Moving Targets: Fighting Resistancein Infections, Cancers, PestsExperiences dealing with resistance in one area of biology can provideinsights in others

Shannon E. Greene and Ann Reid

Antibiotics, antivirals, herbicides, insecticides,and anticancer drugs might seem to have little incommon. However, each serves a similar pur-pose.Whether to cure a disease, control an insectthat spreads disease, or reduce crop losses, theyweaken or eliminate a living entity that is harm-ing humans or their enterprises. The deploy-ment of antibiotics, anticancer drugs, insecti-cides, and other selective chemical agents hashad enormous health and economic benefıts; notfor nothing were early antibiotics called “miracledrugs.”

However, in every case, these “miracle” treat-ments come up against a hard truth. Given timeand opportunity, the organisms we seek to con-trol will develop resistance to the agents deployedagainst them. Resistance is not a new phenome-non, and it did not arise solely because of humaninterventions. Antibiotic resistance in bacteria,for instance, arose as a means for competing spe-cies to dominate one another within ecologicalniches. Mechanisms of resistance in nature rangebroadly and include the degradation of toxiccompounds, expulsion of antibiotics or otherdrugs from cells, or changes to the drug’s cellulartarget sites such that the target can continue tofunction despite the presence of the antibiotics orother drugs.

Each of these cases depends on the same fun-damental process—cells adapting to survive con-tact with an otherwise deadly compound. Theprocess is at work when bacterial pathogensevolve resistance to antibiotics and viruses to an-tiviral drugs, when insects become resistant toinsecticides, and when cancer cells stop respond-ing to anticancer agents. Although the underly-ing phenomenonmay be the same, scientists whodevelop and deploy this broad array of agents donot see themselves as belonging to a single scien-

tifıc community and thus rarely get a chance tolearn from one another.

How Does Resistance Evolve and Spread?

The emergence of resistance provides a straight-forward illustration of evolutionary selectivepressure at work.When a population is subject toa treatment designed to eliminate it, any individ-uals possessing a trait that mitigates the negativeeffects are able to proliferate and eventually be-comedominant.At that point, the treatment is nolonger effective at controlling the target popula-tion.

The odds of a resistance trait preexisting in apopulation depends heavily on genetic diversity –the greater the diversity, the greater the likeli-hood that a variantwith somedegree of resistancealready exists in the population. If resistance isnot already present, the odds that it will evolvedepend on the genetic “flexibility” of that popu-lation. A population’s capacity for generating ge-netic diversity depends on its size, generationtime, and mutation rate. Population size posi-

SUMMARY

➤ Given time and opportunity, the organisms that antibiotics or pesticideshelp to control, or the cancers that anticancer treatments target, willdevelop resistance to the agents that are deployed against them.

➤ Because the molecular mechanisms by which resistance can arise areshared among very different biological species, strategies for counteractingthose resistance mechanisms might also have features in common.

➤ While it may be impossible to prevent resistance, some treatment strategiesmight postpone the inevitable because they prove more difficult for thetarget organisms to “evolve around.”

➤ Taking a “management” instead of an “eradication” approach may provevaluable for physicians dealing with some kinds of cancers and infectiousdiseases.

FEATURE ARTICLE

Microbe—Volume 8, Number 7, 2013 • 279

tively correlates with diversity and, therefore, theprobability of resistance emerging.

Infectious agents, cancer cells, insects, andweeds display a wide range of generation timesand mutation rates (Fig. 1). Some bacterial spe-

cies, such as Staphylococcus aureus, will undergocell division every 30 to 60 minutes, whereasmosquitos have a two-week lifespan. Organismswith shorter generation times can develop resis-tancemore quickly than those with longer gener-

