All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Illinois at Chicago. C57BL/6J (WT, henceforth) and PAR2-null mice breeding pairs (Jackson Laboratory, CT, United States) were bred at the University of Illinois at Chicago. The mice colonies were maintained in a pathogen-free housing facility at the University of Illinois. For the experiments, 6–8-week-old male and female mice were used. Data represent pooled data from both female and male mice as we did not observe any effect of gender on lung injury. The wild-type strain of Pseudomonas aeruginosa (PAO1) constitutively expressing GFP (GFP-PAO1) was used in the study. The expression of GFP was plasmid-based as described previously (Hashimoto et al., 2007). The pharmacological agents NSC23766, Y-27632, and 8-bromoadenosine 3′-, 5′-cyclic monophosphate (8-Br-cAMP) were dissolved in water, whereas H-89 and ESI-09 were dissolved in DMSO.

GFP-tagged PA (GFP-PAO1 strain) was streaked in ampicillin-selective (HiMedia Laboratories LLC, United States) Luria–Bertani (LB) agar plates and incubated overnight at 37°C from the glycerol-preserved vial for the activation of culture. Single colonies were picked and inoculated overnight (∼17 h) in 250 ml LB broth (Amp 100 μg/ml) at 37°C. For the standard plate count method, 1 ml of the bacterial culture from the broth was serially diluted with sterile phosphate buffer up to 1 × 10¹⁰ dilutions. Each dilution plate was spread-plated in ampicillin-selective (100 μg/ml) sheep blood agar plates to get the countable bacterial colonies. The 1 × 10⁶ CFU/25uL was calculated using the following formula: CFU/ml = (no. of colonies x dilution factor)/volume of the culture plate.

Acute lung injury was induced in the WT and PAR2-null mice post PA infection and determined as described previously (Joshi et al., 2020). The mice were anesthetized by intraperitoneal injections of ketamine (100 mg/kg) and xylazine (12 mg/kg). PA (1 × 10⁶ CFU in 50 ul PBS/mouse) was instilled into the lungs through the endotracheal route. Edema was measured in these mice at indicated times by measuring the lung wet–dry ratio. Briefly, the left lobe of the lungs was excised following PA instillation, weighed, and then dried at 55°C. The dry weight was taken after 24 h, and the wet–dry ratio was calculated, which gave a measure of edema, as described previously (Yazbeck et al., 2017).

Bronchoalveolar lavage (BAL) and BAL-AMs were isolated as described previously (Rayees et al., 2019; Joshi et al., 2020). Briefly, after sacrificing the mice, tracheotomy was performed followed by the collection of BAL using an 18-gauge blunt needle. The BAL fluid was poured on plastic dishes, supplied with 10% serum, and the AMs allowed to adhere for 60 min. The non-adherent cells were removed by washing with PBS.

Phagocytosis assays in bone marrow-derived macrophages (BMDMs) or AMs were performed as described previously, with modifications (Ojielo et al., 2003; Descamps et al., 2012; Bastaert et al., 2018; Nagre et al., 2019). BMDMs were isolated as described previously (Rayees et al., 2019) and exposed to PA at MOI (multiplicity of infection) 1:10 in an antibiotic-free media containing 1% FBS for indicated times. The experiment was terminated by washing out PA with PBS. For FACS analysis, for the measurement of gene expression or immunoblotting, the cells were detached from the surface using accutase and processed further as described. For immunofluorescence, the macrophages were fixed in 2% paraformaldehyde for 15 min and stained with phalloidin (1:100) for 1 h and imaged. For AM phagocytosis, BAL-AMs were exposed to PA (MOI, 1:10) for 60 min. The AMs were fixed in 2% paraformaldehyde for 15 min after which cells were permeabilized using 0.1% triton X and stained with phalloidin (1:100) for 1 h in dark. The AMs and BMDMs were imaged on an LSM880 confocal microscope using 63x oil objective. F-actin organization in the form of pseudopods was quantified as the ratio of pseudopods/actin projections toward PA.

For measuring Rac1 activity in AMs by immunofluorescence, we quantified pseudopods, well-known indices of Rac1 activity (Sit and Manser, 2011; Flannagan et al., 2012), by measuring the spatial distribution of Rac1/F-actin fluorescent intensity peaks near cell edges. Briefly, the individual cells with protruding edges expressing fluorescent-labeled Rac1/F-actin were segmented, and the fluorescence intensity was measured using ImageJ.

