Development of Methods for Determining Aflatoxins in

Biological Material

•<pi

LO

Arbetslivsinstitutet

Anders Kussak _________________________________ 1995

Umeå University, Department o f Analytical Chemistry and National Institute o f Occupational Health, Analytical Chemistry Division, Umeå, Sweden

Development of Methods for Determining Aflatoxins in

Biological Material

by

Anders Kussak

Akademisk avhandling

som med tillstånd av rektorsämbetet vid Umeå Universitet för erhållande av Filosofie Doktorsexamen vid Matematisk-Naturvetenskapliga fakulteten, ffamlägges till offentlig granskning vid Kemiska Institutionen, Hörsal C, Naturvetarhuset, fredagen den 29 september 1995, kl. 10.00.

Fakultetsopponent: Dr. Rainer Stephany, Bilthoven, The Netherlands.

Title: Development of Methods for Determining Aflatoxins in Biological Material.

Author:

Abstract:

Key Words:

Anders Kussak, Umeå University, Department of Analytical Chemistry, S-901 87 Umeå and National Institute of Occupational Health, Analytical Chemistry Division, P.O. Box 7654, S-907 13 Umeå, Sweden.

In this thesis, it is shown how aflatoxins can be determined in biological material. The thesis is a summary of five papers.

Aflatoxins are carcinogenic mycotoxins produced by Aspergillus moulds. Methods were developed for the determination of aflatoxins in samples of airborne dust and human urine collected at feed factories. For the dust samples from such agricultural products as copra, cotton seed and maize, methods were developed for the determination of aflatoxins Bi, B2, Gi and G2. For urine samples, methods were developed for analysing the four aflatoxins above that naturally occur in dust, and the metabolites aflatoxins Mi and Qi.

Sample preparation of dust samples included solvent extraction, filtration and immunoaffinity column extraction. Urine samples were cleaned up using immunoaffinity column extraction or solid-phase extraction using ethyl bonded-phase columns. All extractions with these columns were automated by means of a laboratory robot.

Reversed-phase liquid chromatography was used to separate the aflatoxins in the cleaned-up extracts. Detection was performed by fluorescence after post-column derivatization by addition of bromine. Parameters for the derivatization were studied using factorial designs. To confirm the identity of aflatoxins in naturally contaminated airborne dust samples and spiked urine, liquid chromatography was combined with electrospray mass spectrometry.

The detection limits of the aflatoxins in dust samples were in the range 1.8-3.1 ng/g in 10-mg dust samples using fluorescence detection. Aflatoxins were determined in spiked urine down to the 6.8-18 pg/ml level. In naturally contaminated dust of copra and cotton seed, aflatoxins were detected with a content of 9-50 pg/mg of aflatoxin Bi. No aflatoxins could be detected in any urine sample obtained from feed factory workers that were less than 6.8 pg/ml of aflatoxins B|, B2, Gi and G2 and less than 18 pg/ml of aflatoxins Mi and Qi.

Aflatoxins, Mycotoxins, Carcinogen, Airborne dust, Urine, Feed factories, Immunoaffinity extraction, Solid-phase extraction, Post-column derivatization, Electrospray, Mass spectrometry.

ISBN: 91-7191-071-9 Arbetslivsinstitutets tryckeri, Umeå, 1995

Contents

page

1 Papers discussed................................................ 5

2 Introduction..............................................................................72.1 Background.......................................................................72.2 Chemistry of aflatoxins..................................................... 82.3 Effects of aflatoxins.......................................................... 92.4 Occurrence of aflatoxins................................................. 122 .5 Analysis of aflatoxins.......................................................13

3 Aim of the work......................................................................15

4 Sample preparation................................................................174 .1 Solvent extraction............................................................174.2 Solid-phase extraction and clean-up................................184.3 Immunoaffinity extraction and clean-up.......................... 204.4 Automation......................................................................244.5 Stability ............................................................ 254.6 Safety precautions........................................................... 25

5 Analysis...................................................................................275.1 Liquid chromatography...................................................275.2 Fluorescence detection....................................................285.3 Mass spectrometry.......................................................... 31

6 Aflatoxins determined...........................................................356.1 Dust samples.................................................................. 356.2 Biological samples.......................................................... 38

7 Conclusions............................................................................ 41

Acknowledgements........................................................................... 43

References.........................................................................................45

3

1 Papers discussed

This thesis is based on the papers listed below, which will be referred to in the text by their Roman numerals.

I Automated sample clean-up with solid-phase extraction for the determination of aflatoxins in urine by liquid chromatographyKussak A, Andersson B and Andersson K J. Chromatogr., 616 (1993) 235.

II Determination of aflatoxin Qi in urine by automated immunoaffinity column clean-up and liquid chromatographyKussak A, Andersson B and Andersson K J. Chromatogr. B, 656 (1994) 329.

III Determination of aflatoxins in airborne dust from feed factories by automated immunoaflinity column clean-up and liquid chromatographyKussak A, Andersson B and Andersson K J. Chromatogr. A, in press.

IV Immunoaflinity column clean-up for the high-performance liquid chromatographic determination of aflatoxins Bt, B2, Gi, G2, Mi and Qi in urineKussak A, Andersson B and Andersson K J. Chromatogr. B, in press.

V Determination of aflatoxins in dust and urine by liquid chromatography/electrospray ionization tandem mass spectrometryKussak A, Nilsson C-A, Andersson B and Langridge J Rapid Commun. Mass Spectrom., submitted for publication.

2 Introduction

2.1 Background

Aflatoxins are metabolites produced by species of the moulds Aspergillus fla w s and Aspergillus parasiticus. They were discovered in the early 1960s after research following outbreaks called “turkey X” disease, when about 100,000 turkeys died in England. Other animals, such as ducklings, pigs and cattle, were also afflicted, with closely related diseases. The source considered responsible for the disease was groundnut meal infected by mould, from which a toxin could be isolated. The toxin was identified as being derived from the mould Aspergillus fla w s [1, 2], When the structure was determined it was discovered that there were four different aflatoxins produced by the mould. The toxins were named aflatoxins Bi, B2, Gi and G2 from the first letters in Aspergillus fla w s toxin, with the letters B and G from fluorescence in blue or green under ultraviolet light and the number from the retention order on thin-layer chromatographic plates. The historical background of aflatoxins has been described in several textbooks [3, 4],

Since aflatoxins were discovered in connection with outbreaks of disease among animals, several cases of acute poisoning among humans have been observed. One outbreak started in 1974 in several villages in northwestern India, where heavily moulded maize consumed was found o contain levels of up to 15 mg/kg of aflatoxins [4], Besides acute poisoning, strong evidence shows that aflatoxins are one of the factors causing liver cancer in large human populations [5],

7

2.2 Chemistry of aflatoxins

The four major mould-produced aflatoxins are aflatoxins Bi, B2, Gi and G2 [4], These four aflatoxins are a group of difuranocoumarins with similar structures (Figure 1). Aflatoxins B2 and G2 are the dihydro derivatives of aflatoxins Bi and Gi respectively. The double bond of the vinyl ether function of aflatoxins Bi and Gi is important for the cancerogenic effects of the aflatoxins as discussed below, and can be used for derivatization in several ways to enhance fluorescence for detection.

Bi

G i

B'

Figure 1. Structures of aflatoxins Bi, B2, Gi and G2.

8

2.3 Effects of aflatoxins

Aflatoxins are toxic and carcinogenic compounds and have shown mutagenic action in in vitro test systems [5]. Aflatoxin Bi, the most-studied aflatoxin, has been found to be transformed into several metabolites by biotransformation pathways, as reviewed by Eaton el al. [6], The main metabolites that have been found in different species are shown in Figure 2, though not all metabolites have been found in humans.

pH

CH

Aflatoxin

CH

Bi" epoxide

CH

Aflatoxicol

Figure 2. Biotransformations of aflatoxin Bi.

