Pseudomonas aeruginosa is a Gram-negative facultative aerobic bacterium that is ubiquitous in nature and frequently associated with hospital-acquired infections. The National Nosocomial Infections Surveillance System reported 21% of pneumonia, 10% of urinary tract infections, 3% of bloodstream infections, and 13% of ear, eye, nose, and throat infections detected within intensive care units (ICUs) in the United States were caused by P. aeruginosa [1]. P. aeruginosa infection rates have been steadily increasing [2], a phenomenon that may be linked to risk factors including increases in the use of medical devices, chronic disease burden, and number of immunocompromised individuals [3,4,5]. This increased incidence has also been linked to lapses in infection control measures, insufficient environmental testing, and increased prevalence of multi-drug resistant isolates in hospital settings [1, 6]. Consequently, P. aeruginosa is the leading cause of patient mortality in hospital-acquired pneumonia and respiratory failure [3]. It poses a major threat to people with cystic fibrosis and immunocompromised individuals [6]. P. aeruginosa has been designated by the World Health Organization as an ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, P. aeruginosa, and Enterobacter species) pathogen; a group of bacterial pathogens of global health concern for which new antimicrobial development is urgently needed [7].

P. aeruginosa has proven difficult to treat due to its many intrinsic and acquired mechanisms of antibiotic resistance [8]. The global dissemination of high-risk clones of multidrug-resistant or extensively drug-resistant (MDR/XDR) P. aeruginosa poses a significant public health concern [9, 10]. P. aeruginosa bloodstream infections exhibit high morbidity and mortality rates ranging from 43.2% to 58.8%, with MDR strains complicating 16.7% to 28% of cases and contributing to increased mortality [11, 12]. In burn patients, MDR P. aeruginosa is an important cause of death, accounting for 86% of sepsis-related deaths in pediatric burn ICUs, with P. aeruginosa identified as the responsible organism in 64% of cases from 1999 to 2009 [13]. Moreover, therapeutic treatment options are further hindered by the propensity of P. aeruginosa to form multicellular biofilms, which lead to colonization and persistence in unfavorable environmental conditions [3]. Given that there are currently no licensed vaccines against P. aeruginosa, alternative therapeutic avenues must be further explored.

Antibody-based strategies such as passive administration of monoclonal antibodies (mAbs) have proven to be effective for preventing and treating a wide array of infectious diseases [14, 15]. Moreover, with the increase in antibiotic resistance, mAbs are gaining traction in anti-infective drugs. They are becoming more widely accepted as a vital tool in our armamentarium for treating and preventing bacterial infections [16, 17]. Currently, the majority of mAbs used in a clinical setting target bacterial exotoxins; however, the elucidation of new mechanisms of antibody-mediated action against bacteria has led to the identification of novel protein targets on the surface of bacteria that can be used for mAb therapy. Furthermore, the advent of bispecific antibody technology, which involves engineering antibodies with two binding sites, either to two distinct antigens or to two epitopes on the same antigen, has greatly expanded the therapeutic potential of antibodies [18,19,20]. While mAb-based therapies have proven to be safe and effective, there are some drawbacks to this approach, which include the limited half-life of mAbs in circulation, the relatively large doses required (in the range of mg/kg) and for some conditions, the need for repeat dosing, not to mention the significant costs associated with production and manufacturing of mAbs. One way to circumvent some of these challenges is through viral vector-based delivery of mAb genes, which can produce and secrete mAbs into the bloodstream for extended periods [21,22,23].

Adeno-associated virus (AAV) vectors are currently one of the most effective viral vector platforms for in vivo gene therapy due to their strong safety profile, superior transduction efficiency, and ability to mediate sustained transgene expression [24]. AAV-vectored expression of mAb genes, also known as AAV-vectored immunoprophylaxis (AAV-VIP), has been shown to protect against a wide variety of infectious diseases [25,26,27,28] in a range of animal models, including nonhuman primates [29], and is now being tested in two human clinical trials for human immunodeficiency virus. Our AAV-mAb expression platform utilizes a novel, rationally engineered AAV6 mutant capsid, termed AAV6.2FF, which is highly efficacious as a prophylactic against Ebola virus (EBOV) [24], Marburg virus [21, 30], respiratory syncytial virus [31], and Clostridioides Difficile toxin challenge [27], among other infectious diseases.

In this study, we investigated whether AAV-VIP could protect against P. aeruginosa infection in a prophylactic setting. Previously characterized mAbs V2L2MD [32], Cam-003 [33], and the bispecific monoclonal antibody MEDI3902 targeting the Psl exopolysaccharide and PcrV protein in the type III secretion system (T3SS) of P. aeruginosa were vectorized and evaluated in a murine model of lethal P. aeruginosa pneumonia. We demonstrate that AAV-mediated expression of anti-P. aeruginosa mAbs not only prevented morbidity and mortality but also greatly reduced bacterial burden in the lungs and prevented dissemination to other organs, highlighting the potential of AAV-VIP for protection against bacterial infections.

Materials and methods

Construction of AAV vector genomes

All mAb genes were expressed using the bicistronic expression cassette developed by Fang et al. [34] and optimized by Balazs et al. [25], including a CASI promoter, heavy and light chains separated by a furin-2A (F2A) cleavage site, a woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) and an SV40 polyA signal between AAV2 inverted terminal repeats (ITR). Monoclonal monospecific IgG mAbs were synthesized to encode mouse-human codon-optimized human IgG1 (hIgG1) heavy chain and lambda light chain constant domains and the variable heavy and light chain domains of V2L2MD (αPcrV) and Cam-002 (αPsl) (WO2020041540A1). The bispecific mAb, MEDI3902 (WO2017095744A1), contains the mouse-human codon optimized Fab domain (VH and CH1) of αPcrV and the scFv of αPsl linked to the CH2 and CH3 domains of human IgG1. An F2A self-cleaving peptide separated the heavy and light chain domains, the latter comprising the variable light and kappa light chain domains of αPcrV. Due to the large size of the bispecific mAb gene, the WPRE was replaced with the smaller WPRE3 (248 bp) [35] or removed entirely. AAV6.2FF vectors were produced in adherent HEK293 cells (ATCC CRL-1573), affinity purified using heparin columns (Cytiva HiTrap Heparin Column, 17040703), and titered by qPCR as described [36].