FIGURE 1

The emergence and spread of resistance varies across biological systems and depends on key traits of thosesystems. (A) The speed at which resistance evolves positively correlates with mutation rate and rate ofreproduction. A pathogen or pest with a high mutation rate that reproduces quickly, such as HIV or E. coli, willevolve resistance on a much swifter time scale than pigweed, for example, which mutates about three orders ofmagnitude less frequently and also only reproduces annually. (B) However, once resistance has evolved in apopulation, its spread is influenced by the number of progeny produced and how far those progeny can disperse,either in terms of distance or individuals infected. A drug-resistant cancer cannot spread beyond its immediatehuman host. While pigweed may develop resistance more slowly than HIV, the ability of its thousands of seeds todisperse distances of hundreds to thousands of miles makes resistance spread a real threat. Increased globalizationplays a significant role in the spread of resistant pathogens; malaria-laden mosquitos may only travel a distanceof 100m from where they hatched, but infected humans can bring resistant strains to naïve populations. Indeed,in cases of multi-drug resistant tuberculosis, quarantine is the most viable option for preventing resistance spread.

FEATURE ARTICLE

280 • Microbe—Volume 8, Number 7, 2013

ation times. Viruses such as HIV and influenzanot only replicate quickly, but their enzymes re-sponsible for replicating their respective ge-nomes are notoriously error prone, leading tovery high mutation rates, on the order of 10�5,which effectively translates to about one muta-tion per genome replication. Many viruses alsogenerate sequence diversity through recombina-tion or reassortment during replication. Muta-tion rates among bacteria are orders of magni-tude lower, and in eukaryotes that rate dropslower still, ranging from 10�11 to 10�20. More-over, mutation is not the only source of geneticvariability. Bacteria can share entire sets of resis-tance genes through horizontal gene transfer.Additionally, sexual reproduction in fungi, in-sects, and plants provides another means for ge-netic variation.

The molecular mechanisms by which resis-tance can arise are often shared among very dif-ferent biological species. Resistance can occurthrough point mutations: discrete substitutions,deletions, or insertions in the genetic code. Insome instances, a single point mutation can en-gender resistance. In other cases, several muta-tions may be needed. For instance, Plasmodiumfalciparum, one of several parasites responsiblefor causing malaria, acquired seven mutationsbefore it became resistant to the drug chloro-quine.

Gene amplifıcation is another resistancemechanism, one that is associated primarily withcancer, but also accounts for pesticide resistanceamong insects and drug resistance among fungalpathogens. Through this mechanism, the organ-ism or cells increase the overall copy number ofthe target gene such that suffıcient amounts of theencoded protein are available to complete thebiological task, despite some of that protein beinghindered by the drug or pesticide. Genes encod-ing detoxifying enzymesmay also be amplifıed asa way of conferring resistance to drugs or pesti-cides. Amplifıed genes can be found scatteredthroughout genomes, and little is known aboutthe mechanism underlying this mode of resis-tance. Further, it is not knownwhether particularorganisms are better able to amplify genetic ma-terial than others, or whether amplifıcation isspecifıc to particular genomic regions or features.

Pests, pathogens, and cancer cells can adapt tootherwise toxic compounds by metabolizingthem or by pumping toxins from cells throughefflux channels. Drug sequestration is another

effective means to achieve resistance. For exam-ple, some cancers develop resistance to the anti-angiogenic tyrosine kinase inhibitor sunitinib byaccumulating the drug inside lysosomes.

Additionally, target organisms can physicallyavoid the drug or pesticide. Viruses and bacteriacan enter latent phases in which they do not rep-licate, effectively hiding from immune system re-sponses and therapeutic drugs. Many types ofbacteria can grow in dense communities calledbiofılms, which sometimes contain multiple spe-cies, and the interior environments of these bio-fılms are exposed to reduced concentrations ofantibiotics. Similarly, cancerous cells within atumor respond differently to drugs or othertreatments than do their counterparts alongthe boundaries. Understanding three-dimen-sional tumor “ecology” will be an importantdevelopment in oncology and in the explorationof drug resistance in cancer.

Although the ways in which pests and patho-gens develop resistance to treatment vary widely,signifıcant overlaps and similarities are foundacross biological systems, including plants, can-cers in humans and other mammals, insects, andbacteria and viruses. Because such differentbiological systems experience similar types of re-sistance at themolecular and cellular levels, strat-egies for counteracting those resistance mecha-nisms might also have features in common.

