New Antibiotics and New Resistance

Carlos F Amabile-Cuevas. American Scientist. Volume 91, Issue 2. Mar/Apr 2003.

Bacterial infections have been a scourge on humankind for millennia. Plague, tuberculosis, wound infections and typhoid fever have caused historical as well as personal tragedies. No wonder, then, that antibiotics were greeted as miracle drugs. For a few decades the success of antibiotic therapies was remarkable, but enthusiasm for them led to abuses. Observers disregarded the early emergence of resistant bacteria; a number of new antibiotics were still being discovered, suggesting that effective drugs would always be available. With infections deemed under control, pharmaceutical companies lost interest in developing new antibiotics.

After decades of complacency and just 50 years after the first clinical use of an antibiotic, penicillin, the public health threat posed by antibiotic resistance finally gained widespread attention. Resistance made the cover of Time and Newsweek in the early 1990s; now, most people know that antibiotics can fail. Over nearly 20 years, from the early 1980s to the late 1990s, not a single truly new antibiotic was introduced into clinical use. Even now, barely a trickle has reached the market since 1999. Meanwhile, resistance keeps evolving, and drugs are rapidly losing their efficacy, resulting in increased treatment costs, loss of labor time and, of course worst of all, lost lives. My colleagues and I reviewed how bacteria evolve so quickly towards resistance some years ago. Here I will discuss new discoveries on the biology of resistance, as well as efforts to either restrain or circumvent resistant organisms. In the struggle against antibiotic resistance, science is providing useful tools, and physicians are slowly realizing that antibiotics are simultaneously powerful and dangerous drugs. Ultimately, though, we will all need to change the way we deal with bacteria in the coming “post-antibiotic era.”

Ways Bacteria Resist Antibiotics

Antibiotics are compounds that kill or at least inhibit the growth of bacterial cells, without harming the patient. No single antibiotic can kill or inhibit all bacteria. Natural penicillin and macrolides, such as erythromycin, for instance, cannot penetrate into the gut bacterium Escherichia coli and its relatives; only a handful of drugs work against the almost impermeable Mycobacterium tuberculosis, which causes tuberculosis. The intrinsic resistance of bacteria defines the spectrum of each antibiotic; wide-spectrum antibiotics are effective against a variety of germs, whereas narrow-spectrum antibiotics only control a few species. But the antibiotic resistance we normally speak about refers to cases in which organisms that were originally killed by a certain drug suddenly keep growing in its presence. When a concentration of antibiotic safely attainable in the blood and tissues of a patient no longer affects an organism, we say the strain has become resistant.

The first explanation for resistance was that mutations, small changes in the genetic information, of a bacterial cell somehow prevented an antibiotic from acting on it. Certainly, many resistant organisms arose through the acquisition of spontaneous mutations; this is particularly true for germs causing tuberculosis. But, unexpectedly, genes conferring resistance rapidly emerged and accumulated, quickly yielding multi-resistant bacteria-that is, strains resistant to three or more antibiotics. Also, some bacteria were found to have the same resistance genes as those found in species that naturally produce antibiotics. (Most antibiotics are obtained from various species of soil bacteria, which have been producing these compounds for millions of years.)

It became clear that bacteria can exchange genes, a process known as horizontal gene transfer (see “Horizontal Gene Transfer,” July-August 1993). In this way, a mutation conferring antibiotic resistance can be acquired by neighboring bacteria, even if they are very distantly related species. The resistance genes can spread from mutants or even directly from antibiotic—producing species. Furthermore, such genes can accumulate in a single cell, resulting in multi-resistant germs.

Bacteria often carry the resistance genes in small DNA molecules called plasmids, which act as genetic “supplements” to the core genome. Exchanging these supplements is easier than mobilizing genes in the genome, just as it would be easier to borrow a magazine rather than a large, expensive book from a friend. Genes conferring other dangerous traits, such as virulence, are also often found in plasmids.

