In my last post, we took a look at the infamous Séralini study, Long-term toxicity of a Roundup herbicide and a Roundup-tolerant genetically modified maize . The study was poorly conducted and ultimately added little to the scientific understanding of the safety of Monsanto’s Roundup-tolerant NK603 genetically modified (GM) maize.
This article will examine the original Monsanto paper that spawned this GMO melodramatic episode . I'll also be using this post to deliver some broader educational points about the safety, science, and regulations behind genetically modified organisms.
From Lab To Market
To get a GM food to market is a long, drawn-out, expensive process. From the initial discovery of a desirable gene or trait to the point where it can be sold on the market takes an average of 13.1 years (but can range from 7 to 24 years) and costs $136 million . The regulatory step in this process is the longest, taking 8 or more years to complete, due to the fact that there are three US agencies that regulate GMO crops: the FDA, USDA, and EPA .
The USDA, under The Plant Protection Act of 2000, has the authority to conduct mandatory reviews of GM plants to assess their potential impact on current agriculture and the environment [3,4,6]. Data is collected on the new plant by conducting a series of small field trials. In order to pass regulatory review, the USDA requires “a molecular, biochemical, and cellular characterization of the GE plant, along with data on the life cycle, reproductive characteristics, and any expected or unexpected changes from non-engineered plants of the same species” .
The FDA conducts voluntary reviews to assess the safety of GMO consumption for humans and animals. They evaluate the presence of new or increased toxins, allergens, and examine alterations in overall nutritional composition. While this system is still technically voluntary, all currently available GMO’s in the US have undergone this consultation [3,5].
The EPA deals mainly with the health and environmental risks of pesticides. The EPA does not specifically assess the GM food but rather any pesticide, herbicide, insecticide, or fungicide that may be used in tandem, such as the use of Roundup (glyphosate) with Roundup Ready corn [3,8].
Internationally, the Organization for Economic Cooperation Development (OECD) sets additional testing standards, including the requirement of a 90-day animal toxicity study [7,9]. It was this type of trial that was performed in the Monsanto paper we’re about to review. 90-day trials are performed by feeding rodents high amounts of the GE crop, typically up to 1/3 of diet, and then comparing them to a control group eating the closest isogenic, non-GM form of that crop. Typically measured toxicology endpoints include “body weight, food consumption, clinical findings, clinical chemistry parameters (hematology, serum chemistry, and urinalysis), macroscopic observations at necropsy, organ weights, and microscopic findings at histopathological examinations” . To date, two independent reviews of animal feeding trials, including long-term and multi-generational trials, have not found any evidence of toxicological harm in animals fed GE crops [10,11].
Conflicts Of Interest
While there is not a conflict of interest section in this paper there is little need for one, as it shows right in the title the scientists conducting this study are employed by Monsanto.
The Monsanto Study
Onto the paper!
This 2004 study looks at the toxicological effects of Roundup Ready corn (NK603) fed to Sprague Dawley rats over 13 weeks. The GM corn in this study has been modified to resist the herbicide Roundup by introducing the CP4 EPSPS gene sequence into the corn’s genome so it will produce CP4 EPSPS enzymes, which have a low affinity for glyphosate . In plants that do not produce these CP4 EPSPS enzymes, glyphosate will inhibit EPSPS activity and prevent the biosynthesis of 5-enolpyruvyl-shikimate-3-phosphate (EPSPS) proteins. This EPSPS inhibition will halt the production of aromatic amino acids essential for the plants growth; primarily phenylalanine, tyrosine, and tryptophan. In essence, the plant will starve to death.
Additionally, the CP4 EPSPS protein is rapidly digested in mammals, reducing exposure, and has no significant structural similarities to known toxins or allergens [12,17]. However, CP4 EPSPS does share a homology of 7 contiguous amino acids with the dust mite allergen, der p 7 (more on this later) . CP4 EPSPS has also not been shown to be toxic in any other organisms .
It should be noted that the pathway through which Roundup (glyphosate) acts on Roundup Ready plants is not present in mammals (that’s us), fish, birds, reptiles or insects [12,13]. The absence of this pathway in humans in one of the reasons why glyphosate has a relatively low toxicity (LD50 5,108 mg/kg) . Both table salt (LD50 3,000 mg/kg) and caffeine (LD50 192 mg/kg) are more toxic than glyphosate [14,15,16].
But I digress.
400 rats were broken up into 10 groups. Each group had 20 males and 20 females. The animals were divided into groups via stratified randomization so that body weights were not significantly different (P<0.05). The groups comprised of the following:
- Control (40) – 11% Non-GMO corn. Genetically similar to GM corn but lacking CP4 EPSPS trait
- Control (40) – 33% Non-GMO corn. Genetically similar to GM corn but lacking CP4 EPSPS trait
- Roundup Ready Corn (40) – 11% NK603 Corn
- Roundup Ready Corn (40) – 33% NK603 Corn
- Reference Control A (40) – 33% Non-GMO grown in Ohio
- Reference Control B (40) – 33% Non-GMO grown in Iowa
- Reference Control C (40) – 33% Non-GMO grown in Indiana
- Reference Control D (40) – 33% Non-GMO grown in Ohio
- Reference Control E (40) – 33% Non-GMO grown in Colorado
- Reference Control F (40) – 33% Non-GMO grown in Colorado
The incorporation of the reference groups was justified as such in the paper:
“The purpose of these reference controls was to approximate the normal range of responses of rats fed different commercial, non-transgenic corn grain since these data were not available.”