9

By the action of cytochrome P-450 enzymes, the hydroxylated aflatoxins Mi, Pi and Qi are formed from aflatoxin Bi. These are regarded as detoxification products as they are less carcinogenic than aflatoxin Bi. The hydroxylated aflatoxins readily form glucuronide and sulphate conjugates. These conjugates are water-soluble and are excreted in urine or bile together with unconjugated aflatoxins as aflatoxin Mi [7]. Other metabolites transformed from aflatoxin Bi are aflatoxicol by a cytosolic reductase, aflatoxin B^, which may be formed by acids, and aflatoxin Bi- 8,9-epoxide formed by P-450 enzymes.

Aflatoxin Bi derives its carcinogenic effect through conversion to aflatoxin B1-8,9-epoxide by cytochrome P-450 enzymes, and can then react with DNA at the N-7 position of guanine, or derives a toxic effect by forming protein adducts. By reaction with glutathione (GSH) the aflatoxin can be excreted from the body, and is therefore considered to be a detoxification pathway for aflatoxin B1-8,9-epoxide. By hydrolysis of aflatoxin B1-8 ,9-epoxide, aflatoxin-8,9-dihydrodiol is formed, which can form protein adducts by reaction with lysine. These pathways are shown in Figure 3.

Bj -d ihydrodio l

/ / \Protein DNA GSH

adducts adducts conjugate

Figure 3. Metabolic pathways for aflatoxin B1-8,9-epoxide.

10

The carcinogenic and toxic effects of aflatoxin Bi are considered to depend on the competition between these different metabolic pathways, as some pathways activate the aflatoxin to a reactive epoxide and others deactivate the aflatoxin. Aflatoxin Bj is the most toxic of the four, and most research has focused on this aflatoxin. But all four aflatoxins have proved to be acutely toxic, measured as LD$o values, in different species. In ducklings the LD50 values for aflatoxins Bi, B2, Gi and G2 were 0.36, 0.78, 1.70 and 3.44 mg/kg, respectively [8]. As aflatoxin Bi (Figure 2) can be activated to an epoxide and form DNA adducts, it is also presumed to be carcinogenic. Other aflatoxins, such as aflatoxin B2, which is the dihydro derivative of aflatoxin Bi, give a DNA adduct that is identical to that from aflatoxin Bi. One hypothesis is thus that aflatoxin B2 can be converted to aflatoxin Bi [6],

It is considered that the main route of exposure to aflatoxins is ingestion of aflatoxin-contaminated food. Attention has mainly been focused on the causation of liver cancer by aflatoxins, since aflatoxin Bi can be metabolized and activated in this organ. In addition to exposure through diet, other routes of exposure, such as inhalation of aflatoxin-laden dust, may carry a risk of causing other cancers than liver cancer, such as respiratory cancer. In epidemiological studies an increase in incidence of liver cancer has been observed among feed-factory workers in Denmark [9] and grain mill workers in Sweden [10] and respiratory cancer among Dutch oil-press workers [11]. These groups of workers may be exposed to dust containing aflatoxins. Experiments carried out to substantiate the suspicion that aflatoxins may be responsible for respiratory cancer showed that aflatoxin Bi can be activated in extrahepatic tissues, for example, in cells from the upper airways in animals [12, 13], As aflatoxin Bi can be metabolized in the respiratory system, there is a risk of respiratory cancer.

11

2.4 Occurrence of aflatoxins

Aflatoxins are produced by the food-borne moulds Aspergillus flavus and Aspergillus parasiticus. Aflatoxins are secondary metabolites produced by these species. The levels of aflatoxins produced vary between different strains of the moulds and within the mould, that is between the conidia and the sclerotia. Some strains of Aspergillus flavus have been found to produce levels as high as 97 400 ng/g of aflatoxin Bi in conidia while in other strains only low levels or no aflatoxins have been detected [14], Accordingly, the occurrence of Aspergillus moulds does not automatically indicate the presence of aflatoxins.

Several environmental factors, such as temperature, water activity (aw), light intensity, pH and the atmospheric gases, influence aflatoxin production by the moulds [15]. Of these factors, temperature and water activity have been found to be the most important for the production of aflatoxins. The optimum temperature for aflatoxin production is between 25°C and 30°C, but aflatoxins can be produced at temperatures as low as 15°C. The optimum water activity, that is the ratio of water vapour pressure of the substrate to the vapour pressure of pure water at the same temperature and pressure, is 0.95 to 0.99. Aflatoxin production decreases at aw values below 0.85, even if fungal growth can continue without aflatoxin production down to an aw value of 0.78. Biological factors, such as the type of substrate, antifungal agents and the type of species of the mould, are also important for aflatoxin production. Crop damage, either mechanically or by insects, has proved to be another important factor in the development of mould infections. Once a product has been contaminated with aflatoxins, they do not disappear even if mould growth has been interrupted.

Although aflatoxins can be produced in a variety of materials, they mostly occur in peanuts, cotton seed and maize [16]. Accordingly, most of the studies of aflatoxin production have been concentrated on the these products. It is known that aflatoxin production can start at the preharvest stage and then continue during harvesting, drying, storage and processing. Other products where aflatoxins can be be detected are copra [17] and spices [18]. Eggs and milk can be indirectly contaminated with aflatoxins by animals fed on feed containing aflatoxins. Aflatoxins have been mostly detected in products grown in tropical or semitropical areas, but have occurred in feed, at levels of over 400 ng/g, stored by acid treatment in Sweden [19]. Cows fed the aflatoxin-contaminated feed gave milk containing aflatoxin Mi at levels of over 50 pg/g.

12

2.5 Analysis of aflatoxins

Aflatoxins are carcinogenic substances, frequently occurring in foods and feeds. They are therefore routinely analysed in a wide range of agricultural products, and methods have been developed to determine aflatoxins in the ng/g level in these products. Most countries have regulatory limits of between 10 and 20 ng/g of total amounts of aflatoxins in foodstuffs, and lower levels of 0.05 to 0.5 ng/g of aflatoxin Mi in dairy products [20],

A number of techniques have been used for sample preparation and determination. Standard methods, such as those of the Association of Official Analytical Chemists (AOAC) [21 ], developed by the Contamination Bureau (CB) and Best Food (BF), are widely used. In short, the CB method uses liquid-liquid extraction with chloroform and water, column chromatography with silica gel for clean-up and thin-layer chromatography for determination of the aflatoxins. The BF method uses methanol-water extraction, chloroform-water extraction and thin-layer chromatography.

Since the introduction of these AOAC methods, several other methods have been developed, especially different solid-phase extractions in combination with solvent extraction. Liquid chromatography now tends to be used more than thin-layer chromatography [22], Such methods as enzyme-linked immunoassay, that is, a method using antibodies [22], and fluorescence tests using UV light [15], are employed as screening methods for the determination of total amounts of aflatoxins.

13

3 Aim of the work

In epidemiological studies an increased rate of liver cancer has been observed among grain mill workers in Sweden [10], The reason for the increased liver cancer rate may be exposure to aflatoxins from dust of contaminated agricultural commodities.

The level of exposure to aflatoxins can be confirmed by determination of the amount of aflatoxins in airborne dust or by the amount of metabolites excreted in biological material, such as urine. When urine is analysed for aflatoxins, the total level of exposure is determined, that is, the sum of the aflatoxins in the airborne dust aspirated and in the contaminated diet consumed. Methods are available for the determination of aflatoxins in agricultural products, but some additional problems are encountered in the analysis of airborne dust or urine. In the case of dust samples, the main difficulty is that aflatoxins may be present in only trace amounts. Urine samples contain a number of aflatoxin metabolites that need to be analysed. These are formed by biotransformations from the naturally occurring aflatoxins Bj, B2, Gi and G2. To determine the occupational exposure to aflatoxins, other methods for determination than those available for different agricultural products are therefore needed.

The aim of this work was to develop methods for the determination of aflatoxins in human urine and airborne dust. The methods developed will be used to determine the level of exposure to aflatoxins in occupationally exposed groups.

15

4 Sample preparation

Most quantifications of aflatoxins are performed using liquid chromatography or thin-layer chromatography. These methods require preparatory clean-up and concentration of samples. Clean-up is necessary to remove interfering material since the sample often consists of a complex matrix. Concentration is required since the analyte is present only in small amounts. After sample preparation, the sample should be ready for quantification by a chromatographic method.