In addition to specific antibiotic-resistance genes, bacteria have defense mechanisms that prevent the entrance of noxious compounds or that pump them out of the cell. These mechanisms are activated in response to particular chemical signals and can make a population of germs transiently resistant to multiple drugs. In E. coli at least two of these mechanisms are known. The mar regulon, discovered by Stuart Levy’s group at Tufts University, is a set of genes that is activated by a single regulatory protein that leads to multiple antibiotic resistance. The sox regulon, described by Bruce Demple’s team at Harvard, is a somewhat overlapping set of genes that is activated by superoxide radicals. The activating signal for such regulons are very diverse-ranging from salicylate, the active compound of aspirin, to superoxide radicals, released by white blood cells to kill invading bacteria. Ana Fuentes in my lab discovered that even mercury, at concentrations similar to those released from “silver” dental fillings while chewing, can activate the sox system. Once activated, these systems protect bacteria from a number of antibiotics; mutations that keep any of these systems permanently activated result in permanent multi-resistance.

Yet another strategy allows bacteria to survive antibiotics. The cells often grow in complex, multi-layered, multispecies consortia-what one might call “cities of microbes.” These aggregations are called biofilms, a term coined by William Costerton of Montana State University. Small subpopulations within these consortia are able to withstand the presence of antimicrobial agents and can resume growing once the agent is gone. This persistence, rather than actual resistance, is responsible for antibiotic treatment failures, especially when the biofilm is attached to foreign bodies, such as prostheses or catheters. The precise mechanism of this persistence is not well known, but slow diffusion of an antibiotic across the biofilm may give some cells enough time to activate environmental stress responses, such as the mar or sox mechanisms. Also, biofilms might be ideal places for horizontal gene mobilization, allowing resistance genes to be transferred and expressed rapidly. Eliana Drenkard and Frederick M. Ausubel at Harvard Medical School recently identified a protein of Pseudomonas aeruginosa that regulates both biofilm formation and a switch between antibiotic resistance and susceptibility. Fabrizio Delissalde in my lab recently discovered that among strains of P. aeruginosa causing infections in hospitalized patients, the greater the ability to produce biofilm, the fewer the drug-specific resistance genes in the strain. We suppose that the protection from forming biofilms is enough for these bacteria to survive in hospitals, where antibiotics are ubiquitous. In addition, we were surprised to find that biofilm-forming strains are no more likely to carry plasmids than strains that don’t form biofilms, even though gene mobilization has been proven to occur within biofilms and, indeed, conjugation, a powerful gene-mobilizing mechanism, enables the formation of biofilms in E. coli.

New Examples of Resistance

One of the most disturbing new cases of resistance was the recent isolation of a strain of Staphylococcus aureus, also known as “golden staph” for the color of its colonies, that is resistant to vancomycin. This drug is regarded as the final weapon against infections caused by enterococci and staphylococci, two groups of organisms that often cause infections in hospitalized patients. Vancomycin-resistant enterococci are already a public health problem, but until very recently the drug was always effective against the golden staph. Then in July 2002, the Centers for Disease Control and Prevention confirmed the isolation of the first vancomycin-resistant strain of this dangerous germ from a Michigan man undergoing chronic renal dialysis. Some staph strains were already known to have reduced susceptibility to vancomycin, but, reassuringly, experimental attempts to introduce vancomycin— resistance genes into Staphylococcus aureus had failed because the genes became unstable and were quickly lost. However, it seems that after continuous exposure some germs can keep such genes. It is likely that under the enormous pressure applied by the reckless use of antibiotics, the dangerous new strain gained not only the resistance trait but also the capability to somehow retain it.

An interesting sidelight to this dire situation is how vancomycin-resistance genes evolve. A bacterium needs several genes to circumvent the effect of vancomycin; the genes are all arranged together and are activated simultaneously, in what is known as an operon. But genes in the vancomycin-resistance operon seem to come from different sources, as revealed in DNA sequence analysis by Patrice Courvalin of the Institut Pasteur. Enterococci and staphylococci seem to have a particular ability to shuffle and reassemble DNA sequences.

Other antibiotic families are also in jeopardy. Fluoroquinolones are a family of drugs that have been intensely exploited during the last 15 years. One of the first drugs in the group was the now-famous ciprofloxacin, somehow erroneously supposed the only drug effective against the anthrax bacterium Bacillus anthracis during the recent terrorist mailings of that germ. Newer generations of fluoroquinolones are now available and are being used against bacteria, such as Streptococcus pneumoniae, that cause respiratory infections and are increasingly resistant to penicillin and erythromycin. One interesting feature of fluoroquinolones is that resistance to them results mainly from mutations in genes encoding enzymes that the antibiotics target; these genes are difficult to transfer horizontally. But in 2002, John Tran and George Jacoby of Lahey Clinic in Burlington, Massachusetts described the first fluoroquinolone-resistant mechanism caused by a plasmid in gram-negative bacteria such as E. coli. A protein encoded by the plasmid they discovered can protect the target enzyme from the action of the drugs. It is still early to know how common this mechanism is among fluoroquinolone-resistant bacteria and how successful it can be; however, it seems a new path for the horizontal transmission of genes enabling bacterial survival.

