration in urine to dietary FB1 exposure in humans still needs to be investigated. It is well known that FB1 inhibits sphingolipid biosynthesis leading to elevation of sphinganine (Sa) levels as well as the ratio of Sa to sphingosine (So) (Merrill and others 1993; Wang and others 1991; Riley and others 2001). Hence, the Sa:So ratio in blood and urine has been suggested as a useful biomarker for dietary exposure to fumonisins (Qiu and Liu 2001; Cai and others 2007). However, the Sa:So ratio in plasma and urine failed to correlate with dietary exposure of FB1 in humans (Van der Westhuizen and others 2010). Despite this, considering the number of studies in literature promoting these parameters as an effective tool in evaluating toxicity of fumonisins, they may not be excluded as biomarkers without further study.

Deoxynivalenol (DON) DON absorption and clearance in pigs and rodents were rapid, reaching peak plasma concentration within 15 to 30 min after DON administration (Prelusky and others 1988; Pestka and others 2008). The half-life for DON is relatively short in most species, and DON and major metabolites are excreted within 24 to 72 h (Coppock and others 1985; Pestka and Smolinski 2005; Goyarts and D¨anicke 2006). In the study by Meky and others (2003), a total of 37% of the administrated DON was detected in rat urine within 72 h. Absorbed DON can be rapidly metabolized to glucuronide conjugated form (DON-GlcA) in the liver and detected in urine and feces of humans and animals (Gareis and others 1987; Meky and others 2003) whereas it was not accumulated in tissues (Pestka 2010). De-epoxy DON (DON-1) is the main metabolite of DON in most animals (Yoshizawa and others 1986) and is produced by the intestinal microflora of animal (Swanson and Figure 1–Chemical structure of mycotoxins. (A) aflatoxin B1 (AFB1 ), (B) ochratoxin A (OTA), (C) fumonisin B1 (FB1 ), (D) deoxynivalenol (DON), and (E) zearalenone (ZEN).

T: Toxicology & Chemical Food Safety

T2978 Journal of Food Science r Vol. 80, Nr. 12, 2015

Advances in Mycotoxin Research . . .

Zearalenone (ZEN) ZEN is rapidly absorbed from the alimentary canal (Megalla and others 1987), and metabolized in the liver. There are 2 major metabolic pathways for ZEN (Olsen and others 1981; Malekinejad and others 2006). The first pathway is hydroxylation to αzearalenol (α-ZOL) and β-zearalenol (β-ZOL), which is assumed to be catalyzed by 3α- and 3β-hydroxysteroid dehydrogenases. The second pathway is conjugation with glucuronic acid by uridine diphosphate glucuronyl transferase. α-ZOL has the highest affinity to estrogen receptors, followed by ZEN and β-ZOL (Shier and others 2001). ZEN and its metabolites are excreted mainly in the bile and urine within 72 h (Lasztity and others 1989). Biehl and others (1993) also reported that 45% and 22% of the administrated dose of ZEN were excreted in urine and feces of pigs, respectively, within 48 h after administration. Several animal and human studies reported that most of administrated ZEN detected as conjugated metabolites in urine (Mirocha and others 1981; Biehl and others 1993). Although there is no proven biomarker for ZEN, conjugated forms of ZEN metabolites are considered potential biomarkers that could be used in risk assessment.

Risk Assessment During January to June 2004, 317 cases of acute hepatic failure reported and 125 cases of subsequently died during the illness, following a major outbreak of aflatoxicosis in eastern Kenya (CDC 2004; Azziz-Baumgartner and others 2005). These references were linked to aflatoxin poisoning from contaminated corn. Like this, consumption of foods contaminated with mycotoxins leads to not only chronic mycotoxicoses but also acute poisoning that may

result in death. The legislation through risk assessment is in place to control or mycotoxins in food. However, the emergence of mycotoxins and insufficient toxicological data pose challenges in conducting risk assessment and establishing regulations. Risk assessment is a systematic process to characterize the potential of adverse effects resulting from exposure to hazardous agents. It should be noted that the risk assessment is intimately involved in establishing regulatory guidelines and in turn protecting public health. Human health risk assessment is based on a 4-step process: hazard identification, hazard characterization (as known dose-response assessment), exposure assessment, and risk characterization as suggested by the National Research Council (NRC 2009). Hazard identification examines if, and the conditions by which, a certain mycotoxin has the potential to cause a particular adverse health effect or disease. Hazard characterization describes the numerical relationship between level of dietary exposure (dose) and associated adverse effect (response). An exposure assessment involves estimating the frequency, intensity, and duration of ingestion with a mycotoxin. Finally, for risk characterization, the results of the exposure assessment are compared with hazard characterization to indicate the degree of concern. The most important consideration during this process is to address uncertainties particularly in the exposure assessment and hazard characterization (Kuiper-Goodman and others 1994). The calculated level of intake is often expressed in mg of mycotoxin/kg of body weight/day. Additional terms including minimal risk level (MRL), reference dose (RfD), acceptable daily intake (ADI), and PTWI are also commonly used in risk assessment. Human health risk assessment requires data on toxicity and exposure from animals or human study. In the traditional approach, the acceptable daily intake (ADI) for human is derived from the no-observed-adverse-effect level (NOAEL) determined by animal studies. Therefore, a safety factor (SF) should be considered in applying the results from experimental animals (NOAEL) to humans (ADI), that is ADI = NOAEL/SF. SF are numerical values applied to the NOAEL or other effect levels to account for any uncertainty in the data. In general, SF of a 100 is commonly used to accommodate uncertainties between species (×10) and variabilities within the species (×10). The SF may also vary: ranging from 1 for well-characterized chemicals to 1000 for others with incomplete data or for at-risk population. For OTA in example, JECFA (1999) first established PTWI of 112 ng/kg of body weight based on a dose–response study of renal function deterioration in

Figure 2–Illustration of the calculation of the NOAEL/LOAEL (no-observed-adverse-effect level/low-observed-adverse-effect) and BMD (Benchmark dose).

Vol. 80, Nr. 12, 2015 r Journal of Food Science T2979

T: Toxicology & Chemical Food Safety

others 1988; He and others 1992). According to Sundstøl Eriksen and Pettersson (2003), however, DON-1 was undetectable in the human fecal incubation mixture. It indicated that the discrepancy in biotransformation ability of intestinal microflora among species might have a toxicological significance for DON. Hence, DON in urine rather than in blood is considered a more suitable biomarker to assess dietary exposure of DON, as higher concentrations of DON were found to be excreted in urine. In addition to DON, DON-GlcA and DON + DON-GlcA are considered alternative biomarkers in humans. DON-GlcA was identified as a major nontoxic metabolite of DON. However, these biomarkers are yet to be proven with additional studies (Turner and others 2008; Warth and others 2012).

Advances in Mycotoxin Research . . . pigs (Elling 1979; Krogh 1974). According to this, LOAEL was 8 μg/kg of body weight/day. JECFA applied 500 of a combined SF in the calculation. Although this traditional approach serves the purpose, significant pitfalls have been identified, that is (i) the ADI depends entirely on the NOAEL, (ii) the NOAEL is subjected to the species and number of animals, and (iii) the difficulty in selecting the adverse effect with more than one endpoint. Hence, although it may seem similar to ADI, the concept of reference dose (RfD) and SF have been introduced (EPA, 1993). A modifying factor (MF) may also be considered, for example decreased uncertainty (