Recombinant monoclonal antibody production and purification

Recombinant hIgG1 mAbs were produced in HEK293 cells (10 × 10 cm dishes) transfected with AAV-mAb genomes as previously described [21]. Five days post transfection, the supernatant was harvested, and filtered through a 0.45 µm filter, and the antibody was purified using a HiTrap Protein G HP column (17040503). Using a peristaltic pump, the HiTrap Protein G HP column was rinsed with 30 mL of distilled water and equilibrated with 50 mL of binding buffer (20 mM sodium phosphate, pH 7.0). The clarified supernatant was flowed over the column at a rate of 1–2 drops/s. The column was then washed with 50 mL of binding buffer and eluted with 25 mL of elution buffer (0.1 M glycine-HCl, pH 3). Eluate was collected in 5 mL fractions into tubes containing 0.5 mL of neutralizing buffer (1 M Tris-HCl, pH 8). Antibody was pooled and titered using a human IgG (hIgG) ELISA (Abcam 195125, Cambridge, MA, USA) per manufacturer’s instructions.

AAV in vitro transduction assay and flow cytometry

HEK293 cells (5 × 105) were seeded into 6-well tissue culture plates and allowed to adhere overnight. The next day, media was removed and 1 mL of inoculum containing basal DMEM plus AAV6.2FF vectors at an MOI of 5000 was added to the wells in triplicate and incubated at 37 °C, 5% CO2 for 2–4 h with occasional rocking. Two hours later, 1 mL of DMEM + 15% Cosmic Calf serum (Fisher) was added to the wells. At 72 h post transduction, cells were imaged using an inverted fluorescent microscope and subsequently trypsinized and neutralized with DMEM + 7.5% Cosmic Calf serum. Cells were pelleted at 4000 rpm and washed 2× with phosphate-buffered saline (PBS). Cells were resuspended in FACS buffer (2% fetal bovine serum (FBS)-PBS), incubated with 7AAD viability dye, and analyzed for GFP expression using the FACS Canto II (BD Biosciences, ON, Canada). Data was analyzed using FlowJo Software version 10 (FlowJo LLC, Ashland, Oregon, USA).

Animal ethics statement

All experiments involving mice were approved by the University of Guelph Animal Care Committee (AUP #4664) and conducted in accordance with the Canadian Council on Animal Care. Balb/c mice (Strain code 028) were purchased from Charles River Laboratories, and after a one-week acclimation period, AAV vectors were administered to mice at 6 weeks of age.

AAV vector administration

Equal numbers of male and female Balb/c mice were administered 1 × 1011 vector genomes (vg) of AAV6.2FF vector diluted in 1× PBS to a final volume of 40 µL into the gastrocnemius muscle using a 29-gauge needle. Animals were monitored (unblinded), and in-life blood draws were completed by saphenous bleed using EDTA microvettes (Sarstedt Inc, Newton, NC, USA). At the endpoint, mice were anesthetized with isoflurane prior to terminal cardiac puncture or anesthetized with isoflurane followed by CO2 asphyxiation. Mice used in long-term AAV6.2FF kinetics study were anesthetized with isoflurane followed by CO2 asphyxiation, and bronchoalveolar lavages (BAL) were performed as previously described [31].

Enzyme-linked immunosorbent assays (ELISAs)

Commercially available ELISA kits for hIgG (Abcam 195125, Cambridge, MA, USA) were used to determine plasma immunoglobulin concentrations. hIgG concentrations in lavage fluid were expressed as hIgG concentration per mL of lavage sample.

Functional binding of vectorized antibody was determined using half-well 96-well plates (Corning, NY, USA) coated with 1 × 108 CFU/mL of PAO1 diluted in coating buffer (0.1 M carbonate buffer, pH 9.5) overnight at 4 °C. Plates were blocked with 3% BSA in PBS (PBS-B) and for 60 min at 25 °C with gentle rocking. The plates were washed three times with PBS-B, and tenfold antibody dilutions beginning at 1:10 were added to the wells and incubated at 25 °C for 60 min with gentle rocking. The plates were washed as above and goat anti-human HRP conjugated secondary antibody (Thermo Fisher, Waltham, MA, USA) was added at a 1:5000 dilution and incubated for 60 min at 25 °C with gentle rocking. Plates were then washed three times as above and incubated with TMB substrate (Pierce PI34021 Waltham, MA, USA) for 5 min at room temperature. Optical density (OD) values (450 nm) were graphed after subtraction of the mean OD of the negative control.

P. aeruginosa strains and growth conditions

The laboratory reference strains of P. aeruginosa, PAO1, and PA14, were used in this study and were originally provided by Dr. Joseph Lam (University of Guelph, Guelph, ON). The strain ATCC33356 was also sourced from Dr. Joseph Lam, and PA6077 was generously provided by Dr. Joe Harrison (University of Calgary, Calgary, Alberta). Stocks were maintained in glycerol (final concentration 20% v/v) and stored at −80 °C. Isolates were streaked on lysogeny broth agar (Fisher BioReagents, Fair Lawn, NJ) and incubated at 37 °C for 16–24 h. Overnight cultures were grown in lysogeny broth (Fisher BioReagents, Fair Lawn, NJ) incubated at 37 °C with shaking at 200 rpm. Overnight cultures were washed and resuspended in PBS, pH 7.4, prior to challenging mice.

P. aeruginosa challenge model

Six-week-old Balb/c mice (n = 8/group; 4 males and 4 females, randomized at time of arrival, sample size based on previous challenge studies [31]) received 1 × 1011 vg of AAV6.2FF-αPcrV, AAV6.2FF-αPsl, 5 × 1010 vg AAV6.2FF-αPcrV + 5 × 1010 vg AAV6.2FF-αPsl, or 1 × 1011 vg AAV6.2FF-MEDI3902, intramuscularly (IM). Control mice were administered PBS. Twenty-eight days post AAV administration, mice were challenged intranasally (IN) with a previously determined lethal dose of 4.47 × 107 CFU PAO1 or 3.0 × 107 CFU PA14 diluted in PBS to a final volume of 50 µL. Mice were monitored (blinded) every three hours and assessed for clinical signs and were euthanized via isoflurane overdose if endpoint criteria were met (see Fig. S1 for assessment sheet). Endpoint was immediately met if mice lost ≥ 20% of their initial body weight, or if they displayed a combination of any three of the following clinical signs: not responsive, not moving, moderate orbital tightening, hunched posture, loose skin, rapid/shallow breathing, or isolated. Surviving mice were monitored for 21 days post challenge and euthanized via isoflurane overdose.

