In July 2011, a 43-year-old woman walked out of the National Institutes of Health (NIH) Clinical Center in Bethesda, Maryland, after a month of battling a serious bacterial infection. Three weeks later, two more patients tested positive for the same bacterial strain after checking into the clinic. Over the next four months, the pathogen, a multidrug-resistant form of Klebsiella pneumoniae, continued to spread; approximately one clinic patient acquired the infection every week.

Clinicians threw up walls—both physical and chemical—to contain the pathogen. All patients were kept isolated and under surveillance; after the fourth case, infected patients were placed in a separate section of the center and tended to by a dedicated staff using dedicated instruments. Visitors wore caps, gowns, and gloves, surfaces were routinely washed with bleach, and the intensive care unit was regularly gassed with hydrogen peroxide to decontaminate the rooms. Still, patients...

In the last decade, antibiotic resistance has grown from a concern to a crisis. In addition to the deadly incident at the NIH, a multidrug-resistant form of methicillin-resistant Staphylococcus aureus (MRSA) in a UK neonatal unit infected 12 babies in 2011. And just last year, carbapenem-resistant enterobacteria infected seven people and killed two at a Los Angeles, California, hospital. Even when antibiotics do work, they’re not always the best option, as they wipe out beneficial bacteria as well as pathogenic ones, with potentially long-lasting health consequences.

Researchers on the hunt for more-effective therapies that preserve a healthy microbiome are taking a closer look at the many different viruses that attack bacteria. Bacteriophages (literally, “bacteria eaters”) punch holes through the microbes’ outer covering and inject their own genetic material, hijacking the host’s cellular machinery to make viral copies, then burst open the cell with proteins known as lysins, releasing dozens or hundreds of new phages. The cycle continues until there are no bacteria left to slay. Phages are picky eaters that only attack specific types of bacteria, so they’re unlikely to harm the normal microbiome or any human cells. And because phages have coevolved with their bacterial victims for millennia, it’s unlikely that an arms race will lead to resistance. This simple biology has led to renewed interest in the surprisingly long-standing practice of phage therapy: infecting patients with viruses to kill their bacterial foes.

While most research is still in the preclinical phase, a handful of trials are underway, and a growing number of companies are investing in the treatment strategy. Phage therapy is receiving as much attention now as it did in the pre-antibiotic era, when it flourished in spite of the dearth of clinical tests or regulatory oversight at the time. “Bacteriophage therapy will have its day again,” pathologist Catherine Loc-Carrillo of the University of Utah told The Scientist last year. “It sort of had one, before antibiotics came along, but it wasn’t well understood then.”

But with lingering questions about phages, which are often dubbed “viral dark matter” because so little is known about their biology, their use in mainstream medicine still faces many hurdles. And the consequences of moving phage therapies forward without more concrete evidence could be devastating, adds phage biologist Ryland Young of Texas A&M University. “If we have more poor data like we did in the 1920s, it’ll really set applications back in the long term.”

A century of cures

The roots of phage therapy stretch back more than 100 years, even before the discovery of bacteriophages. In 1896, British bacteriologist Ernest Hankin tested water from two Indian rivers, the Ganges and its tributary the Yamuna—locally believed to have curative properties—and found evidence of antibacterial activity. He used porcelain filters to strain the river water, removing bacteria and larger organisms while retaining a suspension that could kill Vibrio cholerae. He suspected some unknown substance or agent in the water played a part in limiting the spread of cholera epidemics in the area. Over the next few years, reports of natural waters with similar antibacterial properties trickled in from Russia and other parts of the world.

The roots of phage therapy stretch back more than 100 years, before the discovery of bacteriophages.

