Can Bacteria Become Resistant to Phages? Mechanisms, Fitness Costs and Therapeutic Trade-Offs
Bacteriophages can kill bacteria that no longer respond to antibiotics, but bacteria are not evolutionarily passive targets. Just as antibiotic exposure selects for antibiotic-resistant populations, bacteriophage infection can select bacteria capable of avoiding viral attack. Phage resistance is therefore one of the central challenges facing therapeutic bacteriophage development.
The comparison with antibiotic resistance is useful, but incomplete. Phages are self-replicating biological entities that must recognize a bacterial cell, attach to an appropriate receptor, inject their genome, overcome intracellular antiviral systems and use the host's molecular machinery to produce new viruses. A bacterium can interfere with almost any of these stages.
Resistance can consequently evolve through many mechanisms, ranging from modification of a surface receptor to sophisticated intracellular immune systems such as restriction–modification and CRISPR-Cas. Some bacterial populations can also enter temporary physiological states that reduce infection without acquiring a permanent genetic resistance mechanism.
Yet phage resistance has an unusual feature that makes it particularly interesting therapeutically. Escaping a bacteriophage is not always free.
If a phage uses an important bacterial structure as its receptor, eliminating or modifying that structure may prevent infection while simultaneously making the bacterium less virulent, less competitive or more susceptible to antibiotics or the immune system. Researchers call these consequences evolutionary or fitness trade-offs.
A study published on October 12, 2026 in Frontiers in Microbiology provides a particularly clear example. Researchers studying an extraintestinal pathogenic Escherichia coli strain found that mutations conferring resistance to a newly isolated phage simultaneously disrupted bacterial surface structures involved in pathogenicity.
The findings reinforce an increasingly important idea in phage therapy: the objective may not always be to prevent bacteria from evolving resistance. In some situations, researchers may instead be able to choose phages that force bacteria to pay a clinically useful price when resistance evolves.
Bacterial resistance usually begins with the first step of infection: adsorption.
A bacteriophage cannot infect every bacterium it encounters. Proteins or other structures on its surface recognize specific molecules exposed by the bacterial cell. These receptors can include lipopolysaccharides, capsular polysaccharides, outer-membrane proteins, pili, flagella and other components of the cell envelope.
Changing one of these structures can prevent the phage from attaching.
Mutations affecting receptor synthesis are consequently among the most frequently observed routes to phage resistance. Bacteria may eliminate the receptor completely, alter its chemical structure, reduce its expression or conceal it beneath extracellular material. Some populations increase capsule or exopolysaccharide production, physically limiting access to receptors underneath.
This strategy can be remarkably effective. If a phage is completely dependent on one receptor and that receptor disappears, the infection cycle can stop before viral genome entry even begins.
But bacterial surface structures rarely exist solely for bacteriophages.
Lipopolysaccharide, capsule, pili, flagella and membrane proteins can contribute to nutrient acquisition, environmental survival, adhesion, biofilm formation, interactions with the immune system, antimicrobial resistance or virulence. Altering them may therefore solve the phage problem while creating another biological disadvantage.
The October 2026 ExPEC study provides a particularly useful demonstration of this principle.
Long Chen, Sixiang Xu, Xiaoyue Li and colleagues characterized the lytic E. coli bacteriophage vB_EcoP_P61425 and then exposed an extraintestinal pathogenic E. coli strain, EC1649, to strong phage selection. Three stable resistant mutants emerged.
Whole-genome analysis identified frameshift mutations affecting wbbL in two mutants and a nonsense mutation in waaG in the third.
Both genes participate in lipopolysaccharide biosynthesis. WbbL is involved in synthesis of the O-antigen component, while WaaG participates in construction of the LPS core. Because LPS forms a major part of the outer surface of Gram-negative bacteria, altering these genes can substantially change what an approaching bacteriophage encounters.
The researchers went beyond simply associating the mutations with resistance. They deleted wbbL and waaG experimentally and produced resistant bacteria, then restored the genes through complementation and recovered phage susceptibility.
The result provides strong evidence that the two genes were directly involved in the phenotype rather than merely being mutations that happened to appear in resistant cells.
