A US government document released inJune 2000 warned of “a growing menace toall people1.” The reference was not to terror-ism or foreign dictators, but to the increas-ing prevalence of antibiotic-resistantmicrobes. Seldom given to tabloid sensa-tionalism, an interagency task force, includ-ing the US Food and Drug Administration(FDA; Rockville, MD, USA), the US Centersfor Disease Control and Prevention (CDC;Atlanta, GA, USA) and the US NationalInstitutes of Health (NIH; Bethesda, MD,USA) concluded that “the world may soonbe faced with previously treatable diseasesthat have again become untreatable, as inthe pre-antibiotic era.” Governments coun-sel better control of antibiotic usage in bothhumans and livestock, while researcherslook ahead for new generations of antibi-otics that can stave off new bugs. But ahandful of companies are looking back intohistory for the answer. They are developingtherapeutics based on bacteriophages, natu-ral viruses that specifically infect—and usu-ally destroy—bacteria (Fig. 1). Whetherthey succeed will depend on their ability toovercome the skepticism of investors, thepreconceptions of the medical professionand regulatory uncertainties concerningthis new class of therapeutics.
Resistance risingAccording to the CDC, methicillin-resistantStaphylococcus aureus (MRSA) accounted fornearly 60% of nosocomial S. aureus infec-tions in 2001—a figure that had nearly dou-bled over the previous decade. Althoughmost MRSA can still be treated with power-ful antibiotics, some superbugs can shrug offeven the strongest drugs. The first reported
case of resistance to Pfizer’s (Groton, CT,USA) Zyvox, the last line of defense againstMRSA, was reported a little more than a yearafter the drug was approved2.
Nor is S. aureus the only problem. TheCDC estimates that in some areas, 30% ofpneumonia caused by Streptococcus pneu-moniae is resistant to penicillin, whereasvirtually all cases were susceptible in the1970s. Vancomycin started failing to keepsome Enterococcus (faecium and faecalis)infections in check in late 1988, necessitat-ing new aggressive combination regimens.
By 1993, according to the NIH, more than10% of hospital-acquired enterococci infec-tions reported to the CDC were attributedto vancomycin-resistant Enterococcus(VRE) faecium. Aventis’s (Strasbourg,France) drug Synercid was introduced in1999 as a new weapon against VRE, butsome resistance was observed before it evenreached the market3.
Mother Nature’s little helperA nearly forgotten therapy may yet reemergeas a savior to this accelerating crisis of
Old dogma, new tricks—21st Century phage therapyKarl Thiel
As antibiotic resistant bacteria threaten a public health crisis, biotechnology is turning to bacteriophages, nature’s tiniest viruses.But can phage therapy overcome its historical baggage?
Karl Thiel is a freelance writer based inPortland, Oregon, USA.
NATURE BIOTECHNOLOGY VOLUME 22 NUMBER 1 JANUARY 2004 31
Viral genome
Tail fiber
Cell wall
Nucleoid
Nucleoid
h
1
2
3
4
5
6
Figure 1 Phage life cycle. (1) Phage attaches to a specific host bacterium and (2) injects its DNA, (3)disrupting the bacterial genome and killing the bacterium, and (4) taking over the bacterial DNA andprotein synthesis machinery to make phage parts. (5) The process culminates with the assembly of newphage, and (6) the lysis of the bacterial cell wall to release a hundred new copies of the input phageinto the environment. (Graphic courtesy of GangaGen, Bangalore, India; Phage image courtesy ofElizabeth Cutter, Evergreen State University, Olympia, WA, USA)
antibiotic resistance, one that has its roots inStalin’s Russia, but which flourished brieflyin the West (Fig. 2). For every bacteriumknown on this planet, there are legions ofbacteriophages—tiny viruses that seek outbacteria and use them as a breeding ground,almost invariably destroying their prokary-otic host in the process. In their single-minded mission, phages ignore every cellbut the strain of bacteria they have evolvedto inhabit. That makes them harmless tomammalian cells and harmless even to non-target bacteria—distinguishing them frombroad-spectrum antibiotics, which, whenthey work, can wipe out beneficial flora inthe intestinal tract along with a troublesomeinfection. This specificity also has its draw-
backs, for in order for the therapy to work,the right match between phage and bacterianeeds to be determined (Box 1).