For biological samples, the general sample preparation steps have been extensively reviewed [23]. For samples consisting of serum, plasma or urine, the most common methods for sample preparation are liquid-liquid extraction, that is, extraction with water and a water-immiscible solvent, or solid-phase extraction. These techniques have also been used for the sample clean-up of aflatoxins in various food products, as described in several recent reviews [15, 22, 24, 25].

Solvent extraction and solid-phase extraction methods are described below for aflatoxin analysis. A description is given of how the methods are used for food and feed, and for samples as airborne dust or urine. In addition to these methods, other methods have been used for aflatoxin clean-up. Supercritical fluid extraction has been used for the clean­up of aflatoxin Bi in food [26], and in settled dust [27], Dialysis has been used for the clean-up of aflatoxin Mi in milk, in combination with immunoaffinity extraction [28].

4.1 Solvent extraction

Extraction of aflatoxins from agricultural products has been carried out with a wide range of solvents. When solid-phase extraction methods were introduced, laborious liquid-liquid extraction methods, such as the AOAC methods, were replaced to a large extent. Solid-phase extraction is used in conjunction with extractions by other solvents, such as methanol- water, acetonitrile-water or acetone-water. This extraction is used to liquidize the sample so that it is ready for a solid-phase extraction clean-up. The solvents used are usually mixed with water to increase the penetration of the solvents into the substrate [29]. The efficiency of extraction of aflatoxins by these solvents has been studied in several articles. The extraction of aflatoxins in palm kernels has been performed with different

17

ratios of acetone-methanol-water [30]; acetone-water (80:20, v/v) extracted most aflatoxins. Methanol-water (80:20, v/v) or acetonitrile-water (90:10, v/v) extracted more aflatoxins in peanuts than other mixtures with a higher percentage of water [31]. And acetone-water (80:20, v/v) extracted more aflatoxins in sorghum and maize than acetonitrile-water (75:25, v/v) [32], The solvent used for extraction is accordingly matrix-dependent. It has also been shown that the extraction efficiency is dependent on the sample-solvent ratio [30, 31],

A method for the determination of aflatoxins in airborne dust samples is presented in Paper III. The method incorporates an acetone-water (85:15, v/v) extraction and an automated immunoaffinity extraction clean­up. The acetone-water mixture was used since it has been employed previously for extracting aflatoxins with good results and since the acetone is easily removed by evaporation [33], This is needed since the immunoaffinity columns used for clean-up do not work in the presence of high levels of acetone, as discussed below.

4.2 Solid-phase extraction and clean-up

In solid-phase extraction the sample is loaded on a column containing a solid adsorbent. The analyte of interest is adsorbed and the unwanted material removed by passing an eluent through the column. The analytes are then eluted with another solvent, collected and analysed. Solid- phase extraction can be divided into columns packed with a wide range of materials, such as Florisil, alumina, silica, chemically modified silica called bonded-phase, and columns containing antibodies which then are called immunoaffinity columns. All these columns are commercially available in disposable columns or cartridges, with packings of usually 100-500 mg of 40-pm particles.

For solid-phase extraction, solid samples need to be extracted with any of the solvents mentioned, so that the aflatoxins are in solution. With liquid biological samples, such as serum, plasma or urine, they can be loaded directly, but previous dilution and centrifugation are often needed as they can be viscous [23], Various solid-phase extraction materials have been used for aflatoxin analysis. Some examples are given below.

For the determination of aflatoxin Mi in milk, octadecyl (Ci8) columns can be used by direct-loading the milk on the columns and then eluting and analysing the aflatoxin [34]. The efficiency of such bonded-phase cartridges as Cis, octyl (C«), ethyl (C2), cyclohexyl (CH) and phenyl (Ph) has

18

been studied for the determination of methanol-water extracts of aflatoxins in maize [35], and the Ph bonded-phase material was found to give the highest recoveries. The widely used AOAC, CB standard method has been compared with a method using Florisil columns for sorghum [36], and a method using Ph bonded-phase columns for maize [37], Both methods were found to be more rapid and their accuracy and precision were equal to or better than the AOAC method. Another way for the sample clean-up was used by Hurst et al. for the determination of aflatoxins in peanuts. An aminopropyl (NH2) column was used for removing interferences. The aflatoxins pass the column and are captured on a Cig column [38], The multi-functional clean-up columns employ the same approach. They are used for clean-up by removing interferences on a column that are packed with solid-phase material containing different functional groups. The aflatoxins pass the column unretained and are then analysed [39],

For the studies in Paper I we used C2-columns for the clean-up of aflatoxins Bi, B2, Gi and G2 in urine. The urine samples were diluted and loaded on the C2-column, which had previously been conditioned with acetonitrile and water. The column was then washed with water containing 10% acetonitrile, and the aflatoxins eluted with dichloromethane. The clean­up procedure was studied for recovery and repeatability at three different levels, 49, 120 and 233 pg/ml. The recoveries were on average 95%, 90%, 93% and 89% for aflatoxins Bi, B2, Gi and G2, respectively, with relative standard deviations of 2-15%. The C2-columns were selected as they resulted in the cleanest extracts when C2, C* and Ph columns were compared with urine samples loaded on the columns (unpublished results). This may be due to the fact that C2 material is more selective, as it is less hydrophobic and the role of the silanol groups on the silica surface is larger [23],

Aflatoxins can be present in various products, and a wide range of solid-phase materials have been used. The techniques can replace the laborious liquid-liquid extraction and column chromatography in the AOAC standard methods. Immunoaffinity columns are increasingly used for sample clean-up, and the technique is therefore discussed separately below.

19

4.3 Immunoaffinity extraction and clean-up

Another type of solid-phase extraction consists in using columns packed with antibodies against aflatoxins, which are attached to a gel support used for immunoaffinity extraction [40], The columns contain antibodies that have been produced to recognize aflatoxins.

When the antibodies are produced, the aflatoxin must be conjugated to a protein, as the aflatoxin molecule is too small to elicit an immunoresponse. Depending on how the aflatoxin is conjugated to the protein, different antibody specificities are obtained. If the aflatoxin is bound to the protein through the cyclopentanone moiety, the antibodies produced recognize aflatoxins that are similar in the dihydro&ran portion. When the aflatoxin is bound to a protein, antibodies can be produced by immunizing an animal to obtain polyclonal antibodies, or be produced by hybridoma technology, to obtain the more specific monoclonal antibodies. Once the antibodies are produced they are bound to a solid-phase support. One common support for binding the antibodies is cyanogen bromide-activated agarose. The support with the antibodies is packed into a column and used for immunoaffinity extraction. The columns usually recognize several aflatoxins, since the aflatoxins are all similar in structure [41].

Liquid samples or sample extracts of methanol-water, acetonitrile- water and acetone-water can be loaded on the columns, but care should be taken not to have too high a percentage of organic solvent in the sample, in which case the aflatoxin-antibody binding is interrupted. After the sample is loaded on the column, the column is washed with water to remove unwanted material. After the columns are washed, the aflatoxins are generally eluted by interrupting the antibody-antigen binding with acetonitrile or methanol [40], The eluted extracts are then ready to be injected onto a liquid chromatograph.

20

We previously developed a method for aflatoxins in urine using a C2-column for clean-up, but to obtain extra confirmation for the determination we also developed methods using immunoaffinity extraction. In Paper II we used immunoaffinity extraction with a column containing antibodies developed against aflatoxin Mi for the clean-up of aflatoxin Qi in urine. The similarity in structures of the aflatoxins made the antibodies interact with aflatoxin Qi. Recovery of aflatoxin Qi was 88% at a level of 50 pg/ml of urine, with a relative standard deviation of 6%. In Paper IV immunoaffinity extraction was used with antibodies developed for the four aflatoxins Bi, B2, Gi and G2. The immunoaffinity column was used for the simultaneous extraction of aflatoxins Bi, B2, Gì, G2, Qi and Mi in urine. Recoveries were studied at the levels 7-73 pg/ml urine of aflatoxins Bi, B2, Gi and G2 and 18-97 pg/ml urine of aflatoxins Mi and Qi. The recoveries were high, with average recoveries between 84% and 106% for spiked urine samples, with relative standard deviations in the range 1-21%. Urine samples were loaded directly on the column with only a previous dilution and the cleaned-up extract was analysed by liquid chromatography, post-column derivatization and fluorescence detection.

