Deceptive reporting of EU-funded feeding study on GM maize

January 25, 2019 by Claire Robinson

GMO90+ study is falsely claimed to show the GM diets had no adverse effects and to refute the Séralini long-term study on GM maize and Roundup.

An EU taxpayer-funded rat feeding study on two GM maize varieties found significant differences in the rats that ate the GM diets. These differences could indicate adverse health impacts, but the authors dismissed them as not biologically relevant, without proper scientific justification.

In fact the relevance of the changes is unknown because the study was too short to measure long-term effects, which can take one to two years to show up, and because it was confined to one generation of rats.

The GMO90+ study tested two types of GM maize, NK603 and MON810, over a 6-month period. It was published in December 2018 in the journal Toxicological Sciences under the self-explanatory title, “The GMO90+ project: lack of evidence for biologically significant effects of genetically modified maize based-diets on Wistar rats after 6-months feeding comparative trial”. The authors reported no adverse health effects that could be attributed to the GM diets tested.

However, on closer examination of the study, this conclusion is scientifically unjustified. Statistically significant differences found in the GM maize-fed animals when compared with animals fed non-GM maize include alterations in the blood and urine biochemistry as well as changes in urine metabolites and hormone levels. Such alterations could lead to ill health.

In addition, differentially expressed genes were found in animals fed a GM diet (see the list below this article for details). Some alterations in gene expression were evidently caused by changes in the GM maize caused by spraying with glyphosate herbicide, since the sprayed NK603 maize had different effects on the rats than the unsprayed NK603 maize.

The authors dismissed findings of differentially expressed genes, for example, in the kidneys of females in groups fed GM NK603 maize compared with females fed the non-GM isogenic maize, when only one gene was involved. This is invalid because a change in the expression of only one gene can lead to far-reaching effects – even the difference between health and disease. For example, a change in the expression of a single gene has been found to be associated with coronary artery disease in a group of people who have the disease, as compared with a group of people who do not have it. In addition, it is well established that mutations in single or a few genes, which either increase or decrease their expression, can contribute to serious illnesses, including cancer.

Biological relevance unknown

The authors commented on the differences found in the GM-fed animals that “no biological relevance could be established owing to the absence of difference in biologically linked variables, dose-response effects or clinical disorders”.

But Dr Michael Antoniou, a London university-based molecular geneticist, said, “The authors’ conclusion of absence of biologically meaningful effects of the GM maize diets is based on an assumption that we know all there is to know about the biology of the animal under test. But this is far from being the case. There is still much that remains to be discovered and understood about how the physiology works in health and disease.”

Dr Antoniou continued, “Based on a relatively short study period such as that employed in this study, it is not possible to state that the statistically significant differences such as in gene expression profiles in various organs seen in treatment groups compared to controls, signify no adverse effects. These changes could indicate that ill health may arise in the long term. Hence the need to extend this study to the entire lifespan of the test animals to show safety or otherwise.”

Give all of the above, it appears that the study should more accurately have been titled, “GMO90+ project: evidence of statistically significant effects from a GM diet; biological relevance unknown”.

Misleading press coverage

An article by the French journalist Sylvestre Huet in the newspaper Le Monde implies that the GMO90+ study, along with other EU-funded GMO research projects, disproves the long-term Séralini study on NK603 maize. This study, led by French researcher Prof Gilles-Eric Séralini, found that rats fed GM NK603 maize over a long-term two-year period developed liver and kidney damage, as well as disruptions in hormones.

Le Monde titled its story, “GMO poisons? The real end of the Séralini affair”. It claimed that the GMO90+ study showed that alarm generated by the media that reported on the Séralini study was fake news.

Differences between GMO90+ and Séralini study

The problem with Le Monde’s approach is that the GMO90+ study (as well as the other EU-funded studies) is in no way comparable with the Séralini study, just as it is not possible to compare apples with oranges. Below are just some of the differences.

1. GMO90+ failed to test long-term effects

Most importantly, the Séralini study lasted for two years, whereas the GMO90+ study lasted only 6 months. While an improvement on the 3 months that industry generally allows for its GMO safety studies, 6 months is too short to identify long-term health effects. It is a mystery as to why the GMO90+ researchers settled for such an inadequate study length, as it is by now well known that short studies show a variety of changes in GM-fed animals, the relevance of which is unclear because researchers are unable to see whether or not the changes develop into more serious illness.

