Introduction

Enabling highly specific drug delivery would be very beneficial in medicine. It would allow doctors to deliver more efficacious therapy using smaller doses and without exposing the patient to the systemic effects of that drug, thus reducing side effects. We are interested in the potential of using bacteria to increase specificity of delivery. Several models for this have already been proposed including utilising bacteria to convert prodrugs into their active compounds (Rubinstein 1990), allowing bacteria to colonise tumours (Gardilk et al 2010) and using 'bacterial ghosts' (Kudela et al 2010). When addressing how GEMs might be used in delivery of drugs, considering how one might control the dosage of drugs being delivered and controlling the colonisation of bacterial vectors within the body is important. Our kill switch aims to give future teams of drug designers a tool with which the can selectively cause bacteria to kill themselves under certain conditions, allowing precise control of bacterial vector colonies in the body.

From the beginning of our kill switch design, we foresaw the application of the project being used within a drug delivery context. As intracellular antimicrobial peptide (AMP) production builds, protegrin-1 would slowly integrate into the bacterial phospholipids bilayer, causing pores to form between the membrane and the extracellular matrix (Lam, 2006). We believe that the pressure difference between the inside and the outside of the cell would cause the intracellular contents of the cell to be exported upon pore formation. This means that along with the cytosol, the potential drug will also be released into the extracellular matrix.

Bacteria as a drug vector?

Why use a non-pathogenic bacterium as a drug vector? Isn’t using a pill a much safer and cost-effective way to absorb chemicals into the body? Capsules have been used since their creation in 1834, as they allow the medicine to be contained within functionally-important media, such as oil, and in pre-measured amounts and combinations to assist with accurate dosing. A capsule simply contains medicine within a stable shell coating, which will dissolve within the body and allow the uptake of the contained chemical into the system. Pills are the current norm within drug delivery, but they are far from the most effective way to deliver medicine to the body.

The advantage that bacterial drug vectors have over pills is the natural specificity that bacteria have towards it preferred ecological niche. Escherichia coli, for instance, move through the body after being ingested and colonize the lower intestines. They have evolved over millions of years, and trillions of replications, to survive within the environment of the gut, evading host defenses, gathering nutrients, and reproducing within the natural parameters of the lower intestine. If there was a drug that needed to be directly delivered to the gut, E. coli may offer a hyper-specific delivery vector that could outshine pill counterparts.

One of the obvious benefits of site-specific drug delivery would be the pronounced reduction in unwanted systemic effects of useful, functional drugs. For instance, catabolic steroids that would benefit a single organ must currently be taken via injection or oral means, resulting in the possibility of a whole range of detrimental side effects including hypertension, weight gain, decreased muscle mass, cataracts, and a decrease in infection sensitivity and response associated with long term use (Noone, 2006). Rather than the site of absorption be into the bloodstream, which spreads the steroids to internal areas where they can cause these side effects, site-specific steroid release would greatly reduce excess steroid concentrations in the blood. Not only that, our cells would need to release a much lower concentration of drug than required in pill form. We would no longer have to flood the bloodstream with a high concentration of drug so the intended site in effected, and this would reduce the severity of, or perhaps even remove altogether, some of the drugs side effects (Noone, 2006). Another example would be the ingestion of antibiotics used to fight other infections within the body. High dose broad-spectrum antibiotics would normally damage gut flora before being absorbed into the bloodstream (Martin, 2003). Our E. coli vectors would be able to release this drug not only in small doses, but to smaller areas within the gut, resulting in far lower collateral damage to symbiotic bacteria. The bacterial ability to move to the site of drug delivery may result in a more effective drug absorbance by the area in need, and less damage to the rest of the body, resulting in a more effective treatment.

E. coli may seem like somewhat of an unimpressive example, as pills release their drug into the same area as E. coli’s colonization specificity: the gut. But other bacteria with different niche specificity could open up entirely new areas of drug absorption. Haemophilus influenzae, whose genome was sequenced in 1995, has adapted niche specificity for the upper respiratory tract (Levine, 1996). Another bacteria of interest, Staphylococcus aureus, is known to colonize mucous membranes (VDH, 2011). Both of these bacteria offer different biospecificities than may be utilized by individual drugs treating different areas of the body, but any non-pathogenic bacteria with a strong affinity for an internal area could be utilized. Site-specific drug delivery in previously inaccessible locations within the body may even lead to changes in molecular drug design. This new accessibility may lead to drugs being further detailed to match their exact site of release, rather than simple absorption through the stomach, with serious potential for increased drug effectiveness.

It is also important to mention that the AMPs providing the functionality of our kill switch could themselves be the “drug” that must be delivered. If the drug needed were for attacking other invasive prokaryotic bacteria in the body, redesigning the promoter and ribosome-binding site of our kill switch would lead to a higher PoPS output, resulting in faster AMP production. In theory, if enough AMPs could be produced before complete membrane disruption, intracellular non-membrane bound AMPs (and, perhaps, membrane-bound AMPs as well) would also be exported from the cell upon pore formation. Using peptides instead of antibiotics to fight other bacterial infections reduces the odds of creating antibiotic resistance. AMPs attack the membrane, the structure of which is highly conserved within evolution; the number of mutations required in a single replication to completely re-structure the bacterial membrane makes the odds of developing AMP resistance near impossible. Not only that, but the AMPs would not attack host eukaryotic cells due to the structure of their membrane (as described in our Kill Switch page), instead binding to only prokaryotic cell membranes, nullifying any potential negative interaction between the drug and the host (Hancock and Sahl, 2006).

