-
Asked by alexrive to Andy, Cathie, Jules, Les, Ricarda on 25 Jun 2012. This question was also asked by mrskwallace, a1lsa, lilygrainger, sophk.
Question: do you think that there may be unkown side affects with GMfood?
- Keywords:
-
Cathie Martin answered on 25 Jun 2012:
This question is based on the premise that genetic modifaction itself has or could have unknown side effects. But genetic modification is simply a method (technological process) for introducing new genes into crops (or other organisms of course). I think that there are enough experimental data available to suggest that ‘transformation’ per se does not have inherent unknown risks. Scientists have been routinely ‘transforming’ bacteria with DNA from other bacteria or other organisms since the 1970s, and many of our medicines have been developed using recombinant DNA technology, for example human insulin for Type 1 diabetics and growth factor for treatment of children who can not produce their own. (Before the advent of recombinant growth factor, people were treated with bovine growth factor, but this practice had to be discontinued because of the high incidence of CJD in such patients). Recombinant rennin is used in many Vegetarian Cheeses available on supermarket shelves.
Transformation of plants has been common in scientific labs since the mid 1980s, and the most common unpredicted effect is the loss of activity of the introduced gene. The scientific reasons for such effects are now understood pretty well. In practical terms, field trials are designed to test whether such gene silencing is likely in transgenic crops under assessment. Any lines with a high propensity for silencing would not be developed further (just like conventional breeding).
In the 1970s when recombinant DNA technology was adopted widely there were very strict regulations governing work with recombinant bacteria, but many of these were relaxed once it became apparent that the bacteria could be contained by good laboratory practice and transformed bacteria did not present unexpected problems to people exposed to them. Of course transfer of DNA between different bacterial species is common in Nature as well.Of course there may be unknown side effects associated with introduction of particular genes into crops. This is why a stringent risk assessment is always undertaken to assess the likelyhood of such events and the potential damage they could cause, before any regulatory approval for sales or dissemination of seed is given. Interestingly, conventional plant breeding which often involves introducing a large number of genes from one species to another, (in interspecific crosses followed by introgression) does not have accompanying regulatory oversight, yet somehow, disasterous unpredictable effects from introducing new varieties developed in this way have not been reported.
-
Andy Stirling answered on 25 Jun 2012:
This question gets to the heart of the matter. No technology that has ever been developed has ever performed entirely without surprise. Nor should this ever be expected. It is not even necessarily a bad thing. Many unexpected adverse effects might be judged to be worth it, given the benefits. And some surprises can turn out to be positive. Take the ‘unexpected side effects’ of the laser, for example (DVDs). Or the world wide web (online shopoing, wikipedia, facebook…).
The point is, that no technology – certainly of the kind or scale envisaged for GM – can be pursued without initially-unknown side effects. Claims that this is not the case are, to put it mildly, seriously misleading. Unexpected outcomes need not of themselves be bad. They are common to all technologies. But where this issue is denied or sidestepped, this is (in itself) is a worrying sign of the kind of attitude that drives the technology. GM advocacy sometimes comes close to this.
And there are more technical issues, to do with the difference between ‘risk’ – and conditions of ‘uncertainty’, ‘ambiguity’ and ‘ignorance’. Put simply, the conventional techniques of ‘risk assessment’ used in the regulation of GM foods (like other consumer products), simply do not – cannot – fully address these deeper and trickier forms of incomplete knowledge.
In even more technical terms, it is quite simply impossible for probabilistic risk assessment to address situations in which the probabilities themselves are not confidently known. This is true of many areas around GM foods. And risk assessment is also unable definitively to resolve a situation in which the problem is disagreement between experts. Where issues are raised by different groups posing contrasting questions or upholding different values (as with GM), risk assessment becomes even less useful.
So, to cut a long story short, there are very strong scientific reasons to predict that – as with any technology – GM will likely pose a number of unexpected implications. This does not mean that GM is necessarily bad or unacceptable. Nor does it mean that other technologies are preferable (since the same basic challenge applies across the board).