Carbapenems, mainly imipenem and meropenem, are two very wide— spectrum antibiotics, used only in hospitals to treat serious infections. Until recently, these drugs were invaluable against infections caused by multi-resistant bacteria, such as opportunistic pathogens, which are often around but typically only seize the chance to cause infection when a patient has a wound, an invasive medical procedure or a weakened immune system; examples of such bacteria are Pseudomonas aeruginosa and Klebsiella pnemoniae. But now isolates of these species are becoming resistant at alarming rates. Some strains have an old, rare enzyme capable of destroying the antibiotic molecule. An increasing number of strains have enzymes that in earlier forms inactivated other, related drugs but have evolved to quell antibiotics that are presently the last hope against some serious hospital infections.

To understand the evolution of resistance, it is worth considering some conceptual proposals. Jack Heinemann of the University of Canterbury in New Zealand, an expert in horizontal gene transfer, proposes that the evolution of resistance strongly depends on the biology of bacterial plasmids. Two decades ago Ken Gerdes, at the Technical University of Denmark, discovered an intricate system that prevents the survival of bacterial cells that spontaneously lose their plasmids. A socalled post-segregational killing (psk) system consists of two genes in the plasmid, one encoding a toxic protein and the other an RNA that prevents the expression of that protein. While the plasmid is in the cell, both genes are transcribed, but the toxic protein is never produced. But when the plasmid is lost, the messenger RNA that encodes the toxic protein outlives the inhibitor (antisense) RNA, so the protein is produced and the cell dies. This system ensures the survival of a plasmid within its bacterial host, since any cells that lose the plasmid will quickly be culled from the population. Heinemann proposes that antibiotics act as an external toxin in an analogous system, where resistance genes provide the antidote. If the plasmid is lost, the antibiotic kills the cell; antibiotics are therefore driving the evolution and spread of plasmids.

It is important to remember that most antibiotics are in no way new to bacteria; Julian Davies at the University of British Columbia even proposes that antibiotics could be some of the oldest biomolecules. Among clinical isolates stored from before the antibiotic era, almost none are resistant to antibiotics. But based on sequence analysis, Miriam Barlow and Barry G. Hall of the University of Rochester have proposed that genes for enzymes that inactivate beta-lactam antibiotics such as penicillin (which inhibit bacterial cell-wall synthesis), or at least enzymes very closely related to them, have been in plasmids for millions of years. It therefore appears likely that interactions between plasmids and antibiotics are much older than humankind itself and that the only thing we did when releasing massive amounts of antibiotics was to make such interactions a much more common phenomenon.

In the field of antibiotic resistance, one deeply disturbing issue concerns how resistance relates to the ability of bacteria to cause disease-that is, their virulence. It’s frightening to think that more resistant bacteria might also be particularly virulent, but we may be creating just such germs. Plasmids containing both resistance and virulence genes have been described since the 1970s. Microbiologists are now finding virulence and resistance genes in other, smaller kinds of mobile genetic elements, such as integrons and gene cassettes, which can rearrange to create dangerous combinations. When these genes are linked, abusing antibiotics can select not only resistant bacteria but also more virulent ones.

It is also becoming apparent that virulence and resistance can often be the outcomes of the same mechanism. Pumps that expel noxious compounds from bacterial cells can detoxify the cells from bile salts, allowing them to survive in the intestinal tract as well as resist antibiotics. Mechanisms that protect bacteria from free-radical molecules generated by immune cells can also protect them from antibacterial drugs. From the epidemiological point of view, a more resistant bacterium will be a more successful invader and will have more chances to spread through contagion, since patients will remain sick for longer periods of time. In this way, resistant bacteria are not only more difficult to control, but also more harmful.

