PcrV in an intranasal adjuvanted tobacco mosaic virus conjugate vaccine mediates protection from Pseudomonas aeruginosa via an early Th1/Th17 skewed localized and systemic immune response

Abstract

Multi-drug-resistant Pseudomonas aeruginosa (PA) infections are a growing problem for at-risk populations, causing thousands of deaths per year, and novel therapies to treat or prevent severe PA infection are desperately needed. We developed a novel intranasal (IN) protein subunit vaccine using tobacco mosaic virus (TMV) as a vaccine delivery platform. Recombinant PcrV (rPcrV) was covalently linked to TMV and delivered IN with the adjuvant curdlan (TMV-PcrV+c). IN delivery of TMV-PcrV+c elicited significantly higher anti-PcrV antibody titers than unadjuvanted TMV-PcrV and rPcrV with or without adjuvant, and it provided significantly greater protection from lethal pneumonic infection with PA compared to all other groups. The protective vaccine formulation elicited anti-PcrV IgA in the lungs, while anti-PcrV IgG1 in the serum was associated with survival after infection. Following infection, TMV-PcrV+c immunization was associated with early IL-17 and IFN-γ secretion in the lungs and later IL-17 secretion in the spleen. This corresponded with a rapid, but transient recruitment of neutrophils to the lungs of vaccinated mice compared to unvaccinated controls, which was followed by an increase in the number of lung macrophages. This is indicative of a Th1/Th17 response mediating localized and systemic inflammation to clear the PA infection while limiting host-derived tissue damage, ultimately enhancing protection. As a vaccine antigen, PcrV did not confer complete protection, so we evaluated OprF and OprI as part of a multivalent vaccine. However, these additional antigens did not enhance protective efficacy despite generating high antibody titers. Overall, TMV-PcrV+c is a promising vaccine candidate that elicits specific antibodies, localized IFN-γ secretion and IL-17 secretion both locally and systemically after PA challenge, phagocyte recruitment to the infection site, and protection against a lethal dose of PA in a pneumonic disease model.

Introduction

Pseudomonas aeruginosa (PA) is a gram-negative facultative anaerobic bacterium that is a source of significant morbidity and mortality as an opportunistic pathogen. One of the ESKAPE pathogens [1,2], PA is a prevalent cause of antibiotic-resistant healthcare-acquired-infections (HAIs) as well as community-acquired infections. Immunocompromised, critically ill, very young, and elderly individuals are at highest risk for severe disease and death [2,3]. In healthcare settings, PA is one of the most common causes of ventilator associated pneumonia, surgical site infection, catheter associated urinary tract infections, and central line blood infections, in part through its ability to form biofilms on medical devices [4]. It is also the most common isolate from chronic pulmonary infections in cystic fibrosis patients [5], and along with Staphylococcus aureus, is one of the most common causes of soft tissue infections and sepsis in burn patients [6,7]. In healthy individuals, it is the most common cause of bacterial keratitis in contact lens users [8].

Antibiotic resistant PA continues to spread worldwide [9]. Multi-drug-resistant (MDR) strains, particularly those resistant to carbapenems, are considered a serious threat by the Centers for Disease Control and Prevention [10] and a critical priority for research and development of new interventions by the World Health Organization [11]. In the face of increasing antibiotic resistance, alternative approaches for prevention and treatment of PA infections, such as vaccination, are being investigated. Multiple experimental vaccines against PA have been developed, but none have yet been approved by the FDA [3]. PA causes a spectrum of mucosal and soft tissue infections; therefore, designing a successful vaccine is challenging and will require an approach that induces optimal protection at multiple sites of pathogen entry [12]. Parenteral vaccination (by intramuscular or subcutaneous injection) rarely induces strong mucosal immunity, but mucosal vaccination has experimentally demonstrated strong protective immunity at both mucosal surfaces and systemic sites [13].

To improve on PA vaccine efficacy, we utilized a novel mucosal vaccination strategy employing tobacco mosaic virus (TMV) as a delivery platform for a protein subunit vaccine. TMV is a plant virus that is immunogenic and inherently safe for human use. It does not infect animal cells and humans are exposed to it regularly, primarily through the food chain, with no known adverse effects. Most humans already have measurable anti-TMV antibody titers independent of their usage of tobacco products [14,15]. TMV is an optimal size to promote uptake by antigen-presenting cells, making it an ideal platform for stimulating a specific, protective adaptive immune response [16–18]. It has been utilized in vaccines against several pathogens including Francisella tularensis, influenza, and SARS-CoV-2 [19–25]. We previously demonstrated that covalently linking pathogen associated antigens to the surface of TMV substantially increases the immunogenicity and protective efficacy of an intranasal (IN) protein subunit vaccine against Yersinia pestis [26]. IN vaccination with a TMV-conjugate vaccine provided complete protection against lethal challenge in a pneumonic plague model with a single vaccine dose [27].

T cells are involved in shaping the antibody response and the activation and recruitment of phagocytes including neutrophils and macrophages that are vital for clearance of an extracellular pathogen like PA. Th1 cellular responses are characterized by the secretion of IFN-γ, which promotes recruitment of monocytes/macrophages to the sites of inflammation, while Th17 responses are characterized by neutrophil recruitment. Both have been shown to play an important role in the control and clearance of PA via phagocytosis [3,28–30].

In the present study we investigated the protective efficacy of mucosally-delivered PA antigens PcrV, OprF, and OprI. We independently covalently linked each of these protein antigens to TMV to generate TMV-PcrV, TMV-OprF, and TMV-OprI. PcrV is the tip protein of the PA type III secretion system (T3SS), a well-known virulence factor, and is required for T3SS activation and translocon assembly [31]. Both OprF and OprI are thought to be important for maintaining PA outer membrane integrity [32] and OprF is multifunctional, playing additional roles in quorum sensing and binding of C3b [33]. An OprF/I fusion protein was investigated as a human vaccine [34] but failed to meet primary endpoints in clinical trials. As OprF and OprI are both integral membrane proteins, we also generated lipid nanoparticles (LNPs) with the proteins adsorbed onto the LNP surface to determine whether vaccine efficacy is affected by protein folding within a membrane. We also assessed the effect of adding the Th1/Th17-skewing [28,35,36] adjuvant curdlan, a 1,3-β-glucan, to the vaccine formula [28,35,36].

Here, we report that TMV-PcrV with curdlan (TMV-PcrV+c) protected against lethal IN challenge with PA. Anti-PcrV IgG1 antibodies were associated with protection from PA challenge, but we did not find evidence that vaccine-induced serum antibodies opsonize PA. TMV-PcrV+c also enhanced IL-17 and IFN-γ secretion in the lungs early after challenge compared to unvaccinated mice and mice vaccinated without adjuvant. Protection was not enhanced by the inclusion of additional PA surface-associated antigens OprF and OprI.

