Anal Bioanal Chem (2004) 378 : 1152–1160 1152 DOI 10.1007/s00216-003-2373-4

REVIEW

Hans P. van Egmond Natural toxins: risks, regulations and the analytical situation in Europe

Received: 25 July 2003 / Revised: 15 October 2003 / Accepted: 27 October 2003 / Published online: 13 December 2003 © Springer-Verlag 2003

Abstract Natural toxins in food and feed are considered pean context are now directed towards significantly im- important food safety issues of growing concern, in par- proving this situation, and techniques such as liquid chro- ticular mycotoxins, phycotoxins and plant toxins. Most matography–mass spectroscopy offer promise in this re- scientific developments have occurred in the past few spect. Both the working group on biotoxins of the Euro- decades in the area of mycotoxins. Formal health risk as- pean Standardization Committee and the network of Na- sessments have been carried out by the Joint Expert Com- tional Reference Laboratories for Marine Biotoxins have mittee on Food Additives of the World Health Organiza- taken up responsibilities here. The plant toxins are a cate- tion and the Food and Agriculture Organization. Limits gory of natural toxins, where the situation is the least de- and regulations for mycotoxins in food and feed have veloped with respect to regulations, validated methods of been established in many countries, including practically analysis and reference materials. Yet, their occurrence in a all European countries. An array of (formally validated) wide range of consumable plant species demands the at- analytical methods and (certified) reference materials have tention of the analytical community. become available. Several European research projects, funded by the European Commission and supported by the Keywords Natural toxin · Mycotoxin · Phycotoxin · European Standardization Committee, have significantly Plant toxin · Analysis · Reference material contributed to this development. Quantitative methods of analysis for mycotoxins often make use of immunoaffin- ity cleanup with liquid chromatographic or gas chromato- Introduction graphic separation techniques in combination with vari- ous types of detectors, including mass spectroscopy. For Over recent years there has been an increased scientific screening purposes (bio)sensor-based techniques are among interest in natural toxins in food and feed. The occurrence the promising newcomers. For the phycotoxins the situa- of these harmful organic compounds of natural origin is of tion is less advanced. Formal risk assessments by authori- growing concern to human health. Some natural toxins tative international bodies have not been carried out. are extremely potent acute toxins (e.g. botulin) or are Methods of analysis, formally validated according to in- strong carcinogens (e.g. aflatoxins). ternationally harmonized protocols, are scarce and animal On the basis of their origin the natural toxins can be di- testing still plays a key role in official methodology. The vided into five main categories: development of the analytical methodology is partly ham- – Mycotoxins: toxins produced by fungi. Most have rela- pered by the limited availability of certain reliable cali- tively small molecular weights, and they are usually brants and reference materials, although this situation is heat-stable. Mycotoxins that are significant in terms of gradually improving. New regulations in the European toxicity and occurrence include the aflatoxins, ochra- Union have increased the pressure to develop and validate toxins, trichothecenes, patulin, fumonisins and zear- chemical methods of analysis. Joint efforts in the Euro- alenone. – Bacterial toxins: toxins produced by bacteria. Many bacterial toxins are proteins, which are not heat-stable. Some well-known bacterial toxins are botulin, Staphy- H. P. van Egmond (✉) lococcus aureus enterotoxin and Bacillus cereus en- Laboratory for Food and Residue Analyses, terotoxin. National Institute for Public Health and the Environment (RIVM), – Phycotoxins: toxins produced by algae that, through P.O. Box 1, 3720 BA Bilthoven, The Netherlands feed chains, end up in fishery products, such as shell- e-mail: [email protected] fish. Some phycotoxins of significant human health 1153 concern include the paralytic shellfish poisoning toxins, a scientific advisory body of the WHO and the Food and the diarrhoeic shellfish poisoning toxins and the am- Agriculture Organization (FAO), has recently evaluated nesic shellfish poisoning toxins. their hazard [2, 3]. JECFA provides a mechanism for as- – Plant toxins (also named phytotoxins): toxins that are sessing the toxicity of additives, veterinary drug residues produced by edible plant species. Some play a role in and contaminants. Safety evaluation of contaminants in- the defence mechanism of plants against attacks of in- corporates various steps in a formal health risk assessment sects and fungi. Examples are the potato glycoalkaloids approach. The 56th session of JECFA was specifically di- and toxins occurring in herbs, such as pyrrolizidine al- rected to mycotoxins. The report addressed several con- kaloids, and anisatin in certain varieties of star anise. cerns about each mycotoxin: explanation of the myco- – Zootoxins: toxins produced by animals, e.g. snakes, toxin, absorption through excretion, toxicological studies scorpions and certain frogs. Zootoxins are generally of and final evaluation. It is interesting to note that, along lesser significance to human health by oral exposure, with the mycotoxin evaluations, JECFA put forward gen- although exceptions occur, such as bufotoxin, a toxin eral considerations on analytical methods, sampling, asso- excreted by Bufo marinus. Licking the head of this toad ciated intake issues and control [2, 3]. In particular, good (a dangerous practice of drug addicts in the Nether- surveillance data should be generated using validated an- lands) leads to hallucinations. alytical methods and reference materials to ensure reliable results. Methods used should be validated through collab- The first three mentioned categories of natural toxins are orative studies of analytical performance; however, the in fact bio-contaminants. They are produced by microor- committee recognized that it may not always be possible ganisms and may contaminate food and food products. to use a validated method due to expense or an official This