Two decades later, another British bacteriologist, Frederick Twort, found a bacteria-killing agent while working with Micrococcus cultures, although he hesitated to hypothesize that it was a virus.2 In the 1910s, French-Canadian microbiologist Felix d’Herelle was testing fecal filtrates from soldiers infected with Shigella, a causative agent of dysentery, when he uncovered evidence to support Twort’s discoveries. After a few days of applying the fecal samples to Shigella cultures, d’Herelle saw kill zones on the culture plates where something had decimated the pathogen. Conjecturing more boldly than Twort, d’Herelle suspected he was observing the work of viruses that infect bacteria, and he coined the term “bacteriophage,” from the Greek word for “to eat,” to describe the disease-fighting agents.3 “In a flash I had understood: what caused my clear spots was, in fact, an invisible microbe, a filterable virus, but a virus parasitic on bacteria,” d’Herelle recalled in a 1948 article.4

D’Herelle conjectured that the sick soldier whose fecal sample had killed the bacteria in culture would probably also recover, thanks to the same microbe-killing virus—and he was right.5 Four years later, in 1919, d’Herelle used these same phages, isolated from the fecal samples of dysentery  patients who’d recovered, to successfully treat children suffering the same infection.

No one had actually seen a phage; it would be another 20 years before scientists captured the earliest electron micrographs of bacteria-infecting viruses.6 But phage therapy began to be used around the world to treat a dizzying array of infections. Belgian researchers reported injecting phages isolated from various sources to cure staphylococcal skin infections; intravenous phage therapy was used to treat cholera in India and streptococcal infections in France; studies in the U.S. reported treating septicemia and meningitis.

The rapid growth of the field was characterized by “an early, enthusiastic period during which claims were excessive and often unrealistic, while at the same time little was understood of the viral nature of phages or their strengths and limitations,” Elizabeth Kutter of Evergreen State College in Washington and colleagues wrote in a 2011 review of phage therapy’s history.7 Despite the fact that some pharmaceutical companies began standardizing and marketing the therapies as early as the 1920s, the US Food and Drug Administration was not overseeing their development, and few were subjected to controlled clinical testing.

In 1934, a three-part JAMA report provided the first objective evaluation of phage therapy. The authors assessed more than 100 studies and concluded that the treatment was only reliable for some staphylococcal infections. Without double-blind trials and clinical research to test its effectiveness and safety, phage therapy fell out of favor in the West. Attention turned to antibiotics, which had been discovered in 1928 and were easier to manufacture and standardize.8 But phage therapy stuck around in many parts of the world, particularly in Eastern Europe, where modern drugs are expensive and often hard to come by.

1910: Wellcome Images; 1915: Obituary Notices of Fellows of the Royal Society; 1917: Pasteur Institute; 1928: The University of Edinburgh Library; 1940: Hans-Wolfgang Ackermann; 1977: Protein Data Bank/2BPA; 2015: Dr. Graham BeardsThese days, centers like the Eliava Institute of Bacteriophages, Microbiology and Virology in the Republic of Georgia offer commercial phage preparations for specific indications, such as MRSA and gastrointestinal infections caused by E. coli and Shigella species. Researchers at the Eliava Institute also mix phages into custom cocktails for many infections. Phages for these mixtures are isolated from some of the same sources that yielded them a century ago: sewage, hospitals, rivers and lakes, and other places where pathogens thrive. Purified isolates from these sources are grown on bacterial cultures in the lab to identify phages that target the pathogen of interest. Years of research at the Eliava Institute have led to a carefully curated library with hundreds of vials of such isolates, from which the scientists prepare their custom combination therapies.

Mzia Kutateladze, the institute’s current director, says she receives a growing number of requests for treatment, including from patients in the U.S. and Western Europe. “They send us clinical materials, either cultures or swabs, before they arrive,” she says. “We first test our commercial products. If they don’t work, we identify phages in the library, prepare and test a final customized product before it’s used in the patients.” Patients can then travel to the clinic for treatment, or the Eliava Institute will send the phages to patients to use on their own.

Kutter suspects that success stories from the Eliava Institute and others will ease acceptance of phage therapy into the modern pharmacopeia. Indeed, retrospective analyses of phage therapies published by these groups are helping researchers understand which particular infections are likely to respond to the treatments. Fueled by these data and a prominent mention in a 2014 National Institute of Allergy and Infectious Diseases (NIAID) report on the agency’s antibacterial resistance program, companies around the world are preparing for a second coming of phage therapy.