The point is, that the usual way that society deals with this, is to withdraw from a technology when and if it is realised that the benefits are outweighed by the impacts. This was the case, for instance, with asbestos… and lead in petrol… and chemicals like PCBs and benzene … and drugs like thalidomide… and the CFCs that disrupt the ozone layer… and (now increasingly) fossil fuels… In each case, our response to learning about initially-unknown side effects was to try to withdraw from the technology and (where we could) reverse the impacts that had occurred.
As living organisms, the particular issue presented in this regard by many GMOs is that they are not as reversible as many of these other examples. This is true, both in social and environmental terms. Changes that become embodied in living genomes are extremely difficult to reverse. And once farmers become dependent on a practice, it can become especially difficult to escape it. This may not be thought a problem, where we are fully confident in a risk assessment – which says these possibilities are negligible. But where we recognise that our knowledge on this score is necessarily incomplete, then we should take this reality correspondingly seriously. Put simply, the science of GM presents us with particular reasons for humility. But it is often treated as if the reverse were true.,
-
Ricarda Steinbrecher answered on 26 Jun 2012:
Yes, that is pretty well established by now. That’s also why all and any GM crops/foods need compositional analysis, laboratory and greenhouse testing, checking, feeding trials, risk assessments, the lot. Yet even with this we cannot necessarily tell whether we checked well enough and long enough, looked for the right thing, and what else might become or turn out to be a problem in the long term. Additionally there are cumulative and synergistic or antagonistic effects, that depend on what else is present in the environment or the food.
Briefly, there has been increasing investigation into the health effects of GM crops through independent research, largely using feeding trials, but also looking at the impacts of exposure to herbicides associated with GM crops, or looking at impacts from inhalation (including pollen, fibre or dust) and touch (eg for Bt cotton workers – Just to explain: Bt cotton is genetically engineered to express an insecticidal endotoxin derived from Bacillus thuringiensis (Bt for short), a soil bacterium used as a pesticide.) The evidence is that there are immune responses and allergenic reactions as well as health impacts due to low-level chronic or sub-chronic toxicity, which can be found for different GM crops and for different animals, including for humans in case of the Bt cotton workers. At times performing a different and new analysis of data originally submitted by the developer led to identification of negative health impacts previously undetected due to different statistical methods. Impacts of herbicides were found both in animals as well as in humans.
A surprising finding was reported by an Australian team in 2005 (Prescott et al.) that investigated the impact of a GM pea on mice. The gene used (alpha amylase inhibitor gene) had been taken from a close relative – the common bean – and nobody expected much of a problem. Yet the mice reacted to the GM pea but not to the bean. Unexpectedly the protein product from the bean gene changed its characteristics in the pea and became immunogenic, leading (for example) to inflammations. Despite having the same sequence, the protein had a different structure in the pea. And even more surprising, the mice did not only react to the GM protein, but also showed an increased immune response to seed storage proteins in the pea, which had not been observed before. This is called ‘immunological cross priming’ or ‘adjuvant effect’, where the presence of one immunogenic compound will cause the immune system to react to other compounds it has previously ignored or tolerated.
Unexpected side effects could also be found for plants genetically modified with genes from within the same species, sometimes referred to as cis-genics, thus indicating that any GM plant/food can result in unknown side effects.
-
Les Firbank answered on 26 Jun 2012:
GM changes the genetic information in a crop using quite precise technology. Conventional crop breeding using less precise technology, such as radiation or chemicals, to create mutations. Both GM and conventional plant breeding can go wrong. The difference is that GM is very tightly regulated, and so any side effects should be seen more easily than with non-GM. In general, I think it’s wonderful just how nutritious and safe the food is in our shops.
-
Julian Little answered on 28 Jun 2012:
Hi Alex, from time to time, you will a headline article that suggests an unknown side effect having been demonstrated with a GM food – Ricarda has illustrated some.