is not the case with the last two categories, which are method being irreconcilable to a particular toxin–matrix inherent components of plants or animals. This article will combination. The committee also stressed the importance be restricted to the mycotoxins, the phycotoxins and the of laboratories participating in interlaboratory comparison plant toxins. These three categories of chemical hazards studies to ensure analytical quality assurance. are characterized by the World Health Organization In Europe various organizations are involved with risk (WHO) as significant sources of food-borne illnesses [1]. assessment of mycotoxins: Of these categories, most research has been carried out on the mycotoxins, which are currently a major food safety – Both the Scientific Committee on Food (SCF) and the issue. Problems with mycotoxins are more severe than Scientific Committee on Animal Nutrition (SCAN) of with the other categories of toxins, due to the widespread the EC have regularly expressed scientific opinions about consumption of goods being contaminated. Various net- risks associated with the occurrence of mycotoxins in works of scientists are active, and the European Commis- food or animal feed, respectively. For example, SCAN sion (EC) is a major sponsor of European mycotoxin re- recently updated its opinion on undesirable substances search projects. Research on phycotoxins (in particular in animal feed [4], which included several mycotoxins the marine biotoxins) is of growing significance in Eu- for which the EC might establish new regulations in the rope where problems regularly lead to closures of shell- near future. fish harvesting areas. Also here networking exists in Eu- – The newly formed Scientific Panel on Contaminants in rope, and EC-funded research activities take place, but to the Food Chain of the recently established European a lesser extent than for the mycotoxins. Investigations on Food Safety Authority (EFSA) is specifically charged plant toxins still seem to have a lower priority in Euro- (among other issues) with mycotoxins. pean food safety research activities, and scientific net- –Efforts to assess exposure, one of the main ingredients working is in its infancy here. of risk assessment, are undertaken within SCOOP (SCOOP: Scientific Cooperation on Questions relating to Food) projects, funded by the EC. SCOOP projects Mycotoxins on mycotoxins are targeted to make the best estimates of intake of these toxins by EU inhabitants. In the 1990s Risks and regulations these activities resulted in a report on the exposure as- sessment of aflatoxins [5]; later SCOOP reports for Mycotoxins are secondary metabolites of fungi, which are several other mycotoxins followed (for ochratoxin A, capable of producing acute toxic, carcinogenic, muta- patulin, trichothecenes, fumonisins and zearalenone) genic, teratogenic, immunotoxic and oestrogenic effects [6, 7, 8]. SCOOP data were used by the SCF for its on animals at the level of exposure. Biological conversion evaluation and advisory work on the risks to public products of mycotoxins are also referred to as mycotox- health arising from dietary exposure to certain myco- ins. Toxic syndromes resulting from the intake of myco- toxins. toxins by man and animals are known to as “mycotoxi- – Some projects in the EC’s fifth Framework Programme coses”. are directed towards determining the risk of certain my- For the mycotoxins currently considered most signifi- cotoxins (e.g. the ongoing project “Ochratoxin A risk as- cant (aflatoxins, ochratoxin A, patulin, fumonisins, zear- sessment: mechanisms of ochratoxin A-induced carcino- alenone and some trichothecenes including deoxynivalenol), genicity as a basis for an improved risk assessment”, the Joint Expert Committee on Food Additives (JECFA), supported by the Quality of Life Programme [9]). 1154 – ILSI-Europe, the European branch of the International ten countries, harmonized regulations or detailed guide- Life Sciences Institute (a non-profit-making, worldwide lines exist for aflatoxins in various foodstuffs, aflatoxin foundation established to advance the understanding of M1 in milk, ochratoxin A in cereals and dried vine fruits, scientific issues relating to nutrition, food safety, toxi- patulin in several fruit products, deoxynivalenol in cereals cology, risk assessment and the environment), has a and cereal products and for aflatoxin B1 in various feed- working group on natural toxins that organizes interna- stuffs [11]. In the future (2004 and following years) a sig- tional symposia on mycotoxins of European concern, nificant further expansion of EU-harmonized mycotoxin e.g. on ochratoxin A in 1996 and on trichothecenes in regulations is expected both for foods and feeds, and lim- 2003. its will be lowered for some mycotoxins. For foods this concerns aflatoxin B1, ochratoxin A and deoxynivalenol The knowledge that mycotoxins can have serious adverse (DON) in infant formulae and follow-up formulae, ochra- effects on man and animals has led many countries to es- toxin A in coffee, wine, beer, spices, grape juice, cocoa tablish regulations which limit the presence of these com- and cocoa products, and several Fusarium-produced my- pounds in food and feed. To date, there are more than 100 cotoxins, i.e. trichothecenes (nivalenol, T-2 and HT-2 tox- countries in the world which have specific limits for my- ins, in addition to DON), fumonisins and zearalenone in cotoxins in foodstuffs and feedstuffs. The total population cereal-based foodstuffs. For feeds, specific limits are to be in these countries represents approximately 90% of the expected for ergot alkaloids, zearalenone, trichothecenes world’s inhabitants [10]. In Europe, practically all nations and ochratoxin A, and possibly fumonisins. were known to have specific mycotoxin regulations in The developments described above, and in particular 2002. Compared to the other continents, Europe has the those on limits and regulations, have dictated and stimu- most extensive and detailed regulations for mycotoxins in lated the analytical community to develop, validate and food and feed (see Figs. 1 and 2). In the EU, now consist- standardize suitable analytical methodology, to develop ing of 15 countries but soon to be expanded with another (certified) reference materials and to establish analytical