Chromatograms of a urine sample cleaned up on an immunoaffinity column are presented in Figure 4. The chromatograms show a urine sample non-hydrolysed, hydrolysed with /^-glucuronidase and spiked with 21 pg/ml of urine with aflatoxins Bi, B2, Gi and G2 and 28 pg/ml of urine with aflatoxins Mi and Qi. The clean chromatograms show the effective clean-up on the immunoaffinity column.

21

A

B

C

20 t (min)105 150

Figure 4. Chromatograms of the same urine sample cleaned up on immunoaffinity columns. (A) Non-hydrolysed. (B) Hydrolysed. (C) Spiked with 21 pg/ml urine of aflatoxins Bi, B2, Gi and G2 and 28 pg/ml urine of aflatoxins Mi and Qi (from Paper IV).

In Paper m we used immunoaffinity extraction for aflatoxins Bi, B2, Gi and G2 in airborne dust in combination with an acetone-water extraction as discussed above. Recoveries were studied by adding standards in the range 62-595 pg of each of the aflatoxins Bi, B2, Gi and G2 to filters. Average recoveries for aflatoxins Bi, B2) Gi and G2 were 97%, 101%, 87% and 84%, respectively, and relative standard deviations 1-27%. After the recovery studies, the method was used with dust samples originating from different agricultural products. With the airborne dust samples as well as the urine samples, the eluted extracts from the immunoaffinity column were concentrated from about 2 ml to 200 pi by evaporation, and later injected onto the liquid chromatograph.

22

Chromatograms of dust samples from copra and cotton seed cleaned up on immunoaffinity columns are shown in Figure 5, together with a spiked filter. Immunoaffinity extraction gives an efficient clean-up both for urine and for dust samples, as shown in Figures 4 and 5. Liquid chromatography is used after the clean-up to separate the different aflatoxins that interact with the antibodies in the column.

Figure 5. Chromatograms (fluorescence detection) of dust samples from (A) copra naturally contaminated with aflatoxins Bi, B2 and Gi,(B) cotton seed naturally contaminated with aflatoxins Bi and B2, and(C) filter spiked with 62 pg of each aflatoxin (from Paper III).

L- UCc_0 5 10 15 t (min)

23

4.4 Automation

Solid-phase extraction methods can be automated with various automation systems, such as Millilab from Waters, Aspec from Gilson or Benchmate from Zymark. These systems are designed for handling the sample and the solvents used for the solid-phase extraction, but can also perform other manipulations such as weighing and injection onto a liquid chromatograph [42], Automation is used to save time and to increase sample throughput, since the systems can work unattended. With solid-phase extraction it is also important that the different samples are handled in a similar way to enhance reproducibility. Another benefit of automation is that it increases safety for laboratory personnel by preventing contact with samples and solvents.

In Paper I, a Millilab work station was used for the automated clean-up of urine on C2-columns, and in Papers II-V for the clean-up of urine and dust on immunoaffinity columns. The time to process each sample was about one hour and the operation could proceed unattended. The Millilab work station uses a probe that moves in the x-y-z directions and is connected to the column by the use of an inflatable tip. The sample and the various solvents are pushed through the column by the use of a syringe. Any volume of solvents can be used, with different flow rates, depending on the syringe volume. Gases can be used for column drying. After the automated clean-up, the samples were manually transferred for evaporation, dilution and liquid chromatography.

Sample clean-up can also be performed in connection with analysis in on-line systems, as described in a review by Brinkman [43], Immunoaffinity extraction has been used in an on-line system, where the antibodies are packed in a precolumn that is connected to a second precolumn of Cig material and then separated on an analytical column [44], The system was developed for aflatoxin Mi in milk but was also used with urine samples spiked with aflatoxin B2. Problems encountered with the on­line system were that when the immunoaffinity column was used repeatedly it tended to lose its efficiency and that the eluate from the column also needed to be diluted on-line with water to be enriched on the second precolumn consisting of Ci» material.

24

4.5 Stability

An important aspect of aflatoxin analysis is the stability of aflatoxins in the solvents used. For aflatoxins and G^, which are the products when pre-column derivatization with trifluoroacetic acid is used (see below), special precautions need to be taken to prevent degradation [22], Even aflatoxins Bi, B2, Gi and G2 have been shown to degrade in various aqueous solvents, such as methanol or acetonitrile [45], It is also important to protect aflatoxin solutions from light [15], In the papers discussed, the standards used for spiking filters and urine samples, and standards for calibration, were prepared daily to minimize the influence of degradation. These standards were diluted from stronger solutions that were kept in a refrigerator at 8°C. In Paper I it was noticed that silylation of glass vials was necessary when evaporation for concentration of aflatoxin solutions was performed. The silylation was important for preventing losses of aflaioxins caused by adsorption on the glass walls.

4.6 Safety precautions

Since aflatoxins are cancerogenic and toxic compounds, safety precautions need to be taken when handling the pure substances. Precautions also need to be taken when handling the producing moulds Aspergillus flavus and Aspergillus parasiticus, or mould-infected material.

In handling pure substances, a fume hood or a glove box should be used, in addition to protective clothing that includes gloves and safety glasses. Detoxification of laboratory wastes and laboratory equipment can be performed with 0.5% sodium hypochlorite solution, followed by 5% acetone [46]. Another detoxification method consists in the use of a basic ethanol solution, as in Papers I-V.

25

5 Analysis

5.1 Liquid chromatography

Liquid chromatographic methods, such as normal-phase chromatography and reversed-phase chromatography, have been used for aflatoxin analysis, and both methods can be used to separate aflatoxins Bi, B2, Gi and G2 [22], Reversed-phase chromatography is now the most-used technique, as it uses aqueous solvents.

In Paper I and V, a 5-pm, 250 x 4.6 mm I D. Cis-column was used to separate aflatoxins Bi, B2, Gi and G2, with a mobile phase of acetonitrile- methanol-water (20:20:60, v/v). In Paper III a 4-pm., 100 x 8 mm I.D. Nova-Pak Phenyl column was used to separate aflatoxins Bi, B2, Gi and G2, and in Paper IV aflatoxins Bi, B2, Gi, G2, Mi and Qi. The mobile phase used in Paper IV consisted of acetonitrile-water (30:70, v/v), where the water contained 1 mM potassium bromide and 1 mM nitric acid used for post­column derivatization. The aflatoxins were separated on the Phenyl column with 150-pl injections, compared with 20-25 pi that was used with the Cu column. The larger injection volume was used to lower the detection limit of the aflatoxins in the samples.

27

5.2 Fluorescence detection

The native fluorescence and the fluorescence after derivatization of the aflatoxins have been used in several ways for detection. Aflatoxins B2 and G2 show stronger fluorescence than Bi and G» in reversed-phase chromatography. By derivatization reactions of aflatoxins Bi and Gi, fluorescence levels can increase to the levels of aflatoxins B2 and G2, as recently reviewed by Kok [47], The fluorescence intensity of aflatoxins Bi and Gi can be increased by pre-column derivatization with trifluoroacetic acid to form aflatoxins B^ and G^ or by post-column derivatization with iodine or bromine. The double bond of the vinyl ether function of aflatoxins Bi and Gi is considered to be the site for the derivatization. The reaction products after derivatization might be formed by the addition of two iodine molecules or two bromine molecules (Figure 6), as proposed by Kok et al. [48], Based upon the characterization by thermospray mass spectrometry of the products formed by reaction with iodine, Holcomb et al. suggested that the products were formed by the addition of an iodine atom and a methoxy group at the furan ring of aflatoxin Bi and Gi [49], Other methods of fluorescence enhancement in derivatization are post-column addition of pyridinium bromide perbromide [18] or /^-cyclodextrin [50],

o c h 3

Aflatoxin B

o c h 3

Figure 6. Reaction products after derivatization with iodine or bromine, as proposed by Kok et al. [48].

Different detection limits with iodine derivatization, reported by Kok [47] are in the range of 2-100 pg injected amounts of aflatoxins Bt and Gi. The detection limits with bromine derivatization used by Kok et al. [48] were 20-40 pg injected amounts, and the same detection limits were attained with iodine derivatization.

