npj Viruses Opens 2026 Call for Papers on Phage Therapy

Phage therapy has spent much of its modern revival suspended between two realities. The first is biological. Bacteriophages can recognise, infect and eliminate bacterial pathogens with a degree of strain level specificity that conventional antibiotics rarely achieve. The second is translational. A phage that produces clear plaques in the laboratory is not automatically a medicine. Between those two realities lies an increasingly complex chain of microbiological qualification, genomic analysis, manufacturing, formulation, clinical decision making, regulation and long term patient monitoring.

A new collection launched by npj Viruses places that translational chain at the centre of its scientific scope. Entitled “Phages as antimicrobial therapy”, the collection is currently open for submissions until 30 November 2026 and is edited by Ruby CY Lin, Sabrina Green and Renee Ng. Rather than limiting the discussion to the antibacterial activity of phages, the collection seeks research capable of explaining how phage biology can be converted into reproducible, quality controlled and clinically accessible therapeutic products.

The timing is significant. Antimicrobial resistance is increasing the clinical importance of alternatives to conventional antibiotics, yet phage therapy still operates through a mixture of compassionate treatment programmes, academic manufacturing centres, early clinical trials and a relatively small number of industrial development pipelines. The field has demonstrated that personalised or engineered phages can sometimes be administered safely in highly complex infections. However, it has not yet established a universally accepted model for moving from bacterial isolation to phage selection, pharmaceutical production and routine treatment within the short time available for a seriously ill patient.

A therapeutic phage is more than a bacterial predator

The precision of phage therapy is frequently reduced to the concept of host specificity. In practice, therapeutic precision involves several biological layers. A phage must first adsorb to a receptor that is sufficiently expressed under conditions resembling the infection environment. It must then inject its genome, escape or overcome intracellular bacterial defence systems, redirect bacterial metabolism, produce viable progeny and complete cell lysis. Failure at any of these stages may create an apparent discrepancy between a standard susceptibility assay and the actual therapeutic response.

Host range therefore cannot be defined adequately by a spot assay alone. Clearing at a high phage concentration may result from productive infection, but it may also reflect lysis from without, local enzymatic activity or abortive infection without meaningful phage amplification. A translationally relevant characterisation strategy should combine efficiency of plating, adsorption measurements, liquid culture killing kinetics, replication parameters, resistance frequency and genomic analysis of both the phage and bacterial host.

The physiological state of the bacterium must also be considered. Bacteria growing in biofilms, mucus, necrotic tissue or oxygen limited compartments may express a different receptor landscape from rapidly growing laboratory cultures. Thick and mature biofilms can restrict phage diffusion, modify bacterial metabolism and generate spatially heterogeneous populations containing susceptible, transiently tolerant and genetically resistant cells. Recent experimental work with skin associated pathogens has shown that activity observed against planktonic bacteria can decline markedly when the same organisms are embedded in mature biofilms, reinforcing the importance of combining biological treatment with drainage, debridement or matrix disrupting approaches when clinically appropriate.

Manufacturing must be treated as part of the science

One of the most important aspects of the npj Viruses collection is its emphasis on manufacturing, purification, formulation and stability. These subjects are sometimes presented as technical steps that follow phage discovery. In reality, they can determine whether a promising phage ever reaches the patient.

Therapeutic development requires controlled phage and bacterial seed stocks, traceable production procedures and assays capable of confirming identity, biological potency and purity. The propagation host must be carefully selected and characterised because it contributes biological material to the crude lysate. Depending on the bacterial species and manufacturing process, purification may need to remove endotoxins, bacterial proteins, nucleic acids, membrane fragments, media components and process related contaminants while preserving infectious phage particles.

The problem becomes more complex when a cocktail contains phages with different structures and physicochemical properties. A purification step that produces a high recovery for one phage may reduce the titre of another. Similarly, exposure to changes in pH, ionic strength, temperature, agitation, freezing, thawing or membrane filtration can affect individual phages differently. Formulation development must therefore consider each component rather than treating all bacteriophages as interchangeable biological particles.

