16 Genetically Modified Foods: Potential Human Health Effects

16 Genetically Modified Foods: PotentialHuman Health Effects

A. Pusztai,1* S. Bardocz1 and S.W.B. Ewen21Formerly of The Rowett Research Institute, Aberdeen, UK; 2Department of

Pathology, University of Aberdeen, Forresterhill, Aberdeen, UK

Introduction

The scope of this review is restricted todata-based considerations about the safety ofgenetically modified (GM) foods of plant ori-gin for health. No opinions unless supportedby experimental results will be discussed.The emphasis will be on papers published inpeer-reviewed journals. A few articles will bementioned from non-peer-reviewed journalsbut only if they influenced the developmentof science-based ideas for the regulatoryprocess. Environmental issues will not bedealt with.

Safety evaluation of whole foods derivedfrom crops with considerable natural varia-bility is more difficult than that of a singlechemical, pharmaceutical or food additive, ordefined mixtures of them. Published results oftests for toxicity and nutritional wholesome-ness of complex foodstuffs are therefore fewand far between. A recent comment in Sciencedescribed this in its title: ‘Health risks ofgenetically modified foods: many opinionsbut few data’ (Domingo, 2000). Even a cursorylook at the list of references of a recent majorreview on food safety issues (Kuiper et al.,2001) showed that most of the publicationsreferred to were non-peer-reviewed institu-tional opinions or envisaged future scientific

and methodological developments for safetyassessments, but were short on actual pub-lished scientific papers on which a reliabledatabase of safety could be founded. Judgingby the absence of published data in peer-reviewed scientific literature, apparently nohuman clinical trials with GM food have everbeen conducted. Most attempts to establishthe safety of GM food have been indirect. Atbest, inferences have been drawn from animaltrials, but the preferred approach is to usecompositional comparisons between the GMfoodstuff and its traditional counterpart. Ifthese results show no significant differences,the two foodstuffs are ‘substantially equiva-lent’, meaning that the GM food is as safe asthe non-GM food. Thus, as the regulation isalmost exclusively dependent on ‘substantialequivalence’, the published results of GMfood analyses and inferences drawn fromthem for health will be examined critically inthis review.

In genetic modification, the intendedgene is incorporated into the genome of a cropusing a vector containing several other genes,including, as a minimum, viral promoters,transcription terminators, antibiotic resis-tance marker genes and reporter genes.Although in GM food safety the role of theintended gene is very important, the potential

©CAB International 2003. Food Safety: Contaminants and Toxins(ed. J.P.F. D’Mello) 347

* E-mail: [email protected]

effects of these other genes need also to betaken into account because other parts of theconstruct or the insertion of the vector couldcontribute substantially to the overall effect(Ewen and Pusztai, 1999a). There is in factsome evidence that some of the other genes ofthe vector may have an effect on safety. This isparticularly so as it is now known that DNAdoes not always break down in the alimentarytract (Schubbert et al., 1994, 1998; Hohlwegand Doerfler, 2001). This opens up the pos-sibility that the antibiotic resistance markergene, in addition to others, may be taken up bybacteria in the digestive tract and contributeto the spreading of antibiotic resistance viahuman gut bacteria. In this context, one poten-tially important observation was that a sub-stantial proportion (6–25%) of a geneticallyengineered plasmid survived a 1-h exposureto human saliva (Mercer et al., 1999). Partiallydegraded plasmid DNA also successfullytransformed Streptococcus gordonii, a bacte-rium that normally lives in the human mouth.Saliva also contains factors which increase theability of bacteria to become transformed bynaked DNA. Therefore, the prospect of theuptake of undegraded or partially degradedvector genes, including the antibiotic resis-tance gene, will have to be seriously consid-ered. However, the main concern in GM foodsafety is what are the direct effects of theexpression of the main intended gene afterits insertion into the plant genome via a geneconstruct. An additional concern is that thismay also cause significant, indirect and unin-tended effects on the expression and function-ality of the plant’s own genes. The number ofcopies of the construct inserted and their loca-tion in the plant genome (pleiotropic effect)are of particular importance in this respect,with the possibility that many unexpectedchanges may occur. This possibility is in factgenerally accepted, and the inadequacy ofthe currently used methods to detect themis frequently acknowledged (Kuiper et al.,2001). Pleiotropic effects always occur withboth conventional cross-breeding and geneticengineering, and their unwanted conse-quences usually are eliminated by empiricallyselecting for the desired trait and discardingthe potentially harmful ones. Some of thesechanges are unpredictable and therefore we

can only compare the known properties andconstituents but cannot look for, or evenless analyse, unknown components. Thisimposes limitations on our selection criteria.Reliance based solely on chemical analysisof macro/micronutrients and known toxinsis at best inadequate and, at worst, danger-ous. More sophisticated analytical methodsneed to be devised, such as mRNAfingerprinting, proteomics and secondarymetabolite profiling (Kuiper et al., 1999).However, and most importantly, there isan urgent need to develop comprehensivetoxicological/nutritional methods to screenfor the unintended potentially deleteriousconsequences for human/animal healthof genetic manipulation to pinpoint theproblems in advance of the incorporation ofthe GM foodstuff into the food chain (Ewenand Pusztai, 1999b). Although some limitedanimal tests have been done, only a few ofthese have been published. However, datafrom some of these studies recently have beenplaced on the Internet. Although they werenot peer reviewed, they were incorporatedinto this review because of their potentialimportance for other scientists.

Non-peer-reviewed Safety Tests onCommercial GM Crops in the

Public Domain

FLAVR SAVR tomatoes

The first example of official safety evalu-ations of a GM crop, Calgene’s FLAVRSAVR tomato, including a 28-day rat feed-ing trial, was commissioned by Calgene forthe Food and Drug Administration (FDA)before its general release. Although thedetails of this study have never been pub-lished properly, because this work had suchan extraordinarily major effect and influ-enced GM food regulation in the USA andelsewhere, there is a compelling need to ana-lyse the methods used and the conclusionsreached. Fortunately, as a result of a courtcase in the USA (Alliance for Bio-Integrityet al. vs. Shalala et al.), most data in the FDA’sfiles are now on the Internet in the public

348 A. Pusztai et al.

domain and can therefore be evaluated(Alliance for Bio-Integrity, 1998).

This GM tomato study shows most ofthe problems which may be encountered inGM food safety evaluation, particularly if, likethe tomato, they are fruits rather than food-stuffs and their protein and energy contentsare insufficient for supporting the growth ofyoung animals. The methods used and resultsobtained in this study are important not onlyfor their own sake but also for their influenceon the process of regulation.

Substantial equivalence

As ‘substantial equivalence’ features soprominently in GM food regulation (Kuiperet al., 2001), including in this GM tomatostudy, there is a need to look more closelyat this concept. This issue has been dealt within some depth by a recent article (Millstoneet al., 1999) in which the problems with thisconcept were highlighted, such as that ‘sub-stantial equivalence’ has never been definedproperly and that there are no legally bindingrules on how to establish it in practice.

Differences in growth conditions canhave a serious impact on composition and,therefore, in the absence of specification of theorigin and the conditions of cultivation ofthe different GM and non-GM samples, strictscientific comparisons cannot be made. Theseare not valid unless the parent line is grownside by side with the GM line. Comparisonswith historical or literary values have onlylimited scientific validity.

‘Substantial equivalence’ is a crude, non-scientific concept. It provides a loophole forthe GM biotechnology companies not to carryout nutritional and toxicological animal teststo establish whether the biological effect ofthe GM crop-based foodstuff is substantiallyequivalent to that of its non-GM counterpart.It therefore allows them to claim that there isno need for biological testing because the GMcrops are similar to their conventional coun-terpart, while on the other hand, because theycontain novel genes from other organism(s),they are patentable. However, unintentionaland unpredictable changes can occur in plantsbecause of the incorporation and positioningof the vector in the plant genome. It cannot

therefore be known which of the hundreds ofcomponents of the GM crop may carry toxicor allergenic properties. As most of theseare unknown, by definition, they cannotbe included in analytical comparisons.Determination of the amounts of protein,carbohydrates, fats and other nutrients canonly be a starting point. The consump-tion of minor and unexpected constituentsof potentially high biological activity mayhave considerable and disproportionatelylarge effects on the digestive tract. Theirpresence, therefore, can only be revealed fromanimal studies, and this makes it imperativethat these are performed with a flawlessdesign and experimentation.

The FLAVR SAVR tomato was pro-duced by ‘antisense’ GM technology. As partof its safety evaluation, it was subjectedto compositional analysis for total protein,vitamins and minerals to establish whetherany unexpected changes in gross fruit compo-sition had arisen as a result of the integrationof the FLAVR SAVR and kanr genes intothe tomato genome. It was claimed that nosignificant changes were found and that thecontents of potentially toxic glycoalkaloids,particularly tomatine, and to a lesser extentsolanine and chaconine, were also similar(Redenbaugh et al., 1992) and therefore thisGM tomato was substantially equivalent toother non-GM tomato lines. However, to sup-plement these, several feeding studies werealso performed by commercial laboratories atthe request of the FDA.

Acute toxicity

First, range-finding, limit acute oral toxicitytests of the processed tomatoes in rats werecarried out by the IIT Research Institute of theLife Sciences Department (Chicago, Illinois,USA). A single dose of the homogenates pre-pared from about 80 g of various GM andcontrol tomatoes, respectively, was adminis-tered (15 ml kg−1) by gavage to groups ofHarlan Sprague–Dawley rats (five male orfemale rats per group) fed ad libitum on ratchow for 14 days to establish whether theGM tomatoes were toxic or not. As claimed,no test substance-related mortalities occurredand increases in mean body weights were not

GM Foods: Potential Human Health Effects 349

significantly different between GM andcontrol groups. However, as the range of therats’ starting weights was unacceptably wide(female rats weighed 131–186 g (± 18%) andmale rats 159–254 g (± 23%)), in such a short(14 day) study with five rats per group,it would have been difficult for significantdifferences to develop. For comparison, onlya few per cent variation in starting weightsis permitted in papers published in high-quality nutritional journals. Thus, the poordesign of this feeding study largely invali-dated the conclusions that GM tomatoeswere not toxic. To supplement these, threemore rat feedings studies of similar designwere carried out by International Researchand Development Corporation (Mattawan,Michigan, USA).

