Loopholes Let Engineered Microbes Reach Millions Without Federal Review

Living genetically engineered microorganisms are no longer confined to laboratories and small field trials. They are now being sprayed on millions of acres of farmland, brushed onto teeth as a probiotic, and swallowed as a hangover-prevention drink, yet according to a policy analysis published in Frontiers in Genetics, none of these products underwent mandatory premarket federal regulatory review in the United States. The article, authored by Jennifer Kuzma, Christopher Cummings, and Khara Grieger of North Carolina State University’s Genetic Engineering and Society Center, examines how three commercially available engineered microbes slipped through jurisdictional cracks in a regulatory framework built for a different era of biotechnology.

For four decades, the United States has governed biotechnology through the Coordinated Framework for the Regulation of Biotechnology, a 1986 policy that divided oversight among the Environmental Protection Agency, the Food and Drug Administration, and the Department of Agriculture by assigning each product to an existing statute. No new biotechnology laws have been passed since. The authors argue that this product-based system, designed largely for static chemical products, leaves living, self-replicating organisms vulnerable to falling between agencies, particularly when a developer’s marketing claims or the taxonomic origin of an engineered gene determine which rules apply.

The first case study concerns PROVEN 40, a product developed by Pivot Bio and launched in 2019. It contains two bacterial species, Kosakonia sacchari and Klebsiella variicola, isolated from maize roots and optimized with CRISPR gene editing so that genes normally downregulating nitrogen fixation were disabled, allowing the bacteria to constitutively fix nitrogen and excrete ammonium for the crop throughout its lifecycle. The company estimates the product can replace more than 30 pounds per acre of the roughly 200 pounds per acre of synthetic fertilizer typically applied, and by 2023 it had been used on 5 million acres in the United States.

Because the edited strains contain no sequences from other species, the product counts as intrageneric rather than intergeneric, and the EPA’s Toxic Substances Control Act regulations cover only intergeneric microorganisms. The regulations defining intergeneric organisms have not been formally revised since 1997. PROVEN 40 is also not a pesticide or plant pest, and FIFRA rules explicitly exclude plant inoculants that enhance nutrient uptake through root systems. The result, the authors conclude, is that the product required no mandatory federal regulatory safety review before marketing. The company states that FDA considered the product under a voluntary consultation regarding detectable microbes in food and feed portions of crops, though the authors could not locate documentation of that consultation online.

The safety picture is complicated by the biology of the organism itself. Klebsiella variicola is an emerging human pathogen of growing medical concern, associated with antibiotic resistance and with respiratory, wound, and urinary tract infections in hospital settings. In the absence of required assessments, the authors reviewed biosafety studies published by Pivot Bio personnel: six studies in rats or rabbits, five of which were short-duration skin, eye, or oral toxicity tests, and a single pathogenicity study using only one exposure route, intratracheal administration in rats, with clearance from tissues reported by day 15 of a 22-day observation period. The authors argue that one pathogenicity study, in one species, via one route and one dose, may be insufficient for a product deployed across millions of acres of a food crop, and they note that no federal surveillance program exists to track long-term persistence, mutation, or health and biogeochemical impacts of these bacteria in the environment.

The second case involves a genetically engineered strain of Streptococcus mutans, the bacterium largely responsible for dental cavities because it produces lactic acid in dental biofilms. Originally developed by Oragenics under scientist Jeffrey Hillman, the strain BCS3-L1 was engineered to replace its lactate dehydrogenase gene with an alcohol dehydrogenase gene from Zymomonas mobilis, so it produces alcohol instead of lactic acid, and it secretes Mutacin-1140, an antibiotic compound that kills competing strains of S. mutans without destroying itself. Early claims that a one-time brushing could indefinitely prevent cavities placed the product under FDA drug jurisdiction, and clinical trials began in 1998. In rats, the strain reduced cavities by roughly 50 percent, and in a small voluntary human trial it colonized participants’ mouths and persisted for 15 years.