The intended route of administration adds another layer of constraint. An intravenous product must satisfy stringent requirements for sterility, endotoxin control, particulate matter and systemic tolerability. An inhaled formulation must preserve phage activity during aerosol generation and passage through the respiratory tract. An oral formulation must protect phages from gastric acidity and digestive conditions. A topical product must remain active in the presence of wound exudate, proteases, biofilm material and variable local pH. Stability is consequently not a single property of a phage genome. It emerges from the interaction between the virion, formulation, container, storage environment and delivery device.

From genomic selection to molecular engineering

Genome sequencing has become indispensable for therapeutic phage qualification. Candidate phages are generally evaluated for genes associated with lysogeny, bacterial virulence, antimicrobial resistance and other functions that could create unacceptable risks. Yet sequence screening is not infallible. Many phage open reading frames remain poorly annotated, and the absence of a recognised undesirable gene does not prove the absence of an undesirable biological function. Genomic analysis must therefore be integrated with phenotypic testing, lifecycle characterisation and assessment of genetic stability.

Molecular engineering expands what can be achieved when naturally isolated phages do not possess the required properties. Engineering may be used to eliminate lysogeny functions, modify receptor binding structures, increase activity against a clinical isolate or introduce antibacterial genetic payloads. A widely discussed clinical example involved a patient with disseminated, drug resistant Mycobacterium abscessus infection who received a three phage combination containing engineered derivatives. The treatment was well tolerated and was associated with objective clinical improvement, illustrating both the potential and the logistical complexity of highly personalised phage development.

Engineering nevertheless creates new questions. A modified phage requires confirmation that the intended genetic change is stable during production and administration. Its phenotype must remain consistent across manufacturing batches. Off target activity, recombination potential and the consequences of introduced genetic material must be evaluated. Regulatory frameworks will also need to distinguish between naturally isolated phages, adapted phages, engineered phages and interchangeable cocktail components without creating procedures so rigid that personalised treatment becomes impossible.

Using bacterial evolution rather than merely reacting to it

Resistance to phages is often presented as a weakness of phage therapy. It is certainly a major risk, but resistance does not always return the bacterium to its original clinical state. When a phage uses a virulence factor, capsule component, pilus, outer membrane protein or efflux associated structure as its receptor, bacterial escape may carry a significant fitness cost.

Experiments with the Pseudomonas aeruginosa phage OMKO1 demonstrated that selection for phage resistance could alter the OprM component of multidrug efflux systems and increase bacterial susceptibility to several antibiotics. This concept, often described as phage steering, reframes the therapeutic objective. The purpose of a phage may not always be to eradicate the entire bacterial population independently. It may instead force the pathogen into an evolutionary state that is less virulent, less fit or more susceptible to another treatment.

Such effects cannot be assumed for every phage and bacterial strain. They must be measured through longitudinal isolation, whole genome sequencing, receptor analysis, antimicrobial susceptibility testing and competitive fitness experiments. The most informative studies will examine not only whether resistance appears, but also which bacterial functions are altered and whether those alterations improve or compromise the overall treatment strategy.

Clinical trials adapted to a replicating antimicrobial

Conventional pharmacology was largely developed for molecules whose concentration declines after administration. Phages do not always follow this model. They may be cleared by the immune system, trapped within tissues or inactivated before reaching their target. Under favourable conditions, however, they may also replicate locally in susceptible bacteria. Their exposure is therefore influenced by the initial dose, bacterial density, spatial distribution of infection, adsorption rate, burst size, immune clearance and the emergence of resistance.

This unusual pharmacology complicates the interpretation of dose and response. A high administered titre does not guarantee sufficient activity at the infection site, while a modest initial dose could theoretically expand in a bacterial population that supports productive infection. Future clinical studies must connect pharmacokinetic observations with microbiological data rather than treating phage concentration and bacterial burden as independent variables.

Patient selection is equally important. Enrolling patients according to bacterial species alone may dilute a therapeutic signal if the infecting isolates differ substantially in susceptibility. Trials may require rapid phage susceptibility testing before inclusion, together with predefined procedures for changing cocktail composition when resistance or bacterial replacement occurs. The design must also account for antibiotics, surgery, drainage and other elements of standard care.