WT and PAR2-null BMDMs were exposed to PA for 45 min. The supernatant was centrifuged and dissolved in 50 μL sterile PBS. The BMDM monolayer was washed twice with PBS and incubated with gentamicin-containing (300 μg/ml) media for 1 h to kill non-internalized bacteria (Plotkowski et al., 1999). Following a PBS wash, the cell pellet was then lysed in 0.1% Triton X-100 (Sigma-Aldrich, United States). The dissolved supernatant and the cell lysate were spread on the blood agar plate, and PA-CFU was counted 24 h later (Almeida et al., 1999; Cavinato et al., 2020).

The BMDMs exposed to PA for indicated time points were washed twice with PBS. The cells were detached from the plates using accutase, centrifuged, and fixed with fixation buffer (1:1 with PBS) (Invitrogen, United States) for 20 min. The cell suspension was washed twice with FACS buffer (0.5% BSA in PBS). All samples were then run on the BD LSR-Fortessa flow cytometer (BD Biosciences, CA, United States). The data were processed using Flow Jo software (TreeStar, Inc., CA, United States).

The BMDMs were exposed to PA at MOI, 1: 10 for indicated times. Cells were lysed using RIPA buffer (Sigma-Aldrich, United States) containing 1% protease inhibitor cocktail. Lysates (120 μg protein) were incubated with PAK-GST protein beads (20 μg) (Cytoskeleton, Inc., United States) for 4 h at 4°C on gentle rotation. The samples were centrifuged, and the pellet containing the beads was washed carefully and dissolved in Laemmli buffer. The total and active Rac1 were assessed by immunoblotting, using the anti-Rac1 monoclonal antibody.

The BMDMs were transfected as described previously (Rayees et al., 2019), using the Amaxa Nucleofactor electroporation system (Lonza). Briefly, the cells were detached from the dishes using accutase and spun down. The cell pellet was suspended in transfection reagent containing cDNA. The cells were slowly mixed with the transfection reagent to apply cDNA to the cells and electroporated. Vector plasmids were used to transfect experimental controls. The electroporated cells were processed for further experiments 48 h after transfection.

The RNA was extracted from AMs and BMDMs using TRIzol (Invitrogen Inc., CA, United States). The quality and integrity of RNA were verified on the NanoDrop spectrometer (Thermo Fisher Scientific, MA, United States). The cDNA was synthesized as per manufacturer’s protocol (High-Capacity cDNA Reverse Transcription Kit, Applied Biosystems), using equal quantities of RNA. Real-time q-PCR was performed for the measurement of gene expression (Applied Biosystems, CA, United States) (Rayees et al., 2015). The primer sequences used in the study were taken from our previously published study (Rayees et al., 2019).

Cationic liposomes were made using a mixture of chloroform, cholesterol, and dimethyl dioctadecyl ammonium bromide, as described previously (Yazbeck et al., 2017; Rayees et al., 2019). The liposomes were filtered through a 0.45-micron filter, and the cDNA was gently added and mixed. The cDNA-loaded liposomes were injected into the mouse lungs through the endotracheal route, followed by PA administration after 24 h. The lungs were either collected for the measurement of the wet–dry ratio or BAL was performed for the bacterial colony count, immunofluorescence, or measurement of gene expression from the extracted AMs.

All values are given as mean ± SD. Statistical analysis was performed using Graph Pad Prism version 7.0 (Graph Pad Software, La Jolla, CA). Multiple groups were compared using One-way ANOVA followed by a paired Student’s t test to assess significance between two groups. The statistics details can be found in the individual figure legends.

We exposed bone-marrow-derived Mɸ (macrophage) isolated from the WT and PAR2-null mice to the GFP-tagged PA (MOI 1:10) to determine the role of PAR2 in regulating phagocytosis of PA by flow cytometry. We found that the mean phagocytic index of WT-Mɸ was 47.2% at 15 min, which increased further to 60.4% at 45 min. In contrast, the mean phagocytic index of PAR2-deficient macrophages was only 11% and 26.5% at 15 and 45 min, respectively (Figures 1A,B). Because phagocytosis requires actin remodeling (Flannagan et al., 2012; Levin et al., 2016), we labeled F-actin using rhodamine-phalloidin to determine the effect of PAR2 deletion on the PA-induced actin reorganization. We observed that the phagocytic index of WT macrophages was ∼55% higher than that in PAR2 null-Mɸ (Figures 1C,D). The formation of membrane protrusions is an important aspect of phagocytosis, which is generated by the actin cytoskeleton in response to a phagocytic target (Flannagan et al., 2012; Levin et al., 2016). We quantified the protrusions as the ratio of pseudopods/actin projections toward PA. We found that WT-Mɸ extended pseudopods (two to four per cell) around PA within 15–45 min (Figure 1E). However, these pseudopods were barely detectable in PAR2-null Mɸ (Figure 1E), recapitulating the aforementioned findings. We next pre-treated the WT or PAR2 null Mɸ with latrunculin A, which is known to block actin polymerization (Morton et al., 2000), followed by PA exposure. We found that latrunculin A reduced phagocytosis of PA in WT-Mɸ to the level seen in PAR2-null Mɸ (Figure 1F, Supplementary Figure S1A). These findings indicate that PAR2-mediated phagocytosis of PA required actin polymerization.