Fig. 1 Mycotoxins regulated in food in Europe (situation 2003). AFT: total aflatoxins B1, B2, G1 and G2; AFB1: aflatoxin B1; AFB1/G1: aflatoxins B1 and G1; AFM1: aflatoxin M1; OTA: ochratoxin A; PAT: pat- ulin; DON: deoxynivalenol; T-2: T-2 toxin; DAS: diace- toxyscirpenol; ZEN: zearale- none; FUM 1/2: fumonisins B1 and B2; FUM 1: fumonisin B1; STE: sterigmatocystin

Fig. 2 Mycotoxins regulated in feed in Europe (situation 2003). Abbreviations: see Fig. 1 1155 quality assurance programmes such as international profi- The foreseen expansion in the European mycotoxin ciency tests. regulatory area (see previous paragraph) demands ongo- ing activities of European research groups to develop and validate suitable methodology. Alternative methods dedi- Analytical methods and reference materials cated to special needs, e.g. the procurement of precise es- timates of exposure of certain groups of the population Current analytical methods for mycotoxins are based on (e.g. babies), and low-cost methods widely applicable for chemical procedures. Test portions are usually extracted rapid screening at an early stage in the food/feed chain with (combinations of) organic solvents and water. Purifi- (e.g. based on TLC or ELISA), will be needed to supple- cation of the extracts to remove lipids and other sub- ment the reference methods mentioned above. There exist stances is often done by leading the extracts through chro- interesting newer developments in mycotoxin methodol- matography columns or pre-packed cartridges. The latter ogy, for which the application would need to be further are commercially available with many types of adsorbents explored, tested for performance and compared with the and in many formats, to suit the needs of the analyst. A existing conventional methods. These newer methods in- purification technique, which is widely applied now in clude the powerful LC–MS(–MS) (mass spectroscopy) mycotoxin methodology, is based on immunoaffinity clean- and GC–MS benchtop techniques, capillary zone electro- up of extracts. Immunoaffinity columns are composed of phoresis, cell tests, biosensors, MIP (molecularly im- monoclonal antibodies, specific for the toxin of interest, printed polymers), fluorescence polarization immunoas- which are immobilized on Sepharose and packed into says, dipstick assays and non-invasive methods based on small columns. These immunoaffinity columns have suf- infrared and acoustic techniques. Worthwhile to read are ficient capacity to clean up heavily contaminated samples, the yearly published reports of the General Referee for and they usually efficiently remove compounds that could Mycotoxins of AOAC International, e.g. the one pub- interfere in the determination, because the antibodies lished in 2003 [16]. These reports provide insight into the specifically recognize the toxin of interest. The purifica- analytical developments in the mycotoxin area and con- tion techniques described can be incorporated in fully au- tain many references for further reading. tomated sample preparation systems that take the sample LC–MS–MS is rapidly gaining ground in routine labo- from the extraction stage through to completion of liquid ratories as a method which enables the simultaneous de- chromatography (LC) in an unattended mode of action. In tection of several mycotoxins in one run. One of the most addition to LC, separation techniques such as thin-layer recent developments includes the application of a multi- chromatography (TLC), gas chromatography (GC) and toxin immunoaffinity column, which allows the determi- enzyme-linked immunosorbent assay (ELISA) are among nation of the aflatoxins, deoxynivalenol, ochratoxin A, zear- the ones most often used in mycotoxin methodology. alenone and the fumonisins in one run [17]. An example Several immunoaffinity-based mycotoxin methods have of a very sensitive GC–MS method is the one recently de- been formally validated in the last five years in European veloped in our laboratory for the determination of de- collaborative studies for various mycotoxin/matrix com- oxynivalenol (DON) and other trichothecenes in toddler binations. These studies were conducted in agreement food. The method allows one to estimate the real exposure with the protocol for the design, conduct and interpreta- of toddlers (age category 9 months to 2 years) to tri- tion of method-performance studies (the “IUPAC/AOAC/ chothecenes, by analysing samples of 24-h duplicate diets ISO harmonized protocol”) [12], and most of them have of toddlers. The method has a limit of detection for DON also been approved by AOAC International, which has of ~0.3 µg/kg, which has been shown to be sufficiently made them widely acceptable as tools for regulatory low to find the first few dozen samples positive for DON analysis. Many of the European methods were standard- [18]. This newly developed method has shown itself to be ized by working group 5 “Biotoxins” of CEN Technical useful for the detection of trichothecenes at very low lev- Committee 275: “Horizontal methods for food” [CEN= els, but quantification of DON at these low, (sub-)µg/kg Comité Européan de Normalisation (European Standard- levels is required only for special studies, because (pro- ization Committee), the European equivalent of ISO (In- posed) legal limits for DON are usually in the range ternational Standardization Organization)]. They are now 500–1000 µg/kg [11]. For rapid screening of cereals for available as reference methods. The CEN work was and is DON other, much faster, methods are to be preferred, e.g. done in response to the needs of the EC to make enforce- ELISAs or methods based on biosensors. An overview of ment possible of the new or proposed EU-harmonized reg- ELISAs for mycotoxins is available through the website ulations (with tolerance limits that are among the lowest of the European EMAN project (EMAN=European My- in the world). CEN has also produced a document that pro- cotoxin Awareness Network) [19]. An example of a re- vides criteria for the performance of mycotoxin methods cently developed biosensor-based method is the one by [13]. This document gives information concerning the per- Tüdos et al. [20] for the selective and quantitative deter- formance of analytical methods to determine mycotoxins mination of DON. This biosensor assay is of the surface in certain matrices that can be expected from experienced plasmon resonance (SPR) type, where DON conjugate is analytical laboratories. Some of these CEN criteria are re- immobilized on the sensor. The analysis results of the op- flected as method-performance requirements in official EU timized biosensor assay and an (in-house validated) legislation, e.g. on aflatoxins and ochratoxin A [14, 15]. LC–MS–MS method, practiced by the Netherlands Food 1156 Inspection Service (Amsterdam, the Netherlands)(see also [17]), were compared for naturally contaminated wheat Phycotoxins samples with DON levels ranging from 200–1400 µg/kg. The results indicated good agreement between the two Risks and regulations techniques, especially for wet-ground samples. Low-cost Phycotoxins are toxic components produced by unicellu- biosensor-based methods are currently developed for lar microalgae. They may enter fishery products through aflatoxin B and ochratoxin A (target costs: approx. 