Twenty-eight-day toxicology/histology study

Of the three studies, the most complete setof data is available for the second. In this,four groups of rats (20 males and 20 femalesper group) fed standard rat chow for 28 dayswere gavaged twice daily with homogenizedtomatoes (15 ml kg−1). Two groups weregiven GM tomatoes, CR3–613 or CR3–623(CR3–623 is the commercial FLAVR SAVRtomato). There were two control groups, oneof which was gavaged with the parent CR3tomato homogenates and a second controlgroup in which the rats were gavaged withwater even though the relevance of thisgroup is somewhat questionable. At therequest of Calgene, an expert panel wasretained (ENVIRON Corp., Arlington, Vir-ginia, USA) to evaluate the data. They con-cluded that gavaging rats with GM tomatopurée resulted in no significant changes inbody weight, food consumption and clinicalchemistry or haematology parameters incomparison with control tomatoes. However,there was a possible treatment-relatedincrease in glandular stomach erosion/necrosis in four out of 20 female rats but nonein the controls or in male rats at the end of the28-day feeding period. The number of fourfemale rats was increased to seven when thehistology slides were re-scored by PATHCO,

an independent pathology working group.This prompted a repeat study in which thedose of the tomato purée was increased bytwofold. Unfortunately, in this study, someof the CR3 control and CR3–623 GM tomatolines were grown at different locations andharvested at different times from those in thesecond experiment. However, this was notregarded as important by the expert paneleven though, when the same tomatoes wereused as in the second experiment, the resultsappeared to show similar tendencies; two outof the 15 females developed stomach glandu-lar erosions with the GM tomatoes, whilenone were found in the control females.However, in a not clearly understandableway, the ENVIRON panel concluded that thelesion of glandular erosion was not related tothe administration of GM tomatoes. Accord-ing to them, such lesions occur spontaneouslyin animals that are stressed or given muco-lytic agents, when food is restricted or whenanimals are restrained in cages, even thoughthese parameters have not been investigatedsystematically. Moreover, none of these cir-cumstances applied, since tomatoes containno mucolytic agents, food was provided adlibitum and the rats were not restrained. Itwas also suggested that, because the lesionswere possibly of short duration, they wereincidental, not related to the test materialand would have healed spontaneously.Unfortunately, none of these assumptionswas confirmed by further experimentation asno samples other than those at the end of the28-day experiment were taken to probe intothe timing and reversibility of the incidenceof the stomach lesions. Clearly, the resultsof these three studies should have promptedmore experimentation to investigate in moredetail the effect of GM tomatoes on stomachhistology and, what is even more important,these studies should have been extended toinclude the possible effects of GM tomatoeson both the small and large intestines.

The red or dark red pin-point lesionspresent in the stomach of female rats whichwere described as necrosis would be termed‘erosion’ in human pathology, which mayhave sequelae, such as life-endangering


haemorrhage. Erosions cannot be termed‘mild’, as unpredictable haemorrhage canoccur in the elderly human, particularly onlow dosage aspirin to prevent thromboticevents, and synergy with transgenic tomatoesmay occur. The assumption that the lesionsare related to stress does not explain the lowincidence in other groups, particularly in thesecond study. The relevance and significanceof gastric erosions in the human may bea matter of life and death in the older agegroups. It has been implied that pathologistsin general might not report such a lesion but,in the present era of vexatious litigation, men-tion would have been made in any humanpathologist’s report to avoid an accusationof negligence. This may not be required in vet-erinary pathology but these rat studies weredone with humans in mind and therefore thepathology findings must be put in this humancontext. It is probably true to suggest thatthese lesions are of short duration, but theserious nature of erosive lesions should notbe trivialized. This is the more serious becauseseven out of 40 rats eating GM tomatoes diedwithin 2 weeks. The nature of these deathswas not specified and the evidence that theywere not related to the ingestion of transgenictomatoes was inconclusive.

In a further development, the ScientificCommittee on Food of the European Commis-sion Directorate C (2000) concurred with theconclusion reached in the US Food and DrugAdministration (1994) memorandum. In theiropinion, although the results showed an unex-plained disparity, they were not supportiveof a substance-related effect of the FLAVRSAVR tomato. However, it is likely thatthe EU Committee may not have seen allthe primary data and their opinion wouldtherefore have been based on incompleteevidence. It is also regrettable that, byascribing the gastric erosions in rats to ‘anartefact of gavage studies’, the EU Committeehas in fact labelled the scientists carrying outthe work as incompetent. As these erosionswere found at the end of a 28-day study dur-ing which 160 rats were gavaged twice dailywith tomatoes, it is unlikely that even poorlytrained workers would not have become more

competent, so as to avoid causing such ananomaly.

Effects on body weight, food intakeand organ weights

The conclusion of the ENVIRON panel thatfeeding rats on GM tomatoes (CR3–623)for 28 days had no effect on weight gain,feed intake and organ weights could not bejustified because the starting weights of therats were so widely different – a range of130–258 g (± 33%) for males and 114–175 g(± 21%) for females – that finding significantdifferences in weight gain, feed intake andorgan weights was not likely. Indeed, weightgains varied between wide limits (102–230 gfor males and 46–127 g for females) in 28days. Even under these conditions, althoughthe average starting weight of the male ratsgavaged with CR3–623 GM tomatoes wasthe highest (148.1 g), their final weight(316.5 g) was the lowest. Accordingly, therats gavaged with GM tomatoes grew theleast of the four groups of rats. The feedintake of the different groups also variedbetween wide limits; 133–203 g for males and102–153 g for females. Not surprisingly, thefeed conversion efficiency (weight gain/totalfeed intake) of female rats on GM tomatoes(0.152) was significantly (P < 0.05) less thanthat (0.167) obtained for female rats oncontrol non-GM CR3 tomatoes.

The large range of starting weight differ-ences also excluded the possibility of findingsignificant differences in the organ weightsof the four groups of rats. The standarddeviations of mean values were very large, insome instances more than 20%. It is the moreremarkable that, even under these conditions,some differences in organ weights werefound, including the testes for males andthe thyroid/parathyroid for females. Findingno significant differences in biochemical,haematology and ophthalmology parametersbetween GM and non-GM tomatoes was notunexpected either, because of the large initialbody weight differences.

Overall, it is regrettable that these rattoxicological feeding studies were poorly


designed, as a great deal of effort, work andmoney must have been spent on them andso much rested on the outcome. The FDA’sconclusion that FLAVR SAVR presentedno more dangers to consumers than ordinarytomatoes does not therefore appear to rest ongood science and evidence which could standup to critical examination. Rather tellingly,the results of these studies have never beenpublished in peer-reviewed journals. Thestudy as described not only raises questionsabout the design, methods and conclusionsfor this study but also whether they couldhave any general validity for other GM foods.In this light, it is the more surprising thatafter these studies the FDA has required nonutritional/toxicological testing of other GMfoods.

Aventis’s Chardon LL herbicide-resistantGM maize

Due to the UK government’s attempt to placeChardon LL seed on the National List, a partof the supporting evidence submitted byAventis contained data on the composition oftwo lines of seed to establish their substantialequivalence to the conventional parent maizeline. The evidence also included the results ofa 14-day rat feeding study. All this is to befound in a file deposited by the Ministry ofAgriculture, Fisheries and Food (MAFF) withthe British Library (British Library File, 1997).

Compositional analysis

In the absence of specifying the originand conditions of cultivation of the differentGM and non-GM samples, strict scientificcomparisons could not be made betweenthem. However, even under these conditions,the composition of T14 and T25 GM maizeexpressing phosphinothricin acetyltrans-ferase enzyme (PAT-PROTEIN) showedmany statistically significant differences infat and carbohydrate contents in comparisonwith non-GM grain samples, and fat, proteinand fibre between silage samples from GMand non-GM maize. Thus, the conclusionthat GM maize is not ‘materially different’

from current commercial varieties cannot beregarded as valid.

Repeated dose oral toxicity (14-day feeding)study in rats

The rationale for this study was to assess thecumulative toxicity of PAT-PROTEIN givento rats in their diet for 14 days and to providea rational basis for toxicological risk assess-ment in man. Although testing of the PAT-PROTEIN can be commended, this study wasno substitute for the nutritional testing of theentire GM plant, seeds, vegetative parts andsilage in all target animal species. Withoutthese, the potentially harmful, unintendedand unpredictable effects of the gene transfer,other components of the vector and geneinsertion (positioning effect) cannot beestablished or excluded.

Unfortunately, as the design of the exper-iment was faulty, it is difficult to draw validconclusions from a feeding study, carried outwith five rats per group, in which the startingweight of the rats varied by more than ±20%(53–82 g for males and 50–74 for females)rather than the usual ±2%. For any differencesto reach significance, they needed to exceed± 20%, and to achieve this in a 14-day studywould have required catastrophic experimen-tal conditions. The five rats per group werenot housed singly and therefore their individ-ual feed intakes could not be monitored eventhough the huge differences in the startingweights should have led to major differencesin the feed intakes of the individual rats.Moreover, the group feed intakes were notmeasured continuously. There were fourgroups of rats (five male/female rats pergroup) in the experiment. However, rats ingroup 1 were fed a different diet (full ratchow) from the other three groups and there-fore group 1 was not appropriate for (stat-istical) comparisons. The diet of the secondgroup contained 5 g kg−1 and the third grouphad 50 g kg−1 PAT-PROTEIN mixed in with45 and 0 g kg−1, respectively, of commercial(SOJAMIN, KLIBA Muhlen AG) low soybeanprotein diet (11% raw protein). The diet of thefourth group contained 50 g kg−1 SOJAMINbut no PAT-PROTEIN. Thus, for statisticalanalysis, the second and third groups ought to


have been compared with rats in the fourthgroup. Curiously, although the main targetorgan of the PAT-PROTEIN fed to rats wasthe digestive tract (and pancreas), the weightsof these were not measured. This is a majorexperimental design fault.

The starting weight and the feed intake ofthe third group (high PAT-PROTEIN) werethe highest, but they ended up with the lowestfinal body weight. This indicated an elevatedmetabolic activity, probably induced by thePAT-PROTEIN. Our analysis of variance(ANOVA) shows that the weight gain forboth male and female rats on the high PAT-PROTEIN diet (group 3: 65.2 and 43.6 g formales and females, respectively) was sig-nificantly (P < 0.05) less than that of eitherthe fourth group (control: 72.8 and 48.8 g formales and females, respectively) or group 2(low PAT-PROTEIN diet: 73.4 and 44.4 gfor males and females, respectively). As PAT-PROTEIN reduced feed conversion efficiency,it is potentially harmful. The conclusion that‘there were no differences which could beattributed to treatment with the test article’was therefore not valid. Similarly, that ‘therewere no changes on … clinical biochemistryand urine analysis after 14 days’ is not valideither as the authors’ own results describeddifferences between the groups in glucose,cholesterol, triglyceride and phospholipidlevels, indicating an increased metabolicfunctional load in the rats. It is unexplainedwhy these differences were dismissed bythe authors as incidental and unrelated tothe treatment. Our ANOVA analysis revealedthat the urine output in rats on the high PAT-PROTEIN diet was significantly (P < 0.05)reduced, indicating treatment-related effects(urine output of 5.4 and 4.4 ml for males andfemales in group 3 vs. 7.1 and 6.5 ml for malesand females, respectively, in control group 4).

