Conventional wisdom holds that when bacteria are battling for survival against antibiotic treatment, resistance to a drug will always give a bug a competitive advantage. Researchers have been looking at ways to deliver antibiotic therapies that will make the odds more equitable, such as using higher doses, combining drugs, and cycling through treatments. But a study from the lab of Roy Kishony, assistant professor of systems biology at HMS, finds that under certain conditions, a combination of drugs could actually handicap resistant strains and give susceptible strains the advantage. This reversal of fortune, detailed in the April 5 Nature, so far is limited to bacteria growing in vitro, but it suggests a new avenue of research for pairing drugs together to counteract resistance in patients.
Measuring Dual Drug Effects
Combining two different antibiotics has long been used as a way to prevent resistance. These combinations have often been described in a qualitative way, and less research has looked systematically at how different drug combinations affect resistant and susceptible bacteria. Kishony’s lab has taken a systems-level approach to studying the evolution of drug resistance, using high-throughput techniques to quickly identify how drugs affect bacterial growth. By finding ways to simplify a complex phenomenon, the lab is able to characterize the effects of drugs on bugs in a more quantitative way.
Drug combinations can be characterized as synergistic, additive, or antagonistic, depending on whether the combined effect of the two drugs is larger than, equal to, or smaller than what would be predicted by their individual activities. A previous study in Kishony’s lab, led by postdoctoral fellow Pamela Yeh, systematically categorized drug interactions between pairs of antibiotics. The study suggested that certain antagonistic drug interactions might create conditions in which resistance to a drug could actually be harmful to bacteria. In this class of interactions, called suppression, the combined effect of the two drugs is even less than the effect of using one drug by itself.
Graduate student Remy Chait explored the effects of suppressive drug combinations on resistant and susceptible bacteria. As a first experiment, he tested the effect of drug combinations on two strains of E. coli, one of which is sensitive to the antibiotic doxycycline and one that is not. The strains carried a bioluminescence gene. Chait used a sensitive detector to measure the light given off by each population as an estimate of their rates of growth. He then varied the concentrations of the two drugs to find the relative concentrations needed to achieve a state of zero growth in the bacteria.
The researchers found that when doxycycline is given in combination with a synergistic drug, erythromycin, the resistant strain requires a much higher level of doxycycline to curb its growth compared to the sensitive strain, and there is no concentration of the two drugs at which the sensitive strain survives and the resistant bacteria die. But when doxycycline is combined with ciprofloxacin, a drug that doxycycline suppresses, there is a level of concentration at which the sensitive strain grows and the resistant strain does not.
The team, which also included graduate student Allison Craney, then looked at how different strains of bacteria fared when they were directly competing with one another in the same environment. The researchers labeled wild-type and resistant strains of bacteria with different fluorescent labels and tracked their growth in response to different concentrations of two drugs. As expected, the resistant strain out-competed the susceptible one when the two drugs were synergistic, but when one drug suppressed the other, there was less selection for resistance overall, and Chait found a region of drug concentration at which the resistant strain lost the battle.
Why is this so? Although the exact mechanism of how the drugs interact is unknown, it appears that when a cell develops resistance to doxycycline, the drug no longer suppresses the effect of ciprofloxacin on the cell. In some cases, this is because doxycycline is literally being pumped out of the cell, but the same effect is seen in bacteria that have acquired different ways of resisting drug effects.
“The mechanisms of resistance—regardless of whether it is a pump or a degradation enzyme or a ribosome protection mechanism that destabilizes binding—in the end they have the same purpose, which is, they decrease the apparent concentration of the antibiotic,” Chait said. “Effectively, the bacterial cell sees two poisons, one of which, the suppressor, is a partial antidote to the other. Taking them together is harmful for all the cells. But those cells resistant to the suppressor also no longer receive its partial benefit, and they do worse at drug concentrations in which the drug’s effects as an antidote outweigh its effects as a poison.”
The study offers the counterintuitive finding that an antibiotic, if paired with another in the right way, can be used to discourage resistance against itself. Kishony said that while suppressive combinations are not common, several do exist. The relationship between doxycycline and ciprofloxacin only goes one way, so pairing them together discourages resistance only against one of the drugs. The team noted that it would be worthwhile to screen for drug pairs that suppress each other, which could potentially generate selection against resistance to both drugs.
Though Kishony’s lab is focused on the basic processes of bacterial evolution, he hopes that the finding will spur other researchers to pursue new treatment strategies. Robert Moellering, the Shields Warren–Mallinckrodt professor of medical research at Beth Israel Deaconess Medical Center, said that current research has focused on using high doses of antibiotics in hopes that in such a highly competitive environment, resistant bugs will prove to be less fit. This study offers a completely new idea. It suggests that in certain circumstances, low doses of drugs could actually discourage resistance more effectively. “This is a novel approach to preventing resistance,” Moellering said. But it is still a long way from being applied in the clinic. The next step, he said, is to screen through more drug combinations at different concentrations and better characterize the phenomenon, and then it could be tested in vivo.