15 € 1 feed chains in the aquatic environment. Phycotoxins may per analysis) within the ongoing EC CRAFT project demonstrate various (mostly acute) toxic effects. Human “MycoSense” [21]. intoxications caused by phycotoxins occur worldwide Many analytical methods for mycotoxins have been through consumption of marine fishery products in which validated through collaborative studies, and a relevant over- the toxins have accumulated. Toxins of freshwater algae view was recently published [22]. However, the availabil- have induced intoxications in farm animals and they can ity of modern and reliable methods of analysis for myco- threaten human health, e.g. through drinking water or toxins is no guarantee of accurate results. Proficiency consumption of freshwater algae. Among the most impor- tests for aflatoxins and ochratoxin A, organized by the In- tant marine phycotoxins are the shellfish poisons and the ternational Agency for Research on Cancer in the past, ciguatera toxins. The latter accumulate in finned fish. Ac- have shown that wide scatter of analytical results is of lit- cording to their toxic effects, the shellfish poisons are usu- tle comfort to those who must either pay for the measure- ally divided into four groups: the paralytic shellfish poi- ments, or base potentially important decisions upon them soning (PSP) toxins, the diarrhoeic shellfish poisoning [23]. Analytical quality assurance (AQA) is needed to (DSP) toxins, the amnesic shellfish poisoning (ASP) tox- demonstrate acceptable analytical performance. For AQA ins, and the neurotoxic shellfish poisoning (NSP) toxins. purposes (certified) reference materials for mycotoxins have Contrary to the mycotoxins, for which chronic effects been developed in EC programmes and proficiency tests are of particular concern, the danger of phycotoxins is for mycotoxins are regularly organized. The (certified) mainly through their acute effects. Because toxin produc- reference materials have been developed since the 1980s tion and accumulation occur only under certain environ- under the auspices of the BCR Programme (BCR: Bureau mental conditions that stimulate algal blooms, marine Communautaire de Référence, or in English, Community fishery products may become incidentally contaminated Bureau of Reference), later known as the Standards, Mea- with (relatively) high concentrations. At present the risk surements and Testing (SMT) Programme. Currently avail- assessment of phycotoxins has not been performed in a able BCR reference materials include aflatoxin M in milk 1 straightforward way. Risk assessments have often been powder, aflatoxin M calibrant, total aflatoxins in peanut 1 based on pragmatic decisions taken during intoxication butter, aflatoxin B in peanut meal, aflatoxin B in feed- 1 1 events with risk management mixed into the process, stuff, ochratoxin A in wheat, DON in maize and wheat, complicating the procedure. The few data that exist about zearalenone in maize and zearalenone calibrant. A project human toxicity of phycotoxins leave the impression that is ongoing on the development of trichothecene calibrants. the margin of safety between exposure at a regulatory The mycotoxin reference materials are currently available limit and human effect levels is often rather small, espe- from the EC’s Joint Research Centre, Institute for Refer- cially for PSP and DSP toxins [25]. Until now formal risk ence Materials and Measurements, Geel, Belgium (see http:// assessments, e.g. by JECFA, have not been carried out. www.irmm.jrc.be). However, such activities would be highly recommended Certified reference materials are relatively expensive, to get insight into the real margins of safety, although it is due to the huge amount of work involved with their de- questionable whether enough reliable toxicological data velopment, and supplies are limited. Therefore laborato- exist, particularly on repeated exposure. If adequate scien- ries are advised to develop their own in-house reference tific (toxicological, epidemiological and occurrence) data materials (homogeneous and stable) for routine use, the were available, risk assessment would need to be per- toxin content of which should be established on the basis formed by applying generally accepted safety or uncer- of the certified materials. tainty factors. Besides the application of (certified) reference materi- The fact that marine phycotoxins pose a threat to human als, regular participation in interlaboratory comparisons, health led a number of countries to establish regulations to such as proficiency testing schemes, is becoming increas- control the presence of these toxins in seafood. Several ingly important as part of AQA measures. A number of dozens of countries have specific limits for marine phyco- proficiency testing schemes for mycotoxins exist at the in- toxins in shellfish and fish [26]. Also, for microcystins (tox- ternational level, including those organized by the Food ins produced by freshwater algae) a guideline limit was re- Analysis Performance Assessment Scheme (FAPAS), op- leased by the WHO for drinking water [27]. In the EU the erated from Europe by the Central Science Laboratory in following harmonized limits and analytical requirements the UK, and those organized by the American Oil exist for ASP toxins [28], DSP toxins and related com- Chemists’ Society (AOCS), operated from the USA. The pounds [29], PSP toxins [30] and ciguatera toxins [31]: experiences of FAPAS with mycotoxin analysis were re- cently published [24]. – For ASP: limit 20 mg/kg; an LC–UV method has to be used. 1157 – For PSP: limit 80 µg/100 g; a biological assay has to be vantages of this assay are the lack of specificity (no dif- used. A chemical assay is allowed for saxitoxin, but if ferentiation between the various components of DSP tox- the results are challenged, the bioassay is the reference ins, whereas in EU legislation specific limits have been method. set for the various toxins, see risks and regulations), the –For DSP and related compounds: various limits are ap- subjectivity of death time of the animals and the main- plied: taining and killing of laboratory animals. In addition, this – Okadaic acid+dynophysis toxins+pectenotoxins: assay is time consuming and expensive, may give false 160 µg okadaic acid equivalents/kg positives because of interferences by other lipids (e.g. free –Yessotoxins: 1 mg yessotoxin equivalents/kg fatty acids) and shows variable results between whole- – Azaspiracids: 160 µg/kg body and hepatopancreas extracts [25]. In the EU a rat bioassay is also officially allowed to be used. This assay is A biological method is to be used, (bio)chemical alterna- based on diarrhoea induction in rats [32]. The (starved) tives are allowed, if formally validated. animals are fed with suspect shellfish tissue (mixed into – For ciguatera toxins: placing on the market of fishery the diet) and observed during 16 h for signs of diarrhoea, products containing ciguatera toxins is not allowed, no consistency of the faeces and food refusal. The method is information about analytical methods is given. at best semi-quantitative and does not detect pectenotox- ins and yessotoxins. The test is still used routinely in the As for the mycotoxins, the problems caused by phycotox- Netherlands [25], as it is less animal-unfriendly than the ins and the existence of specific regulations has led to the mouse assay. development of analytical methodology and (certified) The in vivo tests can in principle be replaced by in reference materials. The progress has been relatively vitro tests, provided that there is analytical proof through slow, however, due to the chronic lack of sufficiently pure convincing validation studies that these alternatives (e.g. materials which can be used as calibrants. Also new phy- physicochemical, functional or immunochemical meth- cotoxins are regularly discovered, e.g. during intoxication ods) fulfil reasonable performance criteria. In the EU, events, which require the immediate attention of the sci- Commission Decision 2002/225/EC (which lays down entific community, without the possibility that extensive maximum levels for specific marine toxins within the analytical experience can be built up and reliable refer- DSP complex and related compounds) [29] permits the ence materials can be developed. use of such alternatives, if these are connected with a val- idation of the methods following an internationally ac- Analytical developments and reference materials cepted protocol. At the time of writing such validated methodology hardly exists. This is partly due to the fact A variety of analytical techniques are available for the de- that, at present, reference standards and reference materi- tection and determination of marine biotoxins, but animal als are not available or are in short supply for all the indi- tests still play a major role, in particular in the control of vidual components of the DSP toxin groups. In addition, PSP and DSP toxin events. In the EU, a mouse bioassay is the conducting of a formal interlaboratory validation study the reference method for PSP determination, and a mouse is accompanied by a considerable investment of man- or rat assay is to be used for the determination of toxins power, time and money. belonging to the DSP complex. For PSP, a chemical method In an attempt to stimulate the development of accepted for saxitoxin may be used in addition to the mouse assay alternatives to the in vivo assays for PSP and DSP toxins according to the relevant EU directive [30], but a problem currently practised in Europe, a recommendation with here is that the PSP complex comprises about 20 different supportive and convincing documentation has recently toxins, and in European waters saxitoxin is often not the been issued by representatives of governmental institu- prevalent PSP toxin. If the results are challenged, the tions of Germany, the UK and the Netherlands to the mouse bioassay has to be used, but a serious limitation members of the ECVAM (European Centre for the Valida- here is that this test is formally not allowed in several Eu- tion of Alternative Methods) Scientific Advisory Com- ropean countries, for animal welfare reasons. The mouse mittee (ESAC) [33]. This has had a positive influence on assay requires the injection of shellfish extract by the in- the contents of the call for project proposals in the 6th EU traperitoneal route. The shellfish extract is crude and ani- Framework Programme for Research and Technological mals suffer severely, as the PSP toxins act as neuromus- Developments on the activities of the EU National Refer- cular blocking agents, causing animal death in 5–7 min. ence Laboratories for Marine Biotoxins (see below). In the mouse assay applied to detect toxins of the DSP Which possibilities, alternative to rodent tests, do exist complex, extracts are also injected intraperitoneally, and to determine marine biotoxins and which European orga- survival time is checked from 24 to 48 h. In this mouse as- nizations can play a significant role in the further devel- say all DSP components are likely to be detected, includ- opment and validation of these methods? Various methods, ing those DSP toxins which do not cause diarrhoea based on different principles, have been explored and pub- (pectenotoxins and yessotoxins) and have an unknown lished. These include techniques such as LC with UV and toxicity for humans. Also other unknown toxin groups ex- fluorescence detectors, LC–MS, capillary zone electro- hibiting ichthyotoxic and haemolytic properties may phoresis, capillary electrochromatography, enzyme activ- cause mortality of mice in this bioassay. The major disad- ity tests, receptor binding assays, cell toxicity tests, solid- 1158 phase immunobead assays and biosensors. LC–MS is cur- http://www.imb.nrc.ca/techdev/crmp_e.html). In Europe rently one of the most interesting techniques which