Engineered living therapeutics

In plain language

Many medicines are molecules that spread through the blood to reach the parts of the body that need them. An engineered living therapeutic works differently: a living microbe, usually a bacterium, is the delivery system. Researchers give the microbe genetic instructions so that it makes a useful protein, senses a signal such as low oxygen or acidity, and releases its product only under chosen conditions.

The hope is that a living cell can make a short-lived medicine close to where it is needed, instead of relying on repeated doses that reach the whole body. The idea is still early. Bacteria can travel to unintended tissues, provoke immune responses and lose their engineered controls, so each candidate needs its own evidence of safety and effect. So far the human evidence is small: one wound-healing study enrolled 36 healthy volunteers, and a small study in adults with phenylketonuria tested an engineered bacterium that lowers a blood amino acid.

Technical explanation

A living therapeutic is built on a chassis strain, a well-characterised microbe such as Limosilactobacillus reuteri or Escherichia coli Nissle 1917. Synthetic biology adds a genetic circuit, which processes inputs such as oxygen, pH or lactate and drives expression of a payload such as a cytokine or an enzyme. The circuit is usually carried on plasmids or inserted into the chromosome. A lysis or kill-switch function can be added to limit how long the organism persists, but such safety features remain largely at the research stage. Each layer matters: the strain shapes colonisation, the circuit sets when the payload is made, and the route of administration determines where the organism travels.

In a phase 1 trial (Öhnstedt and colleagues, 2023), the topical candidate ILP100-Topical, a lactic acid bacterium that expresses CXCL12, was applied at 50 microlitres per wound to 240 induced wounds in 36 healthy volunteers. The trial was adaptive, randomised, double-blind and placebo-controlled, with single and multiple ascending dose parts. In multidose cohorts, 76% (73/96) of treated wounds had healed by day 32, against 59% (57/96) of saline or placebo wounds (p = 0.020). The authors report that time to first registered healing was about six days shorter on average and about ten days shorter at the highest dose.

Mouse data show both the potential and the containment problem. In a 2024 colorectal cancer study (Zhou and colleagues), engineered E. coli Nissle 1917 strains increased signal output 1.8 to 2.3 times in laboratory tests according to the abstract; the results section gives about 1.9-fold and 2.8-fold for two promoters. In a subcutaneous tumour model using intratumoral injection, the results section reports tumour growth inhibition of 42.51% to 54.77% for different inducible strains, while the abstract gives 47% to 52%. Engineered bacteria were also found in the liver and spleen, and plasmid retention in the mouse gut was about 40% and about 30% for the two plasmid configurations tested. In a separate phase 2 open-label study in adults with phenylketonuria (Vockley and colleagues, 2023), the engineered E. coli Nissle 1917 strain SYNB1934 reduced fasting plasma phenylalanine by 40% (95% confidence interval of minus 52 to minus 24) over 14 days. Twenty people enrolled and 15 completed treatment.

What the evidence says

Each statement is typed by the strength of its evidence and linked to the source that supports it.

Limitations and open problems

• The wound study enrolled healthy, non-obese adults up to age 45 with induced wounds. Results may not transfer to people with chronic wounds or other risk factors, and the study was small, with 36 volunteers.
• Mouse tumour results do not establish clinical benefit or safety in people. The two mouse models used different routes: intratumoral injection in one and oral administration in the other.
• Engineered bacteria spread beyond the tumour in mice. They were found in the liver and spleen after intratumoral injection, so containment and clearance need direct testing.
• Genetic circuits can lose function. Plasmid retention in the mouse gut was about 30% to 40%, so a proportion of bacteria may lose the engineered construct.
• Sensing a signal does not guarantee disease specificity, so clinical use would need validated thresholds for the chosen signal.
• The phenylketonuria study was small, open-label and lasted 14 days, so it says little about long-term effect. Live products also need consistent strain identity, genetic stability and viable cell counts, and regulators must assess containment risks.

Possible future uses

Timeline

1. 25 May 2023A phase 1 trial of ILP100-Topical is published in eClinicalMedicine, reporting 240 induced wounds in 36 healthy volunteers. The authors report a safe and well-tolerated profile with no systemic exposure. Sources: [1]
2. 28 Sept 2023A phase 2 open-label study of the engineered E. coli Nissle strain SYNB1934 in adults with phenylketonuria is published in Nature Metabolism, reporting a 40% fall in fasting plasma phenylalanine. Sources: [2]
3. 20 Jan 2024A study of an engineered E. coli Nissle consortium for colorectal cancer is published in npj Biofilms and Microbiomes, including mouse tumour and distribution data. Sources: [3]
4. 23 Oct 2024A review of preclinical and clinical bacteria-based live biotherapeutic products is published in Molecular Systems Biology. Sources: [4]

Futures this could help bring about

How far away?

Bacteria engineered as living medicines

Engineering prototypes · 5 of 7 steps done

Glossary

Live biotherapeutic product

A product containing living microorganisms intended to prevent or treat disease.

Genetic circuit

DNA instructions that process cellular inputs and regulate outputs such as protein production.

Payload

The therapeutic molecule or function produced by the engineered cell.

Lysis

Rupture of a cell, which can release its contents.

Sources

1. peer reviewed paper
   Engineered bacteria to accelerate wound healing: an adaptive, randomised, double-blind, placebo-controlled, first-in-human phase 1 trial
   E Öhnstedt, E Vågesjö, A Fasth et al.. EClinicalMedicine, 25 May 2023. DOI: 10.1016/j.eclinm.2023.102014. Accessed 21 Oct 2026.
2. peer reviewed paper
   Efficacy and safety of a synthetic biotic for treatment of phenylketonuria: a phase 2 clinical trial
   J. Vockley, N. Sondheimer, M. Puurunen et al.. Nature Metabolism, 28 Sept 2023. DOI: 10.1038/s42255-023-00897-6. Accessed 21 Oct 2026.
3. peer reviewed paper
   Enhancing tumor-specific recognition of programmable synthetic bacterial consortium for precision therapy of colorectal cancer
   Tuoyu Zhou, Jingyuan Wu, Haibo Tang et al.. npj Biofilms and Microbiomes, 20 Jan 2024. DOI: 10.1038/s41522-024-00479-8. Accessed 21 Oct 2026.
4. review article
   Bacterial live therapeutics for human diseases
   E. Frutos-Grilo, Y. Ana, J. Gonzalez-de Miguel et al.. Molecular Systems Biology, 23 Oct 2024. DOI: 10.1038/s44320-024-00067-0. Accessed 21 Oct 2026.
5. review article
   Living Engineered Bacterial Therapeutics: Emerging Affordable Precision Interventions
   Rajkamal Srivastava, Cammie F Lesser. Microbial Biotechnology, Nov 2024. DOI: 10.1111/1751-7915.70057. Accessed 21 Oct 2026.

Revision history

3 changes since this profile was created.