New and Semi-New Antibiotics

Pharmaceutical companies are now recovering from nearly 20 years of not looking for new antibacterial compounds. Somehow, during the 1970s and 1980s, the notion that infectious diseases were defeated even diffused into research and development teams. Also, some proposed that the more promising business for the pharmaceutical industry lay in the treatment of chronic diseases, which require continuous medication, rather than infections, which are quickly resolved. In any case, until recently, the only “new” antibiotics were slightly modified forms of old chemicals. Such tinkering gave rise to the third and fourth generations of cephalosporins, originally discovered in the 1950s; the second generation of aminoglycosides, originating from the 1940s; and newer versions of macrolides discovered 50 years ago. The last family developed in the first wave of antibiotic discovery were quinolones, originally patented in the 1960s; their improved derivatives, the fluoroquinolones first came on the market in the early 1980s. This strategy of improving old drugs is still generating many “semi-new” antibiotics (Figure 4). Fortunately, some entirely new families of antibiotics are now also reaching the market.

Among the truly new drugs are oxazolidinones, such as Linezolid and everninomycins. These drugs are aimed at infections caused by grampositive pathogens, such as pneumococci, enterococci and the golden staph and its relatives. Many of these organisms cause infections in hospitalized patients. These paradoxical sorts of diseases, which a patient acquires while being treated for another illness, can be very severe. Drugs against these lifethreatening infections are certainly needed. But infections caused by gut bacteria (most of them gram-negative) are not receiving adequate attention, and resistance among this group is growing continuously.

Drugs for the treatment of tuberculosis are also certainly needed. For many years, almost no effort was made to find new anti-TB drugs for what was considered a vanishing disease, at least in developed countries. But after an outbreak in New York City in the early 1990s, where nearly 10 percent of the cases proved resistant to two or more drugs, TB regained the attention of the public and, one hopes, of pharmaceutical companies. Mycobacterium tuberculosis happens to be a particularly tough germ; it is covered by several layers of cell wall, including one similar to wax, which prevent many drugs from penetrating into the cell. It is further protected from many antibiotics because it grows within immune system cells, and it grows slowly, so treatments need to be maintained for long periods-up to six months on average. Also, it has marketing disadvantages; although TB kills 2 million people a year worldwide, it is mainly a disease of the poor. Until recently, however, the four or five anti-TB agents were still useful, especially when used in combinations of two or three. But as multi-resistance rises, we are running out of options. Although a recent $25 million donation from the Bill and Melinda Gates Foundation and other funds are becoming available for tuberculosis research, the money available pales in comparison to the $350 million typically expended by pharmaceutical companies to develop each new antibiotic.

Toward Rational Use of Antibiotics

Since antibiotic abuse has caused rising antibiotic resistance, the prudent use of antibiotics as an antidote to this trend has been gaining more and more attention. An estimated 90 to 180 million kilograms of antibiotics are used yearly, according to Richard Wise of the City Hospital NHS Trust at Birmingham, U.K. Considering that such an amount would provide roughly 25 billion full treatment courses-four per year for every human being-it seems necessary to cut this number. Some important battles have been won in the fight for rational use. Antibiotic sales are diminishing worldwide although the sale of all other kinds of drugs is increasing. The efforts of the Alliance for the Prudent Use of Antibiotics (APUA) headed by Stuart Levy of Tufts University has won the attention of U.S. lawmakers, and the agricultural use of antibiotics is now facing more obstacles.

Antibiotics have been added to the food of farm animals for decades as “growth promoters.” Some actual growth promotion along with reduced frequencies of infections among animals yield a small profit for farmers, and huge ones for pharmaceutical companies. (About 10 times more antibiotics are used in the United States for agriculture than to treat human infections, according to the Food and Drug Administration.) But it has been well documented that the use of antibiotics leads to resistance among bacteria in animals, and these resistant germs can be transmitted to humans through foodstuff. Therefore, it would be a significant victory to have antibiotics removed from animals’ food. The Preservation of Antibiotics for Human Treatment Act of 2002 was introduced to Congress, and the FDA issued a resolution aiming to limit the use of antibiotics as routine additives to animal feed and water. Also, major chicken producers are significantly reducing the use of antibiotics in healthy animals, as large fast-food companies recently have said they will not buy chickens fed medically important antibiotics. In this regard, the United States is behind Europe; last year the European Court of First Instance upheld a 1998 decision by the EU Council of Ministers to ban the use of several antibiotics in animal feed.