has ap- efforts undertaken by the EC’s SMT programme have led plications in the determination of several of the shellfish to lyophilized mussel reference materials with certified toxins, e.g. the PSP, for which a new type of LC column mass fractions of some of the toxicologically most signif- (hydrophilic interaction) is used [34]. A detailed overview icant PSP toxins [39], which are stored at the IRMM in of the various approaches that are possible for the determi- Geel, but their distribution is severely hampered by the nation of the different shellfish poisoning toxins and the fact that saxitoxin is a schedule 1 chemical in the list of ciguatera toxins, with their advantages and disadvantages, the Chemical Weapon Convention. For the other shellfish is given in the draft FAO Food and Nutrition Paper “Ma- toxins no European activities to develop certified refer- rine biotoxins – a review” [25]. The number of methods ence materials are currently undertaken or foreseen. that have been validated through formal collaborative Despite the positive developments with the activities studies is rather scarce though, for reasons explained undertaken by the NRC, the availability of reference ma- above. This makes it difficult to give an opinion about the terials for phycotoxins remains limited. This has an im- practical usefulness of many of these procedures. pact not only for the analytical activities on phycotoxins, In Europe, two networks of scientists are worth men- but it is also a limiting factor in the development of reli- tioning in respect of joint efforts to advance the develop- able risk assessments for these compounds. ment of analytical methods [35]. Firstly there is the net- work of EU National Reference Laboratories (NRL) for Marine Biotoxins, coordinated by the Community Refer- Plant toxins ence Laboratory (CRL) for Marine Biotoxins in Vigo, cre- ated in 1994. Secondly there is working group 5 “Biotox- Risks and regulations ins” of CEN Technical Committee 275, also active in the area of mycotoxins (see previous section on mycotoxins), Inherent plant toxins are naturally occurring components which took the phycotoxins subject on board in 1999. The in plants that are toxic and/or have negative effects on the work done by this CEN working group follows the same bioavailability of nutrients. Toxins in the latter category pattern as described for the mycotoxins. are often referred to as antinutritional factors or antinutri- Typical examples of CRL-coordinated activities car- ents. Some plant toxins exhibit chronic effects, others acute ried out in the last few years within the NRL network in- effects, or both. Some toxins produced by the plant play a clude the conduct of collaborative studies of methods to role in the defence system against bacteria, fungi, insects, determine ASP and DSP toxins (ongoing in 2003), and in- or other threats. In general, the lack of essential data on tercalibration studies on DSP [36] and PSP toxins (ongo- the toxicological properties and occurrence of inherent ing in 2003). plant toxins does not allow for an adequate safety evalua- A typical example of an NRL-coordinated activity is tion. Some important staple foods form inherent toxins, the organization of a series of proficiency tests among such as potatoes that contain the glycoalkaloids α-sola- Dutch laboratories involved in analysis of marine biotox- nine and α-chaconine. Because of the virtual absence of ins [37].The interaction between members of the EU NRL chronic toxicity data, an adequate no-observed-adverse- network takes place through annual meetings and by effect level (NOAEL) for potato glycoalkaloids has not means of working groups (WGs) hosted by the European been assessed, and a tolerable daily intake (TDI) for hu- Commission. The outcomes of some of the WGs in the mans could not be determined by JECFA [40]. The com- way of conclusions and/or recommendations have been used mittee was only able to recognize that the development of as pre-normative studies to elaborate new regulations. empirical data to support a safe level would require con- The increasing demand for quality assurance has led siderable effort. several organizations to undertake proficiency tests, in ad- Another emerging food safety problem is the toxins oc- dition to those incidentally undertaken as described above. curring in herbs. Herbs are increasingly used, and newer There are now two European organizations starting to un- ways for the consumer to order herbs, such as through the dertake proficiency tests for shellfish toxins on a regular Internet, make it difficult for food inspection services to scale: FAPAS (see section on mycotoxins) and Quasi- keep control over the situation. Examples of toxins occur- meme, operated from the FRS Marine Laboratory in Ab- ring in herbs are anisatin in star anise and the large group erdeen, UK. Both FAPAS and Quasimeme have existed of the pyrrolizidine alkaloids. In the Netherlands in late for a long time, but their activities in the marine biotoxin 2001, a few dozen people got severely ill due to the con- area are just starting for PSP (FAPAS) [38] and ASP sumption of a blended herbal tea, in which star anise was (Quasimeme). It is expected that these programmes will mixed. Instead of the Chinese variety (Illicium verum, expand their activities in the future. containing the desired substance trans-anethole), the Jap- Similarly as for the mycotoxins, efforts have been un- anese variety (Illicium anisatum, containing anisatin, a dertaken to develop (certified) reference materials for sesquiterpene dilactone) was accidentally blended into the phycotoxins. In 2003, certified calibrants and some refer- tea. All victims survived. As a consequence, the EC es- ence matrices were commercially available for domoic acid tablished a decision in 2002 on the import of star anise (ASP toxin), and for some of the PSP and DSP toxins, from Third World countries [41], which should avoid the from the National Research Council (NRC) Canada (see mistakes described. 