The large differences in the startingweight of the rats probably prevented find-ing significant differences in organ weights.However, even under these conditions, ratsfed the high level PAT-PROTEIN diet (thirdgroup) had the lowest liver, thymus andspleen weights of all groups (even thoughthe differences with controls were notsignificant). This is of particular importancebecause the macroscopic findings indicated

thymus foci in 20–40% of the animals fed dietscontaining the PAT-PROTEIN.

In conclusion, the design and executionof this feeding study were poor and, contraryto the authors’ conclusions, the resultsindicated treatment-related effects inducedby PAT-PROTEIN (of unspecified origin).The results therefore could not be takenas evidence that the transfer of its gene intomaize represented no risk for the rat and,by inference, for humans, particularly as nogut histology studies have been completedso far. Finally, a recent publication (Chiteret al., 2000) showed that DNA survived inintact form or slightly fragmented unless theGM maize was heat processed extensively.Therefore, the possibility exists that withunderprocessed maize products humans andanimals might be exposed to the DNA used inthe genetic engineering.

Compositional Studies Published inPeer-reviewed Journals

Herbicide-resistant soybean

Befitting its importance in both human andanimal nutrition, a great deal of attention hasbeen given to the compositional analysis ofherbicide-resistant and other GM soybeans.Several publications appeared in nutritionaland other journals demonstrating the compos-itional ‘substantial equivalence’ of GM andnon-GM soya. Thus, it was claimed that themacronutrient composition of glyphosate-tolerant soybean (GTS) seeds resulting fromthe transformation of conventional soybeanwith a gene encoding 5-enolpyruvylshikimate-3-phosphate synthase from Agrobacterium,to make the soya herbicide resistant, wasequivalent to that of conventional soybeans.This applied equally to GTS unsprayed withglyphosate (Padgette et al., 1996) or sprayedwith this herbicide (Taylor et al., 1999). It wasclaimed that the results of proximate chemi-cal analyses of the contents of crude protein,oil, ash, fibre, carbohydrates and amino acidsof solvent-extracted and toasted or untoastedsoybean meals of unsprayed GTS and controlsoybean had shown that all these lines were


substantially equivalent (Padgette et al.,1996). Similar findings were described forsprayed GTS (Taylor et al., 1999). Althoughthis appeared to be true for most macro-nutrients, several significant differencesbetween GM and control lines, such as inash, fat and carbohydrate contents, were alsofound (Padgette et al., 1996). However, thesewere not regarded as having biologicalsignificance.

A closer inspection of the data in thepapers, however, revealed that the statisticalcomparison of the macronutrients of GMand non-GM lines was not scientifically valid.Instead of comparing their amounts in asufficiently large number of samples of eachindividual GTS with its appropriate individ-ual parent line grown side by side at thesame location and harvested at the same timeto establish whether they were ‘substantiallyequivalent’, what the authors compared wasa large number of different samples from dif-ferent locations and harvest times. As growthconditions have a major influence on seedcomposition, the range of the amounts of con-stituents in the different samples, regardlessof whether they were GM or non-GM, wasso great (±10% or more) that the chancesof finding statistically significant differenceswere unreal. It is possible that from a practicalpoint of view the variation in protein concen-tration of samples of the three lines of between36.8 and 45% would fall into the normal rangeof agronomic variability of soybeans andtherefore may not be of major concern foragronomists. However, this comparison is notstrict enough to establish whether the geneticmodification introduced any unintendedcompositional changes. What is remarkableis that, even with this approach, many sig-nificant changes in macronutrient levels werefound. Thus, the claim of ‘substantial equiva-lence’ of GTS lines with non-GM soybean isnot supported by rigorous scientific evidence.

The potential importance in humanhealth of natural isoflavones, such as genistein,daidzein and coumestrol present in soybeans,is generally recognized. It was, therefore, ofconsiderable interest whether any changesoccurred in these components as a resultof genetic modification. Here the publishedevidence is controversial. Thus, while in

some studies no meaningful differences werereported (Padgette et al., 1996; Taylor et al.,1999), an independent study claimed that GMsoya samples had consistently contained sig-nificantly fewer isoflavones than the parentcultivars (Lappe et al., 1999). In one respect, allauthors agreed, i.e. that the isoflavone contentof soybean seeds showed considerable vari-ability between sites and was dependent onagronomic conditions. However, Lappe et al.(1999) went further and claimed that, whilethe variability of the GM samples was indeedconsiderable, conventional soybeans showedless variation in isoflavone content. As theisoflavone content of soybeans might affecthuman health, there needs to be more aware-ness of potential health problems due tothis variability. While the precise details ofthe changes in isoflavone content on geneticmodification will have to be established inthe future, to ensure clinical consistency, theorigin and the actual phyto-oestrogen levelsin soybean may need to be standardized.

In the study by Padgette et al. (1996),no significant differences were found in thelevels of antinutrients, such as trypsin inhibi-tors, lectin and oligosaccharide flatulencefactors, between solvent-extracted, toasted oruntoasted GM and non-GM soybean seeds.However, the comparisons were made bythe same method as for macronutrients andtherefore the large range of natural variabilityexcluded the possibility of finding significantdifferences. Interestingly, in single soybeanmeal samples of each of the two GTS andparent lines, the trypsin inhibitor (also a majorallergen in soybean) content was substantiallyhigher, by almost 30%, in one of the two GTSlines, with a smaller increase in the other. Notrypsin inhibitor analyses were performedon the protein isolate or protein concentratesamples originating from the meal samples.Although there were other compositionaldifferences in these processed soybeanproducts, it is difficult to decide from singledeterminations whether these were signifi-cantly different or not.

In conclusion, there is insufficientevidence to date to decide whether the com-position of GM and conventional soybeansis equivalent or not. In fact, some data,particularly those for phyto-oestrogens, were


significantly different. Furthermore, becausenot strictly comparable compositional datawere used, the case for equivalence was notproperly established. There is therefore anobvious need for further more critical studies.

GM potatoes

Brief references to GM potatoes, particularlythose expressing Bacillus thuringiensis (Bt)toxin, can be found in non-peer-reviewedbook chapters or other articles. In mostinstances, these contain no data and aretherefore of little scientific value. There aretwo exceptions, one of which is an article onthe safety assessment of GM potatoes expres-sing the soybean glycinin gene (Hashimotoet al., 1999a). However, it is not quite clearwhat the authors wanted to achieve because,at the expression level of glycinin in potatoesof between 12 and 31 mg g−1 total solubleprotein, no significant improvements in theprotein content or amino acid profile couldhave been expected. Indeed, the results in thepaper demonstrated that the total proteincontent of the GM potatoes appeared to besignificantly less than that of the control lineand that no improvement in the essentialamino acid profile was achieved either. Thereappeared to be substantial differences insome vitamins between GM and control lines,and the amounts of both solanine andchaconine increased in the GM lines. It is,therefore, not quite clear why it was claimedby the authors that their GM lines wereequivalent to the parent line and could be uti-lized as safely. The other more recent studyis a conventional compositional analysis ofsome macro- and micronutrients of tubersfrom insect- and virus-resistant potato plants(Rogan et al., 2000) performed by methodswhich currently are accepted by most novelfood regulatory bodies. Although theseshowed some significant differences in anumber of tuber constituents, in the absenceof toxicological/nutritional animal studiesit is difficult to ascertain whether these dif-ferences could have any biological effectson humans/ animals, particularly as theseconventional analyses could not have

revealed the development of any unknownpossible toxic/antinutritive components.Additionally, known antinutrients, such aslectins or enzyme inhibitors, were notincluded in the analysis.

GM rice

GM rice lines expressing the soybean glyciningene have been developed (Momma et al.,1999) by a method similar to that used for GMpotatoes. The glycinin expression level wasbetween 40 and 50 mg glycinin g−1 total riceprotein. The GM rice was claimed to contain20% more protein, but its moisture contentwas less than that of the parent line. How-ever, from the paper, it is not quite clearwhether the increased protein content wasdue to the decreased moisture content of theseeds because it was not specified whetherthe values were expressed for air-dried or fullydried seeds. Thus, most of the arguments inthe discussion of whether the higher proteinlevel was due to the positioning effect of geneinsertion or metabolic interference will haveto await clarification by further work.

GM cotton

Several lines of GM cotton plants havebeen developed using the gene encodingan insecticidal protein from B. thuringiensissubsp. kurstaki. These had increased protec-tion against the major lepidopteran insectpests of cotton. As cottonseed is an importantsource of oil for human consumption, andcottonseed and processed cottonseed mealfor animal feed, extensive analytical workhas been done to establish whether the GMlines were ‘substantially equivalent’ to con-ventional lines (Berberich et al., 1996). Thelevels of protein, fat, carbohydrate, moisture,ash, amino acids and fatty acids in the insect-protected lines were claimed to be compara-ble with those found in commercial varieties.Moreover, the levels of antinutrients suchas gossypol, cyclopropenoid fatty acids andaflatoxin were similar to or less than those inconventional seeds. Thus, the GM varieties


were suggested to be equivalent to conven-tional seeds and just as nutritious. However,the statistics used by the authors were identi-cal to those used with glyphosate-resistantsoya and therefore could be similarly criti-cized. Although the content of known con-stituents fell in between the wide range ofvalues of commercial conventional lines, thisdid not mean that they were compositionallyequivalent, particularly as environmentalstress could have major and unpredictableeffects on antinutrient and toxin levels(Novak and Haslberger, 2000). Thus, withoutanimal experimentation, this approach couldnot reveal whether any new and unknowntoxins/allergens had been created or not.

GM maize

A glyphosate-tolerant (Roundup Ready)maize line GA21 has recently been devel-oped. It was claimed (Sidhu et al., 2000) that,except for a few minor differences, which theauthors think are unlikely to be of biologicalsignificance, the results of compositionalanalyses of proximate, fibre, amino acid, fattyacid and mineral contents of the grain, andproximate, fibre and mineral contents offorage collected from 16 field sites over twogrowing seasons showed that control andGM lines were comparable. The comparisonwas carried out by a method similar to thatdescribed for GTS soya (Padgette et al., 1996)and this may therefore not be scientificallyrigorous enough for the establishment ofsubstantial equivalence.

