Tabitha M. Powledge is a science writer based near Washington, DC. E-mail:
It is the certain fate of all antibacterials to be fought off eventually by the pathogens they target. We need new ways to defeat disease, and we will need them forever.
By next summer, more than 40% of
It is the certain fate of all antibacterials to be fought off eventually by the pathogens they target. The fact that the process is accelerating has been alarming public health officials for some time, especially in the United States. We need new ways to defeat disease, and we will need them forever.
Antibiotics have traditionally been plucked from nature's battleground. For billions of years, tiny organisms have engaged in an arms race, hurling toxic molecules at each other in the struggle to prosper. Nearly all of today's antibiotics are versions of weapons long wielded by microbes and fungi. Chemical synthesis of entirely human-created antibiotics has so far yielded only fluoroquinolones, a group of broad-spectrum antibiotics that includes Cipro, which became famously scarce during the 2001 anthrax scare, and linezolid (trade-named Zyvox), which is effective against some resistant strains of
The usual way to find a new antibiotic has been laborious screening of immense libraries of compounds, natural and otherwise. Some argue that screening chemical libraries is approaching a deadend. There may be diminishing returns from screening, but it's not quite dead yet: in October, researchers at the University of Wisconsin at Madison reported a new class of bacterial RNA polymerase inhibitors with antibiotic potential. They were found by screening for molecules that prevent
Christopher T. Walsh of Harvard Medical School says screening's problem may be simply that libraries aren't good enough. Marine organisms have not been studied well, he points out, and 90% of organisms in the biosphere can't be cultured in standard ways. He says, “We're missing 90% of them every time we go and look in nature.”
Walsh is doing his bit to create new libraries. He and his colleagues have recently employed combinatorial biosynthesis to learn how to use part of the machinery for assembling cyclic peptide antibiotics to control their architecture. The result was a small library of natural product analogs, some of which have improved antibiotic activity against common bacterial pathogens. “There are dozens of such enzymatic domains that in principle one could clone, express, and test with other substrates. I view that as the kind of thing we should do,” he says. For example, Walsh suggests, it is a reasonable approach to second-generation improvement of daptomycin, the antibiotic most recently approved for sale in the United States.
Walsh collaborates with Chaitan Khosla of Stanford University on finding ways to make existing antibiotics better. They are studying biosynthesis of rifamycin, an antibiotic that is increasingly less effective against its prime target, tuberculosis (TB) (see
This 40-year-old Estonian truck driver's TB is resistant to drugs and his right lung was removed three days before this picture was taken. (Photograph by WHO/STB/Colors Magazine/J. Langvad.)
“Basically, what we do is to try and figure out new ways to hijack the biosynthesis of antibiotics in nature so as to modify their structures with the goal of improving them,” Khosla explains. He works with an important class of natural antibiotics called polyketides that have generated dozens of drugs, including erythromycin.
Polyketides are secondary metabolites (which give their producers a competitive advantage in their environment) produced mostly by bacteria and fungi and made by a complex and structurally diverse family of enzymes called polyketide synthases (see the primer by David Hopwood in this issue of
The researchers didn't study the new anthracyclines' effects on bacteria, but Khosla notes that the general principle should apply to other classes of compounds, although the details of how it's implemented will vary from system to system. He says, “The upshot of this paper is that it is now possible to modify a particular methyl group in just about any anthracycline antibiotic.”
Instead of searching for new antibiotics by modifying existing ones, some researchers are trying something completely different—first finding the most vulnerable targets in a bacterium and then designing something that hits one or more of them hard. “You have to understand a helluva lot more about how these little cells work. In fact, we think we understand a lot, but I think we can understand almost everything now that we have all the genomes,” says Lucy Shapiro of Stanford University School of Medicine. While having full genome sequences—more than 100 microbe sequences have been completed—is essential, Shapiro believes that knocking outs genes galore to find out which ones are necessary and going after them all is not a sensible strategy. She observes, “People have been doing that for a while with absolutely no success. That's really going after the problem with a Howitzer instead of with an intelligent approach.”
So instead of screening libraries of existing compounds, Shapiro prefers using structural information about drug targets or their natural ligands to create new drugs, an approach known as rational drug design. And instead of looking at all essential genes in a bacterium and choosing one to target, she and her colleagues look at genetic circuitry that controls the cell cycle, the pathway that coordinates cell growth and differentiation. They have identified key control points, or nodes, in the circuitry for their favorite study subject,
Antibiotic discovery is all chemistry, Shapiro says, which is why she joined with biochemist Stephen J. Benkovic of Pennsylvania State University. They didn't know the structure of CcrM, Benkovic explains, but the literature about other methyltransferases suggested that the adenine molecule, which is the substrate for CcrM within DNA, binds to a specific region of the enzyme.