In Figure 7, the liquid chromatographic system with post-column derivatization is described. Bromine is produced from the bromide- containing effluent, when it passes the electrodes in the cell. The two electrodes are separated by a membrane. After the bromine generation a coil is used for the aflatoxin-bromine reaction. The coil is then coupled to a fluorescence detector.

Pump Injector Column Electrochemical Reactor Detectorcell

Figure 7. Scheme for post-column derivatization by addition of bromine used in Papers I-V.

29

In biological material, a method of increasing the fluorescence detection of the aflatoxin-N7-guanine adduct has been developed [51], The method includes hydrolysis of the guanine adduct with heat and acid to obtain the products guanine and aflatoxin Bi-8,9-dihydrodiol. The detection limit was 39 pg for aflatoxin Bi-8,9-dihydrodiol, which meant a 100-fold increase compared with detection as aflatoxin-N7-guanine adduct. The hydroxylated metabolites, aflatoxins Mi, Pi and Qi, which can be found in biological material have been hydrolysed to more fluorescentic products by the use of trifluoroacetic acid, in the same manner as aflatoxins Bt and Gi [52].

In Papers I-V we used post-column derivatization by addition of bromine. This method was used to derivatize aflatoxins Bi, Gi and Qi in urine and aflatoxins Bi and Gi in airborne dust. In Paper I we studied the post-column derivatization for aflatoxins Bi and Gi using a factorial design. Methods for studying the influence of different variables by factorial designs have been described by Carlson [53]. The derivatization was studied for three parameters, namely reaction time, amount of potassium bromide in the mobile phase and current for generating bromine. Based on only a few experiments, it was seen that the best settings were a reaction time of 6 s, 1 mM of potassium bromide in the mobile phase and 20 pA for bromine generation. These results agree with the results that the reaction is completed within 4 to 8 seconds and that an increase in current for generation over 100 pA decreases the fluorescence [48], In Paper II we studied the post-column derivatization for aflatoxin Qi using a similar factorial design as the one used for aflatoxins Bi and Gi in Paper I. The detection limit using the post-column derivatization was increased 30-fold to the 20-30 pg level of aflatoxins Bi and Gi in Papers I, III, IV and V, which was about the same as the detection limits for aflatoxins B2 and G2. Bromine was selected for derivatization in preference to iodine or trifluoroacetic acid, since the method is easier to perform. All that is needed for derivatization is the addition of bromide and acid in the mobile phase, bromine then being generated in an electrochemical cell.

30

5.3 Mass spectrometry

Although methods are available for quantification of aflatoxins with selective clean-up, such as immunoaffinity extraction and detection by fluorescence, methods are needed to confirm the identity of the substances. A method other than chemical derivatization, using bromine, iodine or trifluoroacetic acid, for the confirmation is mass spectrometry.

Previously, mass spectrometry with negative ion chemical ionization has been used for confirmation of aflatoxin Bi by introducing the sample with gas chromatography [54], The column was a fused silica column of 6 m X 0.2 mm I.D. with an initial temperature of 40°C that was raised to 250°C. With methane as reagent gas, a mass spectrum was obtained from a 60 ng injection. Injection was performed using an on- column injector, which is necessary because of the thermolability of the aflatoxins. Gas chromatography-mass spectrometry have also been used with electron impact for aflatoxins Bi, B2, Gi and G2 with detection limits of 0.3-1 ng injected amounts [55],

Initially, in the work on this thesis, we performed experiments with gas chromatography-mass spectrometry using electron impact or negative ion chemical ionization with ammonia or methane, using 20 ng injections of a standard solution of aflatoxin B2. The detection limits were estimated to be the same or higher than the detection limits previously reported with gas chromatography-mass spectrometry. We then turned to liquid chromatography and fluorescence detection, since the low detection limit attained is needed for such samples as dust and urine. When liquid chromatography was used for the determination there was a need for the coupling liquid chromatography-mass spectrometry.

When capillary gas chromatography can be direct-coupled to a mass spectrometer as the sample components elute as vapour, problems are encountered with the combination liquid chromatography-mass spectrometry. The main problem for the coupling is the large volume of solvents that needs to be removed. One interface that has been used for determination of aflatoxins in peanuts is thermospray [56], The detection limit using selected ion recording of the quasi-molecular ion was, for aflatoxin Bi, 60 pg injected amount. With thermospray a flow of 1 ml/min is used from the liquid chromatograph, and with a heated capillary tube a jet of vapour and droplets are attained. The ionization occurs at a pressure of 2-10 torr, through reaction with a volatile electrolyte present (filament off) or with the help of an electric discharge or an electron beam (filament on) [57], In the work of Hurst et al. [56] ionization was performed by means of 0.1 M

31

ammonium acetate in the mobile phase; with the filament on mode, the sensitivity was increased. After the thermospray ionization the neutral solvent molecules are pumped away and the ionized molecules led into the mass analyser through a cone that is placed vertically to the jet spray. No fragmentations were achieved using the thermospray, and the aflatoxins were therefore monitored for the [M+H]+ ion.

Another method for combining liquid chromatography-mass spectrometry is with atmospheric pressure ionization (API). This interface includes such techniques as electrospray ionization and atmospheric pressure chemical ionization (APCI) [56, 59]. The ionization with API takes place under atmospheric pressure and no problems are therefore encountered with solvent vapours, since they are easily pumped away. The problem is to transport the ionized molecules into the mass spectrometer. With the electrospray interface used in Paper V the ions are transported into the analyser through a sampling cone into an intermediate pressure region and then further to the mass analyser. The ionization is performed through a capillary that is held at about 3-4 kV and a spray of ionized molecules is obtained. With APCI, an aerosol is obtained through a nebulizer and ionization through a corona pin held at approximately 2 kV.

The electrospray interface was used in Paper V with a triple quadrupole mass spectrometer. The instrument was used to determine aflatoxins in naturally contaminated airborne dust and spiked urine samples. Preliminary experiments using standard injections of aflatoxins resulted in approximately 4 times higher intensity of the molecular ion with electrospray than with APCI, and electrospray was therefore used for the continuous experiments. Both techniques give strong [M+H]+ ions. In Paper V the [M+H]+ ion of each aflatoxin was used for the determination with selected ion recording, but also for attaining fragmentations by collision with argon, the fragments being detected by tandem mass spectrometry (MS/MS).

Figure 8 shows the spectra of the [M+H]+ ions of aflatoxins Bi, B2, Gi and G2 after collision with argon, the fragments being detected by MS/MS. The most intense daughter ion from each aflatoxin was used with MS/MS for detection with multiple reaction monitoring (MRM). The detection limit with electrospray was estimated to be 2-8 pg injected amounts of aflatoxins Bi, B2, Gi and G2 using selected ion recording. With MS/MS the detection limit was only a factor 2 higher, as shown in Table 1.

32

241100

B,

zlCLJ aL La

[M+H]+313

°100 120 140 ^60 180 200 220 240 260 280 300 320 340

% 259

• m/z

CM

CD

? i------------------

[M+H]+315

I

n n 1 ón 140 1RO 180 2Ó0 220 240 2!60 280 300 320 340m/z

100243

G i

I [M+H]+

i I I 3?9É g À . A ._,_1 X - J U j — J- - ,---»m/z

%100i 245

.J ..v pM^JApyL . À J , i . , ß JU

[M+H1+331

-m/zj m Må ffUMia^Tfcr y\Éamr um mifMm ■ ■i«.é f \»mm iij i utmn a j ■■■ [ ■ f 1 1 i-----------------1

100 120 140 160 180 200 220 240 260 280 300 320 340

Figure 8. Daughter ion mass spectra of quasi-molecular ions of aflatoxins Bt, B2, Gi and G2 after collision with argon (from Paper V).

33

Compound Electrospray Thermospray

SIR MRM SIR

LOD LOD LOD

(Pg) (pg) (Pg)Aflatoxin Bi 2 4 60

Aflatoxin B2 2 4 40

Aflatoxin Gì 3 5 100

Aflatoxin G2 8 10 100

Table 1. Limit of detection (LOD) with pg amounts injected on column using electrospray mass spectrometry with selected ion recording (SIR) or multiple reaction monitoring (MRM) (from Paper V), and thermospray mass spectrometry using SER as reported by Hurst et al. [56],

The electrospray technique resulted in detection limits about tenfold better than the thermospray technique used by Hurst et al. [56], The low detection limits are needed for samples as airborne dust or urine.