P. aeruginosa biodistribution experiment

Six-week-old Balb/c mice (n = 6/group; 3 males and 3 females) were IM administered 1 × 1011 vg of AAV-mAbs as described above. At 28 days post AAV administration, mice were anesthetized and administered 4.47 × 107 CFU PAO1 or 3.0 × 107 CFU PA14 diluted in PBS to a final volume of 50 µL. At ~18 h post P. aeruginosa challenge, when untreated mice reached endpoint, all mice in the study were euthanized via isoflurane overdose and terminally bled via heart puncture. Terminal bleeds were collected and placed in pre-filled heparin tubes and the lungs, liver, and spleen were harvested. Pieces of tissue ~0.5 × 0.5 cm in size were placed into 1.5 mL screw-cap tubes filled with 200 µL of PBS and homogenized on the Precellys 24 tissue homogenizer (Thermo Fisher, Waltham, MA, USA) at 5000 rpm for 4× three-minute cycles. All samples were kept on ice until plated. Remaining tissues were fixed for histologic analysis. To determine the number of CFU, blood and homogenized tissue samples were added to the first column of a 96-well polystyrene plate (Corning™, Corning, NY). Samples were serially diluted across the plate by transferring 20 µL into 180 µL of PBS to generate dilutions from 100 to 10−11. Each dilution was spot plated (10 µL) in triplicate on Pseudomonas isolation agar (Becton Dickinson, Mississauga, ON). Plates were incubated at 37 °C and colonies were counted after 24 and 48 h.


Tissue samples were fixed in 10% formalin for 48 h, washed three times with 1× PBS, and stored in 70% ethanol. Fixed tissues were paraffin-embedded and 4 μM tissue sections were cut, and hemotoxylin and eosin (H&E) stained. Histopathology of blinded samples was evaluated by a board-certified anatomic pathologist. Histopathology was graded based on a nominal category score multiplied by an area score to capture the extent of the changes, to obtain a final grade that ranged 0–24. The nominal categories included neutrophil clustering in the alveoli and/or bronchioli, interstitial inflammation, hemorrhage, edema around airway or vessels, and presence of bacteria (Table 1). The affected area was semi-quantitatively assessed and scored based on four tiers of increasing area involvement (Table 1).

Table 1 Grading scheme to evaluate the severity of lesions in the lungs of infected mice.


All graphs were generated using GraphPad Prism 9.1 software (La Jolla, CA). Survival curves were generated using the Kaplan–Meyer survival curve and Bonferroni correction. Biodistribution graphs were generated using a normally distributed two-way ANOVA analysis. Randomization was not performed for the animal studies. Samples and animals were not blinded before performing each experiment; however, tissue samples were blinded for the CFU analysis and histopathology and the animal care staff who monitored mice for endpoints were blinded. For histopathology scores, Kruskall–Wallis (non-parametric) test with a Dunn’s multiple comparison test was performed.


Kinetic and functional analysis of AAV6.2FF-mediated expression of P. aeruginosa mAbs V2L2MD (αPcrV) and Cam-003 (αPsl)

mAbs against P. aeruginosa have been validated in preclinical models and in human clinical trials [37]. mAb V2L2MD, derived from PcrV-immunized VelocImmune mice [38], targets the protein PcrV that is known to play an essential role in the transport of multiple virulence factors through the T3SS [39]. Cam-003, which was identified through a whole cell-panning approach with human scFv phage libraries derived from patients with P. aeruginosa infections, targets the exopolysaccharide Psl known to play an important role in host cell attachment and biofilm formation [33]. Cam-003 has also been shown to display potent opsonophagocytic killing activity against P. aeruginosa, as well as to inhibit attachment to epithelial cells [33].

αPcrV and αPsl mAb genes were human-mouse codon optimized and synthesized with human IgG1 heavy chain and human kappa light chain constant domains separated by an F2A self-cleaving peptide (Fig. 1A). Geneblocks were cloned into an AAV genome between AAV2 ITRs and expression confirmed by western blot (Fig. S2A) and hIgG ELISA (Fig. S2B). AAV genomes encoding these mAb genes were packaged into AAV6.2FF and administered to Balb/c mice IM at a dose of 1 × 1011 vg. Western blot analysis of AAV-mediated mAb expression in mouse plasma was evaluated at 25 days post administration (Fig. S2C). hIgG antibody levels in the plasma were monitored for up to 210 days (Fig. 1B). For αPsl, plasma expression levels were similar to what we have observed previously for other mAbs [21, 31], however, αPcrV was expressed at concentrations tenfold higher, and were in the mg/mL range (Fig. 1C). Peak antibody expression was reached around day 28 post AAV administration, with αPsl hIgG expression reaching 279 µg/mL and αPcrV hIgG expression reaching 2818 µg/mL. Long-term expression kinetics were similar to other AAV-expressed mAbs with an increase in serum mAb concentration for the first six weeks after which expression remained relatively constant at an average of 109 µg/mL for αPsl and 1065 µg/mL for αPcrV for up to 210 days. At 522 days post AAV administration, mice in the long-term kinetic study were euthanized and BAL fluid and terminal bleeds were collected. As shown in Fig. 1D, mice sustained high levels of hIgG in their plasma and detectable levels of hIgG in BAL fluid for up to 75 weeks post AAV administration.

Fig. 1: AAV-αPcrV and AAV-αPsl mAb expression kinetics in the blood and bronchoalveolar lavage fluid (BALF) of transduced mice and confirmation of functional binding to P. aeruginosa.
figure 1

A Schematic diagrams of AAV-mAb genomes. Transgenes were expressed under the control of the ubiquitous CASI promoter (CASI) and consisted of the variable heavy chains of either αPsl or αPcrV followed by the human IgG1 heavy chain constant domain, a furin F2A self-cleaving peptide, the variable light chain domains of either αPsl or αPcrV and finally the human Kappa light chain constant domain. Both constructs contained a Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element (WPRE), simian virus 40 polyadenylation signal (SV40 polyA), flanked by two AAV2 inverted terminal repeat (ITR) sequences. B AAV-αPcrV and C AAV6.2FF-αPsl vectors were intramuscularly administered to 6-week-old female Balb/c mice (n = 4/ group) at a dose of 1 × 1011 vg. Blood samples were collected and plasma mAb expression was monitored. At 522 days post AAV administration, mice were euthanized due to old age, and plasma (D) and BALF (E) samples were collected and αPcrV and αPsl mAb concentrations determined using a hIgG ELISA. F Functional binding of AAV-expressed αPcrV and αPsl mAbs to P. aeruginosa as measured via indirect ELISA using PAO1 coated plates.