When a lytic phage enters a bacterium, itgenerally chops up the host DNA and uses itto build copies of itself. A single phage canmake hundreds of progeny, which thenburst out of the bacterium and go off insearch of new hosts, all in the span of about30 minutes (Fig. 1). Thus, a very little bit ofphage can become a lot of phage in a shortperiod of time. Bacteria, limited to ho-humcell division, can’t keep up the pace andquickly get outnumbered. Unless they canlurk in inaccessible locations, they aresought out and destroyed.
And unlike antibiotics, bacteriophages
are ‘living’ organisms. They have beeninfecting bacteria since the beginning of lifeon this planet. Bacteria evolve to resistphage, but phage evolve, too—at an amaz-ing rate that chemists tinkering with newgenerations of antibiotics can never hope toreplicate. That would seem to make phagesthe ultimate antibacterial therapy—lethal,adaptive, highly efficient, safe to humans,all with a few billion years of experienceand Mother Nature working on our side.What’s not to like?
Major obstaclesWell, to listen to critics, quite a bit. For onething, the Eastern European example ofphage therapy in practice does not offermuch help to entrepreneurs hoping to navi-gate US and Western European regulatoryagencies.
“They almost always use cocktails,” saysElizabeth Kutter, a microbiologist atEvergreen State College (Olympia, WA,USA) who has observed many phage treat-ments in Tbilisi (Georgia). A typical off-the-shelf phage therapy for purulent infections,she says, might include 30 phage strainsgoing after five different types of bacteria.And that preparation may vary from hospi-tal to hospital. If that isn’t enough to makeheads spin at Western regulatory agencies,doctors in Eastern Europe will sometimesisolate a novel phage specifically for an indi-vidual patient with a particularly toughinfection. It is a treatment paradigm thatchallenges the regulatory status quo incountries outside of Eastern Europe, to saythe least.
Then there’s the thorny issue of intellec-tual property (IP). Because the concept ofusing phages as therapeutics is almost ahundred years old, it is unpatentable.Entrepreneurs can secure protection for anindividual phage strain they have character-ized, but with an estimated 108 strains ofphage in the biosphere, there’s nothing tostop a would-be competitor from finding adifferent but perhaps equally effectivestrain. That sound you just heard was thewallets of venture capitalists slamming shut!
Indeed, for the handful of intrepid com-panies hoping to develop phage therapeu-tics for Western markets (Table 1), findingfunding has been tough. Most have sur-vived primarily on money from founders,their families and friends, or small groupsof angel investors. The most significantventure capital funding in the field has goneto PhageTech, a Montreal-based companythat has raised $17.1 million. ButPhageTech is essentially using bacterio-
32 VOLUME 22 NUMBER 1 JANUARY 2004 NATURE BIOTECHNOLOGY
Box 1 Phage diagnostics—an early opportunity?
Since specific bacteriophages infect and destroy only a narrow range of bacterial hosts,phage therapy can be effective only if physicians know the bacterial strain they aretargeting and how susceptible it is to the phage therapy of choice. When treating serioushospital-based infections, bacterial typing is a matter of course, but phage susceptibilitytesting is not.
But the need for phage-based diagnostics need not be another obstacle to phagetherapy—it could offer an opportunity, even in guiding the use of conventional antibiotics.The process of working up a clinical specimen, identifying the bacterial species anddetermining antibiotic susceptibility typically takes two to three days, whereas with phage“it might be shorter,” says Stanford University’s Gary Schoolnik.
Indeed, in 1993 William Jacobs, a researcher at Albert Einstein College of Medicine(New York), and colleagues used phage as the core of a rapid diagnostic test forMycobacterium tuberculosis (TB). A simple device called the ‘Bronx Box’ used inexpensivedental X-ray film to detect the glow of luciferase-engineered phage. The glow would appearonly if the TB-specific phage could successfully infect its host, offering a positiveidentification. If that TB strain proved susceptible to a given antibiotic, the glow would beextinguished, since the reproductive source for the phage would disappear. The deviceoffered a way of diagnosing TB without expensive X-ray equipment and shaved about aweek off the time it takes for conventional susceptibility testing of this slow-growingspecies. Both advantages are vital to settings in developing nations where conventional labtechniques are too expensive and time consuming to be commonplace, yet dispensing thewrong antibiotics can lead to increasing resistance problems.