Dangers of Bacterial Drug Vectors

There are, however, dangerous factors involved in using live bacteria as a drug delivery system. The act of releasing the drug from the burst cell also involves releasing the rest of intracellular contents, including chemicals which may indirectly act as stimuli, and could illicit potentially detrimental responses from the body. One concern we’ve raised before is the release of endotoxins imbedded in the phospholipid membrane. While the word “endotoxin” in a general sense can describe any toxin generated within the bacteria, gram-negative endotoxins are more particular. These are found specifically within the outer layer of the phospholipid bilayer and are responsible for the disease-causing aspect of gram-negative toxicity. Release of these endotoxins induce an inflammatory response by the body, causing macrophages and endothelial cells to secrete pro-inflammatory cytokines and nitric oxide (Stewart, 2006). Septic shock is dangerous, and the problem of endotoxin production would have to be addressed before any in vivo testing could occur. This is not, however, beyond the reach of future scientific advances, as the endotoxin molecules are produced by DNA just as malleable as our kill switch. Synthetic biology would need to advance to a point where non-endotoxic bacteria could be engineered before bacterial vectors would be safe for in vivo testing.

There are also complications arising from interactions between both the immune system and the bacteria. Because we are inoculating the body with invasive, colonizing bacteria, how the body and its immune system might react remains an unknown factor. Members of the host’s immune system recognize proteins on the surface of bacteria. This recognition is what allows immune response cells like macrophages to engulf and digest non-host microbes. In order to combat this issue, we would need to synthesize bacteria with either cell-surface proteins that are not recognized by host immune defenses, or with no cell-surface proteins at all. The concern at this point is not whether this is achievable with synthetic biology (as our knowledge of synthetic biology deepens over time, one day this technology will be available), but the risk of passing this genetic sequence to other bacteria via horizontal gene transfer. Conjugation, when one bacteria passes DNA to another, is the only form of horizontal gene transfer involving cell-to-cell contact (Holmes and Jobling, 1996). The physical lack of cell-surface proteins, and subsequently, the state of being completely undetectable by a host’s immune system, is not a safe set of bacterial parameters to be left free-floating within the body. If an invading bacterium conjugates with our GEM and receives this set of genetic changes, its inability to be controlled by the body’s immune system could have devastating consequences for the host. Unless the act of bacterial conjugation can be prevented in vivo, avoiding immune system interference may seem like a tricky task for the future.

Another issue is drug dosage. A convenience of pills is that the amount of medicine in each capsule is constant, reducing issues related to under or overdosing. The fear is that bacteria may overproduce or underproduce their drug, resulting in faulty treatment, or they may over-replicate, resulting in drug overproduction. As the drug will be produced within the cell, the way to preemptively solve these problems would be to characterize the drug-producing gene to know the rate of drug production versus the rate of AMP production, as well as the rate of bacterial replication. If we know the total time the bacteria will spend in the body before the AMPs disrupt their membranes, we can work backwards to find the maximum number of replications the bacteria can undergo in that time, as well as the rate of drug production, to find the total peptide output (and therefore, total drug production) before cell death occurs. In vitro trial and error experiments must be run in order to validate this characterization, but the problem of dosage is not beyond the grasp of modern science.

Our Conclusions

Using live bacteria as a drug delivery vector is a wild, exciting, and as a modern day scientific practice, a rather dangerous idea. It will take many accumulating scientific advancements to reach a level of knowledge that would make this project feasible. However, we believe utilizing the natural biospecificity of each bacterial species to create site-specific drug vectors is the future of drug delivery. Not only that, but being able to utilize bacterial niche specificity will allow for the drugs themselves to be better designed, with an emphasis on site of absorption that may have vast implications for drug efficacy and treatment success. The kill switch we have created is only a small mechanism in the grand design of this delivery system, but it is vitally important to the functionality of that system. With it, we can begin to see an entirely new wave of medical advancement that will better the lives of those in need. It will not be easy, as there are many obstacles to cross along the way, but with time and funding, site-specific bacterial vector drug delivery would be an improvement on the broad-spectrum pills and capsules of our time.

References:

Rubinstein (2006) 'Microbially controlled drug delivery to the colon' Biopharmaceutics & Drug Disposition Volume 11, Issue 6, pages 465–475, August/September 1990Link to paper.

Gardlik R, Fruehauf JH. (2010) 'Bacterial vectors and delivery systems in cancer therapy'Link to paper.

Kudela et al. (2010) 'Bacterial ghosts (BGs)--advanced antigen and drug delivery system' Link to paper.

Hancock, Robert and Sahl, Hans-George. (2009) "Antimicrobial and Host-Defense Peptides as New Anti-Infective Therapeutic Strategies." Nature Biotechnology, Vol. 24, pg. 1551 - 1557 Link to paper.

Holmes, Randall and Jobling, Michael (1996) "Medical Microbiology." Chapter 5. 4th ed., Link to paper.

Lam et al. (2006) “Mechanism of Supported Membrane Disruption by Antimicrobial Peptide Protegrin-1”. The Journal of Physical Chemistry B, Vol. 110, pg. 21282 - 21286. Link to paper.

Levine, Orin et al. (1996) "Generic Protocol for Population-based Surveillance of Haemophilus influenzae Type B." World Health Organization, Geneva. Published . Link to paper.

Martin, E.A. (2003) "Oxford Concise Medical Dictionary." 6th ed.,

Noone Tom. (2006) "An Overview of Steroid Use and its Potential Side Effects." Nursing Times, Vol. 102 No. 17, pg. 24.Link to paper.

NA (2011) "Staphylococcal Toxic Shock Syndrome." Virginia Department of Health. Last updated July, Link to paper.

Stewart, Ian et al. (2006) "Cyanobacterial Lipopolysaccharides and Human Health - A Review." Environmental Health: A Global Access Science Source, Vol. 5. Published March . Link to paper.