But many questions remain. Laboratory studies of phages have typically relied on well-understood model systems, such as E. coli and its viruses; the myriad phages used in cocktails to treat human infections are grossly understudied. And while phages are typically thought to be harmless to human cells, in part because they are ubiquitous in the environment and in the human body, little is known about how phages interact with human cells such as those of the immune system. One recent study, for example, reports that phages may be able to use mucosal surfaces in tissues to enhance their predatory activity.9

“Almost all bacteriophage biology has focused on the paradigm of E. coli phages lambda, T4, and so on,” says Texas A&M’s Young. “Bacteriophages of most major pathogens have not been studied very well. The basic science [of phage therapy] is still way behind where it needs to be.”

Probing dark matter

CLEARING GUT INFECTIONS: Treating infections of the gastrointestinal tract has been one of the most successful uses of phage therapy since its inception. Although antibiotics are the go-to treatment today, concerns about increasing drug resistance and disrupting the microbiome have led researchers to reconsider using bacteria-killing viruses instead.
See full infographic: WEB | PDF
Evaluating a phage’s bacteria-killing activity in the lab is easy: simply apply a known amount of purified phage to a lush lawn of microbes in a petri dish and measure the resulting kill zones. But an infection of human tissue looks nothing like a monoculture of well-fed bacteria smeared across a plate. In diseased tissue, pathogens produce a slew of different proteins and small molecules, often form 3-D biofilms, and exist within complex microbial communities. How phages interact with pathogens in such situations remains unclear.

This lack of understanding has posed obvious hurdles for those interested in developing phage therapies. One issue is whether phages are uniquely suited to treat certain infections. For example, phages may work better than antibiotics when a deep, localized lesion, such as a bone infection, is difficult to access via the bloodstream. “Even when bacteria are inherently sensitive to an antibiotic, the drug often can’t reach the site of infection [in sufficient concentrations],” says Kutter. Phages, on the other hand, can multiply within pathogens once they reach the site, quickly rising to the level of a therapeutic dose.

Another open question is how therapeutic viruses interact with the human immune system, and whether they might cause side effects. At the Ludwik Hirszfeld Institute of Immunology and Experimental Therapy at the Polish Academy of Sciences—where patients receive phage therapy on compassionate-use grounds after they’ve failed to respond to other treatments—researcher Andrzej Górski is sifting through years of clinical data to find answers. In a retrospective analysis of immune responses in 153 people treated with phages between 2008 and 2010, Górski and his colleagues reported that the therapies were well-tolerated in 80 percent of patients.10 Only a small number had to stop treatment because they experienced adverse reactions such as nausea or pain in response to gut treatments, or local reactions to topical phage applications, Górski said at a first-of-its-kind NIAID workshop on phage therapies that convened in Rockville, Maryland, last July.

Górski’s group has also assessed inflammatory markers such as C-reactive protein and white blood cell counts in 37 patients treated with phage for S. aureus bone infections and found that phage therapy lowered the levels of these markers.11 It appeared that the phages dampened the bacteria-triggered inflammation, a modulation that was associated with a positive clinical response in almost half the patients.

Looking more carefully at the immune responses triggered by phage therapy, Górski and his colleagues found that repeated exposure to the same strain of phage could trigger an adaptive immune response, resulting in anti-phage antibodies. At the NIAID workshop he reported that such antibody responses may be higher when using cocktails of phage than when dosing with a single viral strain. But whether such antibodies help or hinder phage therapy is still unclear. In 122 patients and healthy volunteers who received phage orally or in local applications, Górski and his colleagues found that, while several patients produced anti-phage antibodies, the presence of antibodies had little correlation to whether or not the therapy was successful.12 “When you have anti-phage serum activity against phages, that doesn’t necessarily mean the phage isn’t working,” Górski said at the meeting.

Phages may also elicit different immune responses depending on how they are administered—topically, orally, intravenously, or rectally—and on what tissues or organs they target. Lumping every treatment together as “phage therapy” undermines efforts to understand such nuances, says Young. “Phages are treated like they’re all the same, but they can be more different from each other than ants and elephants.”