Phagocytosis is a multi-step process initiated upon sensing pathogens, followed by phagosome formation (Flannagan et al., 2012; Levin et al., 2016). We, therefore, determined the number of colony-forming units (CFUs) in the supernatant or lysates of the WT or PAR2-null Mɸ following PA exposure to evaluate whether the impaired phagocytosis seen later was due to a defect in bacterial internalization by the Mɸ. CFUs were quantified in lysates following gentamycin treatment to remove attached bacteria. We found that the supernatant from PAR2-null Mɸ formed significantly more PA colonies than that from WT-Mɸ (Figure 1G, left), indicating a defective internalization of PA by PAR2-null macrophages as a significant factor in reducing phagocytosis. This conclusion is supported by Figures 1C–E data also. Conversely, PAR2-null Mɸ lysates formed fewer PA colonies than WT-Mɸ (Figure 1G, right).

AMs are phenotypically and genotypically different from BMDMs (Rayees et al., 2020). We thus isolated AMs from WT and PAR2-null lungs and exposed them to PA. We again found that PAR2-null AMs were defective protrusions in forming and PA phagocytosis (Figures 2A–C). Interestingly, PA markedly increased the expression of TNF-α and IL-6 in PAR2-null lung AMs relative to WT-AMs (Figure 2D). The same was true for WT and PAR2-null BMDMs exposed to PA (Supplementary Figure S2B). We also determined IL-1β and MIP-2 expression following PA infection and found a similar increase in their expression in both WT and PAR2-null BMDMs (Supplementary Figure S2B).

Bacterial pneumonia compromises patient survival in acute lung injury (Dreyfuss and Ricard, 2005; Rubenfeld et al., 2005). We instilled PA (1 × 10⁶ CFU/mouse) into WT and PAR2-null mice, and found that 60% of PAR2-null mice died within 64 h, while all WT mice survived ( Figure 2E ). Upon sensing pathogens, AMs induce lung barrier leak and trigger neutrophil influx (Nagre et al., 2019; Rayees et al., 2019). We found that PA (1 × 10⁶ CFU/mouse) induced barrier dysfunction and neutrophil accumulation in the lungs at 16 h in both WT and PAR2-null mice with the same intensity. The WT mice resolved edema formation post 32 h PA infection. However, the surviving PAR2-null mice showed massive edema and increased neutrophil influx (Figures 2F,G). We also isolated BAL from the WT and PAR2-null mice at 64 h and quantified CFUs. PAR2-null lungs failed to clear bacteria after pneumonia since BAL isolated from PAR2-null lungs showed a ∼3-fold increase in CFU relative to WT lung BAL (Figure 2H, Supplementary Figure S2A).

Small GTPases of the Rho family, such as Rho A and Rac1, are known to regulate actin organization as well as the phagocytosis of several bacterial species, including PA (Lee et al., 2000; Kazmierczak and Engel, 2002; Flannagan et al., 2012; Levin et al., 2016). We, therefore, determined if PAR2-induced PA phagocytosis occurs by activating RhoA or Rac1. We pharmacologically inhibited RhoA and Rac1 signaling using specific small molecule inhibitors, after which PA phagocytosis was determined (Routhier et al., 2010; Levay et al., 2013). We found that the inhibition of Rac1 activity using NSC23766 (Levay et al., 2013) reduced the phagocytic index to ∼50% (Figure 3A and Supplementary Figure S3A). By contrast, the inhibition of RhoA signaling using Y-27632 (Routhier et al., 2010) had no effect on PA phagocytosis in WT or PAR2-null Mɸ (Figure 3A, Supplementary Figure S3A). These findings indicate that PAR2 ligation mediates the activation of Rac1, which then induces actin polymerization to promote phagocytosis.