The effect on phage adsorption was substantial. Approximately 96.8% of the phage adsorbed to the parental EC1649 strain under the experimental conditions. Adsorption fell to about 37.7% after deletion of wbbL and 33.4% after deletion of waaG. Restoring the genes returned adsorption toward the original phenotype.
In other words, E. coli had found an effective escape route: alter its LPS and make it substantially harder for the phage to attach.
But the resistant bacteria were no longer biologically identical to their ancestor.
The mutations changed colony and cellular morphology and markedly reduced bacterial adhesion to human bladder epithelial cells. Relative adherence fell to approximately 24.7% in the wbbL deletion strain and 36.1% in the waaG deletion strain, while genetic complementation restored much of the lost phenotype.
The effect extended to an experimental virulence model. In Galleria mellonella larvae, the parental EC1649 strain was highly lethal, with only around 10% of larvae surviving at the end of the experiment. Larvae infected with the spontaneous resistant mutants or the corresponding wbbL and waaG deletion strains showed 100% survival at 72 hours under the conditions tested.
Restoring the genes substantially restored bacterial virulence.
The evolutionary outcome is therefore particularly striking. The bacteria escaped the phage, but the mutations that protected them simultaneously interfered with structures contributing to adhesion and virulence.
This does not mean that phage resistance generally makes pathogenic bacteria harmless.
The study examined one bacterial background, EC1649, interacting with one bacteriophage. Virulence was evaluated partly using epithelial-cell adhesion and an invertebrate Galleria mellonella model rather than a human infection or mammalian therapeutic study. Whether the same mutations would produce clinically useful attenuation during infection remains unknown.
The result should therefore be interpreted as evidence that a particular resistance pathway carries a measurable biological cost, not as proof that allowing phage resistance to evolve is inherently therapeutic.
This distinction is fundamental because bacteria possess many other ways to defeat bacteriophages.
Even if adsorption succeeds, bacteria can prevent viral DNA from entering the cytoplasm. Superinfection exclusion systems, membrane-associated defence proteins and changes affecting genome injection can block infection after attachment.
Once the phage genome enters the cell, an even larger immune arsenal becomes relevant.
Restriction–modification systems recognize particular features of foreign DNA and cleave susceptible phage genomes. CRISPR-Cas systems can use sequence-specific molecular memories of previous infection to identify and destroy related viral genetic material.
The rapid discovery of bacterial antiviral systems during the past decade has revealed that CRISPR-Cas represents only a fraction of bacterial immunity. More than a hundred defence families have now been described, including CBASS, BREX, DISARM, Thoeris, retron-associated systems, Gabija, Zorya, Septu, Druantia and many others.
Their mechanisms are remarkably diverse.
Some attack phage DNA or RNA directly. Others recognize viral proteins or alterations in essential bacterial processes. Some generate intracellular second messengers that activate downstream effectors. Others deplete metabolites, disrupt membranes or shut down translation.
Several systems employ an especially drastic strategy known as abortive infection.
Instead of preserving the infected bacterium, abortive-infection systems prevent production and dissemination of the virus even if doing so severely damages or kills the host cell. From the perspective of a clonal bacterial population, sacrificing one infected cell can be advantageous if it protects neighbouring relatives from an expanding phage epidemic.
Phages are also evolving.
They can modify receptor-binding proteins, encode enzymes that degrade capsules, chemically modify their genomes, produce anti-CRISPR proteins, inhibit restriction systems or carry proteins that neutralize newer bacterial defence pathways.
Phage resistance is therefore not a one-directional endpoint. Bacteria and bacteriophages are engaged in a continuous evolutionary arms race in which each defensive innovation creates selection for a countermeasure.
This is one reason phage cocktails are attractive.
If several phages rely on different bacterial receptors or encounter different intracellular defence mechanisms, a mutation protecting against one virus may not protect against the others. Acquiring simultaneous resistance can require several independent changes and potentially impose larger biological costs.
Cocktail design alone does not guarantee durable activity. If several phages rely on the same receptor, a single bacterial mutation may create cross-resistance to every component. Understanding which receptors each phage uses can therefore be more informative than simply maximizing the number of viruses in a mixture.
The choice of receptor is particularly important when researchers deliberately try to exploit evolutionary trade-offs.