Carl Merril and his colleagues at the NIMH have expanded this concept. In the case of adisease like pneumonia that arises from a variety of bacterial strains, says Merril, “you canhave a whole collection of phages with luciferase in them” distributed in a multiwell plate.Put a little bit of sputum in each well, and you could potentially have a result—bothbacterial identification and phage susceptibility—in twenty to thirty minutes. Merril’slaboratory has filed patents on this approach.
At the Rockefeller University (New York), Vincent Fischetti has developed a method ofbacterial identification that works in mere seconds. This relies not on whole phage but onone of the key enzymes phages produce, lysin. Lysins, which allow phage progeny to breakthrough a bacterial cell wall and escape, are as specific as phages themselves—eachphage strain produces a lysin that works against a narrow host range. And they work almostinstantly. Fischetti has demonstrated how a solution cloudy with bacteria can be cleared bythe right lysin in a matter of seconds (Fig. 3).
Fischetti says the technique can provide a positive identification of anthrax from as fewas 100 spores. And the enzyme destroys germinating anthrax within moments, without theharsh chemicals usually used for environmental cleanup. Fischetti is involved with astartup company called Enzobiotics (Columbia, MD, USA) that seeks to commercialize thetechnology for these and other applications. KT
phages as scouts to find novel targets forconventional small molecule antibiotics,says CEO Pierre Etienne—an approach thatprofessional venture capitalists are morecomfortable with.
Alex Moot, a general partner at SeaflowerVentures (Boston, MA, USA) and aninvestor in PhageTech, says that eventhough his firm concentrates on the high-risk area of startup and early-stage life sci-ence companies, he avoids companieswithout “a proven regulatory pathway thatothers have established.” AlthoughPhageTech will take the tried-and-true pathof small molecule drugs, that very pathwayis up in the air with therapeutic phagepreparations. Moot, predictably, alsoexpressed concern about the lack of strongIP protection for phage therapeutics.
That may sound like a list of tough obsta-cles to overcome, but phage boosters arguethat many of the problems are more aboutperception than reality, and that other tech-nical challenges can be met. And there aremany reasons for phage therapy to succeeddespite the odds. Growing levels of antibi-otic resistance and the exit of major phar-maceutical companies from antibioticdevelopment4 means that physicians mayone day have no choice but to adopt phagetherapy for a growing number of otherwiseuntreatable infections.
The race is onFurthest ahead in the race to bring phage-based therapeutics to the US market isExponential Biotherapies (Port Washington,New York, USA). Founded in 1994, Expo-
nential is the only company among a smallfield of competitors to have completed aclinical trial—a phase 1 study conducted inEurope of a phage treatment for van-comycin-resistant Enterococcus faecium(VREF).
Exponential is planning to put this prod-uct, XpoLysin-EF, into a US phase 2 clinicaltrial near the end of 2004—which wouldmark the first FDA-sanctioned use of aphage preparation in anyone with a knowninfection. What the FDA will require toapprove a product remains an open ques-tion, but Exponential’s founder and CEORichard Carlton has had detailed discus-sions with regulators on the subject.
Early indications are that the agency isproceeding with customary caution.Carlton says he first approached the FDAabout using a panel of several phage as ameans to broaden coverage against VREF.“They said we had to start with just a singlephage strain,” says Carlton.
Carlton still believes the agency may ulti-mately bless clinical trials and products inv-olving phage cocktails as long as the phagesare not too divergent. To get to the next
stage, however, he had to go back and find asingle phage strain he thought could suc-ceed in a clinical trial. Having originally iso-lated a phage active against about 60% ofVREF strains, researchers at Exponentialwere able to find one with 95% coverage, hesays.
But Carlton and most others working inthe field still believe that using multiple-phage preparations will make phage therapymore effective, more practical and easier todevelop. The FDA has traditionally frownedon drug cocktails unless all active compo-nents are proven safe and effective bothindividually and collectively. But RylandYoung, a phage biologist at Texas A&MUniversity (College Station, TX, USA) andthe executive director of research at
GangaGen (Bangalore, India), asserts thatsuch a stance is based on a tradition that hasno relevance to phage therapy. “We’re nottalking about having two antibiotics thatwork by completely different mechanisms.All phages work the same way,” he says. Hethinks the agency will come around, butacknowledges that their failure to acceptcocktails would “greatly hamstring bacterio-phage therapy.”