34

6 Aflatoxins determined

6.1 Dust samples

Since aflatoxins can be produced on a variety of agricultural products, agricultural workers exposed to dust generated from the contaminated products in the working environment may run a health risk. The aflatoxins found in the airborne dust are aflatoxins Bi, B2, Gi and G2, which are the same as those found in the agricultural products. The aflatoxin levels in the airborne dust are most often the same or higher than the levels found in the products generating the dust. This may be due to the presence of conidia containing aflatoxins that are easily generated in the air [14].

In a study, dust was collected during maize harvesting, using high- volume and personal pumps, and the total amounts of airborne aflatoxins were found in levels up to 52 000 ng/g [60], The maize generating the dust contained 1640 ng/g of total aflatoxins. In another study, peanut dust has been found to contain levels of up to 730 ng/g of aflatoxin Bi [61].

The method for the determination of dust samples used in Paper III and IV was acetone-water extraction, immunoaffinity extraction and liquid chromatography, followed by bromine derivatization and fluorescence detection or mass spectrometry. The scheme for the determination are shown in Figure 9. Evaporation was performed after the acetone-water extraction to remove the acetone, and to concentrate the sample to a smaller volume after the immunoaffinity extraction. Silanized glass vials were used during the evaporation steps to prevent loss of aflatoxins. Using this method we reported in Paper III levels up to 53 ng/g of aflatoxin Bi in airborne dust from copra and lower levels of aflatoxins Gi and B2. Aflatoxins could also be detected in cotton seed and maize gluten. The dust samples were collected during unloading of products used for feed production, which is one of the most dusty operations in the factory. The collection of the dust samples was performed with high-volume pumps placed about 2-3 m from the dust source.

35

AirsamplingWeighing

Solvent extraction Concentration

Liquid chromatography/ post-column denvatization

Liquid chromatography/ mass spectrometry

Filtration ImmunoafFinity extraction

Concentration

Figure 9. Scheme for the determination of aflatoxins in airborne dust.

In Paper V, the aflatoxins in the dust samples could be confirmed with mass spectrometry. The dust samples analysed were collected at the same time as those that were analysed in Paper III. With the electrospray mass spectrometry used in Paper V, aflatoxins Bi, B2 and Gi in airborne dust were determined in samples of copra and cotton seed. The method included electrospray mass spectrometry using selected ion recording (SIR) of the [M+H]+ ion and multiple reaction monitoring (MRM) by tandem mass spectrometry to detect daughter ions after collision with argon. The levels of aflatoxin Bi detected in six dust samples using SIR and MRM, compared with a method using fluorescence detection, are presented in Figure 10. The levels detected in settled and airborne dust of aflatoxin Bi were in the range0.2-2 ng/m3 of air.

36

1 2 3 4 5 6

Figure 10. Determination of aflatoxin Bi in dust from copra (samples 1-5) and cotton seed (sample 6), using (□) selected ion recording, (■) multiple ion monitoring and ( a ) fluorescence detection. Samples 1-3 and 6 are airborne dust, samples 4 and 5 are settled dust.

37

6.2 Biological samples

Another way of measuring human exposure to aflatoxins than by the external dose of aflatoxins in dusts, is to measure aflatoxins in biological samples. Aflatoxin detected in biological material, such as serum or urine from humans, has been reviewed in several publications [7, 62, 63]. Detection of aflatoxins can be carried out by measuring the internal dose as unbound aflatoxins or by measuring the effective dose as aflatoxins bound to DNA or proteins.

Aflatoxin Mi was the first aflatoxin detected in human urine. The aflatoxin was detected in urine from Filipinos who had eaten aflatoxin- contaminated peanut butter [64], Aflatoxin Mi has then been found in urine samples from Nigeria [65], from Zimbabwe with an average level of 4.2 ng/ml [66] and from China in the levels 0.03-0.3 ng/ml [67], In addition, aflatoxin Mi has been detected in serum in levels of 5-15 pg/ml [68], Other metabolites, such as aflatoxins Bi, Gi, and aflatoxicol, have also been detected in urine [65]. In several of these studies of biological material, a correlation has been made between the estimated dietary intake of aflatoxin Bi in contaminated food and the excreted aflatoxin Mi in urine [64, 67]; 1- 4% of the estimated consumed aflatoxin Bi has been found to be excreted as aflatoxin Mi.

Aflatoxins may form macromolecule adducts and efforts have been made to detect these in biological samples. The aflatoxin DNA-adduct formed at the N-7 position in guanine is unstable and is therefore released from DNA as an aflatoxin Bi-guanine adduct. This adduct has been detected in urine using immunoaffinity extraction and liquid chromatography with UV detection [69], The detection of the aflatoxin Bi-guanine adduct is an indirect proof of aflatoxin DNA-binding [70], Aflatoxin Bi-guanine has also been observed to be correlated with the estimated dietary intake [69], In serum, the aflatoxin Bi-lysine adduct has been detected [71]. The aflatoxin- albumin adducts were hydrolysed by Pronase to obtain the aflatoxin Bi- lysine adduct, following immunoaffinity extraction and liquid chromatography and fluorescence detection.

In some reports, aflatoxins in urine are quantitated as aflatoxin Bi- equivalents [62], That is when the method for quantitation used does not separate the different aflatoxins, for example, methods such as enzyme- linked immunoassay, when the antibodies used recognize several aflatoxins. In suspected occupationally exposed groups, aflatoxin equivalents have been detected in urine from dockers in levels of 0.5-5 ng/ml [72], However, a control group that was not occupationally exposed excreted a corresponding

38

amount of aflatoxin equivalents. In another study it was found that animal- feed workers had aflatoxin Bi-adducts in serum at levels of 44-100 pg/mg albumin [73],

In paper IV we analysed urine samples from feed factory workers. The urine samples were analysed for aflatoxins Bi, B2, Gi, G2, Mi and Qi, but no aflatoxins could be detected in any of the samples. Urine samples were also hydrolysed to remove possible sulphate and glucuronide conjugates from the aflatoxins, but even then no aflatoxins were detected. The reason for this may be that the workers had not been exposed, as dust and urine samples were not collected at the same time.

The method of analysing the urine samples included dilution and hydrolysation with /^-glucuronidase, immunoaffinity extraction, evaporation for concentration and liquid chromatography, as shown in Figure 11. The liquid chromatography included post-column derivatization with bromine and fluorescence detection.

Immunoaffinity extraction Concentration

10 ml urine sample Dilution

10 ml urine sample Dilution

Hydrolysation

Liquid chromatography/ post-column derivatization

Figure 11. Scheme for the determination of aflatoxins in urine.

39

7 Conclusions

In this thesis, it is shown how aflatoxins can be determined in samples of airborne dust and human urine. The methods can be used to determine occupational exposure to aflatoxins.

For sample preparation of the samples, solvent extraction and solid- phase extraction were used. The solid-phase columns used were ethyl- bonded phase columns and the more selective immunoaffinity columns. By automating the solid-phase extractions with a laboratory robot, reproducible results were achieved.

Liquid chromatography was used with post-column derivatization by addition of bromine, where the parameters for the derivatization were studied using factorial designs. Liquid chromatography was also combined with electrospray mass spectrometry to confirm the identity of the aflatoxins.

The methods developed can be used to determine aflatoxins in urine at the 6.8-18 pg/ml level. In 10-mg dust samples from agricultural products, such as copra and cotton seed, aflatoxins can be determined at the 1.8-3.1 ng/g level.

41

Acknowledgements

This work has been carried out at the Analytical Chemistry Division, National Institute of Occupational Health in Umeå, and I would like to thank all my colleagues for providing a nice atmosphere and good working conditions.

In particular, I would like to thank my supervisor Dr Barbro Andersson, and Professor Kurt Andersson, for their advice and interest in the work.

Professor Anders Cedergren, head of the Department of Analytical Chemistry, Umeå University, for his support.

Dr Hans Malker for the epidemiological studies that initiated this project.

