Personalized, Depolymerase-Armed Bacteriophage Cocktails to Treat Carbapenem-Resistant Acinetobacter baumannii: Rapid Capsule-Typed Host-Range Matching, Phage–Antibiotic Synergy, and an eIND-Ready Translational Path

Project Summary / Abstract

Carbapenem-resistant Acinetobacter baumannii (CRAB) is a top-tier antimicrobial-resistance priority pathogen. As an opportunistic, frequently extensively drug-resistant (XDR) nosocomial organism, it causes ventilator-associated pneumonia, bloodstream infections, and wound, burn, and device-related infections in critically ill ICU patients, often leaving only colistin/polymyxin or sulbactam–durlobactam as last-line therapy — agents constrained by toxicity and emerging resistance. Lytic bacteriophages kill bacteria through a mechanism orthogonal to antibiotics, so carbapenem or colistin resistance does not confer phage resistance. Many A. baumannii phages additionally encode capsule depolymerases that enzymatically strip the protective capsular polysaccharide, dismantle biofilm, and re-sensitize the organism to antibiotics and to host serum/immune killing. The first modern Western phage-therapy success was a disseminated XDR A. baumannii infection cleared with a personalized 9-phage cocktail delivered intravenously and percutaneously, with documented additive activity between the phage cocktail and sub-lethal minocycline (Schooley et al., 2017) — establishing CRAB as the prototype indication for personalized phage cocktails.

This proposal develops and rigorously characterizes personalized, depolymerase-armed phage cocktails for CRAB across three aims. We will (1) build a capsule-typed, genomically vetted phage bank and a rapid host-range matching workflow that nominates multi-phage cocktails against patient isolates within a clinically actionable window; (2) define phage–antibiotic synergy (PAS) and phage-resistance trade-offs in vitro, in biofilm, and in a murine CRAB bloodstream model, prioritizing depolymerase-driven potentiation of colistin/polymyxin B; and (3) establish release specifications, a matching standard operating procedure (SOP), and a prospective compassionate-use case series under FDA single-patient emergency IND (eIND). Each aim carries pre-specified, quantitative go/no-go gates. The work advances the near-term goal of moving CRAB phage therapy from heroic last-resort rescue toward a stocked, antibiogram-style ICU formulary.

Specific Aims

CRAB causes life-threatening, often XDR infections in ICU patients for whom antibiotic options are nearly exhausted. Bacteriophages kill by a mechanism orthogonal to antibiotics, and many A. baumannii phages encode capsule depolymerases that strip capsular polysaccharide, dismantle biofilm, and re-sensitize the organism to antibiotics and immunity. Personalized phage cocktails have already cleared disseminated XDR A. baumannii in compassionate-use settings, with phage activity potentiated by sub-lethal antibiotic (Schooley et al., 2017; Aslam et al., 2020). Yet the field lacks standardized capsule-typed host-range matching, mechanistic characterization of phage–antibiotic synergy specific to CRAB, and a disciplined translational path. We propose three aims to close these gaps.

Aim 1 — Build a capsule-typed CRAB phage bank and a rapid host-range matching workflow. We will assemble and whole-genome–sequence a panel of lytic A. baumannii phages, annotate capsule depolymerases, and confirm absence of integrase, known toxin, and antibiotic-resistance genes. Against a contemporary CRAB clinical-isolate collection, we will capsule (K-) type isolates and run standardized host-range screening (spot assays + quantitative efficiency-of-plating, EOP) to map coverage. A matching algorithm will nominate multi-phage cocktails that maximize coverage while combining complementary receptors to suppress single-step phage-resistant escape. Go/no-go: a 2–4-phage cocktail covers ≥70% of contemporary isolates (EOP ≥0.1) [ILLUSTRATIVE threshold].

Aim 2 — Define phage–antibiotic synergy and phage-resistance trade-offs. Using matched cocktails from Aim 1, we will quantify, in vitro (planktonic + biofilm) and in a murine CRAB bloodstream model, whether depolymerase-armed phages potentiate colistin/polymyxin B and sulbactam–durlobactam, and whether phage-resistant escape forces fitness-costly capsule loss that re-sensitizes the organism to serum and antibiotics. Go/no-go: ≥1 cocktail–antibiotic pair achieves synergy (FIC index ≤0.5 or ≥2-log time-kill advantage over the better single agent) and a statistically significant survival or bacterial-burden benefit in vivo [ILLUSTRATIVE].