To confirm that AAV-expressed mAbs retained their ability to bind to P. aeruginosa, ELISA plates coated with P. aeruginosa were probed with mAbs purified from AAV-αPcrV and AAV-αPsl transduced HEK293 cells. Results showed that the mAbs bound to P. aeruginosa, as demonstrated by the decrease in OD as the mAbs were serially diluted (Fig. 1E). These results show that IM administration of AAV6.2FF mediates robust and sustained expression of functional anti-P. aeruginosa mAbs can be detected in plasma and lung lavage fluid.

Analysis of the impact of the WPRE on AAV-mediated transgene expression in vitro

To avoid the generation of escape mutants and/or to increase the efficacy of mAb therapy, it may be necessary to administer a cocktail of mAbs, as is the case for REGN-EB3, a treatment for EBOV infection [40]. Given the important role of Psl and the T3SS in P. aeruginosa pathogenesis, it was rationalized that a combination of αPsl and αPcrV mAbs could enhance strain and disease coverage. While administration of two single AAV-mAbs each expressing full-length mAbs is possible, vector production costs increase. For this reason, we endeavoured to optimize our AAV-VIP platform for the expression of bispecific mAbs.

Full-length mAbs expressed from our AAV-VIP platform are roughly 2.2 kb, whereas the bispecific P. aeruginosa mAb, MEDI3902, is >3 kb. This increase in size meant that we could not fit the MEDI3902 gene into an AAV genome encoding a CASI promoter and a full-size WPRE, as this would surpass the ~4.7 kb packaging limit [41]. To explore whether it was possible to delete the WPRE or use a smaller version of the WPRE, designated WPRE3, which at 248 bp has been shown to possess 80% of the activity of full-length WPRE (589 bp) [35], we engineered AAV genomes expressing GFP from a CASI promoter and containing either a WPRE, a WPRE3 or no WPRE sequence (Fig. 2A). To quantify differences in GFP expression afforded by the different WPRE-containing vectors, HEK293 cells were transduced and subjected to flow cytometric analysis three days later. As shown in (Fig. 2B), GFP’s mean fluorescence intensity (MFI) did not vary significantly between the three AAV vectors. These results suggest that the WPRE may not impact the efficiency of AAV-mediated transgene expression of small reporter genes like GFP, at least in vitro. Further investigation was warranted to fully elucidate the impact of WPRE on the expression of large transgenes in vivo.

Fig. 2: In vitro analysis of AAV6.2FF vectors expressing AAV-GFP-WPRE, AAV-GFP-WPRE3 or AAV-GFP (without a WPRE).
figure 2

A Schematic of the three GFP expressing AAV genomes tested. Each GFP expression cassette contained a ubiquitous CASI promoter (CASI), followed by the eGFP cDNA, either the full-length WPRE (589 bp), a WPRE3 (248 bp) or no WPRE, and a SV40 polyadenylation signal flanked by AAV2 inverted terminal repeat (ITR) sequences. WPRE, Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element. B In vitro transduction assay comparing AAV-GFP-WPRE, AAV-GFP-WPRE3, or AAV-GFP (without a WPRE). HEK293 cells were transduced at an MOI of 5000 and analyzed by flow cytometry 96 h post transduction. GFP+ cells were quantified by flow cytometry and mean fluorescence intensity graphed. Samples were run in triplicate and graphed relative to the control. Samples were not statistically different.

Optimization of AAV-mediated expression of a P. aeruginosa bispecific mAb

MEDI3902 is a potent bispecific antibody against the bacterial pathogen P. aeruginosa, combining the single-chain variable fragments (scFvs) of both αPsl and αPcrV mAbs [42]. MEDI3902 was encoded into our AAV-VIP platform due to its safety profile established in a Phase I (NCT02255760) dose-escalation study [43], and protection afforded in subjects at risk for P. aeruginosa in a Phase II (NCT02696902) clinical trial [44]. We constructed two AAV-MEDI3902 expressing vectors; one encoding a WPRE3 and one without a WPRE3 (Fig. 3A). Both genomes were engineered to contain the composite CASI promoter, followed by the Fab domain (VH and CH1) of αPcrV and the scFv (VH and VL chains joined together by a flexible peptide linker) of αPsl linked to the CH2 and CH3 domains of human IgG1. An F2A self-cleaving peptide separated the heavy and light chain domains, the latter of which was comprised of the variable light and kappa light chain domains of αPcrV. Both vectors were packaged into AAV6.2FF and IM administered to groups of female Balb/c mice (n = 4). Blood samples were obtained on a weekly basis and the plasma concentration of bispecific mAbs was determined using a quantitative hIgG sandwich ELISA. As shown in Fig. 3B, there were no significant differences in the concentration of MEDI3902 expression from AAV genomes with or without a WPRE3.

Fig. 3: Kinetic analysis and functional binding of AAV-expressed P. aeruginosa bispecific mAb, MEDI3902.
figure 3

A Schematic diagrams of AAV genomes encoding bispecific mAb MEDI3902 with or without a WPRE3 sequence. Transgenes were expressed under the control of the CASI promoter (CASI), consisting of the Fab domain of αPcrV and the scFv of αPsl linked to the CH2 and CH3 domains of human IgG1. An F2A self-cleaving peptide separated the heavy and light chain domains, the latter of which was comprised of the variable light and kappa light chain domains of αPcrV. One construct termed AAV-MEDI3902-WPRE3 contains the minimal WPRE termed WPRE3, while AAV-MEDI3902 does not encode a WPRE. Both AAV genomes contain a SV40 polyA signal and are flanked by AAV2 inverted terminal repeat (ITR) sequences. WPRE3, minimal Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element. B AAV-MEDI3902-WPRE3 and AAV-MEDI3902 were intramuscularly administered to 6-week-old Balb/c mice (n = 4) at a dose of 1 × 1011 vg. Blood samples were collected and plasma mAb expression was monitored for up to 63 days. C Long-term expression kinetics of AAV-MEDI3902-WPRE3 in Balb/c mice administered 1 × 1011 vg IM. D At 260 days post AAV-MEDI3902-WPRE3 administration, mice were euthanized, and plasma and lung lavage samples were collected and quantified using a hIgG ELISA. E Functional binding of AAV-expressed MEDI3902 to P. aeruginosa was confirmed by indirect ELISA using PAO1 coated plates.