One of the best-known examples involves Pseudomonas aeruginosa and the phage OMKO1.
OMKO1 uses OprM, a component of major multidrug efflux systems, during infection. These pumps contribute to the bacterium's ability to expel several antimicrobial compounds. When P. aeruginosa evolves changes that interfere with OMKO1 infection, alteration of the efflux machinery can simultaneously increase susceptibility to certain antibiotics.
The strategy has become known as evolutionary steering.
Rather than choosing a phage only because it kills bacteria rapidly, researchers can potentially choose a phage according to the evolutionary options it leaves available to the bacterium.
Imagine two viruses with similar initial antibacterial activity.
Resistance to the first can be achieved through a mutation carrying little biological cost. Resistance to the second requires modification of a structure essential for virulence, antibiotic resistance or colonization.
The second phage could theoretically be therapeutically preferable even if both appear equally effective in an initial laboratory killing assay.
Capsule-targeting phages offer another example.
The polysaccharide capsule of Klebsiella pneumoniae is both an important virulence factor and a common receptor or barrier involved in phage infection. Studies have repeatedly observed Klebsiella populations escaping capsule-targeting phages by mutating capsule-biosynthesis genes.
In some experimental systems, these resistant variants become less virulent or more vulnerable to immune clearance. Other studies have observed collateral changes in antibiotic resistance or sensitivity to different phages.
But even here the outcome is not universal.
A 2021 study of phage-resistant Klebsiella pneumoniae found that capsule modification did not dramatically reduce every measure of bacterial fitness or protection against innate immunity. Other genetic changes were associated with loss of multidrug-resistance determinants, illustrating that different resistance trajectories can produce very different consequences.
This variability is why “phage resistance reduces virulence” should never be treated as a general rule.
A bacterium may acquire resistance with a major cost, a modest cost or almost no detectable cost under the conditions being measured. The same mutation can also have different consequences depending on the infection environment.
Fitness itself is contextual.
A mutation that slows bacterial growth in nutrient-rich laboratory broth might be neutral inside a biofilm. Loss of a surface receptor may reduce adhesion to one type of host cell while improving survival under another environmental pressure. Increased capsule production may block one phage while simultaneously changing immune evasion, antibiotic penetration or susceptibility to another virus.
Bacteria can also compensate.
An initial resistance mutation may impose a substantial fitness penalty, but subsequent evolution can produce secondary mutations that partially restore growth or virulence while retaining phage resistance. Therapeutically favourable trade-offs observed immediately after resistance evolves may therefore not remain unchanged indefinitely.
Another important distinction is the difference between genetic resistance and phenotypic tolerance.
A bacterium does not necessarily need a new mutation to survive phage exposure.
Phage replication depends heavily on bacterial physiology. Slow-growing, nutrient-limited or stressed cells may support viral replication poorly even though they remain genetically susceptible. Biofilms can create spatial barriers and physiological gradients, while phase variation can reversibly alter expression of receptors.
Recent work in Klebsiella pneumoniae, for example, has demonstrated reversible phenotypic phage resistance associated with downregulation of capsule production. Cells can temporarily shift into an acapsular, less susceptible state without requiring the permanently fixed mutations traditionally associated with resistance.
This distinction matters clinically because a bacterial population surviving treatment might regain phage susceptibility after its physiological state changes.
Measuring resistance therefore requires more than simply noting whether bacteria regrow after phage exposure.
Researchers can use adsorption experiments to determine whether receptor binding has changed, efficiency-of-plating assays to quantify productive infection, genome sequencing to identify mutations, transcriptional analysis to detect regulatory changes and complementation experiments to prove that a candidate gene actually causes the phenotype.
The 2026 ExPEC study is strong in this respect because it combined spontaneous resistance evolution with genome sequencing, targeted gene deletions and genetic complementation. The authors could therefore connect wbbL and waaG directly with altered adsorption and resistance rather than relying only on correlation.
The study also illustrates why understanding the receptor of a therapeutic phage can be useful before treatment.
If resistance emerges, clinicians and researchers may be able to anticipate the likely biological consequences. A mutation affecting an essential virulence-associated structure might create a useful trade-off, whereas loss of a dispensable receptor could provide bacterial escape with little therapeutic benefit.
