Antivirulence Bispecific Monoclonal Antibody-Mediated Protection against Pseudomonas aeruginosa Ventilator-Associated Pneumonia in a Rabbit Model

Abstract

Ventilator-associated pneumonia is an important clinical manifestation of the nosocomial pathogen Pseudomonas aeruginosa. We characterized the correlates of protection with MEDI3902, a bispecific human IgG1 monoclonal antibody that targets the P. aeruginosa type 3 secretion system PcrV protein and the Psl exopolysaccharide, in a rabbit model of ventilator-associated pneumonia using lung-protective, low-tidal-volume mechanical ventilation. Rabbits infused with MEDI3902 prophylactically were protected, whereas those pretreated with irrelevant isotype-matched control IgG (c-IgG) succumbed between 12 and 44 h postinfection (100% survival [8/8 rabbits] versus 0% survival [8/8 rabbits]; P < 0.01 by log rank test). Lungs from rabbits pretreated with c-IgG, but not those pretreated with MEDI3902, had bilateral, multifocal areas of marked necrosis, hemorrhage, neutrophilic inflammatory infiltrate, and diffuse fibrinous edema in alveolar spaces. All rabbits pretreated with c-IgG developed worsening bacteremia that peaked at the time of death, whereas only 38% of rabbits pretreated with MEDI3902 (3/8 rabbits) developed such high-grade bacteremia (two-sided Fisher's exact test, P = 0.026). Biomarkers associated with acute respiratory distress syndrome were evaluated longitudinally in blood samples collected every 2 to 4 h to assess systemic pathophysiological changes in rabbits pretreated with MEDI3902 or c-IgG. Biomarkers were sharply increased or decreased in rabbits pretreated with c-IgG but not those pretreated with MEDI3902, including the ratio of arterial oxygen partial pressure to the fraction of inspired oxygen of <300, hypercapnia or hypocapnia, severe lactic acidosis, leukopenia, and neutropenia. Cytokines and chemokines associated with acute respiratory distress syndrome were significantly downregulated in lungs from rabbits pretreated with MEDI3902, compared with c-IgG. These results suggest that MEDI3902 prophylaxis could have potential clinical utility for decreasing the severity of P. aeruginosa ventilator-associated pneumonia.

Keywords: Pseudomonas aeruginosa; immunotherapy; pneumonia; ventilator-associated pneumonia.

FIG 1

Schematic of the experimental setup for a rabbit model of ventilator-associated pneumonia. Rabbits were intubated and mechanically ventilated with a PIP of 15 cm H₂O, PEEP of 6 cm H₂O, peak inspiratory flow (PIF) of 4 L/min, and FiO₂ of 0.35, with 2.0% isoflurane to maintain general anesthesia. The respiratory rate (RR) was adjusted to 30 to 35 breaths/min to achieve end-tidal CO₂ (EtCO₂) of 35 to 45 mm Hg. The mechanical ventilation parameters resulted in a low tidal volume of 6 to 7 mL/kg and an inspiratory/expiratory ratio (I:E) of 1:2.5. The carotid artery was cannulated for serial blood sampling and arterial blood pressure (ABP) monitoring. The marginal ear vein was cannulated for infusion of Normosol-R with 5% dextrose for fluid maintenance. The patient monitor was used for continuous monitoring of heart rate (HR), electrocardiogram (ECG), rectal temperature, and peripheral capillary oxygen saturation (SpO₂).

FIG 2

MEDI3902 prophylaxis improved survival and reduced bacterial burden in a rabbit model of ventilator-associated pneumonia. Kaplan-Meier survival curves (A and E), LW/BW ratios (B and F), log₁₀CFU per milliliter of blood (C and G), and log₁₀CFU per organ (D and H) for rabbits treated intravenously with 15 mg/kg MEDI3902 or 15 mg/kg c-IgG 24 h before infection with P. aeruginosa strain 6077 and monitored for 36 h (study 1) (A to D) or 60 h (study 2) (E to H) were compared. Bars indicate the medians for all treatment groups. The LW/BW ratio and log₁₀CFU per organ for rabbits pretreated with MEDI3902 were compared to values for rabbits pretreated with c-IgG by nonparametric Mann-Whitney U test.

FIG 3

Histology showed reduced acute lung injury and inflammation for lungs harvested from rabbits pretreated with MEDI3902, compared to those pretreated with c-IgG. Photomicrographs of hematoxylin- and eosin-stained representative lung sections from the rabbit model of ventilator-associated pneumonia are shown for animals pretreated intravenously with 15 mg/kg of c-IgG (A to C) or MEDI3902 (D to F). (A) Prophylaxis with c-IgG results in multifocal areas of pulmonary necrosis and hemorrhage (×2 magnification). (B) Inset, bronchioles contain inflammatory cellular debris (×10 magnification). (C) Inset, alveolar septal necrosis and alveolar spaces are filled with fibrinous edema, hemorrhage, and abundant inflammatory cells (×20 magnification). (D) Prophylaxis with MEDI3902 demonstrated mild to patchy moderate infiltrate of viable inflammatory cells (×2 magnification). (E) Inset, multifocal areas of mild alveolar septal thickening and small aggregates of inflammatory cells within alveolar spaces are evident. Bronchioles are within normal limits (×10 magnification). (F) Inset, alveolar spaces contain primarily heterophils admixed with macrophages (×20 magnification).

FIG 4

Longitudinal analysis of biomarkers associated with ARDS. PaO₂/FiO₂ (A and E), lactate levels (B and F), base excess (C and G), and PaCO₂ (D and H) were determined using arterial blood samples taken every 2 h for the first 24 h and every 4 h thereafter until survivors were euthanized at 36 hpi (study 1) (A to D) or 60 h (study 2) (E to H). Additional parameters, including electrolyte and glucose levels, are shown in Fig. S3 in the supplemental material.

FIG 5

Longitudinal analysis of complete blood counts in rabbits pretreated with MEDI3902 versus c-IgG. Counts of WBCs (A and E), neutrophils (B and F), monocytes (C and G), and platelets (D and H) were determined using arterial blood samples taken every 2 h for the first 24 h and every 4 h thereafter until survivors were euthanized at 36 hpi (study 1) (A to D) or 60 h (study 2) (E to H). Additional hematological parameters are shown in Fig. S4 in the supplemental material.