Going mainstream

TELL-TALE TAILS: Tailed bacteriophages (Caudovirales), which make up the vast majority of known phages, fall into three families: Myoviridae (T4 shown at top), characterized by straight contractile tails; Siphoviridae (λ shown at bottom left), known for flexible non-contractile tails, and Podoviridae (T7 shown at bottom right), which have very short non-contractile tails. Some companies are exploring the use of purified tail proteins as antibacterial agents.R. DUDA/ UNIVERSITY OF PITTSBURGH; P. SERWER/THE UNIVERSITY OF TEXAS HEALTH SCIENCE CENTER AT SAN ANTONIODespite the challenges still facing phage therapy, numerous companies are now looking to bring the treatments to mainstream clinics. (See table below.) In September 2015, researchers in France launched the first multicenter study and clinical trial to evaluate phage therapy. Known as Phagoburn, the project began in 2013 with preclinical work to produce two phage cocktails using methods that met European Medicines Agency (EMA) manufacturing standards. Now, the two treatments—which target burn-wound infections caused by E. coli or Pseudomonas aeruginosa—will be administered to 220 patients and the therapeutic results compared with those in patients treated with silver sulfadiazine, the current go-to drug for such infections.

Like the phage cocktails used by the Eliava Institute, the Phagoburn therapies are a mixture of naturally occurring viruses selected for their ability to target specific bacterial species. The P. aeruginosa cocktail is a mix of 13 phages, according to Patrick Jault, a Phagoburn investigator and chief of the burn treatment center at Percy Military Hospital outside of Paris; the E. coli cocktail contains 12 phages.

A handful of US companies also aim to bring phage preparations to clinical trials. In 2009, the Maryland-based firm Intralytix published the results of its Phase 1 trial for a phage therapy that targets venous leg ulcers in diabetic patients. None of the 40 or so patients who received the phage cocktail had any adverse reactions to the treatment, but the company has not said whether a Phase 2 trial is planned.13 Meanwhile, Richmond, Virginia–based AmpliPhi Biosciences announced in November it was enrolling nine patients to test the safety of a natural phage cocktail intended to treat chronic sinus infections caused by S. aureus.

These and more trials are needed to determine which phage therapies will work best for which indications, and how to scale up the production of those that are successful. “We’ve yet to see how this plays out both in the clinic and when we try to manufacture these at large scales,” says AmpliPhi CEO Scott Salka.

With these hurdles in mind, San Diego–based Synthetic Genomics, founded by synthetic biologist J. Craig Venter, is taking a different approach. Rather than mixing natural phages together, company researchers are attempting to engineer a synthetic virus that combines the properties of multiple phages into a single genome. Such engineered phages are simpler to manufacture than cocktails with dozens of different phages, says Bolyn Hubby, vice president of research and development at Synthetic Genomics. “Not having very large complex cocktails makes these products compatible with current GMP [good manufacturing process] standards.”

If we have more poor data like we did in the 1920s, it’ll really set applications back in the long term.—Ryland Young,
 Texas A&M University

The company is currently applying bioinformatics and viral engineering methods to understand the natural host ranges of various phages, and then inserting genes from other viruses to expand those ranges to include other subtypes of the targeted bacterium. “By iteratively doing this, we can expand a phage’s host range while maintaining specificity so commensal bacteria are unaffected,” says Hubby. Eventually, the researchers plan to layer on additional ammunition, such as potency against biofilms or synergistic interactions with antibiotics.

In addition to phage cocktails and engineered viruses, researchers are also putting phage components to work. New York–based Contrafect and Netherlands-based Micreos, for example, use isolated lysins, the phage enzymes that rip through a bacterial envelope when a virus injects its DNA into a cell or when viral progeny burst out. Others are experimenting with just the phage tail proteins, called tailocins. (See image here.) These parts lack the ability to multiply within a host, so they avoid risks inherent to phages that carry DNA, making such protein therapies potentially much easier to manufacture and bring to market, says Rockefeller University’s Vincent Fischetti, who is a scientific advisor to Contrafect.

But regardless of whether a product is naturally derived or engineered, or a whole virus or just pieces, the bar for success is the same—and it’s a high one. Trials will have to demonstrate that phage therapy, either alone or in conjunction with other drugs, is superior to current standard treatments. “You can’t just show the two arms of a trial are equivalent and convince people to change,” says infectious disease specialist Brad Spellberg of the University of Southern California.