This logic could influence future phage libraries.
Instead of cataloguing candidates exclusively according to which bacterial strains they kill, libraries could increasingly include information about receptor usage, resistance pathways, cross-resistance, effects on antibiotic sensitivity and phenotypes of resistant mutants.
Phage combinations could then be designed partly around evolutionary constraints.
One virus might target a capsule, another a membrane protein and another a structure associated with antimicrobial resistance. The goal would be to reduce the probability that a single mutation protects against the entire treatment while ensuring that common escape routes carry substantial costs.
Phage–antibiotic combinations provide another opportunity.
When phage resistance increases antibiotic susceptibility, simultaneous or sequential antibiotic exposure may eliminate variants selected by the virus. Conversely, antibiotics can sometimes change bacterial physiology in ways that improve phage infection.
But synergy cannot be assumed.
Some antibiotics inhibit bacterial processes that phages require for replication, and the outcome depends on the specific bacterium, phage, drug concentration, timing and physiological conditions. Rational combination therapy therefore requires experimental testing rather than the assumption that two antibacterial agents will automatically work better together.
The immune system provides another potential selective partner.
If phage resistance requires loss of capsule, LPS modification or another structure involved in immune evasion, resistant bacteria might become easier for complement, antibodies or phagocytic cells to eliminate.
Again, the effect is highly system dependent.
Reduced virulence in an insect model does not automatically imply enhanced human immune clearance, and laboratory measurements of adherence or serum susceptibility cannot on their own establish clinical benefit.
This is perhaps the most important limitation when interpreting fitness trade-offs in phage therapy.
Evolutionary steering is scientifically attractive because it turns an apparent weakness of phage therapy into a potential advantage: bacteria evolve resistance, but the phage determines which bacterial traits are placed under selection.
Yet demonstrating a trade-off in vitro is much easier than demonstrating that it improves treatment outcomes in patients.
The October 2026 ExPEC study did not administer phage therapy to humans or demonstrate that emergence of wbbL or waaG mutants improves clinical outcomes. It showed that, under experimental selection, one phage could select resistant bacteria with strongly reduced adsorption, epithelial adhesion and virulence in Galleria mellonella.
That is important mechanistic evidence, but it remains preclinical.
The same caution applies to many celebrated examples of evolutionary steering. OMKO1 and capsule-targeting phages demonstrate that favourable evolutionary constraints can exist, but the magnitude and consistency of these effects need to be evaluated across diverse clinical isolates and infection environments.
Different bacteria exposed to the same phage may even evolve different resistance mechanisms.
One population could mutate the desired receptor. Another might modify LPS, increase extracellular polysaccharide production or activate an unrelated intracellular defence system, bypassing the intended evolutionary trade-off entirely.
Researchers therefore cannot simply choose a receptor-targeting phage and assume that bacteria will evolve along the preferred route.
Predicting resistance trajectories may eventually become an important component of precision phage therapy.
Whole-genome sequencing of the bacterial isolate can identify surface structures and antiviral defence systems. Experimental evolution can reveal likely resistance pathways. Phage susceptibility testing can establish which viruses infect the strain, while antibiotic testing of phage-resistant derivatives can uncover collateral sensitivity.
Together, these approaches could shift therapeutic selection from a simple question — “Which phage kills this bacterium?” — toward a much richer one: “Which phage kills this bacterium while leaving it with the least favourable evolutionary escape routes?”
For students encountering phage therapy for the first time, the answer to whether bacteria can become resistant to bacteriophages is therefore straightforward.
Yes.
Bacteria have been evolving against bacteriophages for billions of years and possess an extraordinary arsenal of mechanisms capable of preventing viral infection. Resistance can arise rapidly and can threaten the durability of treatment.
The more interesting answer begins after that.
Resistance is not always equivalent to therapeutic failure. If the pathway bacteria use to escape a phage simultaneously removes a virulence factor, compromises an antibiotic-resistance mechanism, exposes the cell to immunity or increases susceptibility to another phage, bacterial evolution may create a new vulnerability.
The 2026 ExPEC study provides a striking illustration. Mutations in wbbL and waaG changed LPS, impaired adsorption of phage vB_EcoP_P61425 and produced resistance. The same changes also reduced epithelial adhesion and dramatically attenuated virulence in the experimental model.