When they arise, there used to be a flurry of activity, frequently accompanied by headlines in the newspaper. The study would be looked at by the regulatory authorities and other scientists to see whether the experiments were valid, the results gathered were realistic and/or the conclusions appropriate. In all the cases that I am aware of, the pieces split into two forms – when the results were valid, they referred to experimental GM crops that were nowhere near commercialisation and would not have been allowed to be grown, or they were on commercial crops and were found not to be valid, either for experimental reasons or their interpretation.These days when such research, most newspapers are a little less likely to stick them on the front page – perhaps a case of “crying wolf” being no longer appropriate.
Something that I have used a few times now – the European Commission (which is not known for a pro-GM stance) recently published a report detailing the 120 projects it had funded over the last 25 years with 500 research groups looking at the safety of GM (http://ec.europa.eu/research/biosociety/pdf/a_decade_of_eu-funded_gmo_research.pdf ). It’s conclusion? “That biotechnology, and particularly GMOs, are not per se more risky than, eg conventional plant breeding technologies”.
It may not be a definitive answer, but it does suggest that current GM crops are at least as safe as their non-GM counterparts.
Comments
dingo commented on :
In fact there is a considerable number of animal feeding studies showing unexpected harmful effects from GM foods, including some that are already in the food chain. See Section 3 of this report:http://earthopensource.org/index.php/reports/58-gmo-myths-and-truths
dingo commented on :
@Les Firbank In what proportion of food crops in use today was radiation-induced or chemical-induced mutagenesis used? From what I can see, these methods have been derided (including by a UK govt expert panel) as being too risky and ineffective to reliably produce beneficial mutations. Conventional breeding, sometimes assisted by non-GM biotechnologies like marker assisted selection, appears to be the most effective and most commonly used crop breeding method.
And GM doesn’t seem to be precise, in that the location and effects of the introduced GM gene(s) can’t be controlled or predicted.
Andrew Maynard commented on :
A challenging question and a thoughtful set of answers! To me there are three things that stick out here – how difficult it is to predict the results of tinkering with complex systems; the fact that genetic alterations are designed to replicate in organisms; and the point that it isn’t the genetic modification that raises concerns so much as the substances that are produced as a result of that modification. These together make it tough to predict the consequences of generic modification over long periods.
Does this mean that probabilistic risk assessment runs out of steam here as Andy suggests? I’m not sure it does, as uncertainty can be factored into the analysis. But it does mean that we have to adapt our approaches to assessing risk to situations that are far more complex that “simply” releasing a potentially harmful chemical out into the environment.
That said, if a GM food itself tests safe, there are relatively few ways in which it could turn out to be “not safe” at some later point just because its DNA has been modified
Andy commented on :
It is curious that Andrew Maynard seeks to argue that ‘uncertainty’ can – like ‘risk’ – be resolved by probabilistic methods. He seems to be trying to contest, my own earlier observation that uncertainty is a situation under which we are, by definition, not confident in our knowledge of probabilities.
The usual word we give to conditions under which we can apply probabilities is ‘risk’. ‘Uncertainty’ is a situation where we cannot confidently do this. It is therefore self-evident that probabilistic risk assessment is inadequate fully to resolve a condition of uncertainty. That’s what the words mean!
This is a complex area, further confused by much inconsistent use of terms. As it happens, my own reference to uncertainty as a condition under which probabilities are not applicable follows the oldest and best established technical usages of the words ‘risk’ and ‘uncertainty’ – after the economist Frank Knight in 1920.
But the issue here is not just one of academics tussling over terms. The real issues are about how much authority we assign to expert risk assessment. Behind all this, grave misunderstandings can occur when these words are used carelessly or strategically.
More important than wrangling over definitions of particular words, then, is the serious misrepresentation that occurs where the tricky problem of ‘uncertainty’ (as I define it here) is treated as if it were the more easily-dealt-with state of ‘risk’. What this move does, is effectively deny that there exist any conditions under which probabilistic techniques do not deliver clear ‘science based’ results. This is seriously misleading.
In this light, Andrew’s implication that probabilities can be used to resolve uncertainty, is a pretty important – and problematic – claim. The question for Andrew would be: Does he feel there is any such thing as a situation under which we are not confident in probabilities? If not , why not? If so, this is what I’m calling ‘uncertainty’?
