Failure to demonstrate any harm from Cry proteins on mice: comments on Mezzomo et al., 2013

The B. thuringiensis spore crystal (source: reference at the end of the post)

Bélin Mezzomo and collaborators from the Department of Genetics and Morphology, Institute of Biological Sciences, University of Brasilia, Brasilia/DF, Brazil, have recently published a paper entitled “Hematotoxicity of Bacillus thuringiensis as Spore-crystal Strains Cry1Aa, Cry1Ab, Cry1Ac or Cry2Aa in Swiss Albino Mice”.  The paper lacks important controls and uses very large spore concentrations in a force-feeding experiment using mice. Some comments on the paper can be found below, as well as a link to the paper:http://www.sciencedirect.com/science/article/pii/S1049964409000863

Initial comment

The paper is biased against modern agriculture, in general, as can be easily grasped from the first sentence of the introduction. And, as in an omen, in this very first paragraph the authors come to wrong conclusions that can´t be taken from what they have just commented. This is the global strategy of the paper: the results from the paper may be correct, but the authors’ conclusions are wrong. Indeed, due to a poor experimental design, most of their conclusions can´t be sustained by their results, alone or in combination to other results from the papers cited in the references.

Some comments on the methodology

Bt spores are widely used in agriculture. But they are not genetically modified to produce toxins. The recombinant proteins are produced by GM plants. Therefore, there is no reason to assay GM Bt spores as alias to GM plants.  

The language used in the paper is not clear. The authors claimed to have used spore-crystals. This is not the same as having used purified Cry proteins: to the best of our knowledge, the spore-crystals mentioned here are indeed lyophilized spores for Cry-producing Bt strains, as described in the first paragraph of the Methodology section.

Moreover, the spores have a much higher concentration of toxins that any commercial GM plants and the authors employed unusually high doses in their assays; indeed, doses just above the threshold of 270 mg/kg, used in the assays, showed signs of toxicity and lead to the death of mice, as admitted by the authors themselves; additionally, the BT proteins in these assays are found mixed with all other proteins and molecules, specially peptidoglycans,  forming the spore. Therefore, the assays neither use the right range of toxin concentrations nor rigorously test the toxin, but a complex mixture of different molecules. Finally, at least as food, transgenic grains are cooked. As a conclusion, the results do not allow to extend the results to food safety for human beings.

Surprisingly, the authors do not have a fundamental control: a spore not expressing a toxin or expressing an unrelated toxin. The only controls were water (negative) and cyclophosphamide (positive). The lack of this third control prevents any conclusion on the true effect of the toxins on the animals, since all other spore molecular components could also affect the mice in such unusual high concentration as those directly applied to the digestive system in these ill-designed assays. The authors should be aware that peptidoglycans, a major component of bacterial spores, can be toxic form mammals, as can be seen in any toxicology textbook.

Some comments on the methodology

The hematological effects of the spores are displayed as if they were caused by the proteins but, as commented above, in the absence of a control on all other spore components, it is not possible to attribute a causative relation specifically to Cry proteins.

The table format used to present the hematological effects is confusing. Graphs should be used, instead. A dose-effect is not clear for most spores, either used alone or in combination. The experiments done with the very high dose of 270 mg/kg are misleading, as they are very close to a possible LD50, not informed in the paper for any of the spores used by the authors.

The same remarks are valid for all other tabular results displayed in tables 2 to 4. Table 5 has the micronucleus results for just the two highest doses, without standard deviation values. Results for the controls are also missing. As commented, these doses are far too high to be of any use in a toxicological study.

Comments of the Discussion

The results are discussed as if they really showed that the proteins themselves, and not the spores, were the causative agents of the observed hematological or micronucleus change, what is not correct.  It is not reasonable to use spores instead of pure proteins, since those are readily available in the market.

The lack of clear dose-effect curves, hidden in the inconvenient presentation of all results in tabular form, preclude a conclusion on the toxicity for the smaller doses. The unusual high doses employed here preclude any extension of the results obtained here for the real world, where the doses are orders of magnitude lower.  Moreover, food is prepared from cooked GM grains, which denature proteins.

See also http://cdn.intechopen.com/pdfs/44118/InTech-Microorganisms_in_biological_pest_control_a_review_bacterial_toxin_application_and_effect_of_environmental_factors_.pdf

As a final remark, neither the authors nor anyone else can derive safety conclusions from these results and the maximum the authors advance is the need for new  and hopefully better conducted experiments.

(Paulo Andrade, Dept. Genetics, fed. Univ. Pernambuco, Brasil)