Aim 3 — Establish the translational and regulatory framework for a compassionate-use case series. We will define release specifications (titer, sterility, endotoxin limits), a host-range-matching SOP, and a prospective clinical data-capture protocol, then treat a small number of CRAB-infected patients under FDA single-patient eIND with IRB oversight — capturing safety, microbiologic clearance, clinical response, and serial isolates for emergent-resistance monitoring. Deliverable gate: a finalized, reproducible eIND/matching package usable to power a subsequent controlled trial, independent of accrual rate.

Impact: By coupling rapid capsule-typed matching with mechanistically defined phage–antibiotic synergy and a disciplined eIND pathway, this work lays the evidentiary and operational foundation for CRAB phage therapy to graduate from last-resort rescue toward a stocked ICU formulary.

Significance

CRAB is among the highest-priority antimicrobial-resistance threats recognized by the WHO and CDC. It is an opportunistic, frequently XDR nosocomial pathogen responsible for ventilator-associated pneumonia, bloodstream infections, wound and burn infections, and device-related infections in ICU patients, where therapeutic options are often reduced to colistin/polymyxin or sulbactam–durlobactam. Both last-line strategies face dose-limiting toxicity and emerging resistance, defining an urgent unmet need squarely within NIAID's antimicrobial-resistance remit and directly relevant to combat-wound Acinetobacter of interest to the DoD.

Bacteriophages are uniquely suited to this target for three reasons. First, lytic phages kill through a mechanism orthogonal to antibiotics, so resistance to carbapenems or colistin does not confer phage cross-resistance. Second, many A. baumannii phages encode capsule depolymerases that degrade capsular polysaccharide — both enabling phage adsorption and dismantling biofilm that shields bacteria on catheters, ventilators, and abscess walls. A phage-derived depolymerase (Dpo71) has been shown to strip the A. baumannii capsule, destabilize the outer membrane, potentiate colistin, and re-sensitize MDR strains to serum/host-immune killing (Chen et al., 2022) — the mechanistic crux of our adjuvant strategy. Third, antibiotic potentiation of phage is clinically observed: the index disseminated XDR A. baumannii case documented additive activity between the personalized phage cocktail and sub-lethal minocycline (Schooley et al., 2017).

The clinical proof of concept is established but narrow. The 2016 UC San Diego case used IV and percutaneous personalized 9-phage cocktails (Navy/DoD-derived) to clear a disseminated XDR A. baumannii infection (Schooley et al., 2017). UC San Diego's IPATH subsequently reported its first 10 consecutive IV phage cases, of which 2 were A. baumannii (one full recovery, one uninterpretable owing to comorbidity), with phage resistance emerging in 3 of 10 cases and overcome by introducing new phages, and phage-neutralizing serum noted in 1 case (Aslam et al., 2020). Most recently, a single CRAB-bloodstream-infection phage (vB_AbaP_CV1) combined with polymyxin B/colistin showed checkerboard and time-kill synergy and superior outcomes versus monotherapy in a BALB/c tail-vein (bloodstream) model (Wang et al., 2026), directly validating phage+polymyxin combinations in vivo for CRAB bacteremia.

Despite these advances, CRAB phage therapy remains predominantly investigational/compassionate-use under single-patient eIND, and no controlled efficacy trial specific to CRAB has reported. The field needs standardized capsule-typed host-range matching, rigorous mechanistic characterization of phage–antibiotic synergy, and a disciplined translational path. This proposal addresses each, generating the data and operational infrastructure to advance CRAB phage therapy toward a controlled, formulary-style intervention.

Innovation

This proposal is innovative in four respects, each tied to a specific published precedent.

• Rapid, capsule-typed host-range matching as a standardized workflow. We operationalize K-type–informed matching as a diagnostic-to-therapeutic pipeline returning a candidate cocktail within hours of a positive culture, rather than the ad hoc, weeks-long rescue procedures behind prior compassionate-use successes (Schooley et al., 2017; Aslam et al., 2020).
• Depolymerase-armed phages as the mechanistic centerpiece. We deliberately exploit capsule degradation to dismantle biofilm and drive synergy with colistin/polymyxin, extending the demonstration that a depolymerase strips the capsule, potentiates colistin, and restores serum killing (Chen et al., 2022) from a single isolated enzyme to a banked, phage-delivered, cocktail-scale strategy.
• Phage-resistance evolution as a therapeutic lever. Because phage-resistant escape often forces A. baumannii to shed or alter its capsule at a fitness cost, the design combines complementary-receptor phages with antibiotics to convert resistance into an exploitable trade-off (re-sensitization to serum/colistin; Chen et al., 2022), rather than treating resistance only as a failure mode (Aslam et al., 2020).
• Mechanistic science fused to a disciplined eIND path. Release specifications, the matching SOP, and combination regimens are defined prospectively — anchored to an in vivo phage+polymyxin efficacy benchmark for CRAB bacteremia (Wang et al., 2026) — rather than reconstructed after the fact.