Others have criticized this, saying that the reference groups just add noise to the data and the only comparison that matters is the control to GMO group. As we will see in the results section, a direct comparison was made between the control rats and the Roundup Ready (RR) rats in addition to the reference controls.
All grain samples used were analyzed for nutrient components, pesticide residues, and mycotoxins (mold). The identity of the GMO and Non-GMO corn used in the study was confirmed via polymerase chain reaction (PCR) analysis. In the control and Roundup Ready groups where rats were consuming 11% corn, corn grain was supplied by Purina TestDiet to bring total corn content up to 33% to be comparable to the other diets.
Clinical observations included:
- Twice daily monitoring for mortality and moribundity
- Once daily monitoring for overt signs of toxicity
- Weekly physical examination
- Weekly weigh-ins
- Individual food consumption tracked weekly after baseline food consumption was taken
Blood samples were taken from 10 rats per sex per group. Urine was collected from the same rats used for blood samples. At the end of the 13-week trial, the rats were anesthetized and given a complete pathologic examination. Organs, bones, and tissues were harvested for further testing and analysis.
In no particular order, here were the findings of the paper.
Body Weight And Food Consumption – no statistically significant differences
Clinical Pathology Parameters – no statistically significant differences
Hematology – no statistically significant differences
Serum Chemistry – no statistically significant differences
Urine Chemistry – from the paper:
“There were no differences between the control and treated groups that were considered to be test article related (data not shown). Urine phosphorus and potassium were increased slightly in males from the high dose Roundup Ready corn group when compared to the control group, but were within the mean ±2 standard deviations of the population of reference controls. Values for the male control group were lower than the reference control groups and test groups that contributed to the observed statistical differences.”
I'll note here that I think they should have included a table with this data, as they had done for most of the others.
Organ Weights - from the paper:
“The only statistical difference in organ weight was a slight increase in absolute heart weight of high dose test males (1.98 grams) relative to control (1.78 grams) and reference control males (1.87 grams). However, this difference was within the mean ±2 standard deviations of the reference control population…there was no dose response as the 11% Roundup Ready male heart weight was 1.99 grams, similar to weights for 33% Roundup Ready males, and similar to 2 reference control groups (1.92 and 1.94 grams).”
Pathology results – no statistically significant differences
My biggest beef with all these papers is that they do not release their raw data. I find this absurd. Head on over to AllTrials.net to read more about this issue.
The OECD guidelines do not explicitly call for blinding when performing post-mortem examinations, but it seems like such a simple thing to do in order to eliminate potential bias. Like the Séralini paper, no blinding measures were taken.
Food intake was recorded but not presented. Only data on what the feed contained the test substance was given.
No values for the urine chemistry were given.
Details on the storage conditions of the feed were not provided per EFSA guidelines.
I've seen some bloggers complain that only 10 rats from each group of 20 were used to collect data for hematology and urine chemistry. The OECD Guidelines at the time only recommended the use of 10 out of the 20 rats per group for biochemical analysis . Séralini and this Monsanto study both used the prescribed 10 rats.
No isogenic corn was used as the control. Instead “background genetics representative of the test line [was used that lacked] the nk 603 transgene” . I’m not sure how big of an impact this would have had on the results given the variability we saw in the reference controls.
Speaking of the reference controls, I understand why they included them in the study but I'm not convinced that they needed to be there. Neither the EFSA nor the OECD guidelines call for their use and yet they seem to be pretty standard in GMO 90-day toxicology studies. At least in this paper, no scientific explanation is given for their inclusion.
Oh, and how in the hell am I supposed to decipher these graphs?!
Additional Feeding Studies
I was able to find three additional animal studies looking at the same NK603 strain of corn. The first study was conducted in 800 broiler chickens where the intervention group was fed 63% NK603 GMO corn for 6 weeks. No clinically significant differences were observed in “weight gain, feed intake, feed efficiency, carcass, and quality measures” [18,19]. The second study looked at 304 grower and finisher pigs, where the intervention diets ranged from 68% to 82% NK603 GMO corn, also found no significant differences in growth performance or carcass characteristics . The second looked at 396 feedlot steer being fed 73-79.5% NK603 corn. Again, no statistically significant differences were found between the control and intervention groups .
Via the EFSA’s website:
“Given the level of public interest, EFSA will make all data on genetically modified (GM) maize NK603 publicly available on its website today (14 January, 2013). While the Authority has already made available these data upon specific request on several occasions, any member of the public or scientific community will now be able to examine and utilize the full data sets used in this risk assessment.”