To deny the condition of uncertainty even a name, would be seriously to misrepresent the key issues – not only on GM but many other important areas of science and society. The inconvenient truth is, that there are many situations under which risk assessment is quite simply incapable of delivering definitive answers.
This said, none of this need be prohibilitve in itself for GM. Such dilemmas apply equally to many different technologies – including alternatives to GM. And there are many practical approaches that we can use, instead of risk assessment.
But it is crucial to recognise the real nature of this challenge of uncertainty. It highlights the role for value judgements, democratic accountability and disciplines like the precautionary principle. Above all, it underscores a need for humility about what can and can’t be decided on the basis of science.
Perhaps one of the most disturbing effects of some kinds of advocacy of GM, is the tendency to obscure and subvert honest and rigorous appreciation of the condition and implications of uncertainty. That’s why I’ve taken so much space to try to clarify the issues around this problem here.
reinhard commented on :
Ricarda’s pea story is outdated. The study published in 2005 concluded that the GM pea used in the study, which has never been approved commercially, could trigger allergic reactions in humans and animals. The latest findings of the GMSAFOOD project in 2012 confirmed this finding in mice but found that the equivalent non-GM pea also cause allergic reactions. So there’s no difference between GM pea and non-Gm pea. The paper: Campbell et al. 2011: Comparison of the α-Amylase Inhibitor-1 from Common Bean (Phaseolus vulgaris) Varieties and Transgenic Expression in Other Legumes—Post-Translational Modifications and Immunogenicity. . Agric. Food Chem., 2011, 59 (11), pp 6047–6054 DOI: 10.1021/jf200456j
In general, it can be concluded that very few clearly unexpected effects were observed during the large scale post-release growing of herbicide-tolerant crops and Bt crops within the last 10-15 years. This conclusion is based on a report for the Dutch Biosafety Commission called COGEM (Van den Brink et al. (2010): Inventory of observed unexpected environmental effects of genetically modified crops. Applied Plant Research CGM 2010-08, PPO no. 3250165700). The report can be downloaded here: http://www.cogem.net/showdownload.cfm?objectId=0C938759-1517-64D9-CCC960ED7BF04018&objectType=mark.hive.contentobjects.download.pdf
joseph110 commented on :
It is incorrect to say “Ricarda’s pea story is outdated”. That study (not “story”) found what it found and the results are not contradicted by the subsequent study, which had different findings but also a different protocol (different administration route, plus different beans were used in the first study for a part of the experiment). The researchers of the 2nd study (Campbell et al) themselves say that the different protocols may account for the different findings.
Ricarda commented on :
Dear Reinhard.
The Prescott et al. (2005) “pea story” is neither outdated nor superseded by Campbell et al. (2011).
Firstly, the paper you are referencing does not do what you think it does. It does not compare non-GM pea with GM pea and it does not say that the equivalent non-GM pea also causes allergenic reactions; it only looks at the α-Amylase Inhibitor-1 (αAI) from various sources.
On the contrary, Prescott et al., especially in their 2006 paper, were the only ones comparing and utilising the GM pea (producing the bean derived αAI) and the non-GM parental peas. In fact they developed a “model of food allergy whereby oral consumption of food [pea expressing α-amylase inhibitor-1 (αAI) from the common bean] promotes a T helper cell type 2 (Th2) inflammatory response and predisposes to cutaneous allergic reactions following subsequent food allergen (αAI) exposure.” Basically this is a model of food allergy that is “independent of adjuvant to examine how food allergy sensitization may predispose to and exacerbate cutaneous allergic inflammation”, including acute urticaria and atopic dermatitis (AD).
As can be seen for example in Fig. 4 (Prescott 2006) there is a clear difference in the immune response reactions between αAI-pea and non-GM pea.
The source of the αAI gene was the common bean (Phaseolus vulgaris, variety Tendergreen).
Whist Prescott et al. chose to use the Pinto variety of the common bean in parts of their feeding studies instead of the Tendergreen variety, they also used the purified αAI protein from Tendergreen, in particular in tests regarding skin and tracheal inflammations and cross-priming (see below).