Material and Methods

Bacterial strains and mice

Pseudomonas aeruginosa strains Boston 41501 (ATCC 27853) and PAO1 (ATCC 15692) were cultivated in BBL Trypticase Soy Broth (TSB) (BD 211768) at 37°C. E. coli BL21(DE3) (Sigma CMC0014; Invitrogen C600003) was used for recombinant protein expression and was cultivated in Miller LB broth (Fisher Bioreagents BP1426–2) with 50 μg/mL kanamycin (for rPcrV) or 100 μg/mL carbenicillin (for rOprF and rOprI).

Equal numbers of male and female 5-week-old C57Bl/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME) or Charles River Laboratories (Marbletown, NY) were maintained in the Department of Comparative Medicine at New York Medical College (NYMC). All experiments were approved by the NYMC Institutional Animal Care and Use Committee. The number of mice used per treatment group are indicated for each experiment.

Recombinant protein synthesis

The gene for PcrV was cloned from PA Boston 41501 DNA into the expression plasmid pET-28a using the primers: F-CTGAGGAATCATGCATATGGAAGTCAGAAACCTT, R- CAGTTGCATGCTGAATTCTAGATCGCGCTGAGAATGTC. DNA sequences of OprF and OprI, lacking the putative signal sequence (first 72 and 60 nucleotides, respectively), were synthesized by GENEArt (Thermofisher) and in the expression plasmid pRSET A. All gene sequences were confirmed by Sanger sequencing (Azenta).

Bacteria were grown to OD₆₀₀ = 0.4 and protein expression was induced with 0.75 mM IPTG for 4 hours. All proteins were 6xHis tagged and purified by Nickel-NTA column chromatography using standard techniques with the following modifications. rPcrV was purified under nondenaturing conditions at 4°C in buffers containing 50 mM sodium phosphate and 300 mM NaCl, pH 7.5. rPcrV was dialyzed extensively against PBS. rOprF and rOprI were purified under denaturing conditions with 8M urea in buffers containing 50 mM sodium phosphate (Sigma S-0751; Fisher Scientific S374–500) and 300 mM NaCl (VWR X190–5KG), pH 7.5. For protein refolding, urea was removed through stepwise dialysis into 50 mM sodium phosphate, 300 mM NaCl, pH 7.5 (rOprF) or 50 mM sodium phosphate, 150 mM NaCl, pH 8 (rOprI). For all recombinant proteins, concentration was assessed by BCA assay (Prometheus 18–440) and identity was confirmed by western blot. Endotoxin was not removed from recombinant proteins, but was measured in all vaccine components by ToxinSensor Chromogenic LAL Endotoxin Assay Kit (Genscript, NJ, USA) according to the manufacturer’s instructions. Endotoxin can contribute to immunogenicity, so we confirmed that endotoxin levels were within recommended limits for animal research [37,38]. All vaccine components tested contained less than 5.6 endotoxin units (EU)/ml of endotoxin, or 0.22 EU per dose (Supplementary Table 1). This is well within the maximum recommended level of 20 EU/ml for recombinant subunit vaccines [37].

TMV Conjugation

TMV-PcrV, TMV-OprF, and TMV-OprI conjugates were prepared as described previously with minor modifications [20,26,27]. Briefly, rPcrV, rOprF, and rOprI were mixed in a 1:1 ratio then covalently bonded to lysine-modified and binary ethyleneimine (BEI) inactivated [39] rTMV using EDC-mediated amine conjugation to generate TMV-PcrV, TMV-OprF, and TMV-OprI. EDC concentrations were optimized for each protein to ensure that no “free” protein antigen remained. Conjugate vaccines were dialyzed against PBS overnight and protein concentration was determined by BCA assay. Vaccine aliquots were stored at −20°C until the day of vaccine administration.

LNP preparation

LNPs, or liposomes, were prepared by the film hydration method [40] with some modifications. The lipids ALC-0315 (BP-25498), ALC-0159 (BP-25711), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) (BP-25623), and cholesterol (BP-26125) (BroadPharm, San Diego, CA) were each dissolved in chloroform and combined at the molar ratio 46:2:9.5:42.5 in a round-bottom test tube. Chloroform was evaporated under nitrogen gas and traces of solvent were removed under vacuum. The lipid film was rehydrated in PBS at 60°C. Liposomes were extruded through a 200 nm pore size polycarbonate membrane using a LiposoFast Liposome Factory (Avestin, Sigma Z373400). rOprF and rOprI were adsorbed onto liposomes at 60°C, then liposomes recovered to room temperature and were filtered through Sephadex G75 (Cytiva 17005003) packed Mini Bio-Spin columns (BioRad 7326207) to exclude free protein. Liposomes were analyzed for protein concentration by BCA assay (Prometheus 18–440) and phospholipid concentration by Stewart assay [41]. Liposome size was assessed by ZetaView (Particle Metrix) and ZetaSizer (Malvern Panalytical).

Vaccination, serum collection, and bronchoalveolar lavage (BAL) collection

Vaccines were diluted to 20μg total for TMV conjugate vaccines (10μg protein + 10μg TMV) or 10μg for recombinant protein controls, with 100μg curdlan (InvivoGen CDA-41–01) (where indicated) per dose in 40μl (for IN) or 100μl (for IM) sterile 0.9% saline. Mice were anesthetized with ketamine/xylazine (100mg/10mg/kg), and vaccine was administered by IN droplet using a pipettor or IM injection into the left quadriceps muscle. Blood was obtained from the retro-orbital sinus or the submandibular vein at the indicated timepoints. BAL fluid was collected under anesthesia with two combined lavages of 0.5 mL PBS each.

ELISAs for antibody titers

ELISA was performed using standard techniques as previously described [27] with the following parameters. Proteins were coated at 2μg/ml concentration; blocking buffer was 1% BSA in PBS; secondary antibody was Goat anti-mouse Ig(H+L)-HRP (#1010–05), IgA-HRP (#1040–05), IgG1-HRP (#1070–05), IgG2b-HRP (#1090–05), IgG2c-HRP (#1078–05), or IgG3-HRP (#1100–05) (Southern Biotech, Brimingham AL) diluted 1:5000; substrate was TMB (Seracare 5120–0077); stop solution was 2N H₂SO₄. Absorbance was read at 450 – 570 nm on a Spectramax ABS spectrophotometer (Molecular Diagnostics). Serum antibody concentration was reported as 50% maximal binding titers and were calculated using Prism 7 (Graphpad).