Not phage awayNot all challenges to phage therapy can bereduced to regulatory tradition and histori-cal bias. One part of Exponential’s core IP isbased on another potential problem withphages. Most are cleared very rapidly by thereticuloendothelial system if they don’tquickly find a bacterial host in which tomultiply, meaning that an effective phage,particularly one administered systemically,could potentially be cleared from the bodybefore it ever reaches its target. Carlton,along with National Institute of MentalHealth (NIHM, Bethesda, MD, USA) seniorinvestigator Carl Merril and others, devel-oped a means of selecting phage that aregood at evading the reticuloendothelial sys-tem through a technique called ‘serial pas-sage’—essentially passing a phage ofinterest through a living organism, selectingthose able to get past the body’s filters andrepeating the process until a ‘long-circulat-ing’ phage variant is isolated. If such a tech-nique proves necessary to create an effectivephage therapy, Exponential may have alsoeffectively blocked a lot of competition,because the company holds a patent on thistechnique.
Carlton’s competitors don’t seem wor-ried, however. At Intralytix, a phage com-pany in Gaithersburg, Maryland, AlexanderSulakvelidze, former Director of MolecularMicrobiology for the Georgian NationalCenter for Disease Control, (Tbilisi,Georgia) believes rigorous scientific stan-dards can establish the safety and efficacy ofphage therapy resembling what is practicedin his native Georgia. Like most researcherswho have reviewed the Eastern Europeanliterature on phage therapy and seen it prac-ticed, he believes the approach works butacknowledges that the evidence, while copi-ous, is largely empirical.
However, Sulakvelidze does not believethat it will be necessary to engineer phagesin order to make them acceptable toWestern regulators. “The only reason to doit is for patent protection purposes,” he says.Sulakvelidze, who has been involved in oneinformal discussion with FDA on develop-
NATURE BIOTECHNOLOGY VOLUME 22 NUMBER 1 JANUARY 2004 33
Figure 2 Phage pioneers. In 1923, FrenchCanadian physician Felix d’Herelle, the discovererof bacteriophage (seated) and Sovietmicrobiologist Giorgi Eliava (right) founded aninstitute dedicated to phage research, laternamed Eliava Institute (Tbilisi, Georgia). Whileinitially favored by Stalin, the institute fell onhard times when Eliava was declared an ‘enemyof the people’ and executed. Today, people cometo Eliava from around the world for phage therapyof intractable infections.
Growing levels of antibioticresistance and the exit of majorpharmaceutical companies fromantibiotic development meansthat physicians may one dayhave no choice but to adoptphage therapy for a growingnumber of otherwise untreatableinfections.
ing phage therapies, echoes Carlton in say-ing that the agency does not seem to haveany fundamental problems with the idea ofphage therapy, but has not decided exactlyhow to regulate products. “It’s a learningcurve for them and for us. There is no exist-ing guideline that they can go by.”
Another wayIf the FDA will not accept multiple-phagecocktails without putting each componentthrough its own clinical trial, technologymay have an answer. “What’s really neededhere is to broaden the host range,” saysNIMH’s Merril. “There are more than 20common strains of Streptococcus pneumo-niae that cause pneumonia, so if you wantedto have a universal treatment for pneumo-nia, you’d have to patent all those differentphage and take them all through FDA.Wouldn’t it be nicer if you could make aphage that would infect all 20?”
Some phage enter their chosen host via aspecific enzyme that chops a hole in the cap-
sid coat of bacteria. But Dean Scholl, aresearcher in Merril’s laboratory, discovereda phage that carried more than one suchenzyme in its tail and could thus infect bac-terial strains with two different outer coats5.That has led the group to explore engineer-ing the phage further along these lines.“What stops you from doing three differentenzymes?” asks Merril. “Now you end upwith a phage with a much broader hostrange.” If the FDA proves unwilling to allowresearchers to use phage cocktails, such anapproach may prove necessary.
Specificity and half-life aren’t the onlypotential problems. Because phages multi-ply exponentially, 100-fold or more everygeneration, dose cannot be controlled in acustomary manner. Carlton doesn’t see thatas a problem—he intends to track the phar-macokinetics of phage just as a clinicianwould any small molecule. It’s just that thecharts will look a little odd, with the concen-tration of phage surging during a lytic cycle,diminishing as the bacterial hosts are killed
off and some phage are cleared by the retic-uloendothelial system, then surging anddiminishing again in subsequent cycles, andfinally disappearing when there are no fur-ther bacteria to support propagation andthe remaining phage are cleared from thebody.