We elected to move forward with the WPRE3 containing vector for all subsequent experiments as it was most similar to the vector design used for AAV-αPcrV and AAV-αPsl, which both contain WPREs. We next evaluated the long-term expression kinetics of AAV-expressed MEDI3902. As shown in Fig. 3C, Balb/c mice (n = 4) administered a standard dose of 1 × 1011 vg of AAV-MEDI3902-WPRE3 sustained average plasma mAb expression levels of 23 µg/mL for over 210 days suggesting that despite the rather unusual structure, and potentially immunogenic nature of the bispecific mAb MEDI3902, mice did not mount a robust enough immune response against this foreign protein to result in the elimination of vector transduced cells. When mice were euthanized 260 days post AAV, hIgG was detected in the both the plasma and lung lavage fluid (Fig. 3D). Finally, we demonstrated that AAV-expressed MEDI3902 retained functional binding to P. aeruginosa strain PAO1 using an antibody binding ELISA assay (Fig. 3E).

AAV6.2FF vectorized mAbs provide significant protection against lethal P. aeruginosa challenge

To validate these previously characterized mAbs and their efficacy when expressed from a viral vector, lethal P. aeruginosa pneumonia models using multiple clinically relevant strains (PA6077, ATCC33356, PAO1, and PA14 [33, 45, 46]) were first established (Fig. S3). Based on reports in the literature, three different doses of CFU for each of the four strains of P. aeruginosa evaluated were administered IN to groups of female Balb/c mice (n = 4) with the goal of identifying a dose that resulted in 0 to 20% survival. Mice were assessed for the following clinical signs: non-responsive, not moving, moderate orbital tightening, hunched posture, loose skin, rapid/shallow breathing, or isolated. Endpoint was met if mice lost ≥20% of their initial body weight, or if they displayed any three of the above-listed clinical signs, as per the animal utilization protocol (Fig. S1). Except for P. aeruginosa strain 6077, a 100% lethal dose was established for all strains (Fig. S3). For this study, we elected to move forward with the two most clinically relevant strains of P. aeruginosa tested: PAO1 and PA14 [47]. A challenge dose of 4.47 × 107 CFU for PAO1 conferred 25% survival and we rationalized challenging with 3.0 × 107 CFU PA14 for a predicted survival rate of 25%. Note that female mice were used for the establishment of the lethal model since they are more resistant to P. aeruginosa pneumonia than male mice [48].

For the challenge study, groups of 6-week-old male (n = 4) and female (n = 4) Balb/c mice were IM injected with 1 × 1011 vg of AAV-αPcrV, AAV-αPsl, or AAV-MEDI3902. A fourth treatment group was administered 5 × 1010 vg AAV-αPcrV and 5 × 1010 vg AAV-αPsl into separate leg muscles. This group served as a control for AAV-MEDI3902. Plasma mAb expression was quantified by ELISA two days prior to challenge. As shown in Fig. 4A, B, at 26 days post AAV administration, plasma αPcrV concentrations ranged from 74 to 470 µg/mL, αPsl concentrations ranged from 6 to 85 µg/mL, MEDI3902 concentrations ranged from 3 to 62 µg/mL and the group of mice that received a combination of AAV expressing αPcrV and αPsl had mAb concentrations ranging from 104 to 376 µg/mL. The lighter-shaded data points depict mice that did not survive the P. aeruginosa challenge. Plasma mAb expression levels for male and female mice were graphed and revealed no major sex differences (Fig. S5), similar to what we reported previously [21].

Fig. 4: AAV6.2FF-mAbs confer significant protection in two lethal P. aeruginosa pneumonia challenge models.
figure 4

Groups of male (n = 4) and female (n = 4) 6-week-old Balb/c mice were IM administered 1 × 1011 vg of AAV-αPcrV, AAV-αPsl, AAV-αPcrV + AAV-αPsl (5 × 1010 vg of each vector administered in two separate leg muscles), or AAV-MEDI3902. A, B Plasma concentrations of human IgG were quantified 26 days post AAV administration. Data are represented by the mean ± standard deviation. Data points with lighter shaded centers depict individual mice that did not survive the P. aeruginosa challenge. On day 28 post AAV administration, mice were challenged via the IN route with 4.47 × 107 CFU PA01 or 3.0 × 107 CFU PA14 and monitored for survival (C, D) and weight loss (E, F). Survival of treated groups were compared to the control group using the Mantel–Cox log-rank test and Bonferroni correction, *p < 0.02, **p < 0.003, ****p < 0.001, ns not significant.

Mice were challenged with either 4.47 × 107 CFU PAO1 or 3.0 × 107 CFU PA14 IN and monitored for clinical signs indicative of endpoint (Fig. S1). All control mice reached endpoint by 18 h post infection with PAO1, demonstrating the lethality of this bacterial pathogen at the chosen dose (Fig. 4C). In the AAV-αPcrV treatment group, 100% of the mice survived the challenge, whereas in the AAV-αPcrV + AAV-αPsl and AAV-MEDI3902 treatment groups, 87.5% of mice survived. Only 50% of mice survived the challenge in the AAV-αPsl treatment group. All treatment groups were deemed statistically significant when compared to the control group, with AAV-αPcrV providing the greatest protective efficacy compared to the other groups. The mice that survived challenge displayed mild clinical signs of respiratory distress that subsided within 3 days post challenge. Weight loss was not significant in treatment mice and between 36 and 42 h post challenge all surviving mice either regained ≥100% of their body weight or began to regain initial body weight, while control mice rapidly declined within the first 18 h, leading to endpoint (Fig. 4E).

Sex differences are understood to play a role in the progression of bacterial infections in murine models [48]. Male mice are generally more susceptible to respiratory bacterial diseases and sepsis than females [48], and we observed the same trend in our study. To account for sex differences, we challenged both male (n = 4) and female (n = 4) Balb/c mice per group with PAO1. Kaplan–Meyer survival curves were separated by sex (Fig. S4), impressively demonstrating that AAV-αPcrV, AAV-αPcrV + AAV-αPsl, and AAV-MEDI3902 all conferred 100% survival in female Balb/c mice, with AAV-αPsl conferred 75% survival, and 100% of the control mice succumbing to infection. In the male groups, AAV-αPcrV conferred 100% survival, with the remaining treatment groups conferring between 25 and 75% survival.

