Genetically modified plants as fish feed ingredients

There has been a long and tortuous discussion of Steve Savage’s post  Way Too Much Angst about GMO crops covering many of the usual topics. But among the hundreds of comments there is at last some gold. Fishy smelling gold.
A new paper has been kindly unearthed in the comments by (a commenter): Genetically modified plants as fish feed ingredients by Nini Hedberg Sissener, Monica Sanden, Åshild Krogdahl, Anne-Marie Bakke, Lene Elisabeth Johannessen, and Gro-Ingunn Hemre (2011).
Abstract:

Genetically modified (GM) plants were first grown commercially more than 20 years ago, but their use is still controversial in some parts of the world. Many GM plant varieties are produced in large quantities globally and are approved for use in fish feeds both in Norway and the European Union. European consumers, however, are skeptical to fish produced by means of GM feed ingredients. Concerns have been raised regarding the safety of GM plants, including potential toxicity and (or) allergenicity of the novel protein, potential unintended effects, and risk of horizontal gene transfer to other species. This review will present the current state of knowledge regarding GM plants as fish feed ingredients, focusing on fish performance and health as well as the fate of the GM DNA fragments in the fish, identifying limitations of the current work and areas where further research is needed.

Allan’s comment:

First, I thank (a commenter) for turning us on to a new paper. I’ve downloaded the paper and additionally downloaded many of the references in the paper so that I can check out what the referenced authors are actually saying.
Second, I’m detecting a creeping confusion between hazard, risk, and just plain old homeostatic physiological phenomena that are neither adverse effects, hazards, or risks. Specifically, (a commenter) cites Bodner as saying, ““There is no indication that Bt protein (in transgenic plants or in sprays) is harmful to mammals, birds, or fish…”. He indirectly refutes this generality by referring to the review article by Sissener et al. (2011) and then to Sissener’s summary of some of the papers that she reviewed. The quotes in Sissener’s article does not refute Bodner’s statement. The evidence for this can be drawn directly by examining the referenced author’s own words. For example, the longest section in Sissener et al. (2011) that was quoted by (a commenter) refers to a paper by Sagstad et al. (2007). wherein immune system parameters among others was found to differ in a comparison of reference diet-fed fish, GM corn diet-fed fish, and non GM corn-diet fed fish. However, the question is health effects, which is not addressed by examining immune system biomarkers (I’ll come back to what Sagstad et al. (2007) really showed). A better paper, on which Sagstad is an author, is the one by Hemre et al. 2007 in Aquatic Nutrition, in which they conclude the following (from the abstract) : “Based on the present findings, the conclusions made are: Atlantic salmon smolts fed GM maize (event MON810 ), its near-isogenic parental line and suprex maize (Reference diet), all resulted in high growth rates, ADC and feed utilization. Health, when evaluated by means of mortality (low), normal ranges of blood and plasma parameters, except somewhat elevated ASAT values and minor variations in organ sizes, were considered good in all diet groups. The changes in the glucose transport mechanism and intestinal maltase enzyme activity in the gastrointestinal tract warrant further studies.” If the last statement is worrisome, then consider what the conclusion is based on. Whereas the vast majority of comparisons for the measured parameters were made between three treatment groups: a reference diet control, non GM corn, and GM corn, the latter statement is based on an in vitro study of isolated brush border membrane cells but DID NOT INCLUDE the reference diet fish. This is important, because sometimes there might be a difference between non-GM and GM fed group parameters, but the parameters do not differ significantly between the GM and reference diet group.
Now back to the Sagstad et al. 2007 article and those immune system biomarker results for my third point. Here is where you have to actually look at Table 4 to find out the following. First, the data are expressed as the percentage proportion of several different innate immune system cells in the sampled population of white blood cell populations from the reference diet, non-GM corn, and GM corn diet groups. While it is true that some significant differences in proportions of cells amounting to several percent were found when non-GM corn and GM corn were compared, there were no statistical differences between the reference diet and the GM diet fish. So, what does it mean? I argue it does not mean anything other than observation about homeostatic mechanisms owing to differences in environment. What do I mean by differences in environment? Well, you take 45 fish and isolate them from another group of 45 fish and you’ve automatically changed their environment. Don’t think that fish have personalities and quickly establish dominance hierarchies? I suggest you haven’t observed animals enough. (BTW, I can find no indication of fish sexing prior to group establishment. Lack of sexing could be a confounder, maybe.) The bottom line is that WE EXPECT minor differences in biomarkers between groups, whatever the exposure treatment, because by isolating animals we have changed their environment. The fact that SMALL differences in immune cell proportions were reported by Sagstad, with no statistical difference from a reference diet (that contained an equivalent amount of corn starch as the other diets) tells me there is no consistent physiological effect related directly to treatment. Certainly, these data mean NOTHING regarding health. The important conclusion about health is in the Hemre et al. (2007) paper referenced above.
As an aside, lest you think that Monsanto is hiding something and the regulators don’t know about these studies, consider that Monsanto is credited with providing these researchers with MON 810 corn and its isogenic non GM line. Thus, you can bet that Monsanto had the data and furthermore, as required under FIFRA, had to report those results to the EPA. After all, MON 810 is a registered pesticide (the technical term is a PIP, plant incorporated pesticide). Monsanto was not involved in the GM soya feeding studies because those papers did not use an isogenic line. Differences in cultivars will lead to differences in biomarkers, but even then, the researchers did not see any effects on health, broadly defined for their purposes as effects on growth parameters and mortality (lest you get picky, I’m skipping a lot of details here).
Fourth, the weight of the evidence based on examining more fish GM vs. non GM diet papers shows that the small cadre of authors involved in these studies have more often than not concluded no effects of putting large proportions of plant protein in fish food. And this brings me to the conceptualization of risk in this case. The objective of the research reported in these papers is basically two-fold. First, can you substitute a lot of plant protein for fish meal when rearing fish in aquacultural operations? Second, if the fish health is not affected (remember for this case health = growth parameters, mortality, and proximate nutrient content), will having 15-30% of the plant protein content come from GM soy or corn, affect health? Regulatory toxicology relies on the weight of the evidence, and of the perhaps 20 papers I quickly downloaded following (a commenter’s) remarks, I see no concern expressed by these authors about health. Where specific biomarker parameters were different, the authors call for more study, but what bona fide scientist doesn’t call for more study?
Finally, this forum is a lot more fun when people actually discuss the data and don’t react as if personally threatened by every new bit of information.

Note by Anastasia: (a commenter) was used to replace that commenter’s name. Please visit the comment policy forum if you have any questions about comment use. Thanks!