Nutritional/Toxicological StudiesPublished in Peer-reviewed Journals

Herbicide-resistant soybean

As part of a safety assessment of GTS, thefeeding value, wholesomeness (Hammondet al., 1996) and possible toxicity (Harrisonet al., 1996) of two major GM lines of GTSwere compared with those of the parent line.Processed GTS meal was included in thediets of rats, broiler chickens, catfish and

dairy cows at the same concentrations asin commercial non-GM soybean rations. Ratsand dairy cows were fed these diets for4, broilers for 6 and catfish for 10 weeks. Itwas claimed that in rats, catfish and broilersthe growth and feed conversion efficiency,in catfish the fillet composition, in broilersthe breast muscle and fat pad weights, andin dairy cows milk production and composi-tion, rumen fermentation and digestibilitieswere similar for both GTS and parental lines.According to the authors, these resultsconfirmed that the GTS and parental lineshad similar feeding values.

Rat studies

A critical evaluation of the rat study washampered by the lack of adequate primaryindividual data in the paper. Thus, there wasno full description of the rat diet. It appearsthat the total protein content of the diets wasadjusted to 247 g kg−1 diet to be isonitro-genous with Purina Laboratory Rat Chowby the addition of 24.8 g of GTS and parentsoybean meals, respectively (~10% protein),to a base diet. All comparisons were madewith rats fed commercial Purina Chow.The protein concentration in these diets was,however, appreciably higher than the usual10–16% crude protein and exceeded theprotein requirements of the rat. This extraprotein potentially could have masked anypossible transgene product effects, particu-larly with the raw unprocessed soybean dietsin which the GM meals were incorporatedonly at the level of 50 or 100 g kg−1 of the diet.Thus, these meals only replaced 8.5 and 17%,respectively, of the total protein of 247 g kg−1

diet. In other words, the GM soybean proteinin these meals was diluted by other dietaryproteins by 12- and 6-fold, respectively,producing another possible masking effect.The composition of the control Purina Chowdiet in the ground raw soybean feeding studywas not described. This is important becausethe identity of the raw control soybeansincluded in the Purina Chow control diet wasnot specified.

In the feeding study, four groups of rats(ten males and ten females in each group)singly housed were fed diets containing the


parental line or the GTS lines (40-3-2 or61-67-1) for 28 days. No individual values (ortheir ranges) for feed intake or body weightwere given. The bar diagrams of the combinedbody weight of rats at the end of each weekof the 4-week experiment were rather unin-formative. However, it was observed by theauthors that the Purina Chow-fed male ratsgrew significantly better than the three experi-mental groups fed toasted soybeans (includ-ing the parental line). This was attributed tobetter commercial processing. However, thebar diagrams also indicated that the growthin the group fed with one of the GTS lines(61-67-1) was probably equal to that of thePurina Chow-fed control and, therefore, byinference, these rats also grew significantlybetter than the other two experimental lines(the GTS line 40-3-2 and the parental line).This again underlined the importance ofgiving individual data in papers, withoutwhich it is difficult to assess the results.Similarly, there were no individual data fororgan weights, such as liver, kidneys andtestes. However, it was claimed that thekidney weights of the raw GTS line-fed (andparental control?) male rats were significantlyhigher than those of the controls, while thetestes of the parental line-fed rats were signifi-cantly enlarged. According to the authors, asthese differences were neither dose relatednor only shown by the parental line, they werenot caused by genetic modification. Rathercuriously, the weights of the stomach andintestines, the main target organ in any nutri-tional testing, were not recorded. Observa-tions were not recorded on other organs, andno histology appears to have been done onthese tissues either. The only tissue which wassubjected to microscopy was the pancreas, butthe description of the findings was qualitative.Only minimal to mild lesions were foundand these were claimed to be common toall groups. However, under these conditions,this was not surprising because no pancreatichypertrophy was found. This was probablydue to the effect of the unusually high dietaryprotein concentration, which, as the authorspointed out, masked and/or diluted thebiological effect of the trypsin inhibitors. Thisis of particular concern because the trypsininhibitor content of GTS lines in unprocessed

soybean was significantly higher than in thecontrol line (Padgette et al., 1996).

It is regrettable that the design of thisimportant rat feeding study had such unfortu-nate omissions. It is of particular concern thatno histology was apparently carried out ongut tissue. Thus, more critical work is neededto decide whether the feeding value of GMand non-GM soybeans is equal or not.

Chicken study

The broiler chicken feeding study’s experi-mental design closely followed that of com-mercial practice and therefore the resultsshould only be indicative of the commercialfeeding and production value of the varioussoybean lines. As the data were pooled fromall birds fed on the same diet, it is not easy tosee what, if anything, was the significance ofthe small differences found in the study, suchas the slightly lower body weights, breast andfat pad weights obtained with the GTS lines(particularly with GTS 40-3-2) for the utiliza-tion of GM soybean. It would have beenbetter to measure the nutritional performanceof individual birds (or small groups) fed ondifferent diets and then compare them afterstatistical analysis. In the absence of this, wehave to rely on the authors’ conclusion thatthe design of the experiment gave the upperlimit of differences in weight gain of 3.5% andgain/feed ratio of 2% and that the GTS linesvs. parental line were within this limit. Thus,with this restriction, the feeding value of theGTS lines for broilers was practically equal tothat of the parental soybean line.

Catfish experiment

Catfish are excellent and highly sensitiveindicators for the feeding value of diets. Itwas obvious from the results that, similar tothe findings with rats, one of the GTS lines,61-67-1, was superior to the other lines (GTS40-3-2 and the parental line) in most respects.Thus, fish on GTS line 61-67-1 ate more, hadbetter weight gain and gain/feed ratio andweighed more at the end of the 10-weekstudy than the others, even though the com-position of the fillets from these fish wasnot significantly different. This significant


difference in performance must, therefore,indicate that genetic modification may notbe as reproducible as it has been claimedand that the feeding value and metaboliceffects of GM and parent lines are not always‘substantially equivalent’.

Study on lactating cows

Milk production and composition and per-formance data in the lactating cow studyshowed some significant differences betweencows fed diets containing the different linesof soybean, indicating a lack of ‘substantialequivalence’. In view of these differences,even though we may not at present know alltheir biological/nutritional consequences, itmay be difficult to maintain the view that thefeeding value of the GTS and parent lines isequal, and further work is needed to establishwhether the GTS lines are safe or not forhumans/animals.

Testing of E. coli recombinant gene product

Extensive studies have been carried out toascertain the safety of the gene product,5-enolpyruvylshikimate-3-phosphate syn-thase (CP4 EPSPS), which renders thesoybeans glyphosate resistant (Harrison et al.,1996). Unfortunately, there are some flawsin these experiments, the most important ofwhich is that in the acute gavage studies theauthors did not use the enzyme isolated fromGTS lines but instead that from Escherichiacoli. Although they were at pains to show thatthe EPSPS enzyme samples from the twosources were similar in lack of glycosylation,molecular size, reaction with a polyclonalanti-EPSPS antibody and enzyme assays,these methods do not have sufficient powerto show unequivocally whether they wereidentical. The authors themselves pointedout that post-translational modification ofthe completed polypeptide chains emergingfrom the ribosomes may be done differentlyin two such evolutionarily distinct life formsas higher plants and prokaryotic bacteria.Amidation, acetylation and proteolytic pro-cessing can have such major effects onthe conformation of the protein as to makethese gene products behave differently in the

digestive system. Thus, the use of the E. colirecombinant protein for the acute micegavage studies may invalidate the authors’conclusion that the gene product from soy-bean did not have any toxic effects. Thesestudies must be re-done with the gene prod-uct isolated from the transgenic plant beforethe results could be accepted. In any case, insuch gavage studies, young, rapidly growinganimals must be used to show any distincteffect on growth. As all animal weights wereunchanged in the experiment, the test systemused could not have detected any effectunless the consequences of the gavaginghad been disastrous. Feeding studies withthe gene product in young rapidly growingrodents should be the preferred method forthe demonstration of the deleterious effects.

The other flaw in the experimental designwas the reliance on an in vitro simulatedgastric/intestinal digestion assay, which wasalso carried out with the E. coli recombinantgene product. To obtain physiologically validresults, it would have been necessary to usethe gene product isolated from GM soybeanin an in vivo assay in the rat (or other suitableanimals; see Rubio et al., 1994) or a full feedingtrial. Thus, it has been shown before that thekidney bean (Phaseolus vulgaris) α-amylaseinhibitor is fairly stable to proteolytic degra-dation in the rat gut (Pusztai et al., 1995, 1999),but, when its gene was expressed in peas(Pisum sativum), it was rapidly digested andinactivated in the rat stomach/small intestinein vivo (Pusztai et al., 1999). This may havecontributed to the safety of GM peas forrats and, by inference, possibly for othermonogastric mammals. Thus, in vitro diges-tion assays may have little relevance to thesafety of GM food crops.

In a separate feeding study (Teshimaet al., 2000), the possible harmful effectsof toasted glyphosate-resistant GM soybeanwere investigated at 300 g kg−1 inclusion levelin the diet of rats and mice. After feedingthese animals for 15 weeks, no significantdifferences in nutritional performance, organdevelopment, histopathology of the thymus,liver, spleen, mesenteric lymph nodes, Peyer’spatches and small intestine, and the pro-duction of IgE and IgG humoral antibodiesbetween GM and non-GM line diets were


found. However, as rats grew less than30 g and mice not at all in 15 weeks, theconditions were so unphysiological that novalid conclusions could be drawn from theseexperiments.

GM maize

In a major commercial-scale broiler chickenfeeding study with rations containing trans-genic Event 176-derived Bt maize involving1280 birds (Brake and Vlachos, 1998), it wasclaimed that no statistically significant dif-ferences in survival or bird weights betweenbirds fed diets containing GM maize, Event176, or an isogenic parent maize line werefound. Indeed, birds fed GM maize rationsappeared to have significantly better feedconversion ratios and an improved yieldof breast muscle. However, the authorscautioned against the conclusion that thisenhanced performance could be attributedto the Bt maize per se. It is possible that theresults might have been due to slight differ-ences in the overall composition of the diets.This is reasonable considering the length ofthis study and possible problems of consis-tent diet preparation on a commercial scale.Minor differences in composition such asthe slightly lower protein content of the GMmaize and fat contents of the diets magnifiedto the scale of this trial make the results morerelevant to commercial than to academicscientific studies.

In a poultry feeding study, it was claimedthat the GA21 Roundup Ready maize-baseddiets gave similar performance data ingrowth, feed efficiency and fat pad weightsto diets containing the parental control line(Sidhu et al., 2000). However, this and a similarstudy carried out in Germany with a maizeline expressing PAT-PROTEIN (Flachowskyand Aulrich, 2001) were commercial produc-tion experiments and made little contributionto scientific safety assessment.

In a separate study, maize was geneti-cally modified by the transfer of the gene ofegg white avidin to make the seed resistant tostorage insect pests (Kramer et al., 2000). It wasalso claimed that this GM maize was safe

for mice as apparently, when, instead of abalanced diet, they were fed solely on thiscrop, the mice suffered no ill effects. However,the mice used in the experiment were adultswhich did not grow at all, and therefore theconclusion that the GM maize was safe is, atbest, premature.