Approach

Overarching design & rigor. All in vitro endpoints use biological and technical replicates with pre-registered analysis plans; synergy is defined a priori (fractional inhibitory concentration index ≤0.5; or ≥2-log10 CFU reduction over the more active single agent in time-kill). In vivo studies use predetermined group sizes from power analysis (Aim 2), randomization, and blinded enumeration. Each Aim carries explicit go/no-go gates (Specific Aims). Sex as a biological variable will be incorporated in animal studies. We report negative and isolate-/K-type–dependent results.

Aim 1 — Capsule-typed CRAB phage bank and rapid host-range matching workflow

Rationale. Therapeutic A. baumannii phages are highly strain-specific, frequently binding the capsule (K-type), so cocktails must be matched to a patient's isolate and combined to broaden coverage and suppress escape mutants (Schooley et al., 2017; Aslam et al., 2020). A standardized, genomically vetted bank is the prerequisite for everything downstream.

Experimental design. We will isolate and acquire lytic A. baumannii phages, propagate them on safe production hosts, and perform whole-genome sequencing to (i) annotate capsule-depolymerase genes, (ii) confirm strictly lytic lifestyle, and (iii) exclude integrase, known toxin, and antibiotic-resistance genes. A contemporary CRAB clinical-isolate collection will be capsule (K-) typed (genomic K-locus typing) and screened by standardized spot assays and quantitative EOP to build a phage×isolate coverage matrix. A transparent matching algorithm will nominate 2–4-phage cocktails that maximize coverage while requiring ≥2 distinct receptor classes per cocktail to limit single-step resistance. We will pilot a rapid-turnaround path (isolate in → candidate cocktail out) and benchmark the wall-clock time.

Expected outcomes. A genomically characterized, capsule-typed phage bank; a validated matching workflow; and coverage statistics across contemporary CRAB isolates, including the fraction matchable by a 2–4-phage cocktail and the turnaround time achieved.

Potential pitfalls & alternatives. Some isolates may be unmatchable with the existing bank; we will expand by targeted isolation against gap isolates and, where lytic coverage fails, prioritize depolymerase candidates as enzybiotic adjuvants (Chen et al., 2022). If lytic activity is weak, sub-lethal antibiotic conditioning (Aim 2) may enhance propagation. If K-typing throughput limits screening, we will triage by prevalence-weighted K-types first.

Aim 2 — Phage–antibiotic synergy and phage-resistance trade-offs

Rationale. Capsule depolymerase activity destabilizes the outer membrane and improves colistin binding and serum killing (Chen et al., 2022); sub-lethal antibiotic can potentiate phage activity (Schooley et al., 2017); and phage+polymyxin combinations reduce CRAB burden in vivo (Wang et al., 2026). Combinations therefore plausibly outperform either agent alone. Separately, phage-resistant escape frequently incurs a capsule-associated fitness cost that re-sensitizes the strain to serum and antibiotics — a trade-off worth quantifying.

Experimental design. Using matched cocktails from Aim 1, we will run checkerboard and time-kill assays combining phages with colistin/polymyxin B and with sulbactam–durlobactam against representative CRAB isolates spanning multiple K-types, including static and dynamic biofilm models on abiotic surfaces. We will isolate phage-resistant mutants, characterize capsule status (genotype + phenotype), and assay their susceptibility to serum killing and to antibiotics to test the re-sensitization hypothesis (Chen et al., 2022). Leading combinations will be evaluated in a murine CRAB bloodstream/systemic challenge model, comparing phage alone, antibiotic alone, and the combination on bacterial burden and survival — directly mirroring and extending the BALB/c tail-vein phage+polymyxin benchmark (Wang et al., 2026).

Expected outcomes. Quantitative synergy profiles (FIC, time-kill) identifying combinations that beat monotherapy; evidence that depolymerase-armed phages potentiate polymyxins and disrupt biofilm; and a determination of whether phage resistance forces capsule loss with antibiotic/serum re-sensitization.