Campbell et al. (2011) focuses largely on the identification of structural differences of α-Amylase Inhibitor (α-AI) proteins between different beans and the α-AI from the GM pea, in particular comparing post-translational modifications (eg glycosylation patterns). Using a finer detection method, Campbell could show that structural variation of the α-AI is common and they suggest that the variation(s) found in GM pea fits within the variable range found amongst beans. However, this does not allow the conclusion that any one of the variant proteins is or is not allergenic or immunogenic. This needs testing. Even small differences/variations can have very different effects.
And here is the challenge and the striking difference between the papers. Whilst Prescott et al (2005, 2006) have developed a strong and workable model by using oral exposure (eg feeding trials or administering food directly into the stomach) to test for food allergens and immunogens, Campbell et al only use injections into the body cavity (intra-peritoneal) to sensitize the mice and compare immune responses (here Th1 and Th2 antibody isotype responses). Intra-peritoneal injections are generally used in other contexts, in particular in animal testing of systemic drugs, when the drug is supposed to be delivered throughout the body. It is also used to sensitise the body to a substance, eg the injection with a purified substance is a bit like a vaccine, it warns the immune system. Next time when the same compound comes along the immune system will remember.
Campbell et al. were using the same specific mice (BALB/c) as Prescott. If the authors thought (though wrongly) that Prescott only used pinto bean and failed to use the Tendergreen bean α-AI in the feeding trial and immunological tests, wouldn’t it have been the right response to repeat those trials/tests and add also the Tendergreen bean? This would have been the proper follow up for concerns. Yet they chose to not do this but instead injected the purified food compound (α-AI) into the body cavity. Looking at toxicity or immunogenicity incl. allergenicity of food, it is more appropriate to do so by having the food or food component go down the throat and into the stomach. Intra-peritoneal injections are used to sensitise the body to a substance, which was also utilised by Prescott et al 2005 to create a ‘positive control’ for subsequent subcutaneous and intra-tracheal challenges. For this Prescott used the native Tendergreen bean αAI.
Please also look at Fig. 6 of Prescott 2005, showing that the intra-gastric administration of αAI from peas induces cross-priming of heterogeneous food antigens, here ovalbumin. This was compared to the cross-priming ability of Tendergreen bean αAI. One can see 4-6 fold increases in the mice reacting to ovalbumin when consumed/administered with pea αAI, as compared to when administered with native Tendergreen bean αAI.
Please look at Fig. 7 (Prescott 2005), and notice that the αAI from GM pea had an adjuvant effect (increased immunogenicity) on these major seed storage proteins. This is of particular concern, as what people mostly consume when eating peas is the seed storage proteins (eg. globulin, vicillin-4 and lectin).
None of this – nor of the 2006 paper – stands contradicted or superseded by Campbell et al. 2011.
On the contrary, Campbell et al. even state in the last sentence of their discussion, that their laboratories will compare the immunogenic responses by using oral exposure of mice to common bean (Tendergreen) and GM pea, acknowledging that feeding trials are the appropriate way to test for food allergens, immunogens or toxins, ie that their work needs to be followed up with such trials.
When looking carefully at Prescott’s work and Campbell’s work, Prescott – with the exception of the structural characterisation of αAI – remains valid and consistent.
References:
Prescott VE, Campbell PM, Moore A, Mattes J, Rothenberg ME, Foster PS, Higgins TVJ and Hogan SP (2005). Transgenic expression of bean α-amylase inhibitor in peas results in altered structure and immunogenicity. J Agri Food Chem. 53:9023–9030
Prescott VE, Forbes E, Foster PS, Matthaei K, Hogan SP (2006). Mechanistic analysis of experimental food allergen-induced cutaneous reaction. J Leukoc Biol. 80(2):258–266.
fern commented on :
@amaynard – it may be true that uncertainty can be factored into a risk analysis, but can ambiguity and ignorance also be factored in?