GM peas

Diets containing transgenic peas expressingthe transgene for insecticidal bean α-amylaseinhibitor (~3 g kg−1 peas) at two differentinclusion levels in the diet, 300 or 650 g kg−1,were subjected to nutritional evaluation withrats in a 10-day feeding trial (Pusztai et al.,1999). The nutritional performance of ratsfed GM pea diets was compared with thoseobtained with rats pair-fed iso-proteinicand iso-energetic diets containing parent-linepeas and also lactalbumin diets spiked withisolated bean and pea α-amylase inhibitors,respectively. At 300 g kg−1, but not at 650 gkg−1 inclusion level, the nutritional value ofdiets containing transgenic or parent peaswas not significantly different. Even at650 g kg−1, the difference was small, mainlybecause the transgenically expressed pearecombinant α-amylase inhibitor was quickly(in < 10 min) degraded in the rat digestivetract and therefore its antinutritive effect wasabolished. In contrast, spiking the parentalline pea diet with the stable bean α-amylaseinhibitor reduced its nutritional value(Pusztai et al., 1995, 1999).

In this study, unfortunately, no gut his-tology was done or lymphocyte responsive-ness measured, and therefore one had to relyon the evaluation of nutritional parametersthat are inherently less sensitive in orderto find possible differences in metabolicresponses between GM and conventionalfood components. Although there were signif-icant differences in the development of someorgans, mainly the caecum and pancreas,most organ weights were remarkably similar.At the end of the study, cautious optimismwas expressed that GM peas could be usedin the diets of farm animals, particularly atthe low/moderate levels recommended in


commercial practice and if the progress of theanimals was monitored carefully. However,this relatively short feeding study withmodest objectives cannot at this stage be takenas proof of the safety of GM peas for humanconsumption. There is a need to carry outfurther and more specific risk assessment test-ing procedures, which must be designed anddeveloped with human consumers in mind.It also has to be kept in mind that only oneparticular line of GM peas was tested in whichthe endogenous antinutrient levels wereselected to be similar to those of the parentpeas. In some other GM lines, however, lectinlevels could vary, up or down, by a factor offour. Moreover, in some field pea cultivars,such as Laura, the concentration of trypsininhibitor increased by about 24% and thechymotrypsin by 100%, while the haemag-glutinating activity decreased by a factor offour in the GM line compared with its parent(A. Pusztai, unpublished). This strengthensthe argument that, in the safety assessment ofGM crops, many lines should be included andthat, from the results of a single GM line, noblanket approval should be given to otherlines developed, even if in the transformationthe same vector was used and carried out atthe same time.

GM potatoes

There have been four independent studies ondifferent GM potatoes.

Glycinin-expressing potatoes

The safety of transgenic potatoes expressingthe soybean glycinin gene was evaluated in ashort (4-week) rat feeding study (Hashimotoet al., 1999b). With an interesting experimen-tal design, control rats and the experimentalgroups were fed the same control commercialdiet. However, the rats were also dailyforce-fed by gavage with 2 g of respectivepotato lines kg−1 body weight. The potatoesused were a parental control line and twotransformed lines, one with the glycinin geneand another one with a designed glyciningene (coding for four additional methionines

in the gene product), respectively. However,there were a number of problems with thisstudy. Thus, although no difference ingrowth, feed intake, blood cell count, bloodcomposition and internal organ weightsbetween the groups was found, the uncer-tainty as to whether the animals were fedwith raw or boiled/baked potatoes leaves aquestion mark over the interpretation of theresults.

Bt toxin potatoes

An interesting, mainly histology, study wascarried out on the ileum of mice fed withpotatoes transformed with a B. thuringiensisvar. kurstaki CryI toxin gene. As a control, theeffect of the toxin itself was also investigated(Fares and El-Sayed, 1998). It was shown thatboth the delta-endotoxin and, to a lesserextent, the Bt potato caused villus epithelialcell hypertrophy and multinucleation, dis-rupted microvilli caused mitochondrialdegeneration, increased numbers of lyso-somes and autophagic vacuoles, andincreased activation of crypt Paneth cells. Asa result, it was recommended that ‘thoroughtests of these new types of genetically engi-neered crops must be made to avoid the risksbefore marketing’. Unfortunately, some flawsin the experimental design detract from thestrength of the conclusions. The most impor-tant of these was that, apart from indicatingthat the gene used in the transformationwas the CryI gene from B. thuringiensis var.kurstaki, there was no description of the Btpotatoes. The gene expression level in theGM potato was not given and it was not clearwhether the potatoes in the diet were cookedor raw. Moreover, the amount of the Bt toxinused for supplementing the potatoes withinthe control potato diet was not specifiedeither. This made it impossible to make aquantitative comparison of the effects onthe ileum of the Bt potato with those of thespiked control potato diets. The assumptionthat the ileum is the most important absorp-tive part of the rodent small intestine couldalso be argued against, because 90% of allnutrient absorption in fact occurs in thejejunum. As this was an electron microscopystudy, the fixation of the ileal samples was


not done on well-oriented sections but onchopped up fine tissue pieces, and importantdetail of villus organization was thereforelost. Finally, the delta-endotoxin-inducedhyperplastic changes on ileal villi shouldhave been demonstrated by measuring cellproliferation and mitotic rates in ileal (andjejunal) crypts rather than on the villi. How-ever, despite these shortcomings, this studyhas established once and for all that, in con-trast to general belief, exposure of the mousegut (ileum) to the CryI gene product hascaused profound hypertrophic and hyper-plastic changes in cells of the gut absorptiveepithelium and can lead to mucosal sensitiza-tion (Vazquez Padron et al., 1999, 2000b). Thisshows up the fallacy of drawing comfortingconclusions from in vitro simulated gut prot-eolysis tests. Clearly, concerns about the pos-sible biological consequences of exposure toGM food should be addressed under in vivoconditions because, even if an E. coli productbreaks down in vitro, this does not necessarilymean that the same gene product expressedin the transgenic crop should also breakdown.

GNA GM potatoes

Some of the results of rat feeding studieswith GM potatoes expressing the snowdrop(Galanthus nivalis) bulb lectin (GNA) genewere similar to the results of Fares andEl-Sayed (1998). A part of this work con-cerning the effect of GNA GM potatoes onthe histology of different compartments ofthe rat gut was published (Ewen and Pusztai,1999a). Although this peer-reviewed scien-tific paper was criticized by some, mostof the criticisms were unpublished personalopinions. Moreover, most of the publishedcriticisms (e.g. Kuiper et al., 1999) wereanswered adequately (Ewen and Pusztai,1999b). Some selected results of the nutri-tional/metabolic studies were, against thewishes of the authors, placed on the websiteof The Rowett Research Institute (www.rri.sari.ac.uk), where most of the work was done(Bucksburn, Aberdeen, UK). However, soas not to jeopardize their eventual properpublication, these results will only bementioned briefly.

Young, rapidly growing rats (startingweight of 84 ± 1 g) were strictly pair-fed oniso-proteinic (60 g total protein kg−1 diet, mostof which was from potatoes) and iso-caloricdiets (in contrast to that described in Kuiperet al., 2001) supplemented with vitamins andminerals for 10 days. The test diets containedeither raw or boiled GM potatoes. The controldiets contained the same amount of parental-line potatoes (raw or boiled) alone or supple-mented with GNA at the same concentrationas expressed in the GM potatoes. A positivecontrol group of rats was also included in theexperiment, and these were fed a lactalbumin-based high quality control diet to checkfor any potential problems in rat behaviourand experimental conditions. As part of thenutritional/metabolic evaluation, samplesof stomach, jejunum, ileum, caecum andcolon were taken, fixed and stained withhaematoxylin and eosin for full quantitativehistological evaluation (Figs 16.1, 16.2 and16.3) or reacted with GNA antibody and sub-sequently stained using a PAP (peroxidase–antiperoxidase) method to establish whetherany GNA was bound to the epithelial surface(Fig. 16.4). By measuring the mucosal thick-ness of the stomach and the crypt length of theintestines (Ewen and Pusztai, 1999a), it wasshown that proliferation in the gastric mucosawas in part caused by GNA, the gene product.However, the growth-promoting stimulus onthe small intestine of diets containing GMpotatoes leading to crypt enlargement and apart of the stomach enlargement was not aGNA effect. As shown before and confirmedhere, there was a slight binding of GNA to thesmall intestinal epithelium (Fig. 16.4). How-ever, GNA is not a mitotic lectin and thereforeit did not induce hyperplastic growth in thistissue (Pusztai et al., 1990). Accordingly, thejejunal growth was probably due to someas yet unknown effects of other parts of thegenetic construct used for the transformationor the genetic transformation itself. Hyper-plasia was shown previously by measuringthe increase in crypt length (Ewen andPusztai, 1999a). However, similar resultswere obtained by measuring the increasein crypt cell numbers (Table 16.1) and cryptmitotic figures (not fully significant) in thejejunum of GM potato-fed rats (Table 16.2).


The results suggested that it is possible thatcrypt hyperplasia and an increase in epithelialT lymphocyte infiltration observed with GMpotatoes might also happen with other GMplants that had been developed using thesame or similar genetic vectors and methodof insertion. It is therefore imperative thatthe effects on the gut structure and metabo-lism of all GM crops should be investigatedthoroughly as part of the regulatory processbefore their release into the human food chain.

Potatoes expressing cationic peptide chimeras

Desiree and Russet Burbank potatoes expres-sing N-terminally modified cecropin–melittincationic peptide chimeras and control linepotatoes fed to mice caused severe weightloss. The animals did not grow even aftersupplementing these potatoes with rodentlaboratory chow. According to the authors(Osusky et al., 2000), mice fed with tubersfrom transgenic potatoes were as healthy andvital (sic) as those from the control group,and their faecal pellets were comparable.

However, the severe weight loss seriouslycalled into question the value of the results ofthis poorly designed feeding experiment.

GM tomatoes

Finally, an important study will have to bedescribed even though it was not publishedin a peer-reviewed journal, but the ideasand experiments described had some influ-ence on the development of GM regulation(Noteborn et al., 1995). Thus, a new labora-tory GM tomato line was developed using theB. thuringiensis crystal protein CRYIA(b) genebut, instead of the cauliflower mosaic virus35 S promoter (CaMV 35 S), which is usedin practically all first-generation GM crops, apotentially safer plant promoter was used.Although with this the expression level of theBt toxin was only about 1/20th of that foundwith CaMV 35 S, this might be improvedupon in future. In contrast with most otherstudies with GM crops, there was a


Fig. 16.1. Comparison of the stomach mucosa of rats fed with raw GM potato diet (b) shows markedthickening due to hypertrophy of mucosal cells in comparison with that of rats given the parental line (a)(bar = 100 µm).