1159 The pyrrolizidine alkaloids form a group of 200 differ- ionization detector. The method is able to determine ent compounds, occurring in herbs, herbal teas and com- senkirkine and senecionine at 0.1 and 0.02 µg/ml juice, re- frey. They are a possible under-recognized cause of liver spectively. This limit of determination may not be low cancer. They occur at levels that are too low to produce enough in European situations where legal limits have been acute liver damage but high enough to be of concern as a set at very low levels, such as in the Netherlands (1 ng/g possible long-term cause of cirrhosis and liver failure, es- in herbs). The methodology may need to be further re- pecially in children [42, 43]. Some toxicity data are avail- fined, formally validated and standardized before it can able, but further studies on the exposure and toxicity of become a generalized tool in regulatory analysis. The many pyrrolizidine alkaloids are required to make a mean- chronic lack of reliable analytical standards and reference ingful risk assessment possible. Incidentally, European coun- materials has contributed to a situation where the progress tries (e.g. Germany and the Netherlands) have established in plant toxin analytical methodology is slow. A favourable regulations for pyrrolizidine alkaloids in herbs, but the exception in this regard is the glucosinolates. Glucosino- limits differ significantly. Contrary to the mycotoxins and lates, toxins mainly occurring in the family of the Cru- phycotoxins, there are no efforts yet within the EU to har- cifera, are among the very few plant toxins for which the monize the regulatory area, a situation which is not only analytical methodology is rather advanced, and for which the case for the pyrrolizidine alkaloids, but also general BCR certified reference materials have become available for the other plant toxins. This European situation is not [46], distributed through the IRMM (see also http://www. exceptional. Only few countries in the world have regula- irmm.jrc.be). Another factor hampering analytical progress tions for certain inherent plant toxins. The fact that the is the lack of a European network of food analysts to new EFSA Scientific Panel for Contaminants in the Food specifically deal with plant toxins. This situation was rec- Chain has specifically named the area of “natural toxi- ognized by the CEN working group 5 “Biotoxins” of CEN cants” as one of its fields of attention, may point at a Technical Committee 275. This working group, very ac- growing interest in the EU. tive and successful in the areas of mycotoxins and phyco- The above-mentioned groups of toxins are just exam- toxins (see previous sections), has recently decided to in- ples of plant toxins. There are many more plant toxins of clude the item “plant toxins” in their work programme. concern, but these cannot all be discussed here. Despite this welcome development, it will probably take several more years before the first CEN-standardized method for plant toxins will become available. Analytical developments and reference materials

Inherent plant toxins are quite variable in their chemical Conclusion structures and physicochemical behaviour, so there exist a variety of analytical chemical approaches to determine There is a growing awareness that some natural compounds them. Compared to biocontaminants, such as mycotoxins, in foodstuffs may play an etiological role in or contribute the development of analytical methodology for the plant to the occurrence of human diseases. This has led to an in- toxins has been a rather neglected area, and validated creased interest of the scientific community in the natural methods of analysis are rather scarce. One of the few ex- toxins. For mycotoxins the situation is the most advanced. ceptions is an LC–UV method for potato glycoalkaloids, Several risk assessments have been carried out and de- which was validated in an AOAC collaborative study tailed regulations have been established or are in the [44]. The lack of formally validated methods is partly due process of establishment. Various validated methods and to the fact that only few countries have official regulations certified reference materials have been developed, and for inherent plant toxins, so that there is not much legal much scientific activity is going on in Europe. For the pressure to have adequate analytical methods available. In phycotoxins the situation is less advanced, but some regu- the future, growing interest in the food safety aspects of lations exist; there are promising developments in the area plant resistance may lead to the development of regula- of analysis and reference materials, and active networks tions for certain plant toxins and, inevitably, to the further exist in Europe. For the plant toxins the situation is clearly development of analytical methods to make enforcement the weakest in terms of analytical methods and reference of the regulations possible. materials; regulations are scarce, and European networks The methods that are applied in practice to determine have not come to activity yet. plant toxins are often based on chromatographic proce- dures, such as LC in combination with UV and MS detec- tors, GC in combination with FID and MS detectors, TLC, References immunochemical procedures and, incidentally, NMR. An example of a GC–FID method is that of Wiedenfeld [45], 1. World Health Organization (2002) WHO global strategy for used for the determination of pyrrolizidine alkaloids such as food safety: safer food for better health. Food Safety Pro- senecionine and senkirkine in herbal juices. In the method gramme-2002. World Health Organization, Geneva 2. World Health Organization (2002) Evaluation of certain myco- an extract is purified over a Chromabonddiol® column, toxins in food. Fifty-sixth report of the Joint FAO/WHO Expert and then separated over a diphenyldimethylpolysiloxane Committee on Food Additives, WHO Technical Report Series capillary GC column. Detection takes place with a flame 906. World Health Organization, Geneva 1160 3. Food and Agriculture Organization (2001) Safety evaluation of 25. Van Apeldoorn ME, Van Egmond HP, Speijers GJA (2004) certain mycotoxins in food. Fifty-sixth meeting of the Joint Marine biotoxins: a review. Draft FAO Food and Nutrition Pa- FAO/WHO Expert Committee on Food Additives (JECFA), per. National Institute for Public Health & the Environment, FAO Food and Nutrition Paper 74. Food and Agriculture Orga- Bilthoven nization of the United Nations, Rome 26. Van Egmond HP, Speijers GJA, Van den Top HJ (1992) J Nat 4. European Commission (2003) Mycotoxins. In: Opinion of the Toxins 1:67–85 Scientific Committee on Animal Nutrition on undesirable sub- 27. World Health