Gut Microbiome Genome Editing: Therapeutic Potential and Challenges in Delivery and Safety
Background: From adding bacteria to changing their functions
Genome editing in gut bacteria has shown the potential to alter specific gene functions, but delivering editing tools to target bacteria and ensuring safety remain challenges. A review published in Microorganisms compares editing tools and delivery strategies and summarizes the evidence and limitations for therapeutic applications. It examines existing research; it is not an experimental paper demonstrating the efficacy of a new platform.
Gut microbes contribute to metabolite production and immune regulation. Yet administering beneficial bacteria does not produce the same desired functions in everyone. Probiotics’ persistence and effects in the gut depend on the strain and the host environment, while the strength of evidence for fecal microbiota transplantation (FMT) varies by disease. The review explains that benefits established for recurrent Clostridioides difficile infection cannot simply be extrapolated to ulcerative colitis or metabolic diseases. These limitations underpin efforts to directly regulate functional genes in specific bacteria.
Key findings: Delivery and target selection matter as much as the editing tool
The tools covered in the review include CRISPR-Cas systems, base editing, prime editing, and CRISPR-associated transposases (CAST). The options vary according to the aim, such as cutting target DNA, changing bases, or inserting functional genes. However, editing performance achieved in Escherichia coli (E. coli) is not guaranteed to be reproduced in other gut bacteria. Bacteria differ in their defenses against foreign DNA and their DNA repair capacities.
Strategies that kill pathogens should be distinguished from those that inactivate only their toxin genes. One example described in the review, BACTRINS, used bacterial conjugation to deliver CAST and inactivate the Stx2B gene in Shiga toxin-producing E. coli. In a mouse gut colonization model, the reported insertion rate of the therapeutic genetic element at the target site was 80–90%. However, no statistically significant survival improvement was obtained in a model of severe disease. This result shows that a high gene insertion rate alone is insufficient to establish therapeutic efficacy.
Another example, MetaEdit, is an approach for editing target strains in complex gut microbial communities without first culturing them separately. According to the review, metabolic pathways up to 7.5 kilobases (kb) long were inserted, and diet was used to regulate the relative abundance of Bacteroides carrying an introduced inulin utilization pathway. This suggests the possibility of designing edited strains together with their nutritional environment. There is no basis for interpreting these results as showing a universal increase in the engraftment rate of beneficial bacteria in the human gut mucosa.
Implications and outlook: Connecting preclinical results to human treatment
Delivering editing tools to target bacteria is another challenge. Bacteriophages, viruses that infect bacteria, offer host specificity, but the range of strains they can infect is limited. Gut mucus and bacterial surface barriers also impede delivery. Within the scope examined by the review, in vivo results demonstrating targeted editing of gut commensal bacteria themselves using nanoparticles have not yet been established. Results from editing host inflammatory cells should not be interpreted as bacterial editing outcomes.
Evidence for disease-specific applications remains largely at the proof-of-concept and preclinical animal-study stages. Clinical translation will require assessment of editing durability, off-target changes, competition with resident microbes, and transfer of inserted genes to other bacteria. Artificial intelligence-based design is another future direction, but data bias and generalizability across populations remain constraints. The review compares editing and delivery strategies by therapeutic purpose and outlines questions that need to be addressed before use in humans.
For details, see the full review in Europe PMC and the paper’s bibliographic information and abstract.
The gut microbiome, often termed the human "second genome", profoundly influences host physiology through metabolic interactions, immune modulation, and gut-brain axis signaling. Dysbiosis is implicated in the pathogenesis of obesity, inflammatory bowel disease (IBD), malignancies, and neuropsychiatric disorders. However, traditional gut microbiota interventions, such as probiotic supplementation and fecal microbiota transplantation (FMT), still exhibit significant limitations in precision therapeutics. Probiotic intervention fails to achieve precise regulation at the strain or genetic level, and although FMT demonstrates definitive efficacy against recurrent
Developing microbiome therapeutics starts with distinguishing whether the aim is to eliminate a pathogen, switch off only its toxic function, or add a beneficial metabolic function. This review can inform the selection of editing tools and delivery vehicles for a given aim and the design of preclinical assessments. For example, a candidate that inactivates a toxin gene should be evaluated separately for editing rate, toxin reduction, and improvement in disease measures. For a strain engineered with a nutrient utilization pathway, relevant assessments include diet-dependent changes in relative abundance and persistence after dosing stops. Judging development potential requires evaluating off-target changes and gene transfer risks, without equating high target specificity with a high editing rate or efficacy in patients.