Fig. 16.2. Histology of the jejunum and ileum of rats fed raw GM or parent potato diets. Jejunal cryptlength and cells exhibit marked enlargement after feeding rats GM potato diets for 10 days (b) incomparison with those of rats given parental line potato diets (a). The villus length is similar in both, butintraepithelial lymphocyte cell counts appear to be increased on the GM potato diet. In the ileum, bothcrypts and villi of rats on GM potato diets are elongated (d) in comparison with parent potato-fed rats (c)(bar = 100 µm).


Fig. 16.3. The mucosa of the caecum demonstrates little change. Differences between GM-fed (b) andparent line potato-fed rats (a) are slight, while the colonic mucosa is moderately thickened in GM-fed rats(d) compared with that of rats given the parental line (c) (bar = 50 µm).

commendable attempt to measure the bind-ing of the gene product to the rat gut surfacein vivo rather than using spurious argumentsas to why the gene product should not bind.

Although no in vivo binding was found, thisshould not detract from the significance ofthis initiative because, due to the lack ofavailability of sufficient quantities of Bt toxin


Fig. 16.4. Immunocytochemistry of the jejunum of rats fed raw GM potato diets for 10 days (a) showsmoderate binding of GNA to villus tips (bar = 50 µm). Similar binding of GNA to jejunal villus tips isfound in rats given parent potato diet supplemented with GNA in amounts equivalent to that expressed inthe GM potato (b) (bar = 100 µm). Sections were first treated with anti-GNA rabbit antibody (diluted 1/100),followed by visualization with PAP. Note the strong antibody reactivity of feed particles in the sections.

Diet Parent

Parentvs. parent +

GNA (P)Parent +

GNA

Parent vs.transgenic

(P) Transgenic

Parent +GNA vs.

transgenic(P)

RawBoiledRaw vs. boiled (P)

15.9 (0.5)17.8 (1.1)0.006

0.0370.466

17.0 (0.7)18.2 (0.2)0.003

0.0000.749

20.3 (0.8)18.2 (1.2)0.003

0.0060.769

aThe number of nuclei were counted sequentially on well-oriented haematoxylin and eosin paraffinsections (4 µm). Values represent means (SD) for six rats per treatment; ten crypts per rat were counted.Differences between treatments are significant when P < 0.05 (Student’s t test). The effect of boiling(P = 0.759) is not significant, while that of GNA added or as transgene product (P = 0.019) and the effectof transformation (P = 0.000) are highly significant. The interactions between GNA and cooking(P = 0.043) and between transformation and cooking (P = 0.018) are also significant (multivariantanalysis with Tukey’s test).

Table 16.1. Number of crypt cells in the jejunum of rats fed various potato diets.a

isolated from GM tomatoes, an E. coli recom-binant and potentially less stable form of thegene product was used in the experiment,and its possible degradation in the gut mayhave accounted for the lack of binding. How-ever, Bt toxin was shown byimmunocytochemistry to bind to gut sec-tions, including the caecum and colon, fromhumans and rhesus monkeys in vitro. Unfor-tunately, their short-term toxicity testing inmice (and rabbits) and the in vitro simulatedproteolysis assays were also carried out withthe E. coli recombinant gene product andtherefore their conclusions of finding no toxiceffects may not be valid. Commendably, theauthors carried out a 91-day feeding studywith rats using freeze-dried GM vs. parentline tomatoes, which were included at a 10%level in the diets, but no differences in foodintake or body and organ weights werefound. However, because the Bt toxin expres-sion level in the tomatoes was low, the dailyintake of the gene product(s) by the rats wasalso low. Moreover, as the daily input oftomato proteins was only about 5–6% of thetotal dietary protein intake of the rats, it wassomewhat optimistic to expect any significantchanges in these nutritional parameters. Tohave any reasonable chances to show upsmall differences in the nutritional value ofGM vs. parent line crops, it would have beenimportant to use as high a protein concentra-tion as possible such as that in the 110-dayGM potato feeding study carried out at TheRowett Institute, in which the GM protein inthe diet was diluted only twofold by otherdietary proteins, and this allowed the sig-nificant differences in the growth rates ofrats fed on baked GM potato diets vs. parentpotato diets to show up. In fact, to equalizethe growth rates of the rats on the GMpotatoes to that of the controls, the GM

diet had to be supplemented with an extra12 g lactalbumin kg−1 diet, and this extraprotein gave a quantitative measure of thedifference of the nutritional value betweenGM and non-GM potatoes. Even at thesesimilar growth rates, the weights of someof the rats’ vital organs, such as the gut andparticularly the small intestine, the liver andkidneys, were still significantly different.

There were other omissions in the Bttomato study, the most important of whichwas that no Bt toxin survival was measuredin the gut lumen and no gut histology wasdone to see if there was any Bt toxin bindingto or possible structural changes in the gutepithelium or whether lymphocyte infiltra-tion occurred. This omission is particularlyimportant because later studies showed thatthe similar Bt toxin Cry1Ac could bind togut epithelial cells in mice (Vazquez Padronet al., 2000a,b) and induce mucosal antigenicsensitization (Vazquez Padron et al., 1999,2000a,b). The allergenic potential of Bt toma-toes was not investigated either. However,despite some of its shortcomings, this studyshowed many novel and commendable fea-tures, which, after some improvements, may,hopefully, be incorporated into the generalGM food testing procedures.

Allergenicity

One of the major health concerns with GMfood is its potential to increase allergies inthe human population through the foodchain. The possibility of fatal anaphylaxis insensitized individuals after their unwittingexposure to allergenic proteins in unlabelledGM foodstuffs is a real danger. When agene is transferred from a source of knownallergenic potential, the assessment of theallergenicity of the GM crop is relativelystraightforward. This can be done usingin vitro tests with sera from individualssensitized to the allergen from the originalsource. Similarly, it is relatively easy to assessthe effect of genetic engineering on endoge-nous allergens in crops with some evidenceof allergenicity. With tests such as the radio-allergosorbent test (RAST), RAST inhibition


Diet Parent Parent + GNA Transgenic

RawBoiled

4857

4956

7557

aMitotic cells were expressed per 100 crypts.

Table 16.2. Mitotic numbers per 100 crypts inthe jejunum of rats fed potato diets.a

and immunoblotting, the allergenic potentialof the GM crop is easily measured. There arenow several examples for these, such asthe demonstration of the allergenicity of theBrazil nut 2S seed storage protein in trans-genic soybean (Nordlee et al., 1996) or thecodfish allergy in potatoes genetically engi-neered with cod protein genes to make thepotatoes tolerate cold storage (Bindslev-Jensen and Poulsen, 1997). The claim that inglyphosate-tolerant soybean the introductionof the herbicide resistance gene does notaffect the allergenicity of the soy endogenousallergens is also a good example (Burks andFuchs, 1995). Having shown in a surveillanceprogramme of farm workers before and afterexposure to B. thuringiensis pesticide spraysthat some developed skin sensitization andIgE antibodies to the Bt spore extract and thattwo of them had a positive skin-prick test, itmay now be possible to test for the aller-genicity of Bt toxins engineered into variouscrops (Bernstein et al., 1999). This is all themore important because the Cry1Ac toxinhas now been shown to be a potent oralimmunogen and adjuvant (Vazquez Padronet al., 1999, 2000a,b).

It is much more difficult to assess theallergenicity of GM foods when the gene istransferred from a plant whose allergenicpotential is unknown. Moreover, it is alsopossible that, as a result of the gene transfer orvector insertion, a new allergen is developedor the expression level of a minor allergen isincreased in the GM crop. The gene productcan also have an allergenic adjuvant effect on afood component previously of low allergenicpotential, or some component in the GMfood may have an adjuvant effect on theallergenicity of the transgene product. Unfor-tunately, while there are good animal modelsfor nutritional/toxicological testing, no satis-factory animal models have been developedso far for allergenicity testing (Helm andBurks, 2000). For the time being, only indirectmethods are available for the assessment ofthe allergenic potential of GM foods derivedfrom sources of unknown allergenicity. Thereare a number of recommended approaches tobe followed. A useful preliminary step is toestablish if there are any sequence homologiesin the transgenic protein to any of the about

200 known allergens. If there are, in vitro testsfor IgE reactivity need to be performed. It isthought that the peptide length in the trans-genic protein which is optimally needed forbinding B-cell epitopes requires the presenceof at least eight contiguous identical or similaramino acids. However, the amino acids inthe allergenic epitopes are rarely contiguous.Moreover, the absence of a positive reactionin in vitro testing does not guarantee thatthe transferred protein is not an allergen.In a decision-tree type of indirect approach,the next step is to consider the molecularsize, glycosylation, stability, solubility andisoelectric point of the transgenic protein andcompare them with those of known allergens(O’Neil et al., 1998). Unfortunately, in moststudies to date, the all-important stability ofthe transgenic protein to gut proteolysis isestablished in an in vitro simulated gastric/intestinal system (Astwood et al., 1996;Metcalfe et al., 1996), and this is fundamentallyflawed. The results, therefore, are at best mis-leading and at worst erroneous. Reliance onthe concept that most allergens are abundantproteins is probably also misleading because,for example, Gad c1, the major allergenin codfish, is not a predominant protein(Bindslev-Jensen and Poulsen, 1997).

When the gene responsible for the aller-genicity of a crop is known, its cloning andsequencing open the way for its reductionby antisense RNA strategy. Thus, in rice,the low molecular weight α-amylase/trypsininhibitors are major allergens. A part of thegenomic sequence encoding this protein in anantisense direction was constructed betweenthe promoter of the rice allergen gene andits waxy terminator, and this was introducedinto rice protoplasts. The allergenicity of theregenerated plants was significantly less thanthat of parental wild-type rice (Nakamura andMatsuda, 1996).

In conclusion, allergenicity testingappears to be one of the Achilles heels of GMfood safety. It is clear that, if and when it isknown that the protein gene is derived from asource with a history of allergenicity, there is areasonable certainty that the GM crop will beallergenic. Unfortunately, the reverse is nottrue: the use of a gene from something that isnot allergenic will not guarantee that the GM


crop will not possess allergenicity. In theabsence of new and reliable methods forallergenicity testing, particularly the lack ofgood animal models, at present it is almostimpossible to establish definitely whether anew GM crop is allergenic or not in advance ofits release into the human/animal food/feedchain.