Organization (1998) Guidelines for drinking wa- stances in feed (adopted on 20 February 2003). European Com- ter quality, 2nd edn. Addendum to vol 2. Health criteria and mission, Health & Consumer Protection Directorate-General, other supporting information. World Health Organization, Brussels, pp 6–24 Geneva 5. European Commission (1997) Reports on tasks for scientific 28. European Commission (2002) Commission Decision 2002/ cooperation. Report of experts participating in Task 3.2.1. Risk 226/EC, 15/03/2002, OJ L 75:65–66 assessment of aflatoxins. Report EUR 17526 EN, Directorate- 29. European Commission (2002) Commission Decision 2002/ General for Industry, Office for Official Publications of the Eu- 225/EC, 15/03/2002, OJ L 75:62–63 ropean Communities, Luxembourg 30. European Commission (1991) Council Directive 91/942/EEC, 6. Miraglia M, Brera C (2002) Assessment of dietary intake of 15/07/1991, OJ L 268:1-14 ochratoxin A by the population of EU member states.http:// 31. European Commission (1991) Council Directive 91/943/EEC, europa.eu.int/comm/food/index_en.html. Cited Jan 2002 22/07/1991, OJ L 268:15–34 7. Majerus P, Kapp K (2002) Assessment of dietary intake of pat- 32. Kat, M (1983) Antonie Van Leeuwenhoek 49:417–419 ulin by the population of EU member states.http://europa.eu. 33. Grune B, Tiebach R, Sauer U, Anderson D, de Leeuw W, Har- int/comm/food/index_en.html. Cited Mar 2002 tung T (2003) Letter of 14 February 2003 to the members of 8. Gareis M, Schothorst RC, Vidnes A, Bergsten C, Paulsen B, the 19th ESAC meeting on 28 February 2003. Shellfish toxins Brera C, Miraglia M (2003) SCOOP task 3.2.10. Collection of testing – in urgent need of application of 3Rs occurrence data of Fusarium toxins in food and assessment of 34. Quilliam MA, Hess P, Dell’Aversano C (2001) Recent devel- dietary intake by the population of EU member states. http:// opments in the analysis of phycotoxins by liquid chromatogra- www.europa.eu.int/comm/food/fs/scoop/task3210.pdf phy–mass spectrometry. In: de Koe WJ, Samson RA, Van 9. Mechanisms of ochratoxin A-induced carcinogenicity as a ba- Egmond HP, Gilbert J, Sabino M (eds) Mycotoxins and phyco- sis for an improved risk assessment (2003).http://www.uni- toxins in perspective at the turn of the millennium. WJ de Koe, wuerzburg.de/toxikologie/EU-OTA/OchratoxinA.html Wageningen, pp 383–391 10. Van Egmond HP, Jonker, MA (2003) Mycotoxins and regula- 35. Fernandez ML, Van Egmond HP (2003) European approaches tions. In: Final programme, abstracts of lectures and posters, to marine toxin control and harmonization. In: Proceedings of the Second World mycotoxin forum, Noordwijk aan Zee, 17– the 2nd conference on harmful algal management and mitiga- 18 February 2003, p 18 tion, Qingdao, November 2001 (submitted for publication) 11. Van Egmond HP, Jonker, MA (2004) Worldwide regulations 36. Community Reference Laboratory on Marine Biotoxins (2001) for mycotoxins in food and feed: the situation in 2003. Draft Report on the EU-NRLs intercalibration exercise on DSP de- FAO Food and Nutrition Paper. National Institute for Public termination, April 2001. Community Reference Laboratory on Health & the Environment, Bilthoven Marine Biotoxins, Vigo 12. Horwitz W (1995) Pure Appl Chem 67:331–343 37. Van Egmond HP, Jonker KM, Poelman M, Scherpenisse P, 13. Comité Européen de Normalisation (1999) Food analysis – Stern AG, Wezenbeek P, Bergwerff AA, Van den Top HJ biotoxins – criteria of analytical methods of mycotoxins. CEN (2004) Food Addit Contam (in press) report CR 13505 38. Food Analysis Performance Assessment Scheme (2003) FAPAS 14. European Commission (1998) Commission Directive 98/53/EC, marine toxins pilot study, August 2003. Central Science Labo- 16/07/1998, OJ L 201:93–101 ratory report, Sand Hutton, UK 15. European Commission (2002) Commission Directive 2002/ 39. Van Egmond HP, Mouriño A, Burdaspal PA, Bustos J, 26/EC, 13/03/2002, OJ L 75:38–43 Legarda T, Mesego A, Paulsch WE, Salgado C, Van den Top 16. Trucksess MW, Whitaker TB, Van Egmond HP, Wilson DM, HJ, Boenke A (1998) EUR report 18318. European Commis- Solfrizzo M, Abramson D, Dorner J, Ware GM, Maragos C, sion, Directorate-General Science, Research and Development, Hald B, Sabino M, Eppley RM, Hagler WM (2003) J AOAC Brussels Int 86:129–138 40. World Health Organization (1993) Solanine and chaconine. In: 17. Krol J, Siantar D (2003) Abstract in: Final programme, 117th Toxicological evaluation of certain food additives and naturally AOAC international annual meeting & exposition, Atlanta, occurring toxicants. 39th meeting of the JECFA. WHO Food 14–18 September 2003, AOAC, p 110 Additives Series 30. World Health Organization, Geneva 18. Schothorst RC (2003) (personal communication) 41. European Commission (2002) Commission Decision 2002/ 19. European Mycotoxin Awareness Network (2003) http://www. 75/EC, 01/02/2002, OJ L 33: 31–34 mycotoxins.org 42. Wiedenfeld H (2002) Problems with pyrrolizidine alkaloids. 20. Tüdos AJ, Lucas-van den Bos HR, Stigter ECA (2003) Sub- Background document prepared for working group 5 “Biotox- mitted for publication ins” of CEN Technical Committee 275 21. Mycosens (2003) http://www.alphabvt.com/html/deutsch/myco. 43. Edgar JA (2003) Chem Aust, May 2003 php 44. AOAC International (2000) AOAC official methods of analy- 22. Gilbert J, Anklam E (2002) Trend Anal Chem 21:468–486 sis. AOAC official method 997.13. Glycoalkaloids (α-solanine 23. Friesen MD (1989) Report on the statistical analysis of results and α-chaconine) in potato tubers. In: Trucksess MW (ed) Nat- obtained for the analysis of ochratoxin in wheat flour. Myco- ural toxins. AOAC, Gaithersburg, pp 63–65 toxin sample survey programme. International Agency for Re- 45. Wiedenfeld H (2003) University of Bonn (personal informa- search on Cancer Report EC/92/15–2, Lyon tion) 24. Richard JL, Payne GA, Desjardins AE, Maragos C, Norred 46. Wathelet JP, Marlier M, Severin M, Boenke A, Wagstaffe PJ WP, Pestka JJ, Phillips TD, Van Egmond HP, Vardon PJ, (1995) The measurement of glucosinolates in rapeseeds. In: Whitaker TB, Wood G (2003) Mycotoxins: risks in plant, ani- Van Egmond HP, Visconti A, Boenke A, Speijers GJA (eds) mal, and human systems, CAST Task Force Report 139. Coun- Mycotoxins and toxic plant components, Lisbon, October cil for Agricultural Science and Technology, Ames, pp 101–103 1994. EUR report 16170 EN, Wiley, pp 299–304