Potential pitfalls & alternatives. Synergy may be isolate- and K-type-dependent; we will test across multiple K-types and report the conditions under which synergy holds rather than a single point estimate. If in vivo phage neutralization or pharmacokinetics limit monotherapy efficacy — as anticipated from neutralizing-serum observations in compassionate use (Aslam et al., 2020) — we will emphasize combination regimens where antibiotics suppress residual escape mutants, and adjust dose/route. Animal-model variability will be mitigated by power-based group sizes, randomization, blinding, and predefined humane endpoints.

Aim 3 — Translational and regulatory framework for a compassionate-use case series

Rationale. Clinical CRAB phage therapy remains predominantly investigational/compassionate-use under single-patient eIND (Schooley et al., 2017; Aslam et al., 2020). A disciplined framework — defined specifications plus prospective data capture — is needed to convert anecdotal rescue into trial-enabling evidence.

Experimental design. We will define release specifications (titer, sterility, endotoxin limits), a host-range-matching SOP linking the patient isolate to a selected cocktail, and a clinical data-capture protocol (safety, microbiologic clearance, clinical response). Working with the IRB and FDA, we will prospectively treat a small number of CRAB-infected patients under single-patient eIND, with cocktails matched per Aim 1 and, where appropriate, combined with antibiotics informed by Aim 2 — mirroring the personalized IV/percutaneous approach of the index case (Schooley et al., 2017) and the single-center IV case series reported by IPATH (Aslam et al., 2020). Serial isolates will be banked to monitor microbiologic response and emergent phage resistance, which arose in 3 of 10 prior IV cases (Aslam et al., 2020).

Expected outcomes. A documented, reproducible eIND pathway and matching SOP; preliminary safety and microbiologic/clinical outcome data from a CRAB case series; and a protocol scaffold to power a future controlled trial.

Potential pitfalls & alternatives. Compassionate-use cohorts are heterogeneous, concurrent antibiotics confound attribution, and some outcomes are uninterpretable — as IPATH explicitly observed (Aslam et al., 2020). We will pre-specify microbiologic endpoints, collect serial isolates, and treat clinical response as supportive rather than primary. If clinical accrual is slow, the finalized specifications and SOP remain a stand-alone, trial-de-risking deliverable. Manufacturing constraints will be addressed by prioritizing bank phages with robust, well-characterized production.

Timeline

Year 1 [ILLUSTRATIVE]: Assemble and sequence the phage bank; annotate depolymerases; capsule-type the isolate collection; establish host-range screen (Aim 1). Years 2–3 [ILLUSTRATIVE]: Finalize matching workflow and coverage benchmarking (Aim 1); execute in vitro synergy, biofilm, and resistance-trade-off studies (Aim 2). Years 3–4 [ILLUSTRATIVE]: Murine combination-efficacy studies (Aim 2); finalize release specifications, matching SOP, and eIND/IRB framework (Aim 3). Years 4–5 [ILLUSTRATIVE]: Prospective compassionate-use case series under eIND; analysis and trial-protocol scaffolding (Aim 3). Go/no-go gates are evaluated at the end of Years 1 (Aim 1 coverage), 3 (Aim 2 synergy/in vivo), and 4 (eIND package).

Budget Justification (modular R01-style sketch)

This is a modular R01 request of approximately $250,000 direct costs per year [ILLUSTRATIVE] for 5 years [ILLUSTRATIVE]. Personnel: PI (microbiology/phage biology) at 2.4 person-months [ILLUSTRATIVE]; Co-Investigators in infectious diseases and clinical microbiology (1.2 person-months each [ILLUSTRATIVE]); one postdoctoral scientist and one research technologist (12 person-months each [ILLUSTRATIVE]) for phage isolation, genomics, and assays; partial study-coordinator effort in Years 4–5 [ILLUSTRATIVE] for the case series. Supplies/services: whole-genome sequencing, media and consumables, antibiotics (colistin, polymyxin B, sulbactam–durlobactam), biofilm-assay materials, and endotoxin/sterility testing. Animal costs: per-diems and procedures for the murine CRAB bloodstream model (Aim 2). Other: GMP-aligned production/quality testing for clinical-grade cocktails and regulatory/eIND submission costs (Aim 3). Final figures are placeholders to be set with institutional budgeting.