. Be warned, it’s a 500 megabyte ZIP file and takes a while to download off of their sluggish servers. Unfortunately, they did not release the data on safety tests with glyphosate. They remain confidential, as it is the “general rule for pesticide data” .
A Call For Clearer Guidelines
OECD guidelines need to be updated for animal study models. They were originally designed for use in chemical toxicology tests and are in need of some tweaking to make them better suited for food toxicology tests. The current and future guidelines need to be enforced as well. As they currently stand, application of these rules are not consistent . At the moment, I would give enforcement a B/B+.
The EFSA is also working on procedures for 1-2 year animal feeding studies .
A Side Note on Allergies
Remember 30 paragraphs ago when I mentioned that the CP4 EPSPS protein in NK603 Roundup Ready corn shared a homologous amino acid sequence with the dust mite allergen? Well, I started looking into this issue of GMO’s and increased risk of allergenicity to see if it was a justifiable concern. After writing about 8 pages on the topic, I realized it needed to be its own separate post. Look for that soon.
When you hear people talk about their concerns over GMO safety for human consumption, it usually boils down to something along these lines:
“I’m concerned that GMO’s will introduce some unknown substance/allergen/toxin into the food system.”
“Sure, we can change some of the genes, but we cannot predict the downstream effect it will have on the entire genome.” (AKA the Pleiotropic Effect: where you have multiple effects resulting from a single genetic change).
I believe that those who genuinely hold these reasonable concerns don't understand that GMO is an umbrella term. Your average Joe/Jane probably thinks of a GMO as a plant that has been injected with a gene from another plant or animal. In fact, there are many different methods of genetic modification. Here is a brief overview :
– “Traditional” breeding. Crossing two plants in hopes of creating a new plant with favorable attributes. I.E. higher yield, resistance to cold weather, drought tolerance, etc…
– where whole genomes are duplicated or added.
Examples: apples, strawberries, wheat, and bananas.
– Exposing seeds to chemicals or radiation to damage DNA that hopefully yields desirable traits.
Examples: peppermint, grapefruit, pears, apples, rice, and mint.
Crossing Species Barriers
(Interspecific Hybridization) – crossing two distinct species to create a new one.
Examples: tangelo, plutos, rice, wheat, some apples.
– what is typically thought of as a ‘GMO’. Recombinant DNA (rDNA) molecules are designed to bring genetic sequences to new sources.
Examples: corn, canola, soybeans, cotton, and papaya
So here’s the kicker. When transgenic technology is used to make corn express the CP4 EPSPS protein, an average of 1-3 genes are affected. We know exactly which genes are affected, what they do, and where they sit in the genome. All of the other genetic modification methods change anywhere from 10,000 to 800,000+ genes.
There are people who rant and rave about the evils of transgenic GMO’s. They want them labeled, out of our food system, or banned completely. And yet they never make a peep about other methods of genetic modification; none of which are tested to nearly the degree of transgenic crops before being released into the food system and none of which are labeled.
Those that would see GMO’s continue under their current heavily regulated status should be equally vocal about having those same regulations applied to traditional hybridization, polyploids, mutation breeding, and interspecific hybridization.
Let me be abundantly clear. Transgenic GMO 'breeding' carries no greater health risk than any other breeding method, including traditional breeding. The 8 years and three regulatory agencies that transgenic GMO’s have to go through for approval make them some of the most well studied and arguably safest foods we have available.
In the 20 years that transgenic GMO’s have been available for human consumption there have been no documented cases of harm to health [33,34]. The same cannot be said of traditional breeding methods, which undergo far less scrutiny.
In 1974, the USDA had to recall a potato bread through "traditional" methods called Lenape. It turns out that the Lenape had particularly high levels of an alkaloid called solanine, a natural defense mechanism of the potato. When ingested it induces “vomiting, diarrhea, loss of consciousness, and convulsive twitching” [27,28].
Celery is another example. Using conventional breeding methods, breeders tried to increase insect-resistance by creating a strain that produced higher amounts of psoralens. Grocers and harvesters that came into regular contact with the vegetable experienced cases of dermatitis and severe skin rashes [29,30,31].
I know this has been a grinder of a blog post, so I'll wrap this up with an apropos
quote from the FDA:
“Virtually all breeding techniques have potential to create unexpected, including pleiotropic, effects. For example:
- Mutations unrelated to the desired modification may be induced
- Undesirable traits may be introduced along with the desired traits
- Newly introduced DNA may physically insert into a transcriptionally active site on the chromosome, and may thereby inactivate a host gene or alter control of its expression
Plant breeders, [plant geneticists, and bioengineers] using well established-practices have successfully identified and eliminated plants [and genes] that exhibit unexpected, adverse traits prior to commercial use”  [my additions].
- The introduced gene product or a metabolic product affected by the genetic change may interact with other cellular products to produce a deleterious effect
30 - Ames BN, Gold LS. 1999. Pollution, pesticides and cancer misconceptions. Pp. 18-39 in Fearing Food, J. Morris and R. Bate, eds. Oxford, UK: Butterworth Heinemann