NADPH Oxidase 4-mediated Alveolar Macrophage Recruitment to Lung Attenuates Neutrophilic Inflammation in Staphylococcus aureus Infection

Animals

Male C57BL/6 mice were purchased from Orient Bio (Seongnam, Korea). NOX4 knockout (KO) mice on a C57BL/6 background were bred at the animal facility of Yonsei University. Age-matched, 8- to 12-wk-old mice were used for all experiments. Mice were housed under specific pathogen-free conditions and maintained in a 12 h light-dark cycle until the start of each experiment. All experiments were performed in accordance with the guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC International) and were approved by the Institutional Animal Care and Use Committee (IACUC) of Yonsei University Health System.

Preparation of S. aureus

S. aureus (ATCC 29213) was grown on a Brain Heart Infusion (BHI) agar plate at 37°C overnight. The S. aureus colonies were inoculated into 10 ml of fresh BHI medium and cultured in a shaking incubator at 37°C overnight. The cultured S. aureus was inoculated into fresh BHI medium and grown in the shaking incubator until it reached OD600 values of 1. GFP-tagged S. aureus was cultured on a BHI agar plate with tetracycline (100 µg/ml) at 37°C overnight. GFP-tagged S. aureus was grown in BHI medium containing tetracycline (100 µg/ml) until OD600 values of 1.

Bacterial infection model

Mice were anesthetized by zoletil/rompun mixture. The mice were intranasally instilled with 3×10⁷ colony forming units (CFUs) of S. aureus in 40 µl PBS. Control mice were intranasally instilled with 40 µl PBS. After the infection, mice were sacrificed at the indicated time points (0, 12, 24, 48, 72, and 120 h).

Bronchoalveolar lavage fluid (BALF) and lung tissue sample collection

Mice were sacrificed by administering an overdose of the zoletil/rompun mixture. BALF was obtained by injecting 1 ml of ice-cold PBS into the trachea using a catheter. The aspiration was repeated twice until no further fluid was collected. The collected BALF was centrifuged at 4,000 rpm for 5 min at 4°C. The supernatant was used for cytokine measurement, while the pellet was counted using a hemocytometer and stained for flow cytometry. Through the right ventricle, 10 ml of ice-cold PBS were perfused into the lungs. The left lobe of the lung was fixed with 4% paraformaldehyde for H&E staining, while the right lobes of the lung were extracted and chopped into small pieces. The right lobes of the lung were dissociated with collagenase type II solution (0.005 g collagenase type II and 1 µl DNase in 5 ml PBS per mouse) for 1 h at 37°C. After dissociation, the samples were filtered twice through a 70 µm strainer and then centrifuged at 4,000 rpm for 5 min. The pellet was stained and then isolated using FACS or magnetic-activated cell sorting (MACS).

Histological inflammatory score

H&E staining of the lung section was examined and assessed through BX53 microscope (Olympus Corporation, Tokyo, Japan). The section was analyzed and qualified through random selection. