Entirely artificial medical nanorobots will one day exist to augment or greatly improve on functions presently carried out by cellular machinery. Long before then, however, we will see the widespread use of modified cells and bacteria, altered to form programmable tools such as drug manufactories that travel to where they are needed and take the appropriate actions in response to local circumstances.
A successful microbial diagnostic or therapeutic agent must to able to detect a particular signal with high fidelity, integrate this signal through precise intracellular circuitry, and respond to this signal at the appropriate level. Researchers have recently described genetic tools that allow the commensal bacterial species B. thetaiotaomicron to efficiently perform all three of these functions. Notably, they show that circuits integrating signal detection, genetic memory, and CRISPR interefence function as expected when engineered B. thetaiotaomicron is introduced into the gut microbiome of mice.
In the future, one can imagine the use of these mechanisms to tightly regulate the expression of different genes in a biosynthetic gene cluster for a small molecule therapeutic (e.g., an antibiotic), engineered in a microbiome-derived Bacteroides strain. The in vivo expression of this gene cluster could be controlled by the level of a carbohydrate administered in the diet, or preferably, by a specific small molecule produced by the target pathogen itself. Decoupling of the synthesis and secretion of the small molecule (e.g., to reach an effective local therapeutic dose) can be achieved by putting the export machinery under the control of an inducible circuit that responds only to high intracellular levels of the small molecule, or by engineering a time delay between the synthesis and secretion of the molecule. Once the therapeutic effect has been achieved (e.g., the elimination of a pathogen), CRISPR interference can be used to knock down residual expression of the therapeutic genes or to eliminate the chassis itself by targeting an essential gene. This final step could be triggered by a second signal administered in diet, or by the absence of the pathogen-derived small molecule. This entire series of events could be recorded on memory switches and read through analysis of the Bacteroides genome in host feces, providing timely snapshots of what is happening in vivo.
Although it is still early days for its approval, using engineered commensal microbes to produce therapeutic molecules may be preferred over using oral or systemic drugs for several reasons. First, commensals naturally occupy specific niches in the gastrointestinal tract, allowing drug delivery to a very defined site. Subsequently, the dosage needed to obtain a local therapeutic effect would be much lower than needed if orally administered, and many adverse effects could in turn be eliminated. Second, because the production of a therapeutic molecule can be precisely controlled in engineered bacteria, long-term control of diseases can be achieved using a single organism that produces the drug only when needed. Last, using an engineered bacterium to produce and deliver one or more therapeutic molecules could provide an economical alternative to the costly production, formulation, distribution, and storage of drugs. This is even more applicable in the cases where a drug is specially formulated or administered via intramuscular or subcutaneous injection to avoid degradation in the stomach.