Designing Treatments withResistance in Mind

Evolutionary principles were too often neglectedwhen scientists sought to overwhelm biologywith chemistry in an onslaught of new drugs andpesticides. Thus, their failure to anticipate thatorganisms would develop resistance to antibioticor anticancer treatments as well as to pesticidesled to other, longer-term consequences. While itmay be impossible to prevent resistance, sometreatment strategies might postpone the inevita-ble because they prove more diffıcult for the tar-get organisms to “evolve around.”

One important bit of advice is to make drugsor other agents that are specifıc. Thus, drawingupon structural, enzymatic, and sequence infor-mation about particular targets, it makes sense todesign chemical agents to disrupt specifıc andessential features. Resistance seemsmore likely toarise if drugs interact with flexible targets. Forinstance, several triazine herbicides target the D1

FEATURE ARTICLE

Microbe—Volume 8, Number 7, 2013 • 281

protein of photosystem II (PSII). However, anumber of different, relatively minor changes inD1 render this protein resistant to those herbi-cides.

In contrast, a newer anticancer agent, iman-tinib, which is being used to treat chronic my-eloid leukemia, targets the bcr-abl fusion proteingenerated from the specifıc chromosomal trans-location that instigates this disease. Although re-sistance eventually develops, it is almost alwaysby the same, predictable mutation, whichmade itpossible to develop a specifıc secondary drug thatcan be administered as soon as resistance is de-tected.

Another important piece of advice is to choosethe right target. For example, the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D), whichmim-ics the essential plant signaling molecule auxin,has been usedwidely formore than 70 years. Thissynthetic chemical agent works so well because itinterferes with the extensive network of bindingpartners of auxin. If plants develop resistance to2,4-D, that resistance tends to disrupt nativeauxin interactions and therefore confers a steepfıtness cost. Hence, resistance to 2,4-D proves tobe relatively rare.

Further, if possible, use more than one treat-ment simultaneously. In 1987, initial efforts totreat HIV proved successful in the short term,but the virus quickly developed resistance tothe frontline antiretroviral drug AZT. After theemergence of AZT resistance in a patient, doctorswould introduce a second-line therapy, whichwould continue until secondary resistance evolved.HIV infection management underwent a para-digm shift when reports surfaced of patients pre-scribed just one antiretroviral drug, but takinganother on the side. Indeed, these individualsexhibited undetectable viral loads for much lon-ger than patients treated with a single antiretro-viral agent alone. Despite stories of promisingHIV treatment surfacing in the mid-1980s, insti-tutional barriers prevented combination therapyfrom rapid adoption. Persuading regulatory offı-cials to accept combination therapy requireddemonstrating that each component addedbenefıt. To demonstrate benefıt, “benefıt” had tobe redefıned—in this case, the acceptance by theFDA of viral load as an indicator of disease statusin the 1990s. Now, HIV-infected individuals areroutinely treated with at least three drugs to re-duce viral load andprevent resistance fromdevel-oping to the drugs in that mix.

Despite increasing evidence of the effıcacy ofcombination drug therapies, the practice hasnot been implemented widely beyond patientsbeing treated for HIV or tuberculosis infection.For example, patients with cancers still tend to betreated sequentially with single anticancer drugs,and this practice tends to lead to sequential resis-tance evolution.

However, in a recent clinical trial, patientswith small-cell lung cancer were treated simulta-neously with a vaccine targeting p53 and a third-line chemotherapeutic agent. Neither therapywas particularly effective on its own, but patientsurvival was greatly increased when the twoagents were administered together. This en-hancement comes from placing the cancer in an“evolutionary trap.” In response to the vaccinetargeting p53, the tumors down-regulate p53 ex-pression, but p53 contributes to cell survival inthe face of toxic perturbations. Thus, response toone therapy resulted in sensitivity to a secondtherapy. Despite the success of this particular du-al-treatment strategy, the expense in runningsuch clinical trials along with the challenge offınding multiple targets that are unique to cancercells are formidable barriers to investigating suchcombination therapies across the broad range ofcancer types.