What the field needs now is solid clinical research, says Spellberg, who compared the promise of phage therapy to the elusive Sasquatch. “After 90 years of talking about it, all we have are these grainy videos of some dude in a costume walking through the woods; at this point you want someone to catch Bigfoot and take him to the zoo,” he said. “It’s time for someone to pony up the product and the dough to do the clinical trial.” 

A sampling of firms in the U.S. and Europe that are conducting research on viral treatments for bacterial infections

AmpliPhiRichmond, VirginiaNatural phage cocktails
P. aeruginosa lung infections in cystic fibrosis; S. aureus wound and skin infections; C. difficile gastrointestinal infectionsPhase 1 approved November 2015
ContraFect CorporationYonkers,
New York
Bacteriophage lysinsS. aureus bacteremiaPhase 1 launched  April 2015
EnBiotixCambridge, MassachusettsEngineered phagesStaphylococcal infections of prosthetic jointsPreclinical
EpiBiomeSan Francisco, CaliforniaNatural phage cocktailsE. coli and Shigella dysenteriae diarrheal infections in childrenPreclinical
Natural phages fixed to solid surfacesMRSA wound infectionsPreclinical
IntralytixBaltimore, MarylandNatural phage cocktailsS. aureus, P. aeruginosa, E. coli wound infections; irritable bowel diseasePreclinical
MicreosWageningen, NetherlandsBacteriophage lysinsS. aureus and MRSA skin infectionsPreclinical
NovolyticsWarrington, U.K.Natural phage cocktailsMRSA skin infectionsPreclinical
PherecydesRomainville, FranceNatural phage cocktailsE. coli and P. aeruginosa burn and skin infections; P. aeruginosa respiratory infections; S. aureus bone/joint/prosthetic infectionsPhase 1 launched  September 2015
Synthetic GenomicsSan Diego, CaliforniaEngineered phagesInfections in burn wounds, skin, and cystic fibrosisPreclinical
TechnoPhageLisbon, PortugalNatural phage cocktailsChronic ulcers, respiratory and skin infectionsPreclinical

Jyoti Madhusoodanan is a freelance science writer living in San Jose, California.


  1. E.S. Snitkin et al., “Tracking a hospital outbreak of carbapenem-resistant Klebsiella pneumoniae with whole-genome sequencing,” Sci Transl Med, 4:148ra116, 2012.
  2. A. Sulakvelidze et al., “Bacteriophage therapy,” Antimicrob Agents Chemother, 45:649-59, 2001.
  3. F. d’Hérelle, “Sur un microbe invisible antagoniste des bacilles dysentérique,” CR Acad Sci Paris, 165:373-75, 1917.
  4. F. d’Herelle, “Le bactériophage,” Atomes, 3:399-403, 1948. (Translated in Sci News, 14:44-59, 1949.)
  5. D.H. Duckworth, “Who discovered bacteriophage?” Bacteriol Rev, 40:793–802, 1976.
  6. H.-W. Ackermann, “The first phage electron micrographs,” Bacteriophage, 1:225-27, 2011.
  7. S.T. Abedon et al., “Phage treatment of human infections,” Bacteriophage, 1:66-85, 2011.
  8. M.D. Eaton, S. Bayne-Jones, “Bacteriophage therapy,” JAMA, 103:1769-76, 1847-53, 1934-39, 1934.
  9. J.J. Barr et al., “Subdiffusive motion of bacteriophage in mucosal surfaces increases the frequency of bacterial encounters,” PNAS, 112:13675-80, 2015.
  10. R. Mi?dzybrodzki et al., “Clinical aspects of phage therapy,” Adv Virus Res, 83:73-121, 2012.
  11. R. Mi?dzybrodzki et al., “A retrospective analysis of changes in inflammatory markers in patients treated with bacterial viruses,” Clin Exp Med, 9:303-12, 2009.
  12. M. ?usiak-Szelachowska et al., “Phage neutralization by sera of patients receiving phage therapy,” Viral Immunol, 27:295-304, 2014.
  13. D.D. Rhoads et al., “Bacteriophage therapy of venous leg ulcers in humans: Results of a Phase I safety trial,” J Wound Care, 18:237-38, 240-43, 2009.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?