These encouraging advances have been made despite the gigantic disproportion between the few funds and people devoted to the rational use of antibiotics versus the resources of pharmaceutical companies dedicated to selling more drugs. This is not meant to attack the pharmaceutical industry: They develop new antibiotics at great cost and must sell those antibiotics as much as possible to profit from them. They also face a deadline-the lifespan of patent coverage before others can manufacture and sell the compound without investing in research. Hence, multi-million dollar efforts are conducted to push for clinical and nonclinical use of antibiotics, often resulting in abuse.

Efforts aimed at promoting the rational use of antibiotics rely mostly on academic activities, such as conferences and publications; these have to compete with expensive advertising and other aggressive strategies used to push physicians into prescribing all sort of antibiotics. For a global usage— reduction strategy to work, it is vital that pharmaceutical companies get involved. Researchers must provide evidence that the fast emergence of resistance is not good for business. Some have suggested that extending the lifespan of antibiotic patents will allow pharmaceutical companies more time to sell their products, diminishing the pressure to promote prescriptions. In any case, if rational-use advocates fail to consider the interest of the pharmaceutical industry, they will always have to fight uphill.

In many developing countries there is no need for a physician’s prescription to buy antibiotics. In Mexico City, for instance, around 30 percent of all antibiotics sold at drugstores are sold without a prescription. But banning such sales is not an uncomplicated good; a significant fraction of the population lacks medical services, and self-prescription may be the only way these people can have access to drugs. It is difficult to assess how large the impact of this self-prescription is on increased antibiotic resistance and to weigh that risk against the risk from limiting access to antibiotics among people who lack formal medical care. Also, self-prescription is practiced in developed countries; a recent report from Brandon Goff and coworkers at the Pentagon Clinic found that soldiers often buy and use antibiotics from the fish medication aisle of pet stores, which freely sell many antibiotics, such as erythromycin, kanamycin, penicillin, ampicillin, tetracycline, a variety of sulfonamides, nitrofurazone and metronidazole.

Is Rational Use the Solution?

An enormous amount of evidence shows that the human use of antibiotics created the selective pressure that led to the emergence and early spread of resistance among bacteria. However, it is also clear that once resistance is established in a bacterial population, it won’t disappear easily. This came as a surprise to many researchers, because they believed resistance genes always represent a cost to those bacteria bearing them, one that is too high when antibiotics are not present. However, either the burden of carrying resistance traits is not that high or resistance genes are being maintained through other, essentially unknown mechanisms. In now -classic experiments, Judith E. Bouma and Richard E. Lenski at the University of California, Irvine showed that, in the absence of antibiotics, resistance plasmids and their bacterial hosts co-evolve in such a way that, after several generations, they grow better than a strain that lacks the plasmid or a strain with a new association between plasmid and host. Dan Andersson and his team at the Swedish Institute for Infectious Disease Control and Uppsala University recently showed that additional mutations can compensate for any cost on bacterial fitness imposed by the mutations conferring resistance, without compromising the resistance.

In addition, antibiotic resistance genes are often physically linked to genes encoding other useful traits, as when several different genes are carried on the same plasmid; the selective pressure that favors one of the traits cross-selects for others close to it. Resistance to disinfectants and other toxic compounds is often linked to antibiotic-resistance genes; therefore, such compounds can cross-select for antibiotic-resistant bacteria. Also, some genes provide protection not only from antibiotics but also from other kinds of environmental stress. For example, even an air pollutant like ozone might select for antibiotic-resistant bacteria, as Gabriela Jimenez-Arribas and Veronica Leautaud showed in our lab in collaborative work with Bruce Demple of Harvard.

Appreciating how hard it is to lose resistance is an important part of realizing the real reach of programs that encourage the rational use of antibiotics. Of course, we must move towards rational antibiotic use to prevent the emergence of more resistance genes and more resistant organisms. But rational use will not do away with resistant strains. My lab explored an interesting example of this paradox. The commercial blockade on Cuba, along with the fall of socialism in Eastern Europe, has made antibiotics, particularly newer ones, a very scarce commodity on the island. The use of the drugs has been strictly controlled since 1990. Cuban scientists report that antibiotic resistance among disease-causing bacteria is receding. But my colleagues Javier Diaz-Mejia and Alejandro Carbajal-Saucedo found that the degree of resistance among the harmless bacteria in the mouths of Cubans is about equal to that found in the mouths of Mexicans, even though antibiotics are sold without prescription in Mexico and are also used in agriculture. Since benign bacteria often act as reservoirs of resistance genes that can be transferred to virulent bacteria, it was surprising to find such high frequencies of resistance after 10 years of severe usage reduction in Cuba. This, along with a wealth of other reports, indicates that resistance seldom disappears and that what we can expect from rational-use policies will be the slower emergence of new resistance mechanisms, but not a reversal of the trend nor a solution to the resistance problem. Of course, rational use of old and new drugs is vital, but it must be regarded as just one part of a larger strategy.