Yet another piece of advice for those designingnovel drug treatments is to enlist ecological prin-ciples. Perhaps no approach better highlights thevalue of using evolutionary principles to designand implement a novel pest control strategy thanthe development of resistancemanagement strat-egies for transgenic crops producingBacillus thu-ringiensis (Bt) toxins (Fig. 2). Bt proteins, whichare not harmful to humans and most other non-target organisms, kill many key insect pests. Avariety of crop plants, including corn, soy, andcotton, are genetically engineered to produce Bttoxins, which are ingested by insect pests that tryto feed on the plants, thus reducing the need forother types of insecticides.

Recognizing that target insects would likelydevelop resistance to such Bt crops, scientistsdrew on evolutionary principles to design a strat-egy for managing—and delaying—that resis-tance. The key element of thatmanagement strat-egy is to set aside “refuges” containing host plantsthat do not produce Bt toxins and in which Bt-susceptible insect pests can continue to grow. Therare mutant insect adults that emerge from andare resistant to Bt crops then are free tomate with

FEATURE ARTICLE

282 • Microbe—Volume 8, Number 7, 2013

the abundant Bt-susceptible adults that are feed-ing in the nearby refuges. Because resistance is arecessive trait, the progeny of matings betweenresistant and susceptible adults remain suscepti-

ble to Bt andwill continue to die when feeding onBt crop plants, delaying substantially the emer-gence of full Bt resistance.

Finally, in terms of advice about any new treat-

FIGURE 2

In order to protect agricultural interests from insect pests, key crops such as corn and cotton have been geneticallyengineered to express Bt toxins that target the digestive tract of those pests. (C) Without the Bt toxin, over timethe insects would quickly reproduce and establish widespread and costly crop damage. (A) However, if the entirefield is planted with Bt plants, enormous selective pressure on the insects will favor increased resistance.* If anyresistant individuals are present, they will withstand the toxin and reproduce, eventually enacting crop damage onpar with untreated fields. Because the genetic mutations conferring resistance are often recessive, resistantindividuals need to mate with one another or heterozygous individuals to produce resistant progeny. (B) If awild-type population can be maintained nearby with which the resistant insects can mate, the selective pressureis reduced and the number of resistant insects increases at a slower rate. This can be accomplished by the plantingof refuges, or fields of non-Bt crops fed upon by target pests, near the engineered crops. While a few resistantinsects may remain, the crop losses will ultimately be minimal compared to the alternative scenarios and the spreadof resistance will be reduced. The refuge strategy works best with recessive resistance, as illustrated here.

FEATURE ARTICLE

Microbe—Volume 8, Number 7, 2013 • 283

ment, use it wisely. Improper use can drive thedevelopment of resistance. For instance, antibiot-ics are very widely and carelessly used in China,and levels of resistance there are among the high-est ever measured. In 2001, 89% of hospital-ac-quired infections involving S. aureus in Chinawere resistant to standard antibiotics comparedwith 16% in the United States.

Soon after antibiotics came into wide use astherapeutic agents in the United States and otherdeveloped countries, the drugs were introducedas additives in feedstock for cattle, chickens, andother livestock. This agricultural practice broadlyaffects human health. The wide use of vancomy-cin led to widespread resistance to this importantdrug, and pathogens such asmethicillin-resistantS. aureus (MRSA) are becoming an expandingmedical problem. Both of these clinical chal-lenges can be traced in part to the enormous useof antibiotics as additives to livestock feed. To thisday, there is no coordinated regulation of antibi-otic use, even though the actions of these drugsand paths to the development of resistance areidentical regardless of where they are being used.

While resistance to the insecticides used totreat protective mosquito nettings is not yet ram-pant, this method of malaria control is feared tobe facing imminent failure. Currently all such bednets use similar active ingredients, namely pyre-throid insecticides. These insecticides are alsoused widely in agriculture and as sprays forhomes and other settings, thereby increasing se-lection pressure on mosquitoes and other insectsto develop resistance.

Where To Go from Here

Pathogen, cancer, or pest management strategiesinevitably depend on one’s goal. Drug design andtreatment strategy will be approached differentlyif the sought-after outcome is to eradicate thepathogen, tumor, or pest rather than if one cansettle for maintaining the pathogen, cancer, orpest at a manageable level. If eradication is theonly acceptable goal, the useful lifetime of thedrug or treatment is likely to be reduced becausethe selection pressure will be higher from theoutset.