Lysis to killStill, the potential problems of an exponen-tially replicating therapeutic have led someresearchers to wonder if phage therapywould be less complicated if cell lysisweren’t part of the equation. At GangaGen,one of the newer companies to enter thefield of phage therapy, a research team of 25people in Bangalore, India, is exploringphages engineered to be endolysin deficient.
Most phages reproduce upon entering acell, and in that process kill their host. Theyencode two families of proteins, holins andlysins, that allow the phage progeny to burstthrough the bacterial cell wall and go off insearch of new hosts. But the cell is already
34 VOLUME 22 NUMBER 1 JANUARY 2004 NATURE BIOTECHNOLOGY
Getting a phage-based human therapeutic to market meansnavigating some untested waters at FDA, but an easier commercialpath exists for phage products. Bacterial contamination offoodstuffs is a multi-billion dollar problem affecting an estimated76 million people annually in the US alone, according to the CDC.Bacteriophages could form the basis of products used todecontaminate food-processing plants, livestock and even farmers’fields. Getting such products to market generally means navigatingthe US Department of Agriculture (USDA; Washington, DC, USA) orthe Environmental Protection Agency (EPA; Washington, DC, USA),which have lower regulatory hurdles to product approval. The FDAwould not regulate such phage products unless they are useddirectly on foodstuffs.
Several phage companies are hoping some of these agriculturaland environmental applications will turn out to be low-hanging fruitthat will create a revenue stream to fund research into humantherapeutics. At GangaGen, Janakiraman Ramachandran says hiscompany hopes to have its first product, a phage that killsEscherichia coli 0157:H7 in manure, on the market in about 18months. Manure, used as fertilizer, can contaminate groundwater ifthe manure comes from infected cattle. In May 2000, for instance,over 2,000 people in the farming community of Walkerton, Ontario(Canada) were made ill and seven died as a result of E. coliinfection ultimately linked to a contaminated well.
Manure used as fertilizer is typically collected and liquefiedbefore use; GangaGen is hoping to include an extra step in whichphages that kill E. coli are added to the mix. The exponentialreproduction of phage particles will make it possible to treat largevats of manure with relatively modest amounts of phage. Thecompany has a subsidiary in Ottawa, Ontario, dedicated toagricultural and environmental phage R&D.
Biophage Pharma (Montreal, Quebec, Canada) is likewise
working on phage products to combat E. coli as well asSalmonella typhimurium, says CEO Elie Farah. The companyintends its products to be used in live animals and, in the case ofE. coli, on carcasses before meat processing. That’s particularlyimportant for the processing of ground beef, since meattrimmings from different carcasses are combined as beef isground, and E. coli from one contaminated carcass can be spreadto a large amount of meat.
Intralytix, meanwhile, is concentrating on products that could beused to decontaminate food-processing facilities. In June 2002, thecompany received an experimental use permit from the EPA to testa phage preparation called LMP-102, active against Listeriamonocytogenes, on nonfood contact surfaces. There are about2,500 cases of listeriosis in the United States each year, 20% ofwhich are fatal. Because of the seriousness of the illness, the USDAhas a strict zero-tolerance policy on the bacterium in ready-to-eatfoods; the detection of L. monocytogenes in deli meat in October2002 was responsible for the biggest meat recall in US history.
LMP-102 has been used successfully in pilot tests in a fewfacilities, according to Alexander Sulakvelidze. The company,however, does not have a firm time line for commercialization.Intralytix has also asked FDA for permission to experimentally usethe same preparation directly on foodstuffs, he adds. EvenExponential Biotherapies, furthest ahead in developing humantherapeutics, has a food safety R&D program, with a particular focuson L. monocytogenes and Campylobacter jejuni, a microbecommonly found in chicken. Though not considered deadly, C. jejuniis the leading cause of bacterial diarrhea in the United States.
“One advantage of this approach is that the same phage can beused in cattle, swine, manure and even humans,” says GangaGen’sDavid Martin. “The issue is formulation and dose, but the activeagent is the same.” KT
dead by the time that happens—in terms oflethality, the lysis is just a matter of burningdown the house for good measure.