PA challenge

PA Boston 41501 was grown overnight in a shaking incubator at 37°C in TSB, then seeded 1:50 in a 10 mL culture. The culture was returned to incubator until OD₆₀₀ =~0.7. The culture was pelleted and washed 3 times with PBS, then the concentration was adjusted with sterile 0.9% saline until OD₆₀₀ = 1.25 (~1 × 10⁹ CFU/ml). The pellet was diluted in sterile saline to yield the dosage ~2 × 10⁷ CFU/40μl or ~2 × 10⁸ CFU/40 μl, where indicated. Mice were challenged under anesthesia with ketamine/xylazine (100mg/10mg/kg) by IN droplet in a biosafety cabinet and monitored daily until they succumbed to infection or were euthanized due to severe morbidity (loss of more than 20% of original body weight and/or severe vestibular symptoms). Recovered mice were euthanized by cardiac puncture under anesthesia. Serial dilutions were plated to calculate and report the exact inoculum given for each challenge. We previously determined the LD₅₀ of PA Boston 41501 to be 2 × 10⁶ CFU using the method of Reed and Muench [42]. For sublethal challenge to assess cytokine production and cell recruitment, the mice received 1×10⁶ CFU in 40μl, and were euthanized at the indicated timepoints.

Opsonization analysis by flow cytometry

Serum samples were added to a 96-well plate at 1:10 dilution. 2 × 10⁷ CFU of PA was added to each well to yield a final serum dilution of 1:20. The plate was incubated at 37°C for 1 hour with shaking at 90 rpm, then washed and stained with goat anti-mouse Ig(H+L)-AlexaFluor488 (Invitrogen A-11001). Cells were fixed in 2% PFA and analyzed on a FacsCelesta flow cytometer (BD). Median fluorescence intensity (MFI) was calculated using FlowLogic software (Miltenyi Biotec).

Western blots and T3SS stimulation

To compare the ability of pre- and post-challenge serum to recognize PA antigens, PA was cultured in TSB, pelleted, sonicated, and 8 μg total PA protein was loaded in each lane for separation by SDS-PAGE by standard techniques. rPcrV (300 ng) was loaded in the last lane as a positive control. Proteins were transferred to a PVDF membrane and probed with pooled serum from vaccinated mice collected before or after surviving challenge, diluted 1:1000. Secondary antibody was Goat anti-mouse IgG(H+L)-HRP diluted 1:20000 (Jackson Immunoresearch 115–035-062) and substrate was SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Scientific 34577). T3SS expression was stimulated according to published methods [43]. PAO1 and PA Boston were grown in TSB in the presence or absence of chelated medium treated with Chelex (Sigma C7901), 100 mM monosodium glutamate (Sigma-Aldrich 49621), and 1% glycerol (Apex GB0232) (chel); 0–10 mM nitriloacetic acid (NTA) (Sigma-Aldrich N9877); and 0–5 mM ethyleneglycol-bis-(β-aminoethyl ether) N,N,N’,N’-tetraacetic acid (EGTA) (Sigma E-4378). Pellets were lysed, and 8μg total PA protein was loaded in each lane for separation by SDS-PAGE and probed with pre-challenge pooled sera as stated above.

Lung and spleen cytokine analysis

Mice were vaccinated and then received a sublethal challenge at either day 70 or day 48 after primary immunization. Variation in challenge timing was due to equipment availability in the NYMC vivarium. At timepoints 8, 24, and 48 hours after challenge, mice were euthanized by cardiac puncture under anesthesia. Spleen and right lung lobes were collected in ice-cold PBS. Lungs were digested into single-cell suspensions using a GentleMACS Lung Dissociation Kit (Miltenyi Biotec 130–095-927) according to the manufacturer’s instructions. Spleens were pressed through a 70μm filter, and RBCs were lysed. 2 × 10⁵ cells from each tissue were plated in RPMI supplemented with 10% fetal bovine serum, GlutaMAX (Gibco 35050061), and gentamicin (Gibco 15710–064) in triplicate in 96-well plates, then stimulated with Concanavalin A (5 μg/ml) or 100 μg/mL rPcrV for 72hrs. Supernatants were collected and stored at −80°C. IFN-γ, TNF-α, and IL-17AF levels in supernatants were measured using ELISA kits (Invitrogen 88–7314-88, 88–7324-88, 88–8711-88) according to the manufacturer’s instructions. Experimental replicates were combined for analysis.

Lung phagocyte recruitment analysis

Lungs were collected and single cell suspensions were generated as described above. 2 × 10⁶ cells from each lung were incubated with the antibodies as per the table below. Cells were analyzed on a FacsCelesta (BD) and gated to define populations of neutrophils, macrophages, AMs, DCs, and other leukocytes. MFI of activation markers was calculated using FlowLogic (Miltenyi Biotec).

Antibody	Isotype and clone	Fluorophore	Clone	Manufacturer/catalog
Anti-mouse CD45	N/A	Alexa Fluor 700	30-F11	eBioscience 56-0451-80
Anti-mouse Ly-6G	N/A	FITC	1A8-Ly6g	eBioscience 11-9668-80
Anti-mouse F4/80	Armenian Hamster IgG Clone: eBio299Arm	PE	BM8	eBioscience 12-4801-80
Anti-mouse CD11c	Rat IgG2a K Clone: eBR2a	APC-eFluor 780	N418	eBioscience 47-0114-80
Anti-mouse MHC II I-Ab	Armenian Hamster IgG Clone: eBio299Arm	eFluor 450	AF6-120.1	eBioscience 48-5320-80
Anti-mouse CD80 (B7-1)	Armenian Hamster IgG Clone: eBio299Arm	PerCP-eFluor 710	16-10A1	eBioscience 46-0801-82
Anti-mouse CD86	Rat IgG2a K Clone: eBR2a	APC	GL1	eBioscience 17-0862-81

Statistical analysis

All statistical analyses were performed using Prism 7 or 10 (Graphpad). Values of p < 0.05 were considered statistically significant. Antibody titers and MFI data were compared by One-way ANOVA and Sidak’s multiple comparisons test. Correlation between antibody titers and survival were assessed using Pearson correlation (point-biserial correlation). Pre- and post-challenge opsonization were compared by two-tailed paired t test. Survival curves were compared according to the Log-rank (Mantel-Cox) test. Cytokine secretion and phagocyte recruitment were analyzed by Two-way ANOVA and Tukey’s multiple comparisons test.

Results

PcrV conjugated to TMV protects against lethal infection with PA.

