For bacteria, too, there is strength in numbers.

Many bacteria "act as big organisms," E. Lynn Zechiedrich, assistant professor of molecular virology and microbiology at Baylor College of Medicine, told BioWorld Today. "They are all communicating with each other, and even more, they’re communicating with us and subverting us - as a group, because one [bacterium] could never do it alone."

Bacteria talk to each other via a process known as quorum sensing. Each bacterium secretes signaling molecules into extracellular space and has receptors for those same signaling molecules. The amount of quorum sensing signal a bacterium picks up allows it to estimate how many ilk are in the vicinity; when a critical population density is reached, bacterial growth slows and genes coding for everything from bioluminescence to virulence can be expressed, depending on the bacterial species. Often the bacteria also will stick together in a biofilm, which can be a problem on everything from the lungs of a cystic fibrosis patient to artificial hip joints.

Zechiedrich and her colleagues have published a paper in the Feb. 14, 2006, issue of the Proceedings of the National Academy of Sciences suggesting that quorum sensing signals might not just diffuse out of bacteria. Instead, they may be actively pushed out by bacterial pumps that are known to medical scientists for another reason: They pump antibiotics out of bacteria, rendering them multidrug resistant.

Zechiedrich got into quorum sensing more or less by accident. "We were doing drug resistance," she said. But during genomic studies on which genes are related to multidrug resistance, her lab kept finding that quorum-sensing genes "were regulating our multidrug efflux pumps."

One theory is that bacterial pumps have evolved as a defense mechanism: Bacteria evict antibiotics because they recognize them as toxins. But Zechiedrich said that the pumps are evolutionarily very old and conserved.

"They haven’t changed over all these millions of years - they were around long before people were making the drugs they pump out," she said. In other words, their primary role must be something other than getting rid of antibiotics.

In previous work, the researchers had identified every pump in Escherichia coli that can get rid of fluoroquinolones, a class of antibiotics with 13 members on the market.

Of E. coli’s 40 pumps, three were able to shunt fluoroquinolones out of the cell: acrAB, norE and mdfA. The researchers then constructed E. coli strains that were genetically identical except for the fact that they were lacking one or two of the three pumps.

In the studies now published in PNAS, the scientists compared how E. coli strains lacking one or more pumps behaved in both the rapid-growth phase before quorum-sensing signals have reached critical mass, and the stationary phase, in which they have reached critical mass and bacteria are no longer dividing as rapidly. Cells overexpressing the pumps also were studied.

Bacteria lacking either the acrAB, or the norE pump grew at the same pace as their wild-type cousins in the rapid-growth phase. In the stationary phase, however, both knockouts grew to higher cell density than wild-type strains, consistent with the idea that there was less quorum-sensing signal around. Overproduction of either pump had the opposite effect, that is, cell density was lower than that of wild-type strains.

Removing the third pump, mdfA, had no effect on bacterial growth rates. Zechiedrich said she did not know why, but was "happy" that not all the experiments showed an effect on quorum sensing: "If everything you do has an effect, then you start questioning your assay system."

Conditioned Medium Is Message

The researchers also used conditioned medium from cells expressing more or less acrAB pumps than wild-type. If the pumps are pumping quorum-sensing signals out of the cells - and into the cell culture medium - then there should be an inverse correlation between number pumps and growth of cells treated with conditioned medium: The more pumps, the more quorum-sensing signal and the less cell growth, because cells treated with the medium were getting the message that there were already lots of other cells around. That relationship was indeed what Zechiedrich and her colleagues found.

Gene-deletion experiments showed that the quorum-sensing signal responsible for those effects was not the "classical" quorum-sensing signal autoinducer 2. Zechiedrich said the implications of her work for drug development are twofold. For one thing, drugs that are not transported by the efflux pumps are likely to cause fewer headaches about multidrug resistance in the clinic down the road.

For another, there are a number of drug discovery efforts targeting drug efflux pumps. But if those pumps are central to bacterial communication, any successful attempt at targeting them will probably bring with it a lesson in unintended consequences: "We’re messing with cell communication signals that have been around since the beginning of life," Zechiedrich said. "If we destroy the phones, [bacteria] are going to find another way to communicate, because it’s what has to happen."

It would not be the first time that quorum-sensing bacteria have foiled drug developers. "There are two equally plausible premises," about the effect that targeting quorum sensing would have on resistance, Jeff Stein told BioWorld Today.

Because targeting quorum-sensing mechanisms does not kill bacteria, it could impose milder selective pressure. But it also could apply stronger selective pressure, because "any mutation that enables bacteria to get around your drug would proliferate quite rapidly - because it is not killing them."

Stein, who now is a partner at Sofinnova Ventures, was co-founder and chief scientific officer of Quorex Pharmaceuticals, which was acquired by Pfizer in 2004. At the time of its acquisition, Quorex was no longer pursuing quorum-sensing signals as its target. Why? Because for the target the company had selected, the second equally plausible premise turned out to be true and bacteria developed resistance very rapidly.

Stein stressed that there is no reason to believe that would be the case for every quorum-sensing target, and noted that several biotech firms are pursuing either quorum sensing or other approaches aimed at reducing virulence. But he did note that getting successful quorum sensing drugs onto the market entails educational challenges on top of the usual drug discovery struggles, and even getting the permission to enter the clinic can be difficult.

"How do you convince the FDA that this drug should be in humans when it is not killing bacteria in a Petri dish?" he asked. "It is an uphill battle."