In the PA14 challenge experiment, 75% of control mice succumbed to infection over a 50-h period. Interestingly, treatment with AAV-αPcrV + AAV-αPsl conferred 100% survival, while AAV-MEDI3902 conferred 75% survival, AAV-αPcrV conferred 87.5% survival, and AAV-αPsl conferred 12.5% survival (Fig. 4D). In the female groups, all treatment groups except for AAV-αPsl conferred 100% protection, while the latter conferred 0% survival, with all mice succumbing to infection by 36-h post infection (Fig. S4). In terms of weight loss, the challenge with PA14 led to a more substantial reduction in body weight (Fig. 4F) in comparison to the PAO1 challenge (Fig. 4E). Notably, most AAV-mAb-treated mice returned to their initial body weight within 36–60 h post challenge, while the two surviving control mice did not recover their initial body weight. Similar to the PAO1 challenge, sex-based differences were observed. AAV-αPcrV, AAV-αPcrV + AAV-αPsl, and AAV-MEDI3902 all conferred 100% survival in female Balb/c mice, with the control mice conferring 25% survival and AAV-αPsl conferring 0% survival. In the male groups, AAV-αPcrV + AAV-αPsl conferred 100% survival, AAV-αPcrV conferred 75% survival, AAV-MEDI3902 conferred 50% survival, and the control mice and AAV-αPsl groups conferring 25% survival.

AAV6.2FF vectorized mAbs reduce bacterial burden in the lungs and limit dissemination to other organs following intranasal challenge with P. aeruginosa

We next sought to evaluate how AAV-expressed mAbs reduced bacterial burden in the lungs and prevented dissemination to other organs. To do so, groups of Balb/c mice (n = 3 male and n = 3 female per group) were administered AAV-mAbs at a dose of 1 × 1011 vg IM as was done previously. Twenty-six days later, on day two prior to challenge, plasma hIgG was quantified to confirm mAb expression (Fig. 5A, C). Plasma hIgG levels ranged from 89 to 855 µg/mL in AAV-αPcrV, 30 to 105 µg/mL in AAV-αPsl, 28 to 95 µg/mL in AAV-αPcrV + AAV-αPsl and 16 to 63 µg/mL in AAV6.2FF-MEDI3902 treated mice. On day 28 post AAV, mice were challenged IN with 4.47 × 107 CFU PAO1 or 3 × 107 CFU PA14 as before. At 18 h post challenge, when control mice reached endpoint, blood, lungs, spleen, and liver were harvested and the number of CFU/mg of tissue was determined by plating the clarified supernatant from tissue lysates on Pseudomonas isolation agar. As shown in Fig. 5B, PAO1 biodistribution was markedly reduced in the lungs of all AAV-mAb treated mice compared to the control mice. One female mouse belonging to the AAV-αPsl group showed an elevated level of CFU in the spleen and the liver, which was not observed in any other treatment groups. Control mice averaged 666,346 CFU/mg in the lung, 197 CFU/mg in the liver, 2 CFU/mg in the spleen, and 90,916 CFU/mg in the blood. AAV-αPcrV treated mice averaged 1 CFU/mg in the lung, 21 CFU/mg in the liver, 0.5 CFU/mg in the spleen, and 2 CFU/µL in the blood. AAV-αPsl treated mice averaged 48 CFU/mg in the lung, 110 CFU/mg in the liver, 25 CFU/mg in the spleen, and 3 CFU/µL in the blood. AAV-MEDI3902 and AAV-αPcrV + AAV-αPsl treated mice notably demonstrated an even greater overall reduction in CFU. AAV-αPcrV + AAV-αPsl had 11 CFU/mg in the lungs and no detectable CFU in the liver, spleen, or blood, while AAV-MEDI3902 averaged 15 CFU/mg in the lung, 2 CFU/mg in the liver, 0.9 CFU/mg in the spleen and no detectable CFU in the blood (Fig. 5B). This dramatic decrease in detectable CFU in each organ evaluated suggests that while the monospecific mAbs do confer protection in a lethal P. aeruginosa challenge, co-administration of mAbs targeting different targets on the bacterial and/or use of a bispecific mAb confers greater protection from bacterial dissemination in an acute lethal pneumonia challenge model.

Fig. 5: AAV-αPcrV + AAV-αPsl and AAV-MEDI3902 provide excellent protection in the prevention of bacterial burden in the lungs and dissemination to other organs, while AAV-αPcrV and AAV-αPsl provide moderate protection.
figure 5

Plasma mAb expression levels in Balb/c mice (n = 3 male and n = 3 female per treatment group) IM administered 1 × 1011 vg of AAV-αPcrV, AAV-αPsl, AAV-αPcrV + AAV-αPsl, and AAV-MEDI3902 and quantified using a commercial hIgG quantitative ELISA two days prior to challenge with PAO1 (A) and PA14 (C). On day 28 days post AAV administration, mice were challenged intranasally with a previously determined lethal dose of either PAO1 (4.47 × 107 CFU) (B) or PA14 (3.0 × 107 CFU) (D). Mice were monitored for clinical signs and sacrificed ~18 h post infection for bacterial enumeration in the lungs, liver, spleen, and blood. Data show the CFU per mg of tissue and CFU/500 µL of blood and are represented by the mean ± standard deviation.

In the PA14 biodistribution study, similar trends were observed with respect to average CFU counts in various tissues (Fig. 5D). Control mice showed elevated levels of PA14 CFU in the lung and blood, with negligible CFU detected in the liver and spleen. While most AAV-mAb treatments led to a reduction in CFU in the blood and lung compared to control mice, AAV-MEDI3902 and AAV-αPcrV + AAV-αPsl treated mice demonstrated the greatest reduction in CFU compared to the other treatment groups. The average CFU in the tissues from control mice were as follows: 19,239 CFU/mg in the lung, 21 CFU/mg liver, 90 CFU/mg in the spleen, and 396 CFU/µL in the blood (Fig. 5D). AAV-αPcrV treated mice averaged 1391 CFU/mg in the lung, 13 CFU/mg in the liver, 16 CFU/mg in the spleen and 10 CFU/µL in the blood. AAV-αPsl treated mice averaged 6703 CFU/mg in the lung, 13 CFU/mg in the liver, 116 CFU/mg in the spleen, and 24 CFU/µL in the blood. AAV-αPcrV + AAV-αPsl treated mice averaged 654 CFU/mg in the lung, 2 CFU/mg in the liver, 9 CFU/mg in the spleen, and 2 CFU/µL in the blood. AAV-MEDI3902 treated mice averaged 359 CFU/mg in the lung, 0.4 CFU/mg in the liver, 2 CFU/mg in the spleen, and 1 CFU/µl in the blood.