We first evaluated TMV-PcrV in a pilot study without adjuvant to determine if IN antigen administration using TMV could protect against lethal infection using a mouse model of acute PA pneumonia. Mice were anesthetized and vaccinated by IN administration of 20 μg TMV-PcrV and boosted twice at two-week intervals (Figure 1a). 75% of mice challenged with 8×10⁶ CFU (4 LD₅₀) of PA Boston 41501 survived infection while all mice receiving a 10-fold higher amount (40 LD₅₀) succumbed (Figure 1b). In a second experiment, we vaccinated mice with TMV-PcrV or rPcrV without TMV in the presence of cyclic-di-GMP, a Stimulator of Interferon Genes (STING) agonist [44] that has previously been shown to boost the antibody response to a TMV-conjugated antigen [23]. 67% of TMV-PcrV vaccinated mice and 50% of rPcrV vaccinated mice survived challenge with 1 × 10⁷ CFU (5 LD₅₀) of PA Boston 41501. All mice that received 1 × 10⁸ CFU (50 LD₅₀) of the bacteria succumbed to infection (Figure 1c).

Figure 1: Monovalent TMV-PcrV protects mice against a lethal IN PA challenge.

(a) The vaccine studies include three vaccine doses, serum collection for measuring antibody titers, and a pneumonic challenge with a lethal dose of PA Boston 41501. Serum is also collected from mice that survive challenge. Created in https://BioRender.com. (b and c) Percent survival of mice immunized with (b) TMV-PcrV or (c) either TMV-PcrV or rPcrV + cyclic-di-GMP adjuvant then challenged with one of two lethal PA doses. The mice in (b) received 8 × 10⁷ (40 LD₅₀) or 0.8 × 10⁷ CFU (4 LD₅₀), and the mice in (c) received 1 × 10⁸ (50 LD₅₀) or 1 × 10⁷ CFU (5 LD₅₀). (b-c) n=4–10 mice per group; survival curves compared by Log-rank (Mantel-Cox) test; titers compared by one-way ANOVA and Sidak’s multiple comparisons test; ns: not significant.

Addition of curdlan to TMV-PcrV enhances vaccine efficacy against high dose lethal PA challenge.

As we did not see enhancement of protective efficacy with cyclic-di-GMP (Figure 1b–c), we assessed curdlan as an adjuvant to improve TMV-PcrV vaccine efficacy. Inclusion of curdlan significantly increased total anti-PcrV antibody titers in TMV-PcrV+c vaccinated mice, compared to both TMV-PcrV (p < 0.0001), rPcrV (p < 0.0001), and rPcrV+c (p < 0.0001) vaccinated mice as measured 7 days after each boost (day 21 and day 35). Anti-PcrV antibody was not increased in rPcrV+c vaccinated mice compared to unadjuvanted rPcrV (p = 0.9572, day 21; and p = 0.8437, day 35; Figure 2a–b). These mice were challenged with 4.5 × 10⁷ CFU (22.5 LD₅₀) of PA Boston 41501. Adding curdlan to TMV-PcrV significantly enhanced survival following this challenge: 56% of mice vaccinated with TMV-PcrV+c survived infection, while only 10% of mice that received unadjuvanted TMV-PcrV or rPcrV+c survived. All unadjuvanted rPcrV-vaccinated and unvaccinated mice succumbed to infection (Figure 2c), possibly because this PA challenge dose was 4–5 fold higher than the low dose used in the previous studies (Figure 1b–c). Surviving mice recovered by day 7 after challenge as evidenced by return to pre-challenge weight and activity (Figure 2d).

Figure 2: Adding curdlan to TMV-PcrV boosts humoral immunogenicity and protection against high dose lethal PA challenge.

(a-b) Anti-PcrV total Ig antibody titers in mice vaccinated with TMV-PcrV or rPcrV in the presence of absence of curdlan. Titers were measured at (a) day 21 (7 days after the first boost) and (b) day 35 (7 days after the second boost). (c) Mice vaccinated with TMV-PcrV or rPcrV with or without curdlan were challenged with 4.5 × 10⁷ CFU (22.5 LD₅₀) of PA Boston 41501. (d) Surviving mice recover by 7 days post-challenge. (a-c) n=9–19 mice per group; survival curves compared by Log-rank (Mantel-Cox) test; titers compared by one-way ANOVA and Sidak’s multiple comparisons test. Antibody data are presented as 50% maximal binding titers. (e) Serum from vaccinated mice and unvaccinated controls prior to challenge was incubated with live PA to assess opsonizing ability (Created in https://BioRender.com) against (f) PA Boston 41501 and (g) PAO1. (h-i) Serum from mice that survived a lethal PA Boston challenge was assessed for opsonization of (h) PA Boston 41501 and (i) PAO1. (f,g) n=5 mice per group; MFI compared by one-way ANOVA and Sidak’s multiple comparisons test. (h,i) n=5 mice per group; MFI compared by two-tailed paired t test.

Serum from vaccinated mice collected prior to challenge was then incubated with two PA strains, Boston 41501 and PAO1, and antibody binding was assessed by flow cytometry (Figure 2e). The level of bound antibody was indistinguishable among sera from all vaccinated mouse groups compared to sera from unvaccinated mice for both PA strains analyzed, as similar MFIs were observed among all groups (Figure 2f–g). This indicates either that the method is not of sufficient sensitivity to detect anti-PcrV antibody binding, or that anti-PcrV antibody is non-opsonizing which was been suggested previously by other groups [45–48]. When serum from mice that survived PA Boston 41501 challenge (post-challenge serum) was incubated with PA Boston 41501, MFI was significantly increased compared to pre-challenge sera and unvaccinated mouse serum. A trend towards increased antibody binding was observed for post-challenge antibody binding to PAO1, and values approached statistical significance (p = 0.0867; Figure 2h–i). This indicates the generation of antibodies to other PA Boston epitopes after challenge, which we confirmed by western blot (Supplementary Figure 1a–b). It also suggests that the anti-PA antibody generated, at least in the early stages after infection (14 days post-challenge), is strain specific with limited cross-reactivity to other strains. As PcrV is a component of the T3SS, we also assessed whether culturing PA in media that stimulated T3SS mediator release could enhance serum recognition of the protein. We found that growth in stimulatory medium did not enhance serum-mediated detection of PcrV by western blot (Supplementary Figure 1c).

Anti-PcrV IgG1 antibodies predict survival after PA challenge

As our vaccine is administered IN, we evaluated local antibody production in the respiratory tract by bronchoalveolar lavage (BAL) on day 37, one week after the third boost. Localized antibodies in the lungs, particularly sIgA, are important for protecting the respiratory tract from infection [49]. IN immunization with TMV-PcrV and TMV-PcrV+c resulted in significantly higher anti-PcrV total antibody in BAL fluid (BALF) than unvaccinated mice (Figure 3a). Anti-PcrV IgA was present only in BALF from mice that received TMV-based vaccines. TMV-PcrV and TMV-PcrV+c elicited significantly higher BALF IgA compared to rPcrV, rPcrV+c, and unvaccinated mice (Figure 3b). From these data, we conclude that TMV-PcrV conjugation enhances the production of a localized antibody response when administered via the nose.

