You say yes, I say no. You say stop and I say go, go, go.
Mixed signals can be confusing, as the Beatles’ refrain has it. But when the signals come from antibiotic drug combinations, cells react in surprisingly simple ways, HMS researchers have found. Even when drug pairs affect different genes in a single cell in complex ways, the cell as a whole responds in a manner that’s predictable—an insight that could improve drug design.
So-called combination drug therapy is a staple for treating many infectious diseases. Doctors treating tuberculosis, for example, might prescribe one drug to break down the pathogen’s protective barriers and a second to deliver the knockout punch. But identifying effective combinations for a particular disease has relied on guesswork—and the excruciatingly slow accumulation of data.
Roy Kishony, professor of systems biology at HMS, and Tobias Bollenbach, a postdoctoral fellow in his lab and now an assistant professor at the Institute of Science and Technology Austria, wondered whether there was a better way to explain —and perhaps predict—why some drugs work better together while other pairings are less powerful or even counterproductive. Using a systems approach, Kishony and Bollenbach investigated how, within a living cell, gene expression responds to drug pairings.
“The possibility of predicting how cells respond to multi-drug treatments opens the door to a more rational approach for the design of new drug combinations,” Kishony said.
Average Or Prioritize?
Kishony and Bollenbach measured how the single-celled bacterium E. coli responded when subjected to a combination of two drugs. The effect could be either additive, with the drugs’ combined inhibitory effect equal to the sum of their individual effects, or antagonistic, in which case the drugs have a weaker effect when combined. In either case, the bacterial cell’s response to one of the drugs may prove incompatible with its response to the other. For example, a specific gene in the cell may be “turned off” by drug A but “turned on” by drug B. So how do cells as a whole respond, the researchers wondered, when A says stop and B says go?
In a study published in the May 20 issue of the journal Molecular Cell, Kishony and Bollenbach report that bacterial cells respond in surprisingly simple ways, which can be reasonably predicted by monitoring only a handful of their responses.
When drugs enter a bacterial cell’s environment, the researchers found that the response can be broken down to two components: the first, comprising about 70 percent of the cell’s response, involves processes resulting from the total inhibition of the cell’s growth by the two drugs. In the remaining 30 percent, the cells focus on resolving conflicts that arise when paired antibiotics caused mixed genetic responses. This conflict resolution depends on the nature of the signals sent by particular drug pairs.
Kishony and Bollenbach found that bacterial cells resolve conflicting signals from drug combinations by either “averaging” or “prioritizing.” For a drug pair that is additive, the cell averages the conflicting effects of the two drugs. (For example, when one drug’s effect on the regulation of a gene is a four-fold increase and the other’s is a two-fold decrease, their combined effect on the cell is a two-fold increase.) But for the antagonistic drug pair, the cell responds only to the stronger drug signal, ignoring the other. Particularly surprising was that, no matter the drug pairing, almost all genes within a bacterial cell were in agreement about which conflict-resolution strategy to use and which signal was strongest.
These findings demonstrate that it is possible to quickly predict bacterial responses to combined drugs, simply by measuring just a few aspects of how a cell responds to individual drugs. Thus the most effective combinations can be more easily determined.