Also, why do you put “not safe” in inverted commas in your final sentence but not “safe”? What it means for something to ‘test safe’ is so political and full of value judgements that I would say something can test safe by one person’s standards and turn out ‘not safe’ by anothers. Also of course ‘just’ DNA modification may create effects that are not immediately obvious when testing is done but rather emerge through interactions in complex systems over time (demonstrating our ignorance of asking all the right questions during testing!). For example, the vast majority of GM crops currently in cultivation have a viral promotor as part of their inserted gene cassette, have we really asked and tested what might happen if this promoter becomes a site promoting viral recombination? Ignorance about all the right questions to ask when testing if something is ‘safe’ means there is always the possibility that things may turn out to be ‘not safe’ at a later date, even if the thing we are interested in has ‘just’ had its DNA modified.
Andrew Maynard commented on :
Hi fern,
The inverted commas around “not safe” were simply there because this is an even harder concept to pin down that “safe”, although I take your point that “safe” is ultimately a policy rather than a scientific decision in many cases.
What I really appreciate in your response though is the issue of asking – or not asking – the right questions on risk and safety. This is where more research, better informed and more critical thinking is needed in many areas of risk analysis and management – including GM foods. That said, questions need to be plausible and relevant – it’s very easy to ask questions that can derail a process by raising doubts, by that have little relevance or plausibility.
pheed commented on :
@fern – How would ‘viral recombination’ happen at a promoter site? I don’t understand. As far as I know, a promoter just tells the machinery of the cell ‘start making RNA from the next bit of DNA’. What is viral recombination?
joseph110 commented on :
@pheed I am not sure about the promoter site aspect of yr question, but viral recombination as a result of genetic engineering is explained in section 5.12.4 of this report:
http://www.earthopensource.org/index.php/reports/58-gmo-myths-and-truths
dingo commented on :
Interesting that Cathie Martin cites GM insulin as a GM success. In fact this substance has been plagued with problems and severe side effects since its introduction and some pharmaceutical companies (which had planned to phase out animal insulin) have responded by trying to keep animal insulin available. I am not a diabetic but was alerted to the problems with GM insulin by a diabetic who is also a doctor and clininal researcher. He had gone onto GM insulin himself and suffered horrific side-effects; unfortunately it took months before he was able to firmly identify the GM insulin as the culprit because he was not told that the new insulin he’d been prescribed was different from his customary animal insulin. A return to animal insulin rapidly reversed his problems.
Aspects of this issue are discussed here:
http://www.iddtinternational.org.uk/campaigning/gmvsanimalinsulin.htm
Cathie commented on :
The ‘problems’ with the cloned human insulin resulted from the fact that it was completely pure. Up to the point that cloned human insulin was produced, preparations of insulin from animals were ‘contaminated’ with trace amounts of another homone that could counteract the effects of insulin, so people had got used to having larger injections. When they injected equivalent amounts of the cloned insulin, they effectively overdosed because it did not contain the contaminents. So the ‘problems’ were not because of the cloned human insulin being GM. Your friend should have explained this distinction to you. The vast majority of insulin currently used worldwide is now biosynthetic recombinant “human” insulin or its analogues.
I am a type 1 diabetic and I have been taking cloned human insulin for 20 years, with only positive results.
dingo commented on :
In fact the problem did not seem to be one of dose but of the lack of warnings it gave of ‘hypos’, as mentioned on this website: “The adverse reactions [to GM insulin] affect the ability to satisfactorily control blood glucose levels, the ability to recognise low blood glucose levels [hypoglycaemia] so preventing remedial action being taken so increasing the risk of coma and even death.” He also complained of bizarre personality changes, which he never could explain.
You say that the problems with GM insulin were not due to GM. This remains an unproven hypothesis, not proven by research. But it seems research does underlie the warning given by the diabetics’ website: “‘Human’ insulin was originally thought to have advantages over natural animal insulin but research has since demonstrated that it has no clinical advantages for patients over animal insulin and there is no long-term safety data.”
joseph110 commented on :
For the very latest in unexpected side-effects from GM animal feed (not human food because there’s not much GM human food in the EU), see this story about a major GM company being criminally charged for covering up livestock deaths from GM corn:
http://www.gmwatch.org/index.php?option=com_content&view=article&id=14000