Figure 3: Anti-PcrV IgG1 antibodies predict survival after PA challenge.

(a,b) Bronchoalveolar lavage (BAL) was performed on mice 37 days after the first vaccination (9 days after the second boost). BALF antibody titers were calculated for (a) total anti-PcrV antibodies and (b) anti-PcrV IgA. (c-h) Serum was obtained on day 35 post-vaccination (7 days after the second boost) from all mice and 14 days post-challenge from surviving mice. Anti-PcrV antibody titers were calculated for (c) total antibody; (d) IgG1; (e) IgG2b; (f) IgG2c; (g) IgG3; and (h) IgA. (a-h) n=5–10 mice per group. (a-b) Antibody data are presented as 50% maximal binding titers. Antibody titers were compared by one-way ANOVA and Sidak’s multiple comparisons test.

While we did not detect PA opsonization by anti-PcrV antibodies in vaccinated serum, we did observe that the presence of anti-PcrV antibody was associated with survival from infection: all mice that did not produce detectable anti-PcrV antibody in response to vaccination succumbed to infection (Figure 3c). Despite this, day 35 pre-challenge anti-PcrV total antibody titers only weakly correlated with survival (R² = 0.3460) (Figure 3c). When we evaluated anti-PcrV IgG subclasses individually, we observed that anti-PcrV IgG1 had a stronger correlation with survival (R² = 0.5143). All but two mice with detectable anti-PcrV IgG1 survived infection, while all mice with undetectable anti-PcrV IgG1 succumbed to infection (Figure 3d). Differences in IgG2b and IgG2c antibody titers between surviving and succumbing mice were similar to those for total antibody (Figure 3e–f). Anti-PcrV IgG3 titers showed a trend opposite of IgG1, in which nearly all mice with detectable anti-PcrV IgG3 died after PA challenge (Figure 3g). However, many mice that died and mice that survived also had undetectable IgG3, so it is of little value for predicting protection. There was little to no detectable anti-PcrV IgA in the serum (Figure 3h). Pre-challenge serum from vaccinated mice did not produce detectable antibody against whole lysate, whereas post-challenge serum did (Supplementary Figure 2); though it is not clear if PcrV is present in sufficient concentrations in a whole bacterial lysate to competitively bind to a plate and be detectable by ELISA.

Curdlan enhances the early localized Th17 response after TMV-PcrV vaccination and PA challenge.

T cell responses, particularly Th1 and Th17, have been implicated in protection from PA [28–30,50–52]. We observed greater protection from PA challenge with the use of curdlan (Figure 2c), which skews immune responses toward eliciting a Th1/Th17 cellular response [35,36]. To evaluate whether vaccine-induced T cell responses are also skewed, mice were vaccinated and challenged with a nonlethal dose of PA (9 × 10⁵ CFU or ~0.5 LD₅₀), which allowed us to measure antigen recall responses in cells isolated from lungs and spleens collected 8, 24, and 48 hours post-infection (Figure 4a).

Figure 4: Curdlan enhances early localized and systemic Th17 and Th1 responses after TMV-PcrV vaccination and PA challenge.

(a) After vaccination, serum collection, and a nonlethal PA challenge, right lungs and spleens were collected for ex vivo stimulation to measure cytokine secretion. Created in https://BioRender.com. (b,c) IL-17 secretion from stimulated (b) lung and (c) spleen cells from vaccinated and unvaccinated mice. (d,e) IFN-γ secretion from stimulated (d) lung and (e) spleen cells from vaccinated and unvaccinated mice. (f, g) TNF-α secretion from stimulated (f) lung and (g) spleen cells from vaccinated and unvaccinated mice. (b-g) Data pooled from two independent experiments; n=3 mice per group per experiment; cytokines compared by two-way ANOVA and Tukey’s multiple comparisons test.

Following ex vivo restimulation with rPcrV, lung and spleen cells isolated at all timepoints secreted the cytokines IL-17A/F and IFN-γ, in addition to TNF-α (Figure 4b–g). IL-17A/F secretion was significantly greater in the lung cells of mice vaccinated with TMV-PcrV+c at 8 hours compared to unvaccinated mice. IL-17A/F secretion was also increased from the lungs of TMV-PcrV+c-vaccinated mice at 24 and 48 hours compared to both TMV-PcrV-vaccinated and unvaccinated mice. IL-17A/F was also secreted from lung cells of mice vaccinated with rPcrV+c, but levels did not reach significance (Figure 4b). The use of curdlan during immunization significantly enhanced IL-17A/F secretion from splenocytes at 8 hours in TMV-PcrV+c or rPcrV+c recipient mice. By 48 hours, IL-17A/F secretion declined in all groups except TMV-PcrV+c vaccinated mice, where it remained significantly elevated compared to all other groups (Figure 4c).

Elevated IFN-γ secretion from lung cells was only observed in TMV-PcrV+c vaccinated mice, and only reached significant levels at 24 hours compared to TMV-PcrV and rPcrV vaccinated mice, and unvaccinated mice (Figure 4d), though this higher mean is largely due to a single sample with very high secretion. In contrast, IFN-γ secretion from spleen cells was lower at 24 hours in all vaccinated groups compared to other timepoints. TMV-PcrV+c vaccinated mice secreted significantly more IFN-γ from spleen cells at 48 hours compared to 24 hours, and compared to unvaccinated mice at 48 hours (Figure 4e). rPcrV+c vaccinated mice secreted significantly more IFN-γ from splenocytes at 8 hours compared to 24 hours, though this elevation was also linked to high secretion from a single mouse.

In the lungs, TNF-α secretion trended upward over time for vaccinated and unvaccinated mice, significantly increasing in all groups from 8 hours to 24 hours. In TMV-PcrV, TMV-PcrV+c, and PcrV+c vaccinated mice, levels began to decline after 24 hours, while they continued to increase in rPcrV-vaccinated and unvaccinated mice at 48 hours (Figure 4f). TNF-α secretion was significantly increased at 8 hours for rPcrV+c vaccinated mice compared to unvaccinated mice, and at 24 hours for both TMV-PcrV+c and rPcrV+c mice compared to unvaccinated mice. In the spleen, there were no differences in TNF-α secretion (Figure 4g).

