On the hunt: Voigt tried E. coli. Credit: Martha Bruce Photography

Chris Voigt opens a glass case and delicately lifts a chocolate brown seashell covered with a naturally occurring pattern of uniform white triangles. “From a very simple set of genetic instructions, you get a very complex pattern, see?” He then points to another shell that looks as if it were etched with letter-like symbols. These patterns on shells have inspired the MIT bioengineer. They suggest that genetic systems can contain an almost computer-like logic, and with a few elegant lines of genetic code an organism can be made to behave almost mathematically.

With that logic in mind, Voigt, in collaboration with Chris Anderson and Adam Arkin at the University of California–Berkeley, developed a genetic circuit in Escherichia coli that causes the bacteria to hunt and destroy cancer1. The bacterial cells are programmed to circulate through the bloodstream and detect the low oxygen levels in tumors. Once in contact, the bacteria respond by producing a protein called invasin, which causes them to invade the tumor cells.

Illustrating the reliability of the system, when Voigt and his collaborators recovered the bacteria from the cancer cells, all the bacteria they found had expressed the invasin protein. As a next step, the E. coli could be programmed to release a cytotoxin to kill the cancer.

Theirs is just one of several projects aiming to turn microorganisms into disease fighters that can then be injected into the human body. Other groups have turned to Salmonella for similar cancer-fighting prospects. But, after seven years of working on tumor-invading bacteria, Voigt and his collaborator Chris Anderson, have given up. “It was only six months ago that we said, 'let's not do this anymore',” Anderson says.

Anderson explains that his lab had successfully designed a circuit within the E. coli to target and attack cancer cells, but the immune system slaughtered the bacteria before they could act. To prolong their survival, his lab inserted genes to change the bacteria's surface proteins. “We ended up in highly uncharted territory,” he says, and the gene inserts failed to keep the bacteria alive.

“We discovered so many fundamentally missing abilities. It's a problem that is more difficult than our toolkit can handle,” Anderson says. “At the time, it seemed like you'd be able to make stable devices out of these very noisy [DNA] parts. The result is you really need to understand how all the bits and pieces fit.”

In the aftermath, Anderson has moved onto finding a better way to assess and define which gene parts would work best in a circuit, whereas Voigt has refocused his attention on using synthetic biology to produce small molecules for new—nonliving—medicines.