I was asked for my thoughts on a new paper by Seralini: Cytotoxicity on human cells of Cry1Ab and Cry1Ac Bt insecticidal toxins alone or with a glyphosate-based herbicide. I don't have the patience right now to turn my results into a blog post, so here are my thoughts as I posted them on Facebook:
Well, first thing that pops up is the fact that Dr. Séralini is funded by an anti-biotech organization (Greenpeace). So, there's that. He's also had some sketchy articles in the past that had, to be blunt, huge statistical mistakes and poor conclusions based on those mistakes (we've talked about some of this on Biofortified). This makes me question his ability to do sound science more than his potential bias due to funding. While any paper must be judged on its own merits, the frame of the authors (the PI in this case) does matter.
I was able to find a free PDF of the paper here: http://www.salmone.org/wp-content/uploads/2012/02/seralini-2012.pdf
Hm, the citations are a little suspicious. Most are of Seralini himself, and I notice a few others that have problems (like Rosi-Marshall).
Hm. They cite a Paul 2010, saying digestion is never a complete process and insecticide toxins cannot be fully degraded in vivo" but the abstract of the paper says "The results of the present study indicate that the recombinant Cry1Ab protein from MON810 is increasingly degraded into a small fragment during dairy cow digestion." These two are not the same thing. Hm. http://www.ncbi.nlm.nih.gov/pubmed/19888668
Other papers have shown digestion as well.
So, I've read the whole thing. Here's my conclusion:
I agree with their general idea that it's not scientific to call all Bt the same (this is not something I've ever disputed). The truncated and/or modified proteins do have differing toxicity levels with respect to target insects so it stands to reason that they may have different toxicity levels for non-target organisms or other different properties.
It may be appropriate to test pure protein as done in this study, but I believe it is more valuable to test a transgenic protein in the transgenic plant material or at least to extract the protein from the plant rather than producing in bacteria. Producing the protein in bacteria then trypsinizing may not give the same product as is found in the plant. They didn't sequence the protein to see if is the same as what is produced in a plant with a given construct. So, they may have tested something that didn't exist. We don't know if the protein they have is similar enough to the protein produced in plants to say that this study can be applied to biotech plants.
It may also be appropriate to test the Bt protein in combination with various pesticides that are expected to be used on a given Bt crop. Although, I do question the value in testing a pesticides at LC50 rather than at a realistic exposure level, such as the estimated residue at x time after spraying.
Feeding studies have been and are done for Bt crops. These studies compare plants expressing Bt to isogenic (genetically similar) plants without Bt. The companies do it before releasing the products to the market as part of deregulation risk assessment and independent scientists do additional feeding studies. Many of these studies are done in a "real world" environment or as much as possible for a controlled study. One of my favorites is this one with Mon810 that has the Cry1Ab protein: http://www.ncbi.nlm.nih.gov/pubmed/20579187
Are these whole animal studies more or less valuable than a petri dish study? They are different, and tell us different things. The petri dish studies can only give us a hint of what will happen in a whole organism, and this study complicates the in vitro aspect even more by not using the actual plant-produced protein.
There have not been many in vitro studies of Bt toxicity. One that I found that was not cited by Seralini (that should have been included, in my opinion) used sheep rumen epithelial cells (known to be a good indicator for what will happen in vivo http://www.ncbi.nlm.nih.gov/pubmed/9478009
) and found that even high concentrations of Cry1Ab didn't have an effect on the cells http://www.ncbi.nlm.nih.gov/pubmed/18325653
. I do feel that this study has a flawed literature review, which concerns me greatly. If the authors don't have a firm grasp of the literature, and instead cite papers that are known to be flawed, what does that say about their ability to do good science?
I'm also worried about the sentence "the experiments were repeated at least three times" in the methods. Did they keep doing it until they got the results they wanted? Regardless, for the Bt only part, the experiment showed that the only significant difference was for the highest concentration of Bt, a concentration that wouldn't be found in the real world. The 10 ppm concentration had no difference from the control.
For the Bt+Roundup part, I won't even make conclusions because as I said earlier, that concentration isn't realistic.
Hm. Maybe I should turn all of this research into a blog post. I wish I had more time :( I spent way too long looking at this, although it was interesting.
Probably more than you wanted, but here you go! :)