The recent randomised study of the TP 102 cocktail in diabetic foot ulcers illustrates both the progress and limitations of the current evidence base. Nineteen participants were enrolled, thirteen received the phage product, and no treatment related adverse events were reported. The study produced encouraging microbiological and wound healing observations, but it was not sufficiently powered to establish clinical superiority. Such studies are valuable because they generate practical information about administration, susceptibility testing, safety monitoring and endpoint selection that larger efficacy trials will need.

Clinical evaluation should also follow what happens after the initial treatment period. Relevant measurements may include the persistence of the bacterial pathogen, the appearance of phage resistant variants, changes in antibiotic susceptibility, microbiome disruption, phage neutralising activity and recurrence of infection. Patient experience should not be considered secondary. Access delays, repeated sampling, uncertainty around experimental treatment and coordination between distant laboratories can shape whether a theoretically personalised therapy remains clinically realistic.

Omics, machine learning and the problem of meaningful prediction

The collection explicitly encourages work using metagenomics, transcriptomics, systems biology and machine learning. These approaches could transform phage selection, but only if prediction remains connected to experimentally validated biology.

Metagenomics can expand access to uncultivated viral diversity and reveal candidate phages associated with clinically important bacterial hosts. Transcriptomics can determine how bacterial and phage gene expression changes during infection, antibiotic exposure, biofilm growth or immune pressure. Comparative genomics can identify receptor binding modules and defence evasion functions. Machine learning may help connect genomic features with host range, stability or therapeutic performance.

The central challenge is the quality of the training data. Phage databases contain heterogeneous host annotations, incomplete genomes and susceptibility results generated with different laboratory methods. A model trained on binary spot test data may learn patterns that do not predict productive infection. Computational predictions should therefore be tested against standardised quantitative assays and clinically representative isolate collections. The value of artificial intelligence in phage therapy will depend less on the complexity of an algorithm than on the biological quality and interoperability of the data used to train it.

Making phage therapy accessible beyond exceptional cases

The final transition is not only scientific or regulatory. It is economic. Personalised phage therapy can require bacterial isolation, susceptibility testing, phage screening, sequencing, manufacturing, quality control and clinical approval within a narrow therapeutic window. Maintaining large phage banks and specialised production facilities also generates costs even when individual phages are not immediately required.

Health economic research must determine which parts of this infrastructure should be centralised, which can be distributed through hospital networks and how the costs of rapid personalised production can be reimbursed. The answer may differ between emergency compassionate treatment, fixed cocktails for common pathogens and engineered products intended for conventional marketing authorisation.

Access must also extend beyond a small group of highly connected academic hospitals. Sustainable deployment will require shared quality standards, interoperable phage banks, harmonised susceptibility methods and referral pathways that allow clinicians to identify available phages rapidly. The inclusion of animal health within the collection is similarly important. Phage interventions in veterinary medicine and food production could reduce antibiotic use, but they also raise distinct questions concerning environmental release, resistance ecology, manufacturing scale and economic feasibility.

A collection centred on the difficult middle

The scientific significance of “Phages as antimicrobial therapy” lies in its focus on the difficult middle ground between phage discovery and therapeutic access. The field no longer needs to prove only that bacteriophages can kill bacteria. It must determine how to select them intelligently, manufacture them consistently, administer them effectively, monitor their evolution and make them available within real health systems.

By bringing molecular biology, bioinformatics, pharmaceutical development, clinical research, regulation and health economics into the same collection, npj Viruses is inviting a more mature view of phage therapy. The future of the field will not be secured by one spectacular clinical case, one broad host range cocktail or one regulatory decision. It will depend on whether the global phage community can build a reproducible therapeutic ecosystem in which biological precision is supported by pharmaceutical quality, clinical evidence and equitable access.

The submission deadline for the collection is 30 November 2026. Researchers working on therapeutic phage biology, manufacturing, molecular engineering, formulation, analytical characterisation, clinical development, regulation, patient pathways or health economics can submit original work through the official npj Viruses collection page.

Sources :

The official call and submission information are available on the npj Viruses collection page.

Scientific background is supported by the studies on engineered phages against Mycobacterium abscessus, phage driven antibiotic resensitisation in Pseudomonas aeruginosa, TP 102 in diabetic foot ulcers and phage activity in mature bacterial biofilms.