Conclusions

One has to agree with the opinion expressedin Science (Domingo, 2000) that there aremany opinions but very few data on thepotential health risks of GM foods, eventhough research to exclude such risks shouldhave been carried out before the GM cropswere introduced into the food chain. Ourpresent database is therefore woefullyinadequate. This is clearly seen from a closerscrutiny of the reference lists of recentreviews which contain only a handful of toxi-cological/nutritional and immune studiesof GM food crops published in peer-reviewedscience journals (Ruibal-Mendieta and Lints,1998; Betz et al., 2000; Kuiper et al., 2001;Pusztai, 2001). Moreover, the scientificquality of even what is published is, in mostinstances, not up to the standards that oughtto be expected. In this review, data pub-lished in peer-reviewed and some non-peer-reviewed journals have been examined indetail. However, as our future is claimed tobe dependent on the success or failure of thepromise of genetic modification deliveringGM foods which will be wholesome, plentifuland, most importantly, safe for us all, theemphasis was on strict but fair criticism.

From the results, the conclusion seemsinescapable that the present crude methodof genetic modification has not deliveredGM crops that are predictably safe andwholesome. The promise of a superior secondgeneration of GM crops is still in the future.It is possible that some of the first generationof GM crops may superficially satisfy somecommercial end points, such as their usein broiler chicken production. However, weneed to consider that these GM feed ration-fedanimals eventually will be consumed by

humans, and there is absolutely nothingknown about the potential hazards (if any) forhuman health of this indirect exposure to GMfood. Furthermore, the examples in the papershighlighted some differences even betweensuch crude things as macronutrient composi-tion of GM and conventional lines. It is arguedby some that these differences have littlebiological meaning. However, it was clear thatmost GM and parental line crops would argu-ably fall short of the definition of ‘substantialequivalence’. This crude, poorly defined andunscientific concept outlived its possible pre-vious usefulness. There is an urgent needto come up with novel scientific method-ologies to probe into the compositional,nutritional/toxicological and metabolic dif-ferences between GM and conventional cropsif we want to put this technology on a properscientific foundation and also to allay the fearsof the general public. We need more scienceand not less. For proper safety assessment, ourfirst concern ought to be to establish on acase-by-case basis the impact of componentsof GM foods on the digestive system,its structure and metabolism, because theway our body will respond to GM foods willbe predetermined at this level. According tothe Royal Society (1999), we need ‘to refine theexperimental design of the research done todate’. New ideas were also advocated in theLancet debate (Ewen and Pusztai, 1999b;Kuiper et al., 1999) and at the OECDConference in Edinburgh in February 2000.

Recommendations

Main tasks and methods for safetyassessment of GM crops

1. For compositional analysis and compari-son, the parent and transformed lines must begrown under identical conditions, treated andharvested the same way. In addition to pro-teins, starch, lipids, etc. of the parent and GMlines, their contents of bioactive componentsshould also be compared by novel methods(proteomics, fingerprinting, etc.).2. The stability to degradation by acid orpepsin or other proteases/hydrolases of GM


products, foreign DNA, including the geneconstruct, promoter, antibiotic resistancemarker gene, etc., has to be established in thestomach and intestines of model animalsin vivo. With GM lectins, including Bt toxin,the presence/absence of their epithelialbinding should also be demonstrated byimmunohistology.3. The biological, immunological, hormonalproperties and allergenicity of GM productsmust be established with the GM productisolated from the GM crop, and not withrecombinants from E. coli, as these two mayhave substantially different properties.4. As GM food is unlikely to be highlypoisonous, ‘toxicity’ is an unhelpful conceptand difficult to assay. In contrast, nutritionalstudies in which GM crop-based diets are fedto young growing animals should reveal theirpossible harmful effects on metabolism, organdevelopment, immune/endocrine systemsand gut flora, which together determine thesafety of the GM crop and the development ofthe young into healthy adults.5. For animal testing iso-proteinic and iso-energetic diets need to be formulated in whichmost of the dietary protein is derived from theGM crop. The composition of the control dietsshould be the same as the GM diet but contain-ing the parent line with or without supple-mentation with the isolated gene product atthe same level as expressed in the GM line.Groups of animals (five or more per group), ofsimilar weight, should be pair-fed in short-and long-term experiments. Urine and faecalsamples should be collected for the determi-nation of net protein utilization (NPU), nitro-gen balance and feed utilization ratios. Bloodsamples should be taken before, during andat the end of the experiments for immunestudies (i.e. lymphocyte proliferation assay,Elispot), hormone assays (insulin, cholecysto-kinin, etc.) and for the determination of otherblood constituents. The animals are to beweighed daily and any abnormalitiesobserved. After killing the animals, theirbodies should be dissected, the gut rinsedand its contents saved for further studies(enzymes, GM products, DNA). Sectionsshould be taken for histology, and the wet anddry weights of organs recorded and analysed.

Evaluation

With suitable statistical analyses (ANOVA,multiple comparisons and/or multivariateanalysis), the significance of differences, ifany, in the parameters should be established.

• If differences between animals fed GMand parent line diets indicate that thegenetic modification must have had asignificant effect on utilization andnutritional value, the GM crop cannotbe accepted for inclusion in the human/animal diet.

• If, similarly to the GM diet, the parentline diet spiked with the gene productshows differences, the use of this gene inGM food/feed is not acceptable.

• If negative effects are not observed withthe parent line diet containing the iso-lated gene product, it is likely thatthe harm is caused by the use of the par-ticular construct or by an unwanted orunforeseen effect of the gene insertion onthe genome.

Animal testing is but a first step andnot a substitute for human studies. If thereis no indication of harm to the animals, theresults will have to be validated with humanvolunteers in clinical double-blind, placebo-controlled drug-type tests. Such studies mayhave to go on for considerable lengths of time.It must also be kept in mind that any potentialharm with GM food may be most acute inthe young, elderly and sick, particularly thosesuffering from HIV, hepatitis or other viraldiseases. Many people suffer from allergiesand other disorders of the gastrointestinaltract, and for these the consumption of GMfood may have unforeseen consequences andsome of these may be irreversible. Thus, forthese, the clear labelling of GM food must bemade mandatory.

There is a compelling need to developfurther the concepts of biological testing,particularly for potential long-term effects.Since the GM potato work with male ratsshowed abnormalities in the development oftheir sexual organs, it is imperative that simi-lar experiments should be done with femalerats to be followed by studies of the effects on


reproductive performance of rats (or otheranimals) reared and maintained on GM vs.non-GM diets for several generations.

If there is a general willingness to fundresearch along these or similar lines and theregulators accept the concept of biological/toxicological testing transparently and inclu-sively, the methods are available for the workto start. Following this route, publishing theresults and consulting the public will ensurethat a technology which promised safe andplentiful food will deliver it for us all, and weare confident that if people see that everythinghas been done to establish its safety they willaccept it willingly.

References

Alliance for Bio-Integrity Website (1998) www.bio-integrity.org including Calgene FLAVR SAVRtomato report, pp. 1–604; InternationalResearch and Development Corp. first testreport, pp. 1736–1738; conclusions of theexpert panel regarding the safety of the FLAVRSAVR tomato, ENVIRON, Arlington, Vir-ginia, pp. 2355–2382; four week oral(intubation) toxicity study in rats by IRDC, pp.2895–3000.

Astwood, J.D., Leach, J.N. and Fuchs, R.L. (1996)Stability of food allergens to digestion in vitro.Nature Biotechnology 14, 1269–1273.

Berberich, S.A., Ream, J.E, Jackson, T.L., Wood, R.,Stipanovic, R., Harvey, P., Patzer, S. andFuchs, R.L. (1996) The composition of insect-protected cottonseed is equivalent to that ofconventional cottonseed. Journal of AgriculturalFood Chemistry 44, 365–371.

Bernstein, I.L., Bernstein, J.A., Miller, M., Tierzieva,S., Bernstein, D.I., Lummus, Z., Selgrade, M.K.,Doerfler, D.L. and Seligy, V.L. (1999) Immuneresponses in farm workers after exposure toBacillus thuringiensis pesticides. EnvironmentalHealth Perspectives 107, 575–582.

Betz, F.S., Hammond, B.C. and Fuchs, R.L. (2000)Safety and advantages of Bacillus thuringiensis-protected plants to control pests. RegulatoryToxicology and Pharmacology 32, 156–173.

Bindslev-Jensen, C. and Poulsen, L.K. (1997) Haz-ards of unintentional/intentional introductionof allergens into foods. Allergy 52, 1184–1186.

Brake, J. and Vlachos, D. (1998) Evaluation oftransgenic Event 176 ‘Bt’ corn in broilerchicken. Poultry Science 77, 648–653.

British Library File (1997) Public reference ‘BL SUP1113’ Chardon Hearing Documents, No.10. Available at: www.maff.gov.uk/planth/pvs/chardon/index.htm

Burks, A.W. and Fuchs, R.L. (1995) Assessmentof the endogenous allergens in glyphosate-tolerant and commercial soybean varieties.Journal of Allergy and Clinical Immunology 96,1008–1010.

Chiter, A., Forbes, J.M. and Blair, G.E. (2000) DNAstability in plant tissues: implications forthe possible transfer of genes from geneticallymodified food. FEBS Letters 24098, 1–5.

Domingo, J.L. (2000) Health risks of geneticallymodified foods: many opinions but few data.Science 288, 1748–1749.

European Commission Directorate C (2000) Opin-ion of the Scientific Committee on Food on ‘TheEvaluation of Toxicological Information Rela-ted to the Safety Assessment of GeneticallyModified Tomatoes’. CS/NF/TOM/8 ADD 1REV 3 Final. European Commission, Brussels.

Ewen, S.W.B. and Pusztai, A. (1999a) Effects ofdiets containing genetically modified potatoesexpressing Galanthus nivalis lectin on rat smallintestine. Lancet 354, 1353–1354.

Ewen, S.W.B. and Pusztai, A. (1999b) Authors’reply. Lancet 354, 1727–1728.

Fares, N.H. and El-Sayed, A.K. (1998) Fine structuralchanges in the ileum of mice fed on delta-endotoxin-treated potatoes and transgenicpotatoes. Natural Toxins 6, 219–233.

Flachowsky, G. and Aulrich, K. (2001) Nutritionalassessment of feeds from genetically modifiedorganism. Journal of Animal and Feed Sciences 10,Supplement 1, 181–194.

Hammond, B.G., Vicini, J.L., Hartnell, G.F., Naylor,M.W., Knight, C.D., Robinson, E.H., Fuchs,R.L. and Padgette, S.R. (1996) The feedingvalue of soybeans fed to rats, chickens, catfishand dairy cattle is not altered by geneticincorporation of glyphosate tolerance. Journalof Nutrition 126, 717–727.

Harrison, L.A., Bailey, M.R., Naylor, M.W., Ream,J.E., Hammond, B.G., Nida, D.L., Burnette,B.L., Nickson, T.E., Mitsky, T.A., Taylor, M.L,Fuchs, R.L. and Padgette, S.R. (1996) Theexpressed protein in glyphosate-tolerant soy-bean, 5-enolpyruvylshikimate-3-phosphatesynthase from Agrobacterium sp. strain CP4, israpidly digested in vitro and is not toxic toacutely gavaged mice. Journal of Nutrition 126,728–740.