Lung tissue from the biodistribution experiment was formalin-fixed and paraffin-embedded. Tissues were then sectioned, H&E stained, and evaluated by a board-certified anatomic pathologist. Histopathology was scored based on neutrophil clustering in the alveoli and bronchioli, inflammation, hemorrhage, edema around airway or vessels, the presence of bacteria, and to obtain the final grade, the sum of the nominal scores was then multiplied by a ranked semi-quantitative assessment of the extent of affected lung area (Table 1). No marked qualitative differences were observed between AVV-treated and control mice, with most animals presenting with histopathological changes consistent with an acute bacterial bronchopneumonia, as shown by alveolar fibrin and neutrophilic exudate, interstitial edema (around vessels and bronchi), and multifocal hemorrhage. In the PAO1-challenged mice, the final average histology grades ranged from 5.25 to 14 (median), with AAV-αPcrV having the lowest grade, and the control mice receiving the highest grade (Table 2). The lowest grade in the AAV-αPcrV group was mainly driven by decreased frequency of fibrin and neutrophil exudation compared to the controls and other treatment groups (data not shown). These findings are consistent with the PAO1 survival study results, in which AAV-αPcrV conferred 100% survival in both male and female groups of mice.

Table 2 Average histopathology grade derived from both a nominal tally and lung area affected.

PA14 histology grades ranged from 6 to 8.84, with the control group receiving the lowest score, and AAV-αPcrV + AAV-αPsl receiving the highest grade (Table 2). While these findings are not consistent with our survival data, decreased frequency of hemorrhage was observed in the AAV-mAb treatment groups (data not shown). For both the PAO1 and PA14 biodistribution experiments, a Kruskall–Wallis test with a Dunn’s multiple comparison test was performed on the histopathological grades, and none of the groups were determined to be significantly different.


AAV-vectored antibody expression holds promise as an alternative strategy to passive mAb therapies currently available. Here, we demonstrated that our AAV-VIP platform could successfully prevent acute lethal pneumonia in mice challenged with two different strains of P. aeruginosa. When packaged into our AAV6.2FF capsid, bispecific mAb MEDI3902 performed similarly to a cocktail of two separate AAV-mAbs. To our knowledge, this is also the first time a bispecific mAb has been expressed from an AAV vector for the purposes of immunoprophylaxis.

Mice administered either AAV6.2FF-αPcrV or AAV6.2FF-αPsl demonstrated sustained hIgG expression over the duration of their lifespans. While both mAbs were mouse-human codon optimized and possessed the same heavy and light chain constant domains, αPcrV expressed from AAV6.2FF was an order of magnitude greater than αPcrV and reached plasma hIgG concentrations in the mg/mL range. While it remains unclear as to why αPcrV expresses significantly better in vivo than other mAbs, it is likely one of the reasons why this mAb was selected during the screening of hybridomas [39] and was used as the backbone for the development of Medimmune’s bispecific mAb, MEDI3902. This finding also highlights the importance of evaluating AAV-mediated mAb expression in vivo. Some AAV-expressed mAbs are expressed at very high concentrations while others, albeit rarely, are poorly expressed despite similar cassette design [21].

Despite the log difference in expression levels, both AAV-expressed mAbs were detected in the plasma for over 500 days. Duration of AAV-mAb expression was not appreciably different between monospecific mAbs and bispecific mAbs, potentially suggesting that the unusual structure of the bispecific mAb did not induce an immune response that led to the clearance of AAV-transduced cells. This could be confirmed by quantifying anti-idiotype antibody responses and mapping anti-transgene T-cell responses; however, both assays require the generation of costly reagents and are beyond the scope of this study but are important questions to be addressed in future experiments.

AAV-MEDI3902 expresses a bispecific hIgG1 mAb engineered to contain the Fab domain of αPcrV and the scFv of αPsl linked to the CH2 and CH3 domains of human IgG1 followed by a F2A self-cleaving peptide and light chain domains, the latter of which was comprised of the variable light and kappa light chain domains of αPcrV. Due to the complex nature of P. aeruginosa pathogenesis, it was rationalized that a combination of antigenic targets could maximize the inhibitory properties of the mAb, and by doing so with one AAV vector instead of two would significantly decrease overall vector production costs. Since the gene for MEDI3902 is ~800 bp larger than a standard mAb, it was impossible to encode a full-length WPRE, as is the case for the single mAbs αPcrV and αPsl. WPREs are typically included in lentivectors and AAV genomes to increase transgene expression [49]. Recently, Choi et al. engineered a truncated WPRE termed WPRE3 (248 bp) that retains ~80% of the wild-type WPRE function [35]. To evaluate the importance of the WPRE on AAV6.2FF-mediated transgene expression, we evaluated GFP expression from AAV vectors containing a WPRE, a WPRE3, or no WPRE and it was found that there was no significant difference in the MFI of GFP expression in HEK293 cells that were transduced with these three vectors. Due to the possibility that the WPRE has a limited impact on small reporter genes such as GFP, we engineered AAV genomes encoding MEDI3902 with or without a WPRE3 and evaluated the impact on transgene expression in vivo. As was observed with the GFP experiment, there was no significant difference in plasma mAb concentrations suggesting that a WPRE3 is not required for efficient expression of mAbs from AAV6.2FF transduced muscles. However, it cannot be ruled out that a WPRE might be necessary to induce robust transgene expression in other tissues or when using a different AAV capsid. Additionally, because the transgene expressed from our AAV vector is an antibody, which can be recycled by FcRn receptor binding [50], serum mAb expression levels are achieved by a combination of de novo mAb production from the transduced muscle and mAbs recycled by FcRn receptors. This recycling could potentially mask subtle differences in mAb serum concentrations from vectors with or without a WPRE.

Although it has been reported that pcrV genes and a functional psl operon are present in 99% and 94% of P. aeruginosa clinical isolates [42], we confirmed the functional binding of AAV-expressed mAbs against P. aeruginosa strain PAO1 by performing an indirect binding ELISA. While we were able to confirm that increasing mAb concentrations led to an increase in binding to the target bacteria based on OD readings, a limitation of this assay was the fact that we used whole bacteria as the antigen rather than specific target antigens for αPcrV and αPsl due to lack of available reagents.

AAV-mAbs were detected in the lung following IM administration, a clinically relevant organ as P. aeruginosa establishes persistent and chronic lung infections in immunocompromised individuals [51]. Therefore, we validated these AAV-expressed mAbs in an acute P. aeruginosa pneumonia challenge model using two clinically relevant strains: PAO1 and PA14. PAO1 and PA14 are common reference strains used to assess novel therapeutics and to further understand the biology of P. aeruginosa [47]. While PA14 is a hypervirulent strain in comparison to PAO1, it displays a high level of genomic conservation. PA14 possesses two pathogenicity islands, carrying putative virulence-associated genes which are not conserved in the PAO1 genome [46]. Nevertheless, both PAO1 and PA14 are pathogenic in Balb/c mice, and express the two antigenic targets of interest [47]. To evaluate these AAV-expressed mAbs in a prophylactic setting, mice were challenged 28 days post AAV administration when mAb expression levels started to peak and/or plateau. Pre-challenge mAb expression levels varied but still conferred significant protection, with the exception of αPsl, which in general was not as effective as αPcrV or MEDI3902. We hypothesized that the antigen target may not have been as prevalent or concentrated on the bacteria, or that the concentration of mAb expressed was below the therapeutic threshold, and perhaps increasing the dose of AAV-αPsl could have provided greater protection. Given the large numbers of animals and treatment groups evaluated it was not possible to conduct a dose-response experiment, instead we use our standard dose of 1 × 1011 vg per animal which converts to roughly 5 × 1012 vg/kg.