But that trade-off belongs to this particular phage–bacterium interaction. It should not become a universal claim about phage resistance.
Some resistant bacteria pay substantial biological costs. Others do not. Some become easier to treat with antibiotics or alternative phages. Others acquire resistance without an obvious therapeutic weakness. Temporary physiological tolerance adds another layer of complexity.
The future of phage therapy may therefore depend less on trying to stop evolution entirely than on learning how to anticipate and manipulate it.
A well-designed therapeutic phage would not merely kill today's bacterial population. Ideally, it would also constrain the populations that evolution is likely to create tomorrow.
Sources :
Chen L., Xu S., Li X., Li P., Su M., Liu P., Fang Y. & Zhang W. — “Spontaneous wbbL and waaG mutations selected by phage vB_EcoP_P61425 confer resistance and attenuate virulence in ExPEC.” Frontiers in Microbiology, published October 12, 2026. The study identifies wbbL and waaG mutations as causes of reduced phage adsorption and resistance and documents reduced epithelial adhesion and virulence in Galleria mellonella. Frontiers
https://doi.org/10.3389/fmicb.2026.1953743
Georjon H. & Bernheim A. — “The highly diverse antiphage defence systems of bacteria.” Nature Reviews Microbiology 21, 686–700 (2023). Overview of the rapidly expanding diversity of bacterial antiviral defence mechanisms. Nature
https://doi.org/10.1038/s41579-023-00934-x
Saxton D.S. & Laub M.T. — “How bacterial immune systems sense phage infection.” Nature Reviews Microbiology, published October 2, 2026. Reviews how bacterial immune systems detect phage nucleic acids, viral proteins and infection-induced perturbations. Nature
https://doi.org/10.1038/s41579-026-01376-x
Kortright K.E., Chan B.K., Koff J.L. & Turner P.E. — “Phage Therapy: A Renewed Approach to Combat Antibiotic-Resistant Bacteria.” Related experimental work from Turner and colleagues established the OMKO1 concept, in which phage resistance involving OprM can increase antibiotic susceptibility. The original experimental study is: Chan B.K. et al., Scientific Reports 6, 26717 (2016). PubMed Central (PMC)
https://doi.org/10.1038/srep26717
Fujiki J., Nakamura K., Nakamura T. & Iwano H. — “Fitness Trade-Offs between Phage and Antibiotic Sensitivity in Phage-Resistant Variants: Molecular Action and Insights into Clinical Applications for Phage Therapy.” International Journal of Molecular Sciences 24, 15628 (2023). Reviews evidence for virulence and antibiotic-sensitivity trade-offs while emphasizing their strain- and mechanism-specific nature. PubMed Central (PMC)
https://doi.org/10.3390/ijms242115628
Majkowska-Skrobek G. et al. — “The evolutionary trade-offs in phage-resistant Klebsiella pneumoniae entail cross-phage sensitization and loss of multidrug resistance.” Environmental Microbiology (2021). Demonstrates that resistance outcomes can include cross-phage effects and loss of antimicrobial-resistance determinants while also illustrating that not every measured fitness or virulence characteristic necessarily declines. PubMed
https://doi.org/10.1111/1462-2920.15476
Song L. et al. — “Phage Selective Pressure Reduces Virulence of Hypervirulent Klebsiella pneumoniae Through Mutation of the wzc Gene.” Frontiers in Microbiology 12, 739319 (2021). Capsule-related resistance mutations were associated with reduced virulence in experimental models. PubMed
https://doi.org/10.3389/fmicb.2021.739319
Reversible phenotypic resistance to phage infection via capsule downregulation in Klebsiella pneumoniae (2025). Demonstrates that reduced phage susceptibility can also emerge through reversible regulation rather than permanent mutation. PubMed Central (PMC)
Annual Review of Virology — “Developing Phage Therapy That Overcomes the Evolution of Bacterial Resistance.” Reviews two complementary strategies: reducing the evolution of resistance and deliberately steering resistant populations toward clinically favourable phenotypes. annualreviews.org
.png){kind=logo}
Comments
Post a Comment