Life is a series of risk-benefit analyses. With every decision—from trying a new toothpaste to choosing a career—we decide if the benefits are worth the risks. Does the possibility of whiter teeth outweigh the risk of lower cavity protection? Does a high salary outweigh the risk of burnout from long hours?
These are personal choices, but there are some risk assessments that we have to make as a species. The changes we make to the DNA of plants, animals, and humans can be passed on ad infinitum, fundamentally altering the flora and fauna of the Earth. The use of genetic modification on food crops and in medicine also raises questions about health risks.
As the technology used for genetic modification evolves rapidly, so does the conversation on acceptable risk. New techniques, broadly known as “gene editing,” are poised to take hold of America’s food and agricultural industry. About 5 percent of U.S. canola on the market is already made using these techniques. And scientists in China, the United Kingdom, and Sweden are testing them on human embryos, something never done with the older techniques.
They are billed as the lowest risk way to manipulate DNA and gain all the benefits, like creating mushrooms that don’t turn brown or soybean oil that’s lower in trans fat—or, in the case of humans, repairing disease-causing genes.
But some scientists and consumer advocates who have long been concerned about traditional genetically modified organisms (GMOs) are equally concerned about these new kinds of altered organisms.
The older technologies involve inserting genes from foreign organisms into a plant’s DNA to give it a desired trait. For example, a gene from a bacterium was inserted into a soybean plant to make it herbicide-resistant. The process of “inserting” the genes is imprecise; one method involves attaching the desired genes to tiny metal balls and shooting them into plants’ cells.
The new technologies, on the other hand, use molecular tools that are designed to specifically target the desired part of the DNA. They don’t require the use of genes from other species, but can simply cut out an undesirable gene or make other rearrangements to the genome.
Biotech companies using these technologies hope that this will make all the difference to consumers wary of so-called “frankenfoods,” GMOs made with a patchwork of DNA from multiple species that is unlikely to occur in nature.
The advanced precision is one of the new techniques’ greatest assets, decreasing the risk of making additional, unintended changes to the genome. But studies from researchers in Germany, Switzerland, and China, among others, have shown the new techniques can still have off-target effects.
Off Target
It is hard to detect these unintended effects, according to Guillermo Montoya, a biologist at the University of Copenhagen. Sequencing the entire genome to look for problems is costly and technically difficult, he said via email. It is especially difficult to find off-target effects that happen less frequently.
Current methods for detection rely on probability, not 100 percent certainty. For example, a method may have a high chance of detecting an off-target effect that happens about 40 percent of the time, but a very small chance of finding one that happens only 10 percent of the time.
Montoya wrote in a 2016 study published in the peer-reviewed journal Bioessays, “Talen and Crispr-Cas9 [two of the new techniques] are vastly used in genome editing; however, none of them has perfect DNA recognition specificity, so possible breaks can occur on other DNA sites in the genome.
“This off-target effect can introduce undesired changes in sequences of the genome with unpredictable consequences for cells, organs, organisms, and even environments.”
Shengdar Q. Tsai, a genetic engineering expert at St. Jude’s Children’s Research Hospital in Memphis, Tennessee, has also noted the problem of low-frequency off-target effects. He wrote in a 2014 article in the peer-reviewed journal Cell Stem Cell: “Clearly, an unbiased, genome-wide method that is also sensitive enough to identify even lower frequency off-target effects is required. ... This is critically important because unintended, off-target modifications in cell populations can lead to unexpected functional consequences in both research and therapeutic contexts, where functional consequences of even low-frequency mutations can be of significant concern.”
One of Talen’s creators, Dan Voytas, said that he has not found any unintended changes in the food crops he has worked on as chief scientist for biotech company Calyxt. The company has developed several food crops it is hoping to start selling to farmers in the next few years, including reduced-gluten wheat and a canola low in saturated fat.
After designing molecular tools to target and snip out particular genes, Voytas’s team looked at selected parts of the genome for off-target effects, in places that the molecular tools could easily have mistaken for the target areas. His team has not found any off-target effects in these places, but they have not checked the DNA in its entirety.
Researchers at Osnabrück University in Germany also reported that off-target effects are rare with Talen. It is far less prone to off-target effects than Crispr-Cas9. But they did note in their article, published in March n the peer-reviewed journal Plant Methods, that the use of Talen to create a rockcress (Arabidopsis) plant resulted in the deletion of three genes other than the ones intended. This “seemed to have occurred spontaneously,” they wrote.