2. GMO90+ used a different rat strain

The GMO90+ researchers used a different strain of rat – the Wistar to Séralini’s Sprague-Dawley – which is much less sensitive to the kinds of mammary tumours observed in the GM maize- and Roundup-dosed animals in the Séralini experiment.

3. Roundup effects were not tested in GMO90+

In a crucial difference from the Séralini study, the GMO90+ study failed to test the effects of glyphosate herbicide, commonly sold as Roundup, which is used on GM NK603 maize. The Séralini study researchers carefully controlled chemical inputs during cultivation of the test and control maize as well as the base diet. As a result no glyphosate was detected in the base or control diets. Thus the researchers were able to test for the effects of glyphosate herbicide by including a “Roundup only” group, which included a very low dose below EU permitted levels for glyphosate in drinking water.

Even the lowest dose of Roundup was found to cause liver and kidney damage in the rats, a finding that was later confirmed in separate experiments in which “omics” molecular analyses of the rat tissues were performed. The “omics” analyses showed that the Roundup-dosed rats suffered from non-alcoholic fatty liver disease.

In the GMO90+ study, all the diets were equally contaminated with glyphosate. Thus the effects of the chemical could not be tested, apart from indirectly, in terms of changes in the NK603 maize caused by the glyphosate herbicide spraying, as tested in the NK603 + glyphosate group.

Such changes were present and did have biological effects on the rats, as evidenced by the fact that glyphosate-sprayed NK603 maize had different effects on the rats’ physiology than unsprayed NK603 maize.

4. GM plants and non-GM isogenic relatives were grown in different locations in GMO90+

An important failing in the GMO90+ study is that the GM plants and their non-GM isogenic relatives were grown in completely different locations, adding “noise” from different environmental conditions to the data that will mask any effect of the genetic modification. In contrast, in the Séralini study, the GM maize tested and its near-isogenic non-GM relative were grown at the same time in the same location.

5. Statistically significant differences were dismissed by the GMO90+ authors

In the Séralini study, statistically significant differences in rats fed NK603 maize and Roundup herbicide were not dismissed but were taken seriously and followed up with further experiments, which confirmed and amplified some of the findings – for example, the “omics” analyses that revealed non-alcoholic fatty liver disease in the Roundup-dosed rats.

In the GMO90+ study, in contrast, statistically significant differences in GM-fed groups of animals were repeatedly dismissed as not biologically relevant.

Researcher provides table of differences

Dr Nicolas Defarge, who was one of Séralini’s co-researchers on the two-year study and is now a researcher in environmental sciences at ETH Zurich, Switzerland, has prepared a table setting out the major differences between the two studies and noting the significant differences that were dismissed by the GMO90+ study authors. We are pleased to publish the table, with Dr Defarge’s permission, on GMWatch.

Dr Defarge commented on the GMO90+ results:

“All the diets (including the control diet) used in the GMO90+ study were contaminated by residues of glyphosate-based herbicides (glyphosate and its main metabolite, aminomethyl phosphonic acid AMPA) at about the same level, globally less than 75 µg/kg. This is far above the lowest dose of Roundup in water (0.1 µg/L) given to the treated rats in the Séralini et al. study and in the followup ‘omics’ studies by Mesnage et al.

“I consider that the main risk factor of herbicide-tolerant GMOs arises from the herbicide residues they contain, before the risk due to the genetic transformation (that may exist and hence should be assessed case-by-case, according to the Cartagena Protocol). In this perspective, it appears that the GMO90+ study lacks a real negative control since the control diet contained similar amounts of glyphosate-based herbicides as the diets containing the Roundup-tolerant GMO given to the treated rats.

“This flaw questions any conclusion on innocuity drawn in the GMO90+ study because any biochemical difference between treated and control rats may be hidden by the adverse effects probably triggered by Roundup residues in the control group.”

Other EU-funded studies also don’t show GMOs are safe

The Le Monde article cites other EU-funded research projects that it claims show that the GMO(s) tested are not toxic. However, these projects have not produced any peer-reviewed published data that shows this and on the contrary, some data indicates that the GMOs tested may be unsafe.