Neutrophils are believed to be critically important for the clearance of PA lung infections. We therefore examined the recruitment of neutrophils to the lungs by flow cytometry using the same model as above (Supplementary Figure 3a). Neutrophils were rapidly recruited and represented a large percentage of the total CD45+ cells in the lungs of all infected mice at 8 hours post-challenge. Levels declined significantly with time in all groups, but were significantly lower in TMV-PcrV, TMV-PcrV+c, and PcrV+c vaccinated mice as compared to rPcrV-vaccinated and unvaccinated mice at 48 hours (Supplementary Figure 3b). This complemented observed changes in TNF-α expression with time: TNF-α continued to increase in the lungs of rPcrV vaccinated and unvaccinated mice at 48 hours, whereas it decreased in all other groups (Figure 4f). By comparison, F4/80+ cells (macrophages and monocytes) were at their lowest detected levels at 8 hours post-challenge and significantly increased with time, indicating that the macrophage/monocyte recruitment was slower than neutrophils (Supplementary Figure 3c). At 8 hours, the majority of macrophages in TMV-vaccinated mice had an alveolar macrophage (AM) phenotype, which was significantly higher than unvaccinated mice and moderately higher than rPcrV vaccinated mice (Supplementary Figure 3g). Levels of AMs decreased with time corresponding to increases in total macrophages/monocytes, likely indicating an influx of inflammatory macrophages/monocytes diluting the AM population. Antigen presentation markers were highest on macrophages/monocytes at 8 hours post-infection and decreased with time, which may correspond to activated APCs leaving the lung to the draining LNs via the lymph (Supplementary Figure 3d–f; 3h–j). In contrast, DCs were largely unaffected during the early stages of infection (Supplementary Figure 3k–n).

OprF and OprI do not contribute to protection in combination with PcrV or alone.

PcrV, as a single antigen vaccine, induced protection from lethal infection in 56%–75% of mice vaccinated depending on the dose of challenge bacteria used and the presence of adjuvant (Figures 1b, 1c, 2c). In parallel to evaluating the use of curdlan as an adjuvant, we also evaluated the inclusion of additional TMV-conjugated vaccine antigens to increase protective efficacy. We immunized mice with an unadjuvanted trivalent vaccine containing PcrV with additional antigens OprF and OprI. TMV-PcrV + TMV-OprF + TMV-OprI (TMV-V/F/I) or unconjugated recombinant proteins (rV/F/I) were mixed and administered as described above (Figure 1a). Similar to TMV-PcrV in the absence of curdlan (Figure 1b–c), 70% of vaccinated mice challenged with 1 × 10⁷ CFU (5 LD₅₀) of PA Boston 41501 survived infection, but all mice given 1 × 10⁸ CFU (50 LD₅₀) succumbed to challenge. Equivalent survival after 5x LD₅₀ challenge was observed in mice that received TMV-V/F/I and rV/F/I (Figure 5a), and all surviving mice returned to their original body weight by day 8 post infection (Figure 5b). Anti-vaccine antibodies from day 35 demonstrated a trend (p = 0.1530) towards increased anti-PcrV antibody in TMV-V/F/I vaccinated mice over rV/F/I vaccination, while anti-OprF antibody titers were significantly higher in rV/F/I vaccinated mice compared to TMV-V/F/I vaccinated mice. TMV-V/F/I and rV/F/I vaccines induced similar anti-OprI antibody titers (Figure 5c). The trivalent TMV-V/F/I vaccine (Figure 5a) did not appear to offer any improvement over the monovalent TMV-PcrV vaccine (Figure 1b), as both elicited similar levels of survival.

Figure 5: OprF and OprI do not contribute to protection in combination with PcrV or alone.

(a) Mice were vaccinated with unadjuvanted trivalent TMV-V/F/I vaccine and challenged with PA Boston 41501 (5 or 50 LD₅₀). (b) Weight loss tracking during challenge. (c) Titers for anti-PcrV, anti-OprF, and anti-OprI total antibody. (d) Anti-OprF and (e) anti-OprI antibody titers after three doses of bivalent TMV-OprF/I+c, LNP-OprF/I+c, or rOprF/I+c given IN or IM. (f) Anti-OprF IgA levels in BALF. (g) Mice were vaccinated, boosted, and challenged with 3 × 10⁷ CFU (15 LD₅₀) PA Boston 41501. (h,i) Serum collected from vaccinated and unvaccinated mice prioer to challenge was incubated with live PA to assess opsonizing ability of (h) PA Boston or (i) PAO1. (a-c) n=9–10 mice per group; survival curves compared by Log-rank (Mantel-Cox) test; titers compared by one-way ANOVA and Sidak’s multiple comparisons test. (d-i) n=7–8 mice per group; survival curves compared by Log-rank (Mantel-Cox) test; titers and MFI compared by one-way ANOVA and Sidak’s multiple comparisons test. Antibody data are presented as 50% maximal binding titers.

The decreased level of anti-OprF antibody titers in TMV-V/F/I immunized mice as compared to rV/F/I immunized mice suggested the possibility that the integral membrane protein was not folded properly on the surface of TMV. To address this, we generated lipid nanoparticles (LNPs), and independently adsorbed rOprF and rOprI onto them, generating LNP-OprF and LNP-OprI. We included OprI in the LNPs because it is also an integral membrane protein. To evaluate the contribution of OprF and OprI to the trivalent PcrV vaccine, we evaluated the protective efficacy of TMV-OprF + TMV-OprI (TMV-OprF/I) and LNP-OprF + LNP-OprI (LNP-OprF/I) independently of PcrV. We included curdlan as we had observed increased protective efficacy in the PcrV vaccine (Figure 2c).

TMV-OprF/I+c or LNP-OprF/I+c or rOprF/I+c were administered either IN or IM, followed by challenge with 3 × 10⁷ CFU (15 LD₅₀) PA Boston 41501. Overall, IM administration of vaccine significantly increased antibody titers compared to IN administration. IM LNP-OprF/I+c and rOprF/I+c produced significantly higher anti-OprF antibody titers than IM TMV-OprF/I+c, while IN administration resulted in titers that were similar among the groups (Figure 5d). For OprI, the opposite was observed: IM TMV-OprF/I+c elicited higher anti-OprI titers than IM LNP-OprF/I or IM rOprF/I (Figure 5e). Mice vaccinated with OprF (without OprI) had low to undetectable anti-vaccine BALF IgA for both IN and IM routes of administration, with absorbance values so low they precluded calculating 50% maximal binding titers (Figure 5f). More importantly, none of the vaccine formulations or administration routes induced protective immunity against lethal challenge; vaccinated mice succumbed to infection at the same rate as unvaccinated mice (Figure 5g). As with PcrV (Figure 2e–g), we also evaluated if anti-OprF and OprI antibodies in sera were able to bind directly to live PA. Incubation of PA Boston 41501 with serum from IM rOprF/I+c vaccinated mice showed a trend (p = 0.0658) towards increased antibody binding compared to serum from unvaccinated mice. Incubation of serum antibodies with PAO1 did not demonstrate any pathogen-specific binding trends (Figure 5h–i). These data suggest that OprF and OprI, either delivered as TMV conjugates or in lipid nanoparticles, are insufficient on their own to protect mice against a lethal PA infection with Boston 41501, and that these two antigens likely did not contribute to (or enhance) the protective efficacy of recombinant proteins when administered with rPcrV in the trivalent vaccine (Figure 5a). In contrast, monovalent TMV-PcrV+c induced a protective humoral and cellular immune response against PA Boston 41501 (Figures 2–4).

