Looking at a scary future
July 19, 2010 § Leave a comment
Roy Kishony pointed me to this fascinating but worrying paper (D’Costa et al. 2006. Sampling the antibiotic resistome Science. 311 374-7), which has apparently influenced the work of the Kishony lab quite a bit.
The authors wanted to know what kinds of antibiotic resistance genes exist in the wild. Often it turns out that the antibiotic-resistant strains that arise in the clinic didn’t evolve their resistance from scratch. A good deal of drug resistance comes from genes that originally evolved to protect an antibiotic-producing bacterium from killing itself with its own poison, which are acquired by the pathogenic bacterium via horizontal gene transfer. So, the authors argue, finding out what pre-formed resistance strategies are available might well help us to predict — and perhaps prepare to defend against — the problems of the future. They selected 21 commonly used antibiotics, with origins that ranged from completely natural, through semi-synthetic, to entirely synthetic, and looked to see whether they could find strains that are resistant to them in bacteria from natural soil samples.
The bacteria D’Costa et al. picked to look at were strains of Streptomyces, the source of over half the antibiotics in clinical use today. Because Streptomyces strains produce so many antibiotics, they reasoned that these would also be the most likely bacteria to contain genes that might convey antibiotic resistance. They isolated 480 strains of appropriate morphology from a variety of soil samples, checked that they were indeed Streptomyces by sequencing 16S RNA, and screened them for resistance.
The horrifying thing about this study is that the average Streptomyces strain is resistant to 8 of the 21 antibiotics tested. Some of them are resistant to 15/21. Though a few antibiotics did very well against the panel of strains — Gentamicin, for example, was effective against essentially all the strains — others did unexpectedly poorly. Daptomycin, a lipopeptide that is thought to insert itself into bacterial cell walls (made by local company Cubist Pharmaceuticals) showed almost no efficacy against the panel. This was a total surprise (although perhaps it shouldn’t have been, since daptomycin is made by a Streptomyces — Streptomyces roseosporus), and it implies that we need to watch daptomycin very carefully for signs of developing resistance. Two additional nasty discoveries were apparently novel mechanisms of resistance to vancomycin, traditionally the drug of last resort in gram-positive infections, and telithromycin, which has also been useful against bacteria that are resistant to other drugs.
The authors helpfully point out that the number of worrying drug-resistant genes they’ve found in the natural environment is bound to be just the tip of a very large iceberg. The strains they were able to study are only the ones that were easily cultured in the lab; and their exploration of resistance mechanisms was far from exhaustive, since they only looked at resistance to high concentrations of antibiotics. We have to assume that unless something radical changes, almost every antibiotic available to us today will sooner or later lose its efficacy, because of transfer of resistance genes from bacteria like these that started out resistant — or, of course, the evolution of de novo resistance mechanisms.
A very bleak prospect. I am not looking forward to a world without effective antibiotics at all.
But cheer up. (Well, cheer up slightly.) Roy looks at these findings from a different and more hopeful angle. Yes, most of D’Costa et al.’s strains were resistant to several antibiotics, and almost all the antibiotics (18/21) can be detoxified and/or evaded by several strains. But another remarkable finding, if you look at the authors’ results from a slightly different angle, is that not every strain has taken up every available resistance gene. Why would this be? Admittedly the bacterial strains come from several different places, but still, if a resistance gene is useful you would think it would spread pretty far in the population, as it does in the clinic. Maybe there is something stopping the spread.
The Kishony lab has been pondering on the evolutionary questions raised by this paper for a while. What kinds of selection pressures would select for drug-sensitive bacteria instead of drug-resistant ones, and why would these pressures be more important in the soil than in the clinic? One possible answer is described in a recent paper (Palmer et al. 2010 Chemical decay of an antibiotic inverts selection for resistance. Nat Chem Biol. 6 105-7. PMID: 20081825), which shows that the breakdown products of an antibiotic, tetracycline (another Streptomyces product), can last much longer than does the antibiotic itself, and can invert the evolutionary advantage enjoyed by tetracycline-resistant bacteria. A probable mechanism is that in the absence of tetracycline itself breakdown products can induce costly, and unnecessary, production of the proteins responsible for resistance, putting resistant bacteria at a disadvantage relative to their sensitive cousins. Because the breakdown products last longer than tetracycline, the overall effect is selection for sensitive bacteria. (In the clinic this mechanism is probably irrelevant, since tetracycline and its breakdown products are all flushed out through the kidneys together). This, Palmer et al. say, “may help explain the puzzling coexistence of sensitive and resistant strains in natural environments.” And perhaps some of the insights from this work will also help us to find ways to help slow or prevent the spread of resistance — just as D’Costa et al. hoped.
D’Costa VM, McGrann KM, Hughes DW, & Wright GD (2006). Sampling the antibiotic resistome. Science (New York, N.Y.), 311 (5759), 374-7 PMID: 16424339
Palmer, A., Angelino, E., & Kishony, R. (2010). Chemical decay of an antibiotic inverts selection for resistance Nature Chemical Biology, 6 (3), 244-244 DOI: 10.1038/nchembio0310-244a