“We realized that this last function is notessential for containing infectivity,” saysGangaGen founder and CEO JanakiramanRamachandran. Researchers at the companyreplaced the lysin gene in the T4 coliphagewith a green fluorescent protein gene, anddetermined that the engineered strain killedEscherichia coli but did not release progenyfrom the cells.
Ramachandran argues that this approachoffers a number of advantages. For onething, there is no sudden release of endo-toxin that could result from the rapid lysis ofbacteria and potentially become a deadlycomplication. In addition, it is easier to con-trol dose. “I suspect that the regulatory agen-
cies are going to be much more comfortablewith a linearly dosed therapeutic than withone that you can’t control because it expo-nentially replicates,” says David Martin,chairman of GangaGen’s board of directors.
But Evergreen’s Kutter argues that theexponential replication of phage allowsthem to pierce progressively deeper into tis-sue to root out deep-seated infections. Shealso suggests that lysin-deficient phages“might” be attractive in some circum-stances, such as infections of the gut, butpoints out that such phages will be in a racefor survival against the reticuloendothelialsystem, whereas natural, replication-compe-tent phages can mount an effective attackagainst a localized area of infection oncethey get a foothold, even if most of the ini-tial dose is cleared away.
On the other hand, Martin notes that inmany infections, an initial dose should beadequate and that with a wild-type phage,“you’ve generated a huge excess of phagefrom the last lysis,” which could theoreti-cally prove immunogenic and which apatient would potentially be “dumping backinto the environment.”
This last issue springs from the fact thatphages actively swap around DNA amongbacteria. Although most phage reproducedirectly upon entering a bacterial cell, some(temperate phage) become dormant, eitherintegrating into the host chromosome orsimply living in plasmids within the cell,going into a destructive lytic cycle later.When the phage become active again, theycan shuffle genes around through a processcalled specialized transduction. Even lytic
NATURE BIOTECHNOLOGY VOLUME 22 NUMBER 1 JANUARY 2004 35
Table 1 Select therapeutic phage companies
Company (Location) Focus Web Founded
Biophage Pharma Cancer, infection/inflammation and immune modulation, http://www.biophage.com/ 1995
(Montreal, Canada) E. coli as well as Salmonella typhimurium in livestock
and carcasses
Enzobiotics/New Horizons Lysin enzymes derived from phage as diagnostic agents and http://www.nhdiag.com/ 2003
Diagnostics topical therapeutics
(Columbia, MD, USA)
Exponential Biotherapies Phase 1 clinical trial completed of phage against VRE bacteria; http://www.expobio.com/ 1994
(Port Washington, NY, USA) also MRSA phage and biodefense
GangaGen Nontemperate, nonlytic phage for human therapeutic use; http://www.gangagen.com/ 2000
phage can be mispackaged with bacterialDNA through a similar process called gener-alized transduction. Phage can carry outrandom pieces of the bacterial genome,potentially including resistance or patho-genic toxin genes, and introduce them toother bacteria. Ultimately, some phagewould end up in the sewer where they couldtheoretically introduce toxin genes to other-wise benign bacteria in the environment.
“You could argue that those are bad geneswith which you are populating the environ-ment,” says Young. Preventing phage fromlysing out of cells could limit the problem,since dead bacteria trapping the phageshould eventually be chewed up bymacrophages, but nothing is likely to elimi-nate the problem altogether.
What regulators will construe as saferemains for now another unanswered ques-tion. Young, for one, is quick to note thatDNA transduction is going on all over theenvironment anyway—at any moment onplanet Earth there are literally tons of DNAbeing swapped around by phages. “Unlessyou’re completely compulsive, it doesn’t makea whole lot of sense to me to worry abouttransduction,” he says. Nevertheless, Ganga-Gen and all its competitors are seeking outphages that are lytic, not lysogenic (integratetheir genetic material into bacteria), andrarely engage in generalized transduction.
Phage zombiesA conceptually related approach is beingpursued by a research team at the MedicalUniversity of South Carolina (Charleston,SC, USA), working in conjunction withHexal Gentech, the research and develop-ment arm of German pharmaceutical com-pany Hexal (Holzkirchen).
Working with the filamentous coliphage
M13, the researchers created a sort of zombiephage—a regular phage body that still seeksout a specific microbial host (in this case, E.coli) but that has had its head emptied of theusual DNA necessary for replication. Itinstead injects only a lethal protein system,killing the host cell but not leading to thelysis of the cell or new phage production.