Discussion

MDR PA is a significant cause of morbidity and mortality from HAIs. The continuing spread of MDR PA, particularly outside of western nations where antibiotic stewardship programs are lacking, requires the development of alternative approaches to limit the impact of this public health threat. Vaccination has the potential to eliminate the threat of MDR PA in all healthcare settings; however, an effective vaccine has remained elusive. TMV is an ideal platform because unlike some vaccine platforms, prior exposure to TMV does not impair subsequent immune responses to TMV conjugates [15,22]. It is important to note that TMV is a delivery platform and not a viral vector. Adenoviral and other viral vectors encode vaccine antigens, so the vectors must overcome any pre-existing immunity to infect host cells, which then express the vaccine antigen. TMV is intended to be internalized like an immunogen, as TMV conjugate vaccines already contain the protein antigen. Pre-existing antibodies in this case may assist in Fc mediated uptake rather than neutralizing entry [15,22]. We chose PcrV as the focus of these studies, as we have previously demonstrated that the equivalent T3SS tip protein in Y. pestis, LcrV, is an efficacious vaccine antigen when fused to TMV to improve antigen delivery [26,27]. An anti-PcrV monoclonal antibody (mAb) has demonstrated some efficacy in reducing bacterial burden in patients with PA VAP [45] and a bispecific mAb against PcrV and Psl polysaccharide enhanced neutrophil killing of PA in bronchiectasis patients [53], indicating that PcrV has potential as a human therapeutic target. Other groups have also recently developed PcrV-containing vaccines that protected against PA in preclinical studies [54–56].

In the present study, we observed protective efficacy of PcrV as a vaccine antigen when it is conjugated to TMV (Figure 1b–c). Anti-PcrV serum antibody titers (Figure 2a–b), local cytokine production (Figure 4), and protection against PA Boston 41501 challenge (Figure 2c) were enhanced with the addition of curdlan compared to vaccination with TMV-PcrV alone. We attempted to determine if these antibodies might contribute to clearance of the bacteria through opsonization of live PA. We observed a lack of PA opsonization from pre-challenge antibodies in the serum of vaccinated mice, regardless of curdlan use (Figure 2f–g). Serum from mice vaccinated against PcrV also did not bind to PA lysate in an ELISA format (Supplementary Figure 2). However, it is not clear that PcrV would be a dominant plate coating antigen when using whole bacterial lysate to coat plates, and we do not have a commercial anti-PcrV antibody available to us to test this. Thus, we evaluated binding to the whole bacteria by flow cytometry. These results corroborate previous data from other groups that were evaluating the role of the T3SS in pathogenesis. They found that antibodies against PcrV functioned through inhibition of T3SS function rather than opsonizing/bactericidal activity [45–48]. We did observe a correlation between IgG1 antibodies and survival, indicating a role for antibody in protection. IgG1 is not typically associated with either complement mediated killing or enhancement of phagocytosis [57,58], thus a role in T3SS neutralization is fitting and the target of future studies. Anti-OprF and OprI antibodies showed a small amount of opsonization in our assay, but only from the group with the highest antibody titers (Figure 5d,e,h). Thus, the method we are using may not be capable of detecting antibody binding when antigens on the bacterial surface are limited. By comparison, the assay does detect antibody binding in serum from post-challenge surviving mice (Figure 2h), which exhibited a much more diverse antibody repertoire (Supplementary Figure 1a–b). For respiratory and other mucosal infections, SIgA is an important antibody that can neutralize pathogens, preventing colonization. In this study, we found that TMV-PcrV vaccines with and without curdlan elicited anti-PcrV IgA in BALF, suggesting that conjugation to TMV may mediate protection by skewing the humoral response toward IgA production (Figure 3b).

Previous studies have demonstrated that Th1 and Th17 responses are important for controlling PA infections [28,29,52,59], so we assessed the effect of adding the Th1/Th17-skewing [28,35,36] adjuvant curdlan to the vaccine formula. In this study, we utilized a nonlethal challenge model to identify differences in cytokine responses to PA in vaccinated compared to unvaccinated mice, as all unvaccinated mice succumb rapidly to lethal infection. We observed that mice treated with TMV-PcrV+c saw consistent early rises in localized secretion of IL-17A/F (Figure 4b). TMV-PcrV+c immunized mice showed a sharp increase in systemic (spleen) IL-17A/F secretion 48 hours post-challenge in contrast to all other vaccine formulas (Figure 4c). This could indicate a distribution of IL-17A/F producing cells in these mice that follow dissemination of the bacteria, demonstrating that this IN vaccine can produce appropriate immune responses in distal tissues. However, it is notable that mostly TMV-PcrV+c immunized mice survived the 22.5 LD₅₀ lethal challenge (Figure 2c). Taken together, these results indicate that the TMV-PcrV+c vaccine has the potential to provide protection against a mucosal PA infection by eliciting PA-specific immune responses at both the site of infection (the respiratory tract) and distal sites of dissemination. It also raises the possibility that the IN vaccine could be protective against nonmucosal PA infections, though this needs to be confirmed using other disease models.

Neutrophils were a high proportion of leukocytes in the lung at 8 hours post-challenge in all vaccinated and unvaccinated mice. Cell numbers decreased over time in all groups but remained significantly higher in rPcrV and unvaccinated mice compared to other groups (Supplementary Figure 3b). These mice were also the only groups to see a continuing increase in TNF-α from lung cells at 48 hours, while other vaccine groups started seeing levels drop by 48 hours (Figure 4f). This speaks to a more intense inflammatory environment in the lungs of these unprotected mice (Figure 2c), which could contribute to further lung damage. Macrophage responses proceeded at a similar rate in all groups by comparison, indicating a dominant early Th17/PMN inflammatory response, followed by a slower influx of monocytes/macrophages. It should be noted that the cytokine response evaluated in figure 4 was an anti-vaccine recall response and not a measure of total IL-17, IFN-γ, or TNF-α production in the lungs.