The Search for New Drugs

Until recently, the search for new antibiotics was performed using 40-yearold strategies: Compounds from several sources-from crude extracts of plants, animals and bacteria to synthetic molecules-were tested for the ability to inhibit the growth of selected microorganisms. Those compounds that showed some potential were further analyzed to isolate the active component and to look into its stability and toxicity. The few promising molecules or their slightly modified forms were then brought into pre-clinical and then clinical trials. The mechanism of action and other important details of their interactions with bacteria and humans have often been discovered only after the drugs come into use. This strategy, undoubtedly successful in many ways, missed important candidates that were unstable or toxic, or just did not penetrate to their targets in bacterial cells. Such problems could have been fixed with chemical manipulations. Also, only compounds that killed or completely inhibited bacteria growth were detected; this screening would miss those compounds able to limit the ability of virulent bacteria to harm us without damaging the bacteria.

During the late 1980s and early 1990s, a new approach to drug discovery became available-chemical design. Databases filled with the chemical structure of known drugs and their biological activities could be used to design new candidates. Such designs were synthesized in chemical labs and then tested for the expected biological activity. Although this would seem a more directed search, trying to incorporate additional data, such as the pharmacological behavior and toxicity of known compounds, made this a formidable task.

Now, as the full genome sequences of a growing number of pathogenic bacteria are becoming available, another process is starting to become possible. DNA and protein sequences can reveal potential drug targets in virulent bacteria. These targets need to be essential for bacterial survival or virulence and, simultaneously, absent in the genomes of humans and other mammals. New computational methods are also helping to infer the function of proteins solely from the genome sequence that codes for them. Selected targets need to be validated in the lab, and then an easy, controlled assay must be developed to screen for compounds that inhibit the protein’s specific activity. Candidate compounds can be modified to improve their penetration and stability, or to diminish their toxicity, if any The search can be directed toward wide-spectrum agents by finding targets shared by a large number of bacterial species or toward narrow-spectrum ones by looking into targets found only in a few organisms. Advances in miniaturization and robotics allow thousands of compounds to be screened in a few days or hours using minute quantities of experimental compounds. Meanwhile, combinatorial chemistry can provide a larger number of candidate drugs much faster. New antibacterial agents that arose partially from these strategies are on their way to clinical trials; inhibitors of peptide deformylase enzymes, essential to life for most bacteria but absent or not vital to humans, are being studied by Versicolor and British Biotech; and ACP-reductase, involved in the synthesis of bacterial fatty acids, is an attractive target to GlaxoSmithKline.

But perhaps the very way we fight infection should be reconsidered. As in other aspects of our social behavior, we identify sometimes-annoying creatures as mortal enemies and are determined to annihilate them. Even when these efforts prove futile, we insist on the approach, as in the fad of including disinfectants in numerous household products. The abuse of antibiotics is another instance of this ill-conceived strategy. A more promising avenue may be the idea of inhibiting virulence instead of killing germs outright. This is an old idea, but new technologies such as microarray profiling, are making this a more feasible goal. Back in the 1970s, compounds that could inhibit the adhesion of virulent bacteria to a tissue, believed to be the first step in infection, yielded inconsistent results. Now, new technologies can detect specific genes that are switched on during infection and which are essential for a successful infection. The products of these genes are ideal targets for new anti-infection drugs. These drugs might have important advantages compared to current antibiotics. Since they do not kill bacteria, selection of resistant strains could be much slower. Also, they will only affect virulent bacteria, diminishing the risk of selecting resistance among normal flora, which, in turn, can transfer those genes to pathogenic germs later.

