Pseudomonas aeruginosa rugose small-colony variants evade host clearance, are hyper-inflammatory, and persist in multiple host environments

Abstract

Pseudomonas aeruginosa causes devastating infections in immunocompromised individuals. Once established, P. aeruginosa infections become incredibly difficult to treat due to the development of antibiotic tolerant, aggregated communities known as biofilms. A hyper-biofilm forming clinical variant of P. aeruginosa, known as a rugose small-colony variant (RSCV), is frequently isolated from chronic infections and is correlated with poor clinical outcome. The development of these mutants during infection suggests a selective advantage for this phenotype, but it remains unclear how this phenotype promotes persistence. While prior studies suggest RSCVs could survive by evading the host immune response, our study reveals infection with the RSCV, PAO1ΔwspF, stimulated an extensive inflammatory response that caused significant damage to the surrounding host tissue. In both a chronic wound model and acute pulmonary model of infection, we observed increased bacterial burden, host tissue damage, and a robust neutrophil response during RSCV infection. Given the essential role of neutrophils in P. aeruginosa-mediated disease, we investigated the impact of the RSCV phenotype on neutrophil function. The RSCV phenotype promoted phagocytic evasion and stimulated neutrophil reactive oxygen species (ROS) production. We also demonstrate that bacterial aggregation and TLR-mediated pro-inflammatory cytokine production contribute to the immune response to RSCVs. Additionally, RSCVs exhibited enhanced tolerance to neutrophil-produced antimicrobials including H2O2 and the antimicrobial peptide LL-37. Collectively, these data indicate RSCVs elicit a robust but ineffective neutrophil response that causes significant host tissue damage. This study provides new insight on RSCV persistence, and indicates this variant may have a critical role in the recurring tissue damage often associated with chronic infections.

Fig 1. RSCVs persist during a model for chronic burn wound infection and inhibit wound healing.

Porcine burn injuries were infected after 3 days with PAO1 or PAO1ΔwspF. A) Wound biopsies were quantified for CFUs. B and C) CLSM was used to assess bacterial burden. P. aeruginosa was stained with Alexa Fluor 488 (green) and host tissue with DAPI (blue). Area of green fluorescence was quantified. Scale bars indicate 35μm. D and E) Wound strips were H&E stained to assess re-epithelization 35 d.p.i. The distance between epithelial tongues (ET) in H&E stained tissue sections was measured. Due to H&E size constraints, images in panel D were generated by merging scans from two different slides containing half of the biopsy. Data presented as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001.

Fig 2. RSCVs evade neutrophil phagocytosis.

A) Neutrophil phagocytosis was assessed following infection (MOI 1:50) using CLSM. Neutrophils were stained with DAPI (blue), extracellular P. aeruginosa was stained with Alexa Fluor 488 (green), and neutrophil internalized P. aeruginosa was stained with Alexa Fluor 647 (red). White arrows indicate neutrophils containing phagocytosed P. aeruginosa. B) Number of neutrophils containing phagocytosed P. aeruginosa was quantified and the ratio of internalization determined. C) Neutrophil phagocytosis was assessed with or without serum opsonization by quantifying the population containing GFP producing P. aeruginosa following infection (MOI 1:50). Data is presented as mean ± SEM. *p<0.05, **p<0.01.

Fig 3. RSCVs stimulate neutrophil ROS production.

A) Neutrophils were infected with log phase bacteria (MOI 1:50) or treated with PMA in the presence of a luminol reporter. Luminescence was measured for 60min (right) and the area under curve (AUC) was calculated and normalized to the PMA response (left). RLU images are representative of the ROS response from a single donors neutrophils measured in triplicate, while AUC data was collected using neutrophils from at least 3 different donors. B and C) ROS response to the CF clinical RSCV isolates CF127 and CF39s. Data is presented as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001.

Fig 4. Bacterial aggregation promotes neutrophil ROS production, but exopolysaccharides alone are not sufficient.

A) ePS was purified from PAO1ΔpelP_BAD-psl, PAO1ΔpslP_BAD-pel, or PAO1ΔpelΔpsl and total carbohydrate was quantified by phenol sulfuric acid assay. PAO1 levels or 10-times PAO1 levels of ePS was added to human neutrophils and the ROS response was measured with a luminol reporter. B) PAO1ΔpslΔpel/pHERD20T and PAO1ΔpslΔpel/pCdrAB were grown to log phase in the presence of 1% arabinose leading to the formation of CdrA-mediated aggregates in the strain containing pCdrAB. Primary human neutrophils were infected with bacteria at an MOI (1:50). C) Neutrophils were treated with supernatant collected from cultures of PAO1/pCdrAB or PAO1ΔcdrA/pHERD20T, and ROS was quantified. The AUC of the ROS response over 1h was calculated and normalized by CFUs of the inoculation culture to ensure identical cell numbers regardless of aggregation.

Fig 5. PAO1ΔwspF exhibits tolerance to neutrophil antimicrobial products.

Log phase cultures of PAO1 or PAO1ΔwspF were treated with A) LL-37, B) H₂O₂ and C) HOCl at the labeled concentrations for 15min. CFUs were quantified before and after treatment and log fold killing determined. Data presented as mean ± SEM. ***p<0.001.

Fig 6. D-PAO1ΔwspF is hyper-inflammatory and persists during acute pulmonary infection.

Mice were infected with 10⁸ bacteria via intranasal instillation. PAO1ΔwspF aggregates were disrupted with a 22G needle prior to inoculation as indicated (D-PAO1ΔwspF). A) Lungs were homogenized and CFUs quantified. B) Lungs were H&E stained 24 h.p.i. and damage was scored as minimal, moderate, or severe. C) During PAO1ΔwspF infection, large bacterial aggregates (black dashed area) were observed 2 h.p.i in the bronchiole spaces, and after 8h neutrophil plugs (yellow dashed area) were observed in the bronchioles. Scale bars indicates 20μm. D) IL-1β and E) IL-6 in lung homogenate was measured using ELISA. Data presented as mean ± SEM. *p<0.05, **p<0.01.

Fig 7. D-PAO1ΔwspF pulmonary infection leads to severe tissue damage and neutrophil infiltration compared to PAO1 infection.

Mouse lungs 24 h.p.i. were stained with H&E to assess neutrophil infiltration and lung damage based on pulmonary architecture, necrosis, and suppurative inflammation. Images are representative of lungs treated or infected with A) PBS, B) PAO1, C) PAO1ΔwspF, D) D-PAO1ΔwspF. Scale bars indicate 200μm.

Fig 8. PAO1ΔwspF induces pro-inflammatory cytokine production in a TLR-dependent manner.

NR-9456 wild type and NR-15632 MyD88^-/-/TRIF^-/- mouse macrophages were infected with PAO1 or PAO1ΔwspF for 4h. A) IL-1β and B) IL-6 was measured in cell supernatants via ELISA. ** p< 0.01. Data presented as mean ± SEM.