Imagine a bacterium that, when injected into the bloodstream, would travel to the site of a tumor, insert itself into the cancer cell, and then produce a cancer-killing compound. That’s exactly what scientists at the University of California, Berkeley (UCB) and University of California, San Francisco (UCSF) have set out to do.

Traditional cancer therapies are limited for two key reasons: little of the drug actually reaches the tumor and the drug is toxic to both cancerous and healthy tissues. Bacteria, however, have the potential to precisely target cells. “In a way, bacteria are the ultimate in smart drugs,” says George Church, a geneticist at Harvard Medical School in Boston (he was not involved in the current work, but will collaborate on the project in the future). “It’s hard to pack a lot of intelligence into a small molecule or protein; but bacteria can have sensors and actuators and can drill into a cell, like a submarine.”

To build a cancer-killing bacterium, biologists must create organisms that can perform a series of complicated functions – namely, when in the bloodstream, they have to sense and respond to the tumor environment. Once inside the tumor, the bacteria must infiltrate the cancer cell, and then – and only then – start producing a tumor-killing toxin. The researchers plan to engineer such super-organisms by co-opting parts from different types of bacteria and inserting them into Escherichia coli, a bacterium commonly used in research.

Tumor tissue has unique characteristics, including lower oxygen and higher lactic acid concentrations than surrounding tissue. To create a bacterium that can sense a tumor, Christopher Anderson, a postdoctoral researcher at UCB and UCSF, and colleagues took an oxygen sensor from E. coli and linked it to a special protein, called invasin, from another type of bacteria, which allows the organism to invade cancer cells. In a paper published earlier this year in the Journal of Molecular Biology, the researchers showed in a test tube that the engineered bacterium selectively invades tumor cells.

Anderson and colleagues are now working on making the system even more specific. To ensure that the bacteria invade only tumor cells, they will create a genetic mechanism that allows the invasin protein to be expressed only when two conditions are met, such as when both the oxygen and lactic acid concentrations are at a certain level. Essentially, it’s a genetic version of what’s known in engineering terms as an AND gate – a regulatory circuit that’s turned on only if two conditions are met.

“By using multiple cues, we can garner a great deal of specificity,” says Adam Arkin, a bioengineer at UCB and the Lawrence Berkeley National Laboratory, a TR100 recipient in 1999, and one of the senior scientists on the project. “After the bacteria sense the cues, they turn on the rest of the apparatus to do the job.”

The last step, which the biologists haven’t yet started, will be to engineer the bacteria to produce an enzyme that converts a drug precursor into a cancer-killing compound. The patient would take a pill containing the precursor, which would be converted into the active drug only inside the bacterially infected cancer cells.

“The bacteria are engineered to sense cells and then invade – that’s a unique demonstration of what you can do by programming cells,” says James Collins, a biomedical engineer at Boston University (who was not involved in the project).

Anderson and colleagues still face many technical and safety hurdles before they have a viable bacterial cancer treatment. For example, the human immune system attacks bacteria when they’re injected into the bloodstream. So the researchers engineered the bacteria to produce a special lipid coating that makes the organism invisible to part of the immune system.

The team is also designing a series of safeguards into the system. Patients could develop sepsis, for example, if the bacteria reproduced too quickly in the blood. Previous research has shown that bacteria need iron from human blood to grow and reproduce. So the scientists deleted a certain gene that allows the organisms to extract iron from the blood, which should drastically limit bacterial growth inside the body.

While the idea of bacterial therapies may sound highly experimental – and even frightening – it has existed for at least a century, and it’s already been put into practice. The Bacillus Calmette-Guerin (BCG) vaccine, a preventative treatment for tuberculosis made from a live but weakened version of the tuberculosis bacterium, is also effective in treating bladder cancer when injected directly into the bladder. Use of this therapy, however, is limited to cancers lining the surface of the bladder, because the bacteria can easily access these cancer cells.

In addition, some strains of bacteria have a natural affinity for tumor cells, a quality that scientists have tried to take advantage of when designing new therapies. Vion Pharmaceuticals, based in New Haven, CT, is developing a cancer treatment centered on the salmonella bacteria, which have this property. Anderson’s bacteria work differently, though: they actively target tumor cells through the invasin protein, rather than relying on the bacteria’s natural mechanism, and therefore may be less sensitive to differences in tumor types.

Still, the latest work is likely to require a lot of scrutiny from the Food and Drug Administration. Anderson emphasizes that, at least at first, these bacteria would be given only to people who’ve failed to benefit from other treatments.