Others have demonstrated the efficacy of TMV as a platform to present epitopes from OprF to protect against PA infection in preclinical studies [24]. Further, OprF-OprI fusion proteins given IN and IM have been evaluated in phase 1/2 and 2/3 clinical trials, where they demonstrated immunogenicity but failed to prevent mortality [34,60]. We suggest that the IM vaccine was suboptimal because it was administered parenterally but intended to protect against respiratory infection. In this study, we evaluated full-length OprF and OprI conjugated to TMV or formulated in a LNP and administered either as part of a trivalent vaccine with PcrV or independently in the presence of curdlan. The inclusion of rOprF and rOprI with rPcrV increased protective efficacy of the trivalent recombinant protein vaccine to that of its TMV-conjugated counterpart, but apparently did not enhance protective efficacy beyond TMV-PcrV on its own. In the absence of PcrV, OprF and OprI were not protective, despite inducing strong antibody responses. It is unclear as to why OprF and OprI were not protective at all, despite several preclinical studies demonstrating their efficacy [2]. It is possible that anti-OprF and OprI antibodies and cellular responses are not protective against the strain of PA we utilized for most of our studies, Boston 41501, or that whole antigen vaccines do not drive a neutralizing response, and peptide vaccines may be more effective [24]. This highlights a need for multivalent vaccines for protection against multiple different strains.

There were some limitations in this study. First, we only demonstrated protection against one strain of PA. We were unable to demonstrate binding of antibodies to another strain, PAO1, a different strain commonly used in laboratories, but we also did not see direct binding to Boston 41501. Interestingly, serum collected from vaccinated mice that survived a PA Boston 41501 challenge (14 days after challenge) showed a trend toward opsonizing PAO1 better than serum collected before challenge (Figure 2i), suggesting that cross-strain protection by an optimized vaccine is possible. However, given the strain diversity of disease-causing PA, a broadly protective vaccine will likely require several target antigens. PcrV alone appears protective for at least some PA strains (our work and [54–56]), and it can serve as a base for further multi-antigen vaccine development. A further limitation is that cytokine secretion, phagocyte recruitment, and protection cannot be assessed in the same animals because unprotected mice succumb to disease rapidly. Rapid death is believed to result from acute lung injury leading to rapid dissemination [61]. Thus, vaccine-induced cytokine secretion and phagocyte recruitment cannot be correlated with vaccine-mediated protection. In our PA infection model, minor variations in bacterial growth cause variability in the final challenge dose, which is determined by plating dilutions of the challenge dose. We determined the LD₅₀ of PA Boston 41501 to be 2 × 10⁶ CFU, and targeted doses of 10x and 100x LD₅₀ for our experiments. Regardless of treatment, no mice survive a dose of 50–100 LD₅₀. In this study, challenge doses varied from 4–5 LD₅₀, which did not differentiate between the TMV platform and recombinant antigen alone (Figure 1c), to 22.5 LD₅₀, which showed a significant enhancement of protective efficacy by TMV-PcrV+c after curdlan was added (Figure 2c). As we observed this dosing variability and associated changes in mortality from experiment to experiment, we chose to report the actual calculated challenge dose rather than ascribe a generic label of our target dose, which was 10 LD₅₀.

Taken together, the survival data, antibody titers, observed cytokine responses after rPcrV restimulation of ex vivo lung and spleen cells, and phagocyte recruitment to the lungs suggest that our PcrV vaccine on a TMV platform with the adjuvant curdlan (TMV-PcrV+c) induces humoral and cellular immune responses that are specific to PA antigens and effective at protecting mice from lethal infection. This vaccine may serve as a base antigen for the formulation of a multi-antigen vaccine that could provide broad protection against multiple strains of PA, and, with further study, could protect against both mucosal PA infections and diseases caused by PA infections in other tissues.

Supplementary Material

1

Supplementary Figure 1. Vaccinated serum does not recognize native PcrV by western blot. (a-b) Sonicate from PAO1 (O1), PA Boston 41501 (Bos) and PA14 (14); and rPcrV (rV) was separated by SDS-PAGE, transferred to a PVDF membrane, and probed with pooled vaccinated sera collected (a) before challenge and (b) from the same mice after surviving PA Boston 41501 challenge. (c) PAO1 and PA Boston were grown in TSB in the presence or absence of chelated medium (Chel.); nitriloacetic acid (NTA); and ethyleneglycol-bis-(β-aminoethyl ether) N,N,N’,N’-tetraacetic acid (EGTA). Lysates were separated by SDS-PAGE and transferred to a PVDF membrane, then probed with pooled serum from vaccinated mice collected before challenge (day 35). Chel.: TSB treated with Chelex, monosodium glutamate, and glycerol. Very faint bands corresponding to PcrV may be visible in lysates. We do not have a positive control antibody for PcrV available to us.

Supplementary Figure 2: Vaccination alone does not induce anti-PA antibodies. ELISA plates were coated with PA lysate to assess pre-challenge and post-challenge serum anti-PA antibody titers of (a) total antibody, (b) IgG1, and (c) IgA. (a-c) n=5–10 mice per group.

Supplementary Figure 3: Vaccination alters recruitment of phagocytes to the site of PA infection. (a) Gating strategy. Cells were first gated on CD45, then debris was eliminated using FSC and SSC. Neutrophils were identified as Ly-6G⁺ F4/80⁻, macrophages were identified as F4/80⁺, AMs were the CD11c⁺ subset of macrophages, and DCs were the CD11c⁺ subset of F4/80⁻ Ly-6G⁻ cells. (b,c,g,k) Percentages of (b) neutrophils, (c) total macrophages, (g) AMs, and (k) DCs. Antigen presentation capability and activation of cells was measured by (d,h,l) MHC class II, (e,i,m) CD80, and (f,j.n) CD86 expression in all macrophages, AMs, and DCs, respectively. (b-n) n=3 mice per group; groups compared by two-way ANOVA and Tukey’s multiple comparisons test.

Supplementary Table 1. Endotoxin levels in endotoxin units (EU)/ml and EU/dose for each vaccine component used in this study.

Highlights.

• IN TMV-PcrV is up to 75% protective against a lethal challenge with PA Boston 41501

• Addition of the adjuvant curdlan, but not antigens OprF and OprI, improves protective efficacy

• Presence of non-opsonizing anti-PcrV IgG1 antibody is associated with protection

• Vaccination induces a disseminated Th1/Th17 response starting at the infection site

• Vaccination limits prolonged neutrophil infiltration at the infection site