Were you among the thousands who saw last week’s Atlantic piece on “The Very Real Danger of Genetically Modified Foods”? Food writer Ari LeVaux sought to use recent research findings on the biology of digesting plant materials to argue for an overhaul of regulations for genetically modified (GM) food in the United States. The scientific missteps in his article and nonexistent link between the study he cites and any specific danger from GM foods led a number of science writers to crack knuckles and get to rebutting.
Unfortunately, by the time these correctives had circulated, the Atlantic’s readers had shared the original piece 11,000 times on Facebook alone. A danger that was not “very real” at all had morphed into an alleged GM time bomb, one that LeVaux asserted would “blast a major hole” in arguments for maintaining the regulatory status quo. A more careful look at the science in question instead provides an argument for having the three federal agencies with GM food oversight—the Food and Drug Administration, Environmental Protection Agency, and Department of Agriculture—retain the flexible, evidence-based framework already in place.
The bomb components came from a Cell Research paper reporting identification of tiny plant molecules in human blood. We harbor plenty of human versions of these molecules, called microRNAs or miRNAs for short, which regulate protein production. No one had ever reported finding plant miRNAs in people, until Lin Zhang and colleagues at Nanjing University looked carefully at people who had consumed ordinary cooked rice and other plant foods. According to their study, molecules specific to some of these plant foods ended up in human blood and tissues. And the plant miRNAs weren’t sitting there doing nothing: In other tests, one of the types they found was shown to inhibit production of a protein that normally removes “bad” cholesterol, or LDL, from the blood.
The rice results, simply stated, show an effect of one miRNA from one non-GM plant on one protein in live mice and in cultured human liver cancer cells. The findings still require confirmation but open the door to interesting questions about the effects of natural dietary miRNAs on human health. Stirring in the complexities and emotions associated with GM foods, however, ignited a social-media chain reaction, and corrections couldn’t outpace the light-speed transmission of misinformation. LeVaux rewrote the piece with some fixes in place, and the new version now appears on AlterNet and at the Atlantic. (You can read the original here.)
Too bad the rapid-acting, social-media fix wasn’t more effective. LeVaux’ second try addressed some of the errors in how he described miRNAs and added some caveats about confirming results, but he retained much of his initial argument, maintaining that his critics had shot wide. As one of those critics, I demurred. In addition to stressing scientific inaccuracies, the critiques highlighted how off-target LeVaux was in trying to link the Zhang results to GM plant foods in particular, instead of all plant foods. No one has altered miRNAs in existing GM plants, so in that respect, at least, they’re no different from non-GM plants.
Yet his rewrite still argues this link and still shoots wide of the study’s real implications. LeVaux focuses on “substantial equivalence,” the risk assessment tenet that if testing indicates a GM food has the same essential features as a non-GM food—the same protein profile and nutritional content, for example—then it should fall under standard food regulations. The corollary, of course, is that non-equivalence requires more testing.
LeVaux says the miRNA findings compel a framework overhaul, asserting that they reveal a way that GM foods could influence human health. This work does not show that, as it involved non-GM foods and a class of molecule that hasn’t been altered in existing GM foods. Either way, dispensing with substantial equivalence would be a mistake. Biotech innovation seems to move almost as quickly as social-media misinformation, and substantial equivalence remains a flexible framework for this breakneck pace. A creaky, detailed series of federal regulations would undoubtedly lag behind the science.
Another option, the precautionary principle, while potentially flexible, functions in a context of “what if” instead of evidence, rendering it toothless and inconsistent. The Obama administration recently applied this principle in its unprecedented override of FDA approval of over-the-counter sales of Plan B, the “morning after” pill, asserting that drug makers couldn’t “conclusively establish” its safety for 11-year-old girls. This conclusion illustrates a main fault of the principle: It doesn’t balance uncertainty against known risks, which are substantial for 11-year-olds who become pregnant.
Frameworks aside, pending further evaluation, these new findings suggest that humans and other animals have lived with ingested plant miRNAs for a long time. No evidence exists of altered miRNAs in genetically modified plants, so the GM plants that are now on the market won’t affect that longstanding relationship.
In trying to link miRNAs and GM foods, LeVaux seems to confuse this type of small RNA with another kind that also regulates protein production in our cells. Some GM plants have been engineered to produce “small interfering RNA,” or siRNA. In the short-lived 1994 GM product the Flavr Savr tomato, for example, engineered siRNAs specifically blocked production of a protein that normally hastens tomato ripening.
LeVaux seeks to emphasize the dangers of one with examples of the other. Some plants have been engineered to create siRNAs or longer RNAs that silence toxin-processing genes in insect consumers. But these are perfectly matched to their insect targets, and no studies have identified any such GM plant molecules in human tissues. Further, siRNAs and miRNAs are not simply interchangeable. They can have similar effects, but cells process and use them differently.
Some groups—with no industry affiliation, thus far—do research artificial miRNA technology to alter plant characteristics, although its use in crop plants remains speculative. If artificial miRNA GM foods become possible, would the Zhang data “blast a hole” in substantial equivalence? No. Substantial equivalence would dictate further testing.
Speaking of testing, LeVaux asks, “A tomato with fish genes? … That, to me, is a new plant and it should be tested. We shouldn’t have to figure out if it’s poisonous or allergenic the old fashioned way, especially in light of how new-fangled the science is.” (Next, he’ll be yelling at scientists to get off of his lawn.) But substantial equivalence doesn’t mean “untested.” Researchers did develop a tomato that contained an inserted Arctic flounder gene to render the plant cold tolerant. But the tomato foundered in tests of its cold tolerance. Had it not failed, according to one expert review, “additional environmental and safety tests would have been conducted” to demonstrate substantial equivalence—as was done for the Flavr Savr tomato. As that review notes, all current GM products in the U.S. marketplace have undergone a full regulatory review.
Such tests aren’t full-scale clinical trials on human volunteers, but they do assess how toxic or allergenic GM foods might be in ways that are similar to nonclinical safety assessments for pharmaceuticals. In addition, thanks to technological advances in the “–omics” fields that allow comparisons of genomes, proteins, and metabolites between organisms, these tests can now incorporate comparisons of the molecular profiles of GM and non-GM versions to identify any distinctive differences that might warrant further investigation.
Largely overlooked in all of the GM alarmism are the real implications of the Zhang miRNA study. Human versions of these tiny molecules play many roles in development and disease, and now it seems that some plant miRNAs can survive our digestion and enter our bloodstream. What influence, if any, do they have on our health? That’s the question that shouldn’t get lost amidst the shrapnel of LeVaux’s social-media bombshell.