Mr Rajesh Kumar for checking the English of the manuscripts.

Dr Carl-Axel Nilsson for advice on how to handle statistics.

Mrs Margaret Rhèn and others at the Department for their collaboration.

Mrs Ninni Broms-Dahlgren at the Library for ordering literature and references, Mrs Margaretha Karlsson for administrative work and Mr Bo Seliman for printing the thesis.

I also wish to thank my parents for their support.

43

References

1. Van der Zijden A S M , Blanche Koelensmid W A A, Boldingh J, Barrett C B, Ord W O and Philp J. Aspergillus flavus and turkey X disease. Isolation in crystalline form of a toxin responsible for turkey X disease. Nature, 195 (1962), 1060-1062.

2. Nesbitt B F, O'Kelly J, Sargeant K and Sheridan A. Aspergillus flavus and turkey X disease. Toxic metabolites of Aspergillus flavus. Nature, 195 (1962), 1062-1063.

3. Dvorackova I. Aflatoxins and Human Health, CRC Press, Boca Raton, FL, 1990.

4. Betina V. Mycotoxins, Bioactive Molecules, Vol. 9, Elsevier, Amsterdam, 1989.

5. Hall A J, Wild C P. Epidemiology of aflatoxin-related disease. In Eaton D L and Groopman J D (Eds). The Toxicology o f Aflatoxins, Academic Press, San Diego, CA, 1994, 233-258.

6. Eaton D L, Ramsdell H S and Neal G E. Biotransformation of aflatoxins. In Eaton D L and Groopman J D (Eds). The Toxicology o f Aflatoxins, Academic Press, San Diego, CA, 1994, 45-72.

7. Groopman J D. Molecular dosimetry methods for assessing human aflatoxin exposures. In Eaton D L and Groopman J D (Eds). The Toxicology o f Aflatoxins, Academic Press, San Diego, CA, 1994, 259- 279.

8. Roebuck B D and Maxuitenko Y Y. Biochemical mechanisms and biological implications of the toxicity of aflatoxins as related to aflatoxin carcinogenesis. In Eaton D L and Groopman J D (Eds). The Toxicology o f Aflatoxins, Academic Press, San Diego, CA, 1994, 27-43.

9. Olsen J H, Dragsted L and Autrup H. Cancer risk and occupational exposure to aflatoxins in Denmark. Br. J. Cancer, 58 (1988), 392- 396.

10. McLaughlin J K, Malker HSR, Malker B K, Stone B J, Ericsson J L E, Blot W J, Weiner J A and Fraumeni J F Jr. Registry-based analysis of occupational risks for primary liver cancer in Sweden. Cancer Res., 47(1987), 287-291.

11. Hayes R B, van Nieuwenhuize J P, Raatgever J W and ten Kate F J W. Aflatoxin exposures in the industrial setting, an epidemiological study of mortality. Food Chem. Toxicol., 22 (1984), 39-43.

45

12. Ball R W, Huie J M and Coulombe R A Jr. Comparative activation of aflatoxin Bi by mammalian pulmonary tissues. Toxicol. Lett., 75 (1995), 119-125.

13. Larsson P, Pettersson H and Tjälve H. Metabolism of aflatoxin Bi in the bovine olfactory mucosa. Carcinogenesis, 10 (1989), 1113-1118.

14. Wicklow D T and Shotwell O L. IntrafUngal distribution of aflatoxins among conidia and sclerotia of Aspergillus flavus and Aspergillus parasiticus. Can. J. Microbiol., 29 (1983), 1-5.

15. Ellis W O, Smith J P, Simpson B K and Oldham J H. Aflatoxins in food: occurrence, biosynthesis, effects on organisms, detection, and methods of control. Crit. Rev. Food Sci. Nutr., 30 (1991), 403-439.

16. Wilson D M and Payne G A. Factors affecting Aspergillus flavus group infection and aflatoxin contamination of crops. In Eaton D L and Groopman J D (Eds). The Toxicology o f Aflatoxins, Academic Press, San Diego, CA, 1994, 309-325.

17. Stoloff L. Aflatoxins-an overview. In Rodricks J V, Hesseltine C W and Mehlman M A (Eds). Mycotoxins in Human and Animal Health. Pathotox, Park Forest South, IL, 1977, 7-28.

18. Gamer R C, Whattam M M, Taylor P J L and Stow M W. Analysis of United Kingdom purchased spices for aflatoxins using an immunoaffinity column clean-up procedure followed by high- performance liquid chromatographic analysis and post-column derivatisation with pyridinium bromide perbromide. J. Chromatogr., 648 (1993), 485-490.

19. Pettersson H, Holmberg T, Larsson K and Kaspersson A. Aflatoxins in acid-treated grain in Sweden and occurrence of aflatoxin Mi in milk. J. Sci. FoodAgric., 48 (1989), 411-420.

20. Van Egmond H P. Current situation on regulations for mycotoxins. Overview of tolerances and status of standard methods of sampling and analysis. FoodAddit. Contam., 6 (1989), 139-188.

21. AO AC. Natural poisons. Official Methods o f Analysis. Association of Official Analytical Chemists, Arlington, VA, 1984, ch. 26.

22. Holcomb M, Wilson D M, Trucksess M W and Thompson H C Jr. Determination of aflatoxins in food products by chromatography. J. Chromatogr., 624 (1992), 341-352.

23. McDowall R D. Sample preparation for biomedical analysis. J. Chromatogr., 492 (1989), 3-58.

24. Scott P M. Recent developments in methods of analysis for mycotoxins in foodstuffs. Trends Anal. Chem., 12 (1993), 373-381.

46

25. Betina V. Thin-layer chromatography of mycotoxins. In Betina V (Ed). Chromatography o f Mycotoxins. Elsevier, Amsterdam, 1993, 141-251.

26. Engelhardt H and Haas P. Possibilities and limitations of SFE in theextraction of aflatoxin B1 from food matrices. J. Chromatogr. Sci., 31(1993), 13-19.

27. Selim M I and Tsuei M-H. Development and optimization of a supercritical fluid extraction method for the analysis of aflatoxin Bi in grain dust. Am. Ind. Hyg. Assoc. J., 54 (1993), 135-141.

28. Faijam A, van de Merbel N C, Nieman A A, Lingeman H and Brinkman U A Th. Determination of aflatoxin Ml using a dialysis- based immunoaffinity sample pretreatment system coupled on-line to liquid chromatography. Reusable immunoaffinity columns. J. Chromatogr., 589 (1992), 141-149.

29. Betina V. Sampling, sample preparation, extraction and clean-up. In Betina V (Ed). Chromatography o f Mycotoxins. Elsevier, Amsterdam, 1993, 3-11.

30. Nawaz S, Coker R D and Haswell S J. Development and evaluation of analytical methodology for the determination of aflatoxins in palm kernels. Analyst, 117 (1992), 67-74.

31. Cole R J and Domer J W. Extraction of aflatoxins from naturally contaminated peanuts with different solvents and solvent/peanut ratios. J. Assoc. Off. Anal. Chem., 77 (1994), 1509-1511.

32. Bradbum N, Coker R D and Blunden G. A comparative study of solvent extraction efficiency and the performance of immunoaffinity and solid phase columns on the determination of aflatoxin Bi. Food Chem., 52(1995), 179-185.

33. Werner G. Bestimmung der Aflatoxine in fiittermitteln nach selektiver reinigung an einer immunoaffinitätssäule, Agribiol. Res., 44 (1991), 289-298.

34. Carisano A and Della Torre G. Sensitive reversed-phase high- performance liquid chromatographic determination of aflatoxin Mi in dry milk. J. Chromatogr., 355 (1986), 340-344.

35. Tomlins K I, Jewers K and Coker R D. Evaluation of non-polar bonded-phases for the clean-up of maize extracts prior to aflatoxin assay by HPTLC. Chromatographia, 27 (1989), 67-70.

47

36. Jewers K, John A E and Blunden G. Development of rapid methods for the analysis of aflatoxin in agricultural products. Part 4.Assessment of suitability of the Florisil and CB clean-up procedures for determining aflatoxin levels in sorghum by bidirectional HPTLC. Chromatographia, 27 (1989), 617-621.