Despite the promise of this approach, assessing which genes are suitable targets and even designing the clinical tests for their efficacy are proving to be enormous challenges. Virulence genes are often switched off in ordinary laboratory conditions. Furthermore, inhibiting virulence might be a powerful way to prevent an infection, but not to get rid of an already established pathogen. For these strategies to work, a number of problems must be solved. Testing of the efficacy of such non-lethal drugs will require entirely new techniques, since it won’t be possible to simply see if bacteria grow or not. Drugs that inhibit only the virulence trait will require specific tests that are more difficult to set up and more expensive. Also, since virulence inhibitors are likely to have a narrow spectrum, affecting only a few related organisms, it would be necessary to improve methods of ascertaining the types of infection in a patient, a task that now often takes a couple of days.

Genome sequencing efforts can also fuel the search for vaccines. Preventing infectious disease is often much better than trying to fight an established illness. The sequencing of the immense genome of Neisseria meningitidis, which causes meningitis, in 2000 allowed the identification of candidates for a vaccine against this organism. Although a vaccine has yet to be developed from such efforts, the approach seems very promising where others have failed.

My colleagues and I have taken another approach-searching for compounds that circumvent resistance. This is hardly a new approach; the compounds clavulanic acid and sulbactam have been successfully used for some years to inhibit the enzymes that resistant bacteria use to destroy some antibiotics. One idea for fighting antibiotic resistance has been to target the plasmids that contain the resistance genes. Based on old reports that ascorbic acid (vitamin C) can suppress the replication of the bacterial viruses known as bacteriophages and knowing that the DNA in bacteriophages is similar to plasmid DNA, we tested the anti-plasmid activity of ascorbic acid—with good results. Looking into compounds that have similar activities, we have found a handful of interesting drug candidates. This approach may only be useful for a few plasmid-bacteria combinations, but even a narrow—spectrum drug could prove valuable when treating critically ill patients. Also, as some virulence determinants exclusively reside on plasmids, affecting their expression or stability could be beneficial. For instance, the main difference between dangerous Bacillus anthracis, the causative agent of anthrax, and the almost-innocuous Bacillus cereus, a soil bacterium, is a virulence plasmid. This plasmid could make an extraordinary target for research.

Education of Health-Care Workers

Along with regulating antibiotic use and searching for new drugs and even new strategies to fight infections, it is necessary to improve the education of health-care personnel. As much as half of medically prescribed antibiotics are unnecessary, a remnant of the deeply wrong notion that antibiotics are drugs that, if not always beneficial, are at least not harmful. Perhaps on the level of a single patient, this view can still hold, since most antibiotics have few side effects, but from the public health perspective antibiotic abuse is extremely dangerous and also expensive, since the consequences of antibiotic resistance are estimated to cost $4 billion to $5 billion annually in the United States, according to the FDA. Many physicians prescribe antibiotics with a justin-case philosophy, for example doling out “preventive” antibiotics for travelers’ diarrhea. In many hospitals, thirdgeneration cephalosporins are predominantly used in emergency rooms when the existence of infections has not even been established. Many such interventions are of dubious efficacy and certainly pose an additional pressure favoring resistant bacteria.

Physicians and medical students do not often fully realize the intricacies of evolution and natural selection, let alone the powerful mechanisms that bacteria have to face aggression and stress. This is completely understandable, since doctors already have to deal with hundreds of organs, diseases, drugs and procedures. But in order to sway medical personnel toward the rational use of antibiotics, it is imperative to teach them the basics of resistance from an evolutionary point of view. Ideally, such education should be as appealing as the pharmaceutical ads used to push doctors towards antibiotic prescription. My colleague Isabel NivonBolan designed a card game that explains dynamically how resistance genes are acquired or activated and how antibiotics and other agents can favor resistant organisms. This game does not require previous knowledge of bacteria, infections or antibiotics and is actually fun as well as illuminating, as we’ve heard from medical and biology students. Approaches like this can transform the attitude of both lay people and physicians toward antibiotics.

We must assume that the war against bacteria, as it was conceived during the antibiotic era, is already lost. We must necessarily move into a post-antibiotic era. As we do so, we should adjust our attitude against these much older and much more abundant organisms that share the planet with us. We know now how tough bacteria are as enemies, but we have new data on the molecular mechanisms we can use to tame them. Let’s hope we all-patients, physicians, researchers and pharmaceutical companies-quickly learn the lessons of the lost battle.