To further corroborate these findings, we inhibited Rac1 activity in WT-Mɸ, which also prevented pseudopod formation and PA phagocytosis (Figures 3B–D and Supplementary Figure S3B-C). We next measured the Rac1 activity in WT and PAR2-null BMDMs following PA exposure. We observed that Rac1 activity was barely detectable in naïve WT-Mɸ. PA exposure increased Rac1 activity by ∼5-fold within 15 min, consistent with robust PA phagocytosis seen in Figure 1. Rac1 activity declined at 45 min (Figures 3E,F). Interestingly, PA only modestly increased Rac1 activity in PAR2-null Mɸ. Thus, Rac1 activity was significantly lower in PAR2-null Mɸ at 15 and 45 min than in WT-Mɸ.

We also addressed the possibility that loss of PAR2 in AMs induced their apoptosis following infection. Thus, we determined the expression of caspase 11 and PARP1 (two well-known markers of apoptosis), (Chaitanya et al., 2010; Huang et al., 2019). In WT and PAR2-null macrophages following PA infection. We failed to detect any observable differences in the expression of these proteins between the two groups (Supplementary Figure S3D). We also performed annexin V staining, another marker of cell death, (Bossy-Wetzel and Green, 2000) in WT or PAR2-null macrophages after PA infection. We again did not observe any significant apoptosis of macrophages at the indicated time (Supplementary Figure S3E).

We have recently shown that the ligation of PAR2 with thrombin in macrophages induces cAMP generation (Rayees et al., 2019). cAMP is known to regulate Rac1GTPase activity (Schlegel and Waschke, 2013; Aslam et al., 2014). Thus, we tested the possibility that PAR2 stimulates cAMP generation, which activates Rac1 to mediate PA phagocytosis. Interestingly, PA increased cAMP levels by 2-fold within 15 min and 4-fold at 45 min in WT-Mɸ. PA failed to increase cAMP levels in PAR2-null Mɸ above baseline (Figure 4A).

Next, we pre-treated PAR2-null Mɸ with cell-permeable cAMP (8-Br-cAMP) to test the prediction that augmenting cAMP levels in PAR2-null Mɸ should rescue Rac1 activity and phagocytosis. We found that treatment of PAR2-null Mɸ with the cAMP analog restored the Rac1 activity (Figures 4B,C) and phagocytosis (Figures 4D–F) to the level seen in WT Mɸ. These findings demonstrate that PA activates PAR2, which induces cAMP generation to activate Rac1.

cAMP activates PKA and EPAC to signal Rac1 activation (Cheng et al., 2008; Birukova et al., 2010; Bachmann et al., 2013). Thus, we inhibited PKA and/or EPAC activity in WT-Mɸ and found that inhibition of PKA but not EPAC reduced phagocytosis significantly (Figures 4G–I).

Based on our observation that constitutively active Rac1 restored PA phagocytosis in AMs lacking PAR2 expression, we surmised that rescuing constitutive Rac1 would clear infection and thereby mitigate lung injury. Indeed, we observed that constitutively active Rac1 (CA-Rac1) mutant restored Rac1 activity in PAR2-null BMDMs (Supplementary Figure S4A, B). Thus, we delivered vector or YFP-tagged CA-Rac1 in mouse lungs, using liposomes. After 48 h, we injected unlabeled PA and determined the protrusive activity of Rac1 in AMs 3 h later using Rac1 and actin antibodies. We found that CA-Rac1-receiving AMs formed numerous pseudopods upon PA infection than WT-AMs (Supplementary Figure S4D, E). Importantly, CA-Rac1-receiving PAR2-null mice resolved lung edema (Figures 5A,B). Moreover, CA-Rac1 delayed the mortality of PAR2-null mice from pneumonia such that 70% of PAR2-null mice survived at 120 h post bacterial challenge than PAR2-null mice receiving control vector (Figure 5C). Survived PAR2-null mice receiving CA-Rac1 also showed decreased CFU counts (Figure 5D, S4C). However, the restoration of Rac1 activity failed to reduce the expression of pro-inflammatory cytokines to the levels observed in WT-AMs post PA challenge (Figure 5E), indicating that the augmented expression of TNF-α and IL-6 in PA exposed PAR2-null AM (Figure 5E) is independent of Rac1 activation and hence not related to phagocytosis in the present study model.