In 2004, the FDA imposed additional constraints on the trials, citing concerns about person-to-person transmission, increased alcohol production in the mouth, changes to the oral microbiome, and possible reversion to the wild type. The agency required the strain to be dependent on an amino acid not typically found in food, so it would die outside the trial, and restricted trial subjects to denture wearers so the GEM could be removed if problems arose. Recruitment failed; the company sought participants under 55 with full dentures and found only one when the FDA demanded more than 100. The effort collapsed by 2014, and the patent expired. Then in 2023, Lantern Bioworks acquired the rights and launched the product as Lumina Probiotic. In spring 2024, the company removed cavity-prevention and disease-treatment claims from its marketing after its lawyer advised repositioning the GEM as a cosmetic, a category for which the FDA has no premarket approval authority, only post-market recall and, since the Modernization of Cosmetics Regulation Act of 2022, mandatory recall, product listing, and adverse event reporting. A product once deemed to require rigorous clinical trials thus entered the market without any premarket federal review.

The third case is ZBiotics’ pre-alcohol probiotic, an engineered strain of Bacillus subtilis called ZB183 that constitutively expresses an acetaldehyde dehydrogenase enzyme from Cupriavidus necator to break down acetaldehyde, a toxic byproduct of alcohol metabolism, in the gut. More than 15 million shots of the liquid product had been sold as of July 1, 2026. By framing the product as a functional probiotic food ingredient rather than a hangover treatment, the company avoided the drug approval pathway and instead used the self-affirmed GRAS route, under which a company can determine its own ingredient is generally recognized as safe without notifying the FDA at all. The product does not appear on FDA’s GRAS notification database. The company did publish a 90-day rat toxicity study at one, ten, and one hundred times the expected commercial dose, which found no adverse effects across the measured endpoints, and convened an expert panel of unknown composition to review it. The authors commend this transparency but note that human trials that could have detected effects like the nausea, bloating, and cramps some Amazon reviewers report are not required, and no federal post-market monitoring is mandated.

Across all three cases, the authors identify a common structural problem: jurisdiction turns on legacy statutory categories, the source of engineered genetic material, and commercial claims rather than on the biological characteristics and exposure contexts that actually drive risk. A microorganism’s capacity to survive, reproduce, exchange genes horizontally, colonize hosts, and spread beyond its intended use distinguishes it from the static chemical products the statutes were written for. The authors also flag a looming challenge from synthetic biology: wholly synthetic DNA sequences and xenobiological components may not fit the intergeneric definition at all, and highly engineered intrageneric organisms could have substantially altered risk profiles while remaining formally unregulated. A recent National Academies report similarly warns that synthetic cells with non-traditional biochemistries may challenge existing detection and monitoring methods.

The article proposes a set of reforms that could largely be implemented without new legislation, alongside some that would require congressional action. These include baseline federal notification before any commercial use or intentional environmental release of a living GEM, replacing the intergeneric trigger with a tiered framework based on the organism, trait, use, exposure, persistence, and scale; a unified federal entry point to assign jurisdiction; mandatory FDA notification for GRAS determinations on living GEM food ingredients, a change the FDA itself proposed in a rule issued for public comment on August 11, 2026; a public registry of released GEMs with disclosed funding and authorship of safety studies; proportionate post-market monitoring for products released at scale or intended to persist; and clear labeling so consumers know a product contains a living engineered microorganism. The authors stress that oversight should remain proportionate to risk, preserving expedited pathways for well-characterized, low-exposure products, while ensuring that no living GEM escapes federal review simply because of how it was built or how it was marketed.

Subject of Research: Regulatory gaps in U.S. oversight of living genetically engineered microorganisms in consumer and environmental products

Article Title: Regulatory review and governance gaps for genetically engineered microbes

Article References: Kuzma, J., Cummings, C., & Grieger, K. (2026). Regulatory review and governance gaps for genetically engineered microbes. Frontiers in Genetics, 17, Article 2008399. https://doi.org/10.3389/fgene.2026.2008399

Image Credits: AI Generated

DOI: 10.3389/fgene.2026.2008399

Keywords: genetically engineered microbes, biotechnology regulation, Coordinated Framework, EPA, FDA, TSCA, GRAS, synthetic biology, nitrogen-fixing bacteria, oral probiotic, post-market monitoring, regulatory gaps