Michael Schmidt, a professor at the uni-versity and member of the research team,theorizes that the potential for bacterialresistance could be reduced by introducingmultiple killing systems into the engineeredphage, making it harder for the bacteria tomutate around their demise. And unlikeother phage, filamentous phage do not havea strict limit on the amount of genetic infor-mation they can encapsulate, which hasendeared them to many genetic engineersover the years.
The potential challenge is that filamen-tous phage are thought to infect only bacte-ria with a pilus, a spear-like protein tuberequired for the phage to gain entry. Not allbacterial species, nor all strains within com-mon species, have a pilus—meaning thatfilamentous phage may not always be thebest choice of vector for antimicrobialagents.
Nevertheless, says Schmidt, “we anticipatethat with appropriate engineering, weshould be able to extend the host range ofour system.” In any case, the researchershave also done work with the lethal proteinsystem in P1, a nonfilamentous coliphage,and Schmidt believes the system can beadapted to other kinds of phage.
Waiting for the rulesThe FDA will ultimately set the hurdles thatcompanies must jump to put therapeuticson the US market, but for now those hurdles
remain mostly undetermined. Indeed, somephage companies plan to do an end runaround some regulatory obstacles by seek-ing first to commercialize phage productsfor agricultural applications, where regula-tions are less stringent (see Box 2).
GangaGen’s Young, for one, thinks thatcompanies in the field should take a proac-tive approach in determining the standardsthat phage therapies must satisfy. The firststep toward that goal may be a ‘PhageSummit’ that Young is helping to organize,scheduled to take place next August in KeyBiscayne, Florida, USA.
“I’m hoping we can pound out some com-mon standards,” he says. In the absence of aunified and scientifically grounded frontamong the companies, the FDA may tendtoward its traditional standards, which couldprove difficult for the nascent industry.
But even without success in the majorWestern markets, phage therapy could stillscore a huge victory. Gary Schoolnik, pro-fessor of Medicine, Microbiology andImmunology at Stanford University (Stan-ford, CA, USA) and chair of GangaGen’s sci-entific advisory board, notes that onefurther advantage of phages is that they canbe produced very cheaply. That could makethem ideal for treating infections in devel-oping countries—particularly enteric infec-tions like cholera, where Schoolnik notes apreparation could be “comparatively crudeand still be effective.” Indeed, the very obsta-cles to commercializing phage in developedmarkets, like patentability, could actually bean advantage to developing nations thatwant to gain a little know-how and go it ontheir own.
And problems in developing countriescould spread to the developed world ifphage therapy fails. “I don’t want my grand-children to live in a world that is effectivelyin a pre-antibiotic era,” says Kutter.
1. US Centers for Disease Control, the Food and DrugAdministration and the US National Institutes ofHealth. A Public Health Action Plan to CombatAntimicrobial Resistance (CDC, FDA & NIH, June,2000) (http://www.cdc.gov/drugresistance/action-plan/aractionplan.pdf).
2. S. Tslodras et al. Linezolid resistance in a clinical iso-late of Staphylococcus aureus. Lancet 358,207–208 (2001).
3. G.M. Eliopoulos, et al. Characterization of van-comycin-resistant Enterococcus faecium isolatesfrom the United States and their susceptibility invitro to dalfopristin-quinupristin. AntimicrobialAgents Chemother. 42, 1088–1092 (1998).
4. Fox, J. Concerns raised over declining antiinfectiveR&D. Nat. Biotechnol. 21, 1255–1256 (2003).
5. Scholl, D. et al. Bacteriophage K1-5 encodes two dif-ferent tail fiber proteins, allowing it to infect andreplicate on both K1 and K5 strains of Escherichiacoli. J. Virol. 75, 2509–2515 (2001).
36 VOLUME 22 NUMBER 1 JANUARY 2004 NATURE BIOTECHNOLOGY
a
d e
b cFigure 3 Lysis to kill. (a–c)Electron micrographs showphage enzyme wiping out acolony of Bacillus cereus: ahealthy colony of bacteria (a);the bacteria 1 minute afterenzyme treatment (b); andbacteria 15 minutes aftertreatment (c). (d) Lysinpunches holes in bacterialouter cell wall, causing it toexplode. (e) A singlebacterium begins to rupture,its inner membrane spillingthrough an enzyme-inducedhole. (Image courtesy ofVincent Fischetti, RockefellerUniversity, New York, NY.)