37. Bradbum N, Coker R D and Jewers K. A comparative study of phenyl bonded-phase, CB and Römer clean-up procedures for determining aflatoxin levels in maize by bi-directional HPTLC. Chromatographia, 29(1990), 177-181.

38. Hurst W J, Snyder K P and Martin R A Jr. Determination of aflatoxins in peanut products using disposable bonded-phase columns and post­column reaction detection. J. Chromatogr., 409 (1987), 413-418.

39. Wilson T J and Römer T R. Use of the mycosep multifunctional cleanup for liquid chromatographic determination of aflatoxins in agricultural products. J. Assoc. Off. Anal. Chem., 74 (1991), 951-956.

40. Candlish A A G and Stimson W H. Emerging techniques: immunoaffinity chromatography. In Betina V (ed). Chromatography o f mycotoxins. Elsevier, Amsterdam, 1993, 99-123.

41. Chu F S. Development of antibodies against aflatoxins. In Eaton D L and Groopman J D (Eds). Toxicology o f aflatoxins. Academic Press, San Diego, CA, 1994, 451-490.

42. Majors R E. Automation of solid-phase extraction. LC-GC Int., 6(1993), 346-350.

43. Brinkman U A Th. On-line sample treatment for or via column liquid chromatography. J. Chromatogr. A, 665 (1994), 217-231.

44. Faijam A, van de Merbel N C, Lingeman H, Frei R W and Brinkman U A Th. Non-selective desorption of immuno precolumns coupled on­line with column liquid chromatography: determination of aflatoxins. Int. J. Environ. Anal. Chem, 45 (1991), 73-87.

45. Beaver R W. Degradation of aflatoxins in common HPLC solvents. J. High Resolut. Chromatogr., 13 (1990), 833-835.

46. Castegnaro M, Friesen M, Michelon J and Walker E A. Problems related to the use of sodium hypochlorite in the detoxification of aflatoxin Bi. Am. Ind. Hyg. Assoc., 42 (1981), 398-401.

47. Kok W Th. Derivatization reactions for the determination of aflatoxins by liquid chromatography with fluorescence detection. J. Chromatogr. B, 659(1994), 127-137.

48

48. Kok W Th, van Neer Th C H, Traag W A and Tuinstra L G M Th. Determination of aflatoxins in cattle feed by liquid chromatography and post-column derivatization with electrochemical generated bromine. J. Chromatogr., 367 (1986), 231-236.

49. Holcomb M, Korfmacher W A and Thompson H C Jr.Characterization of iodine derivatives of aflatoxin Bi and Gi by thermospray mass spectrometry. J. Anal. Toxicol., 15 (1991), 289- 292.

50. Francis 0 J Jr, Kirschenheuter G P, Ware G M, Carman A S and Kuan S S. /?-Cyclodextrin post-column fluorescence enhancement of aflatoxins for reverse-phase liquid chromatographic determination in com. J. Assoc. Off. Anal. Chem., 71 (1988), 725-728.

51. Weaver V M and Groopman J D. Fluorescence quantification of aflatoxin N7-guanine adducts. Cancer Epidemiol., Biomarkers Prev., 3(1994), 669-674.

52. Orti D L, Grainger J, Ashley D L and Hill R H Jr. Chromatographic and spectroscopic properties of hemiacetals of aflatoxin and sterigmatocystin metabolites. J. Chromatogr., 462 (1989), 269-279.

53. Carlson R. Design and Optimization in Organic Synthesis, Elsevier, Amsterdam, 1992.

54. Trucksess M W, Brumley W C and Nesheim S. Rapid quantitation and confirmation of aflatoxins in com and peanut butter, using a disposable silica gel column, thin layer chromatography, and gas chromatography/mass spectrometry. J. Assoc. O ff Anal. Chem., 67 (1984), 973-975.

55. Goto T, Matsui M and Kitsuwa T. Analysis of Aspergillus mycotoxins by gas chromatography using fused silica capillary column. Maikotokishin, 31 (1990), 43-47.

56. Hurst W J, Martin R A Jr and Vestal C H. The use of HPLC/thermospray MS for the confirmation of aflatoxins in peanuts.J. Liq. Chromatogr., 14 (1991), 2541-2550.

57. Garcia J F and Barcelo D. An overview of LC-MS interfacing systems with selected applications. J. High Resolut. Chromatogr., 16 (1993), 633-641.

58. Bajic S, Doerge D R, Lowes S and Preece S. APCI: a technique for routine LC-MS. Intern. Chromatogr. Lab., 13 (1993), 4-11.

59. Bruins A P. Atmospheric-pressure-ionization mass spectrometry. I. Instrumentation and ionization techniques. Trends Anal. Chem., 13(1994), 37-43.

49

60. Burg W R and Shotwell O L. Aflatoxin levels in airborne dust generated from contaminated com during harvest and at an elevator in 1980. J. Assoc. Off. Anal. Chem., 67 (1984), 309-312.

61. Sorensson W G, Jones W, Simpson J and Davidsson JI. Aflatoxin in respirable airborne peanut dust. J. Toxicol. Environ. Health, 14 (1984), 525-533.

62. Autrup H and Autrup J L. Human eposure to aflatoxins-biological monitoring. In Bhatnagar D, Lillehoj E B and Arora D K (Eds). Handbook o f Applied Mycology, volume 5, Marcel Dekker, New York, NY, 1992, 213-230.

63. Yourtee D M and Kirk-Yourtee C L. Aflatoxin detoxification in humans laboratory, field and clinical perspectives. J. Toxicol., Toxin Rev., 8(1989), 3-18.

64. Campbell T C, Caedo J P Jr, Bulatao-Jayme J, Salamat L and Engel R W. Aflatoxin Mi in human urine. Nature, 227 (1970), 403-404.

65. Bean T A, Yourtee D M, Akande B and Ogunlewe J. Aflatoxin metabolites in the urine of Nigerians comparison of chromatographic methods. J. Toxicol., Toxin Rev., 8 (1989), 43-52.

66. Nyathi C B, Mutiro C F, Hasler J A and Chetsanga C J. A survey of urinary aflatoxin in Zimbabwe. Int. J. Epidemiol., 16 (1987), 516-519.

67. Zhu J-Q, Zhang L-S, Hu X, Xiao Y, Chen J-S, Xu Y-C, Fremy J and Chu F S. Correlation of dietary aflatoxin Bi levels with excretion of aflatoxin Mi in human urine. Cancer Res., 47 (1987), 1848-1852.

68. Okumura H, Kawamura O, Kishimoto S, Hasegawa A, Shrestha S M, Okuda K, Obata H, Okuda H, Haruki K, Uchida T, Ogasawara Y and Ueno Y. Aflatoxin Mi in Nepalese sera, quantified by combination of monoclonal antibody immunoaffinity chromatography and enzyme- linked immunosorbent assay. Carcinogenesis, 14 (1993), 1233-1235.

69. Groopman J D, Hall A J, Whittle H, Hudson G J, Wogan G N, Montesano R and Wild C P. Molecular dosimetry of aflatoxin-N7- guanine in human urine obtained in the Gambia, West Africa. Cancer Epidemiol., BiomarkersPrev., 1 (1992), 221-227.

70. Autrup H, Seremet T, Wakhisi J and Wasunna A. Aflatoxin exposure measured by urinary excretion of aflatoxin Bi-guanine adduct and hepatitis B virus infection in areas with different liver cancer incidence in Kenya. Cancer Res., 47 (1987), 3430-3433.

71. Sabbioni G, Ambs S, Wogan G N and Groopman J D. The aflatoxin- lysine adduct quantified by high-performance liquid chromatography from human serum albumin samples. Carcinogenesis, 11 (1990), 2063-2066.

50

72. Dragsted L O, Larsen A I, Bull I and Autrup H. Aflatoksinudskillelse hos erhvervseksponerede havnarbejdere. Ugskr Laeger, 147(1985), 4148-4150.

73. Autrup J L, Schmidt J, Seremet T and Autrup H. Determination of exposure to aflatoxins among Danish workers in animal-feed production through the analysis of aflatoxin Bi adducts to serum albumin. Scand. J. Work Environ. Health, 17 (1991), 436-440.

51