Vertebrate Animals

Animal work is proposed in Aim 2. We will use a murine CRAB bloodstream/systemic challenge model to compare phage alone, antibiotic alone, and combination regimens on bacterial burden and survival. Justification: in vivo testing is required to establish phage–antibiotic synergy and pharmacologic relevance not captured in vitro; a BALB/c tail-vein CRAB model has demonstrated superior outcomes for phage+polymyxin B combinations versus monotherapy (Wang et al., 2026), providing a validated benchmark for our design. (Mechanistic depolymerase/colistin re-sensitization data to date derive from a Galleria mellonella invertebrate model and in vitro serum-killing assays (Chen et al., 2022); the murine model is therefore needed to extend those findings to a mammalian host.) Group sizes will be the minimum required for statistical rigor, set by formal power analysis, with randomization, blinded enumeration, predefined humane endpoints, and analgesia/anesthesia and euthanasia per approved IACUC protocol. Specific species/strain, numbers, and endpoints will be finalized with the IACUC. Approximate animal numbers and group sizes are [ILLUSTRATIVE] until protocol approval.

Human Subjects / Clinical Trial

Aim 3 involves human subjects in a small prospective compassionate-use case series. Investigational phage cocktails will be administered under the FDA single-patient emergency IND (eIND) pathway — the current regulatory route for individualized phage therapy in the United States — with full IRB oversight, informed consent, and predefined safety and microbiologic/clinical endpoints. This personalized, isolate-matched approach follows the index UC San Diego case (Schooley et al., 2017) and the IPATH single-center IV case series (Aslam et al., 2020). Enrollment is anticipated to be a small number of CRAB-infected patients [ILLUSTRATIVE]; because phage resistance emerged in 3 of 10 prior IV cases and treatment failure can occur despite in vitro susceptibility (Aslam et al., 2020), serial isolates will be collected to monitor microbiologic response and emergent resistance, and concurrent-antibiotic confounding will be documented in analysis. This work is intended to inform, not substitute for, a future controlled trial.

Team & Environment

• Principal Investigator [NAME, INSTITUTION] — phage biology/microbiology; overall direction, bank development, and genomics (Aim 1).
• Co-Investigator, Infectious Diseases [NAME, INSTITUTION] — clinical lead for eIND/IRB strategy and the case series (Aim 3).
• Co-Investigator, Clinical Microbiology [NAME, INSTITUTION] — capsule typing, host-range screening, susceptibility testing.
• Co-Investigator, Antimicrobial Pharmacology / In Vivo Models [NAME, INSTITUTION] — synergy assays and murine efficacy (Aim 2).
• Collaborators / Resources [TO FILL] — established US phage-therapy infrastructure and DoD-derived phage expertise (e.g., IPATH at UC San Diego; Walter Reed Army Institute of Research / Naval Medical Research Center phage program) and a clinical-grade phage-banking partner. Environment: BSL-2 microbiology and genomics cores, an AAALAC-accredited animal facility, a CLIA clinical-microbiology laboratory, and institutional IRB/IACUC and regulatory-affairs support.

Preliminary data: this proposal is written as concept-stage; where the assembled team holds pilot phage-isolation, K-typing, or synergy data, it should be inserted to strengthen feasibility for each Aim [ILLUSTRATIVE].

References

1. Schooley RT, Biswas B, Gill JJ, et al. Development and Use of Personalized Bacteriophage-Based Therapeutic Cocktails To Treat a Patient with a Disseminated Resistant Acinetobacter baumannii Infection. Antimicrobial Agents and Chemotherapy. 2017;61(10):e00954-17. PMID: 28807909. https://pubmed.ncbi.nlm.nih.gov/28807909/
2. Aslam S, Lampley E, Wooten D, et al. Lessons Learned From the First 10 Consecutive Cases of Intravenous Bacteriophage Therapy to Treat Multidrug-Resistant Bacterial Infections at a Single Center in the United States. Open Forum Infectious Diseases. 2020;7(9):ofaa389. DOI: 10.1093/ofid/ofaa389. https://doi.org/10.1093/ofid/ofaa389
3. Chen X, Liu M, Zhang P, et al. Phage-Derived Depolymerase as an Antibiotic Adjuvant Against Multidrug-Resistant Acinetobacter baumannii. Frontiers in Microbiology. 2022;13:845500. PMID: 35401491. DOI: 10.3389/fmicb.2022.845500. https://pubmed.ncbi.nlm.nih.gov/35401491/
4. Wang L, Wang F, Yuan Z, Liu Y, Jing Y, Xing J. Carbapenem-resistant Acinetobacter baumannii bloodstream infections and specific phages: isolation, analysis and application. Frontiers in Cellular and Infection Microbiology. 2026;16:1774993. PMID: 41994204. DOI: 10.3389/fcimb.2026.1774993. https://pubmed.ncbi.nlm.nih.gov/41994204/