Some food products are being made with Crispr-Cas9, such as a sweet corn by DuPont that is expected to be available to U.S. growers in the next five years. Agrochemical giant Monsanto announced in January that it will be using Crispr-Cas9 and its sister technology, Crispr-Cpf1, to create new crops. Cibus, a biotech company based in California, uses another new technique called the Rapid Trait Development System. Cibus was the first company to launch a product using one of these new techniques commercially, beginning to sell its SU Canola seeds to farmers in 2014.
‘It’s Hard to Say There’s Zero Risk to Anything’
While there are risks to these new technologies, there are also risks to technologies that have long been used in agriculture, said Richard Amasino, a professor of biochemistry and genetics at the University of Wisconsin–Madison who served on a committee assembled by the National Academy of Sciences (NAS) to assess the future of genetically engineered crops.
“If you ask the question, ‘Could it possibly create something harmful?’, well, yes, any process that results in a change of DNA, including conventional plant breeding, could, in principle, create something harmful,” he said.
He explained that even conventional breeding for desired traits can create unintended effects, like increased allergenicity or toxicity. “It’s hard to say there’s zero risk to anything,” he said.
But Amasino thinks the degree of precision makes these new techniques safe in a broad sense and preferable to previous methods. He thinks the risk is low and the potential benefits are high.
Voytas similarly commented on the benefits: “Almost all of the products that we’re making have a direct consumer benefit—healthier soybean oil …, a wheat product lower in gluten and higher in fiber. We’re hopeful that the consumer will see that biotechnology can be used to address consumer needs and perhaps that will influence acceptance. Whereas in the past, agricultural biotechnology has mostly benefited the farmer and the production system—[creating traits like] herbicide tolerance and pathogen resistance.”
Megan Hochstrasser earned her doctorate researching Crispr in the lab of its creator, Jennifer Doudna. Hochstrasser explained the difference between mutagenesis, an older, commonly used GM technique, and Crispr-Cas9. She said it’s comparable to the difference between Boggle and Scrabble.
“[Mutagenesis means] taking the existing DNA and shaking things up, almost like a game of Boggle, where you end up getting letters and maybe making a nice word, maybe not. Maybe you have changes somewhere else that you don’t know about.”
She continued: “Crispr, I would say, is closer to Scrabble ... where you can choose the precise sequence of letters you want. So even if there are occasionally some off-target effects, it’s still monumentally different from the previous approaches.”
Since all previous breeding methods, from selective breeding to genetic engineering, have been about changing DNA and have had unintended effects, Hochstrasser feels Crispr is preferable for use in agriculture because it is more precise.
Its use in humans concerns her more. She is worried people might even use it for enhancing or creating aesthetically pleasing traits rather than preventing disease.
Another concern raised by many is the increased risk of off-target effects if these techniques are combined—if scientists try to create more than one change in the genome. For example, the 2017 NAS report titled “Preparing for Future Products of Biotechnology” reads, “The magnitude of risk might change as the synergistic effects of multiple genetic changes could lead to unintended effects in the biochemistry of crops (affecting nutrients, immunogens, phytohormones, or toxicants).”
That report also said that since risk assessments of biotechnology products use qualitative language and don’t give probabilities of risk, NAS was unable to quantify the risks. It suggests that assessments should begin to show these probabilities, such as, for example, how much more likely these random effects are to occur with Talen and similar technologies than with random mutations in nature.
The novelty of these techniques has also raised concerns. A joint statement by Greenpeace and other advocacy groups issued in February said, “Given that many of the techniques are new, it is not yet possible to fully evaluate the potential for adverse effects.”
Megan Westgate, executive director of the Non-GMO Project, which provides verification and labeling for non-GMO products, said via email: “GMOs, including the products of these new technologies, have not been adequately tested—no long-term feeding studies have been conducted.”
Aside from the risks related directly to off-target effects and human health, the USDA’s organic advisory board has discussed secondary effects of concern. In a recommendation it published in November last year, it listed some problems with GMOs in general, noting that these concerns also apply to the new breed of GM crops: the altered nutritional profiles of GM crops, the displacement of small-scale farmers, and the decline of diversity and soil fertility from the use of herbicides.