The researchers designed adenine-like molecules that would bind to CcrM and then developed inhibitors. Benkovic says, “We already knew what kind of structure we wanted, and we simply fine-tuned it.” They worked their way through 1,000 inhibitor candidates, ending up with a small subset—no more than about 20—that not only inhibited CcrM, but also killed
And not only inoffensive
More recently, Shapiro reports, they have demonstrated efficacy against rats infected with anthrax or multidrug-resistant
Benkovic points out that theirs is an entirely new class of compounds, small molecular weight compounds that can be made in a few steps. He says, “They don't look like the normal antibiotic, so that's why I think they're fairly unique.” The basic research was done under a grant from the Defense Advanced Research Projects Agency (DARPA), the United States Department of Defense's (DOD) central research and development organization, and once the researchers realized they wanted to develop drugs against three agents that have been considered bioterrorism threats —
The most radical approach to new antibiotics may be the resurrection of an old idea: bacteriophage therapy (see
Negative stain electron micrograph of the gamma phage from which the PlyG lytic enzyme was cloned for use to control
Phage were the model organisms of choice for genetics research in the 1930s and 1940s, but became less fashionable as research tools when investigators moved on to eukaryotes. A few held on, like Ry Young of Texas A&M University, who has made phage-induced cell lysis his life's work. “The cell is basically genetically dead as soon as the phage goes in there, but it will keep living as sort of an infected zombie for as long as the phage wants it to, with virus particles accumulating inside the cell,” he explains. “Only when the phage is ready and has decided that it's the right time will it pull the trigger. And the cell blows up.” The freed phage then spew forth to infect new cells.
Antibiotic resistance has led to new interest in phage therapy by several small biotech companies. Young continues basic research at Texas A&M, but has also joined one of them, GangaGen, providing bacteriophage expertise to its labs.
Phage do kill pathogenic bacteria effectively, and they do it without penetrating human cells, which they can't even recognize. So what is keeping phage therapy out of the clinic? Problems that some doubt can be overcome.
Because bacteria develop resistance to phage rapidly, phage therapy companies will need to direct cocktails against a single pathogen, according to Vincent Fischetti at The Rockefeller University. Phage are also antigenic, and the antibodies they stimulate will neutralize their effects during subsequent treatment, he says. But the chief problem appears to be regulatory—regulatory in the political, rather than the genetic, sense. When bacteriophage package their DNA, they occasionally include varying amounts of their hosts' DNA, too. This miscellany, Fischetti points out, is likely to make the Food and Drug Administration unhappy. “Phage normally are very fragile, their tails break, so lot-to-lot homogeneity could be a problem too,” he adds. “So even though it will work, I think they'll have an uphill battle.” Phage may well enter agricultural or veterinary use, he predicts, but are probably not going to be available to patients in the United States any time soon.
Fischetti chose a different approach to phage therapy. It does not rely on phage themselves, but on enzymes that phage produce to smash their way out of their host bacteria so they can infect new hosts. He and his colleagues employ these enzymes externally to kill bacteria. He reports, “We now have enzymes that will kill
The enzymes can be loaded into a nasal spray that wipes out pathogens such as
Fischetti and his colleagues have moved on to using the enzymes systemically to wipe out
Up to now the enzymes must make contact with bacteria to kill, but Fischetti is hoping that a new generation of engineered enzymes will be able to kill pathogens inside cells too. A second disadvantage is that they are effective only against gram-positive bacteria, although that group includes many vicious pathogens.
But phage enzymes seem to offer one very big advantage: resistance to them has yet to develop. Fischetti says, “We've tried very hard to identify resistant bacteria, but so far we haven't found resistant organisms in all three of the enzymes we're working with. It appears to be a very rare event, much rarer than resistance to antibiotics.” Fischetti cautions against expecting that gladsome state to last forever, but he points out that even if widespread resistance takes the same 40 or 50 years that antibiotics required to become significantly resistant, phage enzymes could buy researchers decades for inventing other approaches.
There is no shortage of ideas for unearthing new antibiotic candidates. Why are they so slow to enter medical practice? The bottleneck, researchers agree, lies in the development process of turning them into effective therapies. Several researchers blame the big pharmaceutical companies that got so big by leading the way to new drugs for battling infectious disease, but in recent years have dropped out. Fischetti complains, “These are the big companies that have the money to develop antiinfectives, but they leave it to small biotech companies, and it's not going to happen as rapidly as it should. I think it's really unconscionable for these big companies to drop the ball because it's not going to be a billion-dollar market for them and that's what they're looking for.”
Half a billion at least, says Francis Tally, a big pharmaceuticals veteran who is now chief scientific officer at Cubist Pharmaceuticals, a biotech company located in Lexington, Massachusetts. According to Tally, Cubist produced daptomycin, approved in September 2003, by licensing it from Eli Lilly, which shelved the new compound after concluding its potential market was only $250 million.
But, Tally argues, the size of the market is not the only barrier to new antibiotics. Combinatorial chemistry and the genomics revolution have simply not delivered on their early promise. “The pipeline is very dry,” he says. “There's been a real lag at the basic research level.”
“Antibiotic discovery is hard,” Shapiro says. “It's a huge long process to get a decent antibiotic.” Walsh agrees. “It's easier to find inhibitors of particular enzymes for particular processes—and a very long road to convert that into something for development.”
In the meantime, there is a rising clamor to slow down the rate at which bacteria develop resistance. Doctors are exhorted to cut back on prescribing antibiotics and decline to prescribe for viral diseases, which antibiotics can't combat, even when their patients badger them.
But even if antibiotic consumption slowed, we will still need new antibiotics. “I always say it's not a matter of if, it's only a matter of when,” says Walsh. “There will always be a need for new antibiotics because the clock starts ticking on the useful lifetime of any antibiotic once you start to use it. That cannot be argued.”
Defense Advanced Research Projects Agency
Department of Defense
tuberculosis