  • The MARLON project included Gijs Kleter and Esther Kok, insiders from the agbiotech industry-funded lobby group ILSI, on the research team. In fact, ILSI insiders oversaw and managed the EU’s GMO risk assessment development process from the very beginning of the importation of GMOs. The MARLON project did not actually test the safety of any GM food by feeding it to animals. Instead, it focused on developing surveillance methods to look at the effects of GM feed on animals. It also argued that animal feeding studies longer than the 90 days favoured by the industry were unnecessary.
  • The GRACE project did carry out a 90-day rat feeding study with GM MON810 maize that claimed to find no adverse effects from the GM diet. Once again ILSI insiders Kok and Kleter were involved. In fact the study showed that the GM maize diet caused an decrease in the total serum protein concentration and pancreas weight and an increase in blood glucose levels – effects that the authors claimed were toxicologically irrelevant. The German research organization Testbiotech said the authors’ dismissal of these effects was “not justified”.
  • G-TwYST studied the effects of feeding GM NK603 maize to rats over a two-year period, as did Séralini. However, there the similarity ends, since this study is in no way a replication of the Séralini experiment. It used a different strain of rat – the Wistar, which is not as sensitive to tumour-promoting effects as Sprague-Dawley rats, the strain that Séralini used and which is also used in the US National Toxicology Program, as well as by Monsanto, to study cancer-causing and other long-term toxic effects. Also unlike the Séralini study, G-TwYST failed to test the effects of Roundup alone. It also failed to test the maize as it is normally grown. Under real-life growing conditions, the maize is sprayed more frequently and in higher doses than in this study.

Nonetheless, the GM-fed animals in G-TwYST did show a significantly higher body weight than the non-GM-fed animals, an effect that was not mentioned by the researchers when they gave a press conference ahead of publishing their results in a peer-reviewed journal. Peer-reviewed publication still has not happened at the time of writing this article.

At the press conference, the researchers claimed there were “no toxicologically relevant effects” of the GM maize. However, given the history of deception around EU-funded studies on GMOs and the common abuse of phrases like “no toxicologically/biologically relevant effects”, we should wait for the results to be published before taking their claims at face value.

The way that these studies were written up and reported is heavily politically oriented. Rather than just reporting the results, the researchers used them to question the need for any feeding trials at all with GM foods prior to commercialisation. They dismissed statistically significant effects of the GM diet in the test animals as not biologically or toxicologically relevant, thus enabling potentially “slow-poison” GMOs to remain on the market and putting human and animal health at risk.


SIDEBAR:  GMO90+: Differentially expressed genes in rats fed a GM diet

The following are among the effects that can be seen in Table 6 of the published paper:

  • In males at the 180-day time point, 822 genes were differentially expressed in the kidneys of males fed 33% MON810 maize, compared with males fed the non-GM isogenic variety. This effect was dose-dependent, as males fed 11% GM maize also had more differentially expressed genes than those fed the non-GM isogenic, but fewer than the 33% group.
  • In the kidneys of males fed NK603 maize treated with glyphosate herbicide at the 180-day time point, the number of differentially expressed genes was very high (3,658), when compared with the males fed NK603 maize not treated with glyphosate herbicide. This shows there was an effect on the maize from the treatment with glyphosate herbicide, which in turn affected the animals’ gene expression. No direct effect from the glyphosate herbicide on the animals’ gene expression could be identified due to the fact that all diets in the experiment were equally contaminated with glyphosate, including the base and control diets.

The authors dismiss this result because it was not correlated to major metabolic modifications in blood plasma and urine, or to changes in a major signalling pathway. This is invalid as there may be another type of effect on health that is not signified by the parameters the authors looked at.

The data on urine metabolite modifications has not yet been published (presumably it will be published at the time of print publication rather than the “first online” version that is available currently).

  • In the liver of females at the 90-day time point, there was one differentially expressed gene in the 33% NK603 group compared with the 33% NK603 sprayed with glyphosate group (an indirect “glyphosate effect”). Similarly one gene in the liver was differentially expressed at the 90-day time point in female rats fed 33% MON810 compared with those fed the non-GMO isogenic maize (a “GM effect”). Two genes were differentially expressed at the 180-day point in the liver of females fed 11% MON810 maize compared with those fed non-GMO isogenic maize (a “GM effect”). Eleven genes were differentially expressed in the liver of males fed 11% NK603 maize compared with males fed non-GM isogenic maize (a “GM effect”).

The authors wrongly dismiss findings of one differentially expressed gene at the 90-day time point when comparing the kidneys of females fed 11% NK603 maize to the non-GM isogenic maize-fed group; and the females fed 33% NK603 compared with the non-GM isogenic-fed group, on the grounds that the number of such differentially expressed genes is “very low”. This is invalid because as explained above, a change in the expression of only one gene can produce significant effects on health.


  • This article first appeared on the GM Watch website and is reproduced here with permission