Hashimoto, W., Momma, K., Katsube, T., Ohkawa,Y., Ishige, T., Kito, M., Utsumi, S. and Murata,K. (1999a) Safety assessment of geneticallyengineered potatoes with designed soybean


glycinin: compositional analyses of the potatotubers and digestibility of the newly expressedprotein in transgenic potatoes. Journal of Scienceof Food and Agriculture 79, 1607–1612.

Hashimoto, W., Momma, K., Yoon, H.J., Ozawa, S.,Ohkawa, Y., Ishige, T., Kito, M., Utsumi, S.and Murata, K. (1999b) Safety assessment oftransgenic potatoes with soybean glycinin byfeeding studies in rats. Bioscience BiotechnologyBiochemistry 63, 1942–1946.

Helm, R.M. and Burks, A.W. (2000) Mechanisms offood allergy. Current Opinion in Immunology 12,647–653.

Hohlweg, U. and Doerfler, W. (2001) On the fate ofplant and other foreign genes upon the uptakein food or after intramuscular injection in mice.Molecular and General Genetics 265, 225–233.

Kramer, K.J., Morgan, T.D., Throne, J.E., Dowell,F.E., Bailey, M. and Howard, J.A. (2000) Trans-genic avidin maize is resistant to storage insectpests. Nature Biotechnology 18, 670–674.

Kuiper, H.A., Noteborn, H.P.J.M. and Peijnenburg,A.A.C.M. (1999) Adequacy of methods for test-ing the safety of genetically modified foods.Lancet 354, 1315–1316.

Kuiper, A.H., Kleter, G.A., Noteborn, H.P.J.M. andKok, E.J. (2001) Assessment of the food safetyissues related to genetically modified foods.Plant Journal 27, 503–528.

Lappe, M.A., Bailey, E.B., Childress, C. and Setchell,K.D.R. (1999) Alterations in clinically impor-tant phyto-oestrogens in genetically modified,herbicide-tolerant soybeans. Journal of MedicalFood 1, 241–145.

Mercer, D.K., Scott, K.P., Bruce-Johnson, W.A.,Glover, L.A. and Flint, H.J. (1999) Fate of freeDNA and transformation of oral bacteriumStreptococcus gordonii DL1 plasmid DNA inhuman saliva. Applied and EnvironmentalMicrobiology 65, 6–10.

Metcalfe, D.D., Astwood, J.D., Townsend, R.,Sampson, H.A., Taylor, S.L. and Fuchs, R.L.(1996) Assessment of the allergenic potentialof foods derived from genetically engineeredcrop plants. Critical Reviews in Food Science andNutrition 36 (Supplement), S165–S186.

Millstone, E., Brunner, E. and Mayer, S. (1999)Beyond substantial equivalence. Nature 401,525–526.

Momma, K., Hashimoto, W., Ozawa, S., Kawai, S.,Katsube, T., Takaiwa, F., Kito, M., Utsumi, S.and Murata, K. (1999) Quality and safetyevaluation of genetically engineered rice withsoybean glycinin: analyses of the grain compo-sition and digestibility of glycinin in transgenicrice. Bioscience Biotechnology Biochemistry 63,314–318.

Nakamura, R. and Matsuda, T. (1996) Rice allergenicprotein and molecular–genetic approach forhypoallergenic rice. Bioscience BiotechnologyBiochemistry 60, 1215–1221.

Nordlee, J.A., Taylor, S.L., Townsend, J.A. andThomas, L.A. (1996) Identification of a Brazilnut allergen in transgenic soybean. NewEngland Journal of Medicine 334, 688–692.

Noteborn, H.P.J.M., Bienenmann-Ploum, M.E.,van den Berg, J.H.J., Alink, G.M., Zolla, L.,Raynaerts, A., Pensa, M. and Kuiper, H.A.(1995) Safety assessment of the Bacillus thurin-giensis insecticidal crystal protein CRYIA(b)expressed in transgenic tomatoes. In: Engel,K.H., Takeoka, G.R. and Teranishi, R. (eds)ACS Symposium Series 605 Genetically ModifiedFoods – Safety Issues. American ChemicalSociety, Washington, DC, pp. 135–147.

Novak, W.K. and Haslberger, A.G. (2000) Substan-tial equivalence of antinutrients and inherentplant toxins in genetically modified novelfoods. Food and Chemical Toxicology 38, 473–483.

O’Neil, C., Reese, G. and Lehrer, S.B. (1998) Aller-genic potential of recombinant food proteins.Allergy and Clinical Immunology International 10,5–9.

Osusky, M., Zhoul, G., Osuska, L., Hancock, R.E.,Kay, W.W. and Misra, S. (2000) Transgenicplants expressing cationic peptide chimerasexhibit broad-spectrum resistance to phytopa-thogens. Nature Biotechnology 18, 1162–1166.

Padgette, S.R., Taylor, N.B., Nida, D.L., Bailey, M.R.,MacDonald, J., Holden, L.R. and Fuchs, R.L.(1996) The composition of glyphosate-tolerantsoybean seeds is equivalent to that of con-ventional soybeans. Journal of Nutrition 126,702–716.

Pusztai, A. (2001) Genetically modified foods:are they a risk to human/animal health?Available at: www.actionbioscience.org/biotech/pusztai.html

Pusztai, A., Ewen, S.W.B., Grant, G., Peumans, W.J.,van Damme, E.J.M., Rubio, L. and Bardocz, S.(1990) Relationship between survival andbinding of plant lectins during small intestinalpassage and their effectiveness as growthfactors. Digestion, 46 (Supplement 2), 308–316.

Pusztai, A., Grant, G., Duguid, T., Brown, D.S.,Peumans, W.J., Van Damme, E.J.M. andBardocz, S. (1995) Inhibition of starch digestionby α-amylase inhibitor reduces the efficiencyof utilization of dietary proteins and lipids andretards the growth of rats. Journal of Nutrition125, 1554–1562.

Pusztai, A., Grant, G., Bardocz, S., Alonso, R.,Chrispeels, M.J., Schroeder, H.E., Tabe, L.M.and Higgins, T.J.V. (1999) Expression of


the insecticidal bean alpha-amylase inhibitortransgene has minimal detrimental effecton the nutritional value of peas fed to ratsat 30% of the diet. Journal of Nutrition 129,1597–1603.

Redenbaugh, K., Hatt, W., Martineau, B., Kramer,M., Sheehy, R., Sanders, R., Houck, C. andEmlay, D. (1992) Safety Assessment of GeneticallyEngineered Fruits and Vegetables: a Case Study ofthe FLAVR SAVR Tomato. CRC Press Inc.,Boca Raton, Florida.

Rogan, G.J., Bookout, J.T., Duncan, D.R., Fuchs, R.L.,Lavrik, P.B., Love, S.L., Mueth, M., Olson, T.,Owens, E.D., Raymond, P.J. and Zalewski, J.(2000) Compositional analysis of tubers frominsect and virus resistant potato plants. Journalof Agricultural and Food Chemistry 48,5936–5945.

Royal Society (1999) Review of Data on PossibleToxicity of GM Potatoes. Royal Society London.

Rubio, L.A., Grant, G., Caballé, C., Martinez-Aragon, A. and Pusztai, A. (1994) High in vivo(rat) digestibility of faba bean (Vicia faba), lupin(Lupinus angustifolius) and soya bean (Glycinemax) soluble globulins. Journal of Food Scienceand Agriculture 66, 289–292.

Ruibal-Mendieta, N.L. and Lints, F.A. (1998) Noveland transgenic food crops: overview ofscientific versus public perception. TransgenicResearch 7, 379–386.

Schubbert, R., Lettmann, C. and Doerfler, W. (1994)Ingested foreign (phage M13) DNA survivestransiently in the gastrointestinal tract andenters the blood stream of mice. Molecular andGeneral Genetics 242, 495–504.

Schubbert, R., Hohlweg, U., Renz, D. and Doerfler,W. (1998) On the fate of orally ingested foreignDNA in mice: chromosomal association andplacental transmission in the fetus. Molecularand General Genetics 259, 569–576.

Sidhu, R.S., Hammond, B.G., Fuchs, R.L., Mutz, J.N.,Holden, L.R., George, B. and Olson, T. (2000)Glyphosate-tolerant corn: the composition and

feeding value of grain from glyphosate toler-ant corn is equivalent to that of conventionalcorn (Zea mays L.). Journal of Agricultural andFood Chemistry 48, 2305–2312.

Taylor, N.B., Fuchs, R.L., MacDonald, J.,Shariff, A.B. and Padgette, S.R. (1999)Compositional analysis of glyphosate-tolerantsoybeans treated with glyphosate. Journal ofAgriculture and Food Chemistry 47, 4469–4473.

Teshima, R., Akiyama, H., Okunuki, H., Sakushima,J., Goda, Y., Onodera, H., Sawada, J. andToyoda, M. (2000) Effect of GM and non-GMsoybeans on the immune system of BN rats andB10A mice. Journal of the Food Hygiene Society ofJapan 41, 188–193.

Vazquez Padron, R.I., Moreno Fierros, L., NeriBazan, L., De la Riva, G.A. and Lopez Revilla,R. (1999) Intragastric and intraperitonealadministration of Cry1Ac protoxin from Bacil-lus thuringiensis induces systemic and mucosalantibody responses in mice. Life Sciences 64,1897–1912.

Vazquez Padron, R.I., Moreno-Fierros, L.,Neri-Bazan, L., Martinez-Gil, A.F., de la Riva,G.A. and Lopez Revilla, R. (2000a) Character-ization of the mucosal and sytemic immuneresponse induced by Cry1Ac protein fromBacillus thuringiensis HD 73 in mice. BrazilianJournal of Medical and Biological Research 33,147–155.

Vazquez Padron, R.I., Gonzalez Cabrera, J., GarciaTovar, C., Neri Bazan, L., Lopez Revilla, R.,Hernandez, M., Morena Fierros, L. and De laRiva, G.A. (2000b) Cry1Ac protoxin fromBacillus thuringiensis sp. kurstaki HD73 binds tosurface proteins in the mouse small intestine.Biochemical and Biophysical Research Communi-cations 271, 54–58.

US Food and Drug Administration (1994)Memorandum dated 17 May. Summary ofconsultation with Calgene Inc., concerningFLAVR SAVR tomatoes. Available at:www.bio-integrity.org


16 Genetically Modified Foods: Potential Human Health Effects

Documents

Transcript of 16 Genetically Modified Foods: Potential Human Health Effects