Overall, AAV-αPcrV conferred 100% survival in the PAO1 challenge model, while AAV-αPcrV + AAV-αPsl conferred 100% survival in the PA14 challenge. In both challenge models the bispecific mAb MEDI3902 performed either similarly or identical to the protection provided by AAV-αPcrV + AAV-αPsl. Note that the apparent reduction in function of the AAV-αPcrV when combined with AAV-αPsl in the PAO1 challenge model could be due to the fact that when these AAV-mAbs were administered as a single treatment, they were given at a dose of 1 × 1011 vg, but when they were administered in combination, they were administered at 5 × 1010 vg per AAV-mAb. Therefore, the mice in the combination group should have half as much αPcrV as the mice in the single AAV-αPcrV group, which might explain the slight reduction in survival from 100% in the AAV-αPcrV group to 87.5% in the combination group.

Among the top-performing vectors, the lowest therapeutic plasma antibody concentration that conferred survival ranged from 74 to 219 µg/mL for AAV-αPcrV, 3 to 45 µg/mL for AAV-MEDI3902, and 104 to 320 µg/mL for AAV-αPcrV + AAV-αPsl for both PAO1 and PA14 challenges. In the PAO1 challenge, the minimum therapeutic plasma antibody concentration required for survival was 18 µg/mL for AAV-αPsl, while in the PA14 challenge the minimum concentration required was 86 µg/mL. These findings, taken together with the significant weight loss observed in PA14-challenged mice in comparison to PAO1-challenged mice highlights the major difference in virulence among these two strains.

Due to the integral role Psl plays in the establishment of chronic P. aeruginosa infections and biofilm formation, we predict that this antigenic target may be expressed at higher levels in chronic infections, therefore it is possible that the therapeutic benefit of αPsl would be greater in a model of chronic P. aeruginosa infection. However, it may be possible to increase the therapeutic benefit of AAV-αPsl by simply increasing the dose of AAV administered. Dose-response studies must be conducted to determine the optimal dose of AAV-αPsl.

We observed that sex differences play a major role in overall survival outcomes following IN challenge with P. aeruginosa. While female treatment groups had a greater survival outcome than their male counterparts, AAV-MEDI3902 conferred 100% survival in both PAO1 and PA14 challenges in female mice. In comparison, in male mice, it conferred only 75% and 50% survival in PAO1 and PA14 challenges, respectively. These results highlight the difference in lethality of P. aeruginosa induced pneumonia in male and female mice, and correlate well with previous reports in which male mice displayed increased levels of pro-inflammatory cytokines during early stages of P. aeruginosa infection compared to females, leading to an increase in susceptibility to respiratory and systemic bacterial infections [52]. The sex differences observed in the acute P. aeruginosa pneumonia model differ from that of other challenge models, such as influenza, where female mice display increased susceptibility to infection [53, 54]. These findings highlight the need to include animals of both sexes in infectious disease research studies.

To determine the impact of anti-P. aeruginosa AAV-mAbs on bacterial growth and dissemination, we quantified CFU in various tissues at an acute time point. Having confirmed the presence of the AAV-expressed mAbs in the plasma and lung lavage fluid, we anticipated that this would afford some protection against bacterial dissemination. The untreated mice for these experiments reached endpoint slightly earlier (18 h vs 20–24 h) than anticipated, so to standardize the time post-challenge for the quantification of CFU, all treatment groups were euthanized when control mice reached endpoint rather than the desired 48-h time point. Nevertheless, we did detect a marked reduction in bacterial burden in the lungs and blood of all AAV-mAb treated groups, which was evident even at 18 h post P. aeruginosa challenge. Importantly, the overall reduction in CFU suggested that the AAV-expressed mAbs protected against bacterial dissemination to other organs, which plays a major role in the progression of disease leading to dysfunction of multiple organs, bacteremia, and death. In the PAO1 experiment, one female mouse belonging to the AAV-αPsl group showed increased CFU levels in the liver, spleen, and blood, which was not consistent with our other groups. We anticipate this was due to biological variability, but the study would need to be repeated to confirm. Overall, AAV-MEDI3902 and AAV-αPcrV + AAV-αPsl demonstrated the greatest overall reduction in CFU compared to the control groups, highlighting the benefit of targeting multiple antigens simultaneously.

Histological findings were not significantly different with respect to overall grades, but AAV-mAb treatment did lead to decreased levels of hemorrhage in the PAO1-challenged groups. In the PAO1 challenge, AAV-αPcrV-treated mice showed a reduction in fibrin exudate and fewer neutrophils compared to the controls and other treatment groups. The overall average grade (Table 2) tended to be lower than the AAV-αPcrV treatment group, despite not reaching statistical significance. It appears there was no protection from pathology in the PA14-challenged mice. As an acute and lethal challenge model was used for this study, there is a strong likelihood that any differences in pathology may not have been quantifiable due to the high-challenge dose used. Additionally, if the experiment had been terminated at a later timepoint, e.g., 24–48 h post challenge, or a chronic model with a lower bacterial dose was tested, it is possible that we may have observed more substantial differences in pathological findings between treatment groups. Other limitations include the potential for inconsistent sampling of the lungs, which likely also contributed to the variability of the observed lesions.

Previous work from our group has demonstrated that AAV-VIP can be used to protect against challenge with both viral infections [21, 31, 55, 56] and bacterial toxin challenge [27]. Here we endeavored to explore the application of AAV-VIP for other bacterial targets: serotype-independent proteins and polysaccharides which play a major role in P. aeruginosa pathogenesis. Here, we show that this AAV-mAb platform holds promise as an alternative or adjunct treatment for P. aeruginosa infections. Future experiments are warranted and will aim to evaluate the protective efficacy of anti-P. aeruginosa AAV-mAbs in a therapeutic model, and in a persistent chronic model of P. aeruginosa infection.