In treating patients with cancer or infectiousdiseases, the priority is survival, which almostalways equates with eradicating the infectiousagent or the cancer. HIV is a notable exception—cures are sought but long-term management is

widely accepted as a triumph. The agriculturalindustry is farmore likely to adopt amanagementstrategy in dealing with insect pests andweeds. Ingeneral, farmers are willing to accept lower yieldsto extend the lifetime of insecticides and herbi-cides. Conceivably, adopting management ap-proaches for human infectious diseases beyondHIVmay result in better outcomes both in termsof patient lifespan and the slowing of drug resis-tance.

The development of resistance to drugs or pes-ticides is an evolutionary phenomenon. Hence,taking selective pressures into account from theoutset might enable physicians or farmers toavoid or delay some of the pitfalls that can arisewhen those forces are not considered during ini-tial treatment plans. Reconceptualizing cancer asa disease as not necessarily to be cured, but to bemanaged, may help in implementing treatmentsto minimize drug resistance and prolong life.Treatment strategies in medicine well could im-prove by incorporating components from agri-culture and vice versa. Taking a “management”rather than an “eradication” approachmay provevaluable for physicians dealing with some kindsof cancers and infectious diseases. Further gainsmay be had with a paradigm shift away fromsequential drug treatment plans towards combi-nation therapy approaches.

Molecular and genetic analyses are bringing agreater, in-depth understanding of pathogensand cancers, and are leading towardmore specifıcmeans of combating them. Taking into accountthe ecological roles of those species will greatlyenhance the available toolkit. Drug developerscan benefıt from considering how different envi-ronments might be manipulated or exploited tocontrol disease outbreaks or pest infestations. Forexample, applying ecological approaches to thecomplex malaria infection cycle is revealing newpoints at which to intervene, either within thehuman host or when it is being carried by mos-quitoes, which can be targeted at different stagesin their life cycle.

Another ecological factor to consider is thespatial organization of target organisms in spe-cifıc environments. When targets are subjectedto different doses of a drug, for example, somewill be killed while others may be only slowed orunharmed; these differences affect the rate atwhich resistance evolves in the survivors. There-fore, deepening our understanding of how bio-fılm and tumor architecture affects drug dosing

FEATURE ARTICLE

284 • Microbe—Volume 8, Number 7, 2013

within cell communities could lead to better andmore effective means of delivering drugs topathogens in biofılms or cancer cells in tumors.

Evolutionary principles teach us that drug re-sistance is an inevitable challenge to face. Lessonslearned across biology, however, provide prom-ising avenues to slow the evolution of resistance.Including ecological and evolutionary principles

when selecting targets, designing drugs, and im-plementing treatment plans will strengthen ourcapacity to treat infectious dis-eases and cancersand to manage agricultural pests.

Shannon E. Greene is the Colloquium Fellow in the AmericanAcademy of Microbiology and Ann Reid is the Director of theAmerican Academy of Microbiology, Washington, D.C.

Color Atlas ofMedical

BacteriologySECOND EDITION

Authors: Luis M. de la Maza

Marie T. PezzloJanet T. Shigei

Grace L. TanEllena M. Peterson

2013. 366 pages, color illustrations, index.Hardcover ISBN 978-1-55581-475-5

List Price: $169.95ASM Member Price $135.95

“The Color Atlas of Medical Bacteriology is a good reference book that is fairly complete with wonderful images. It would be a welcome addition to the bookshelves of any teaching institution or clinical microbiology laboratory.”

– Clinical Infectious Diseases review of the first edition, written in 2006

A unique visual reference for the diagnostic microbiology laboratory

Conceived by a team of authors with decades of classroom and laboratory experience.

A valuable illustrative supplement for lectures and laboratory presentations, this easy-to-use atlas was written for laboratorians, clinicians, students, and anyone interested in the field of diagnostic medical bacteriology.

estore.asm.org/press

FEATURE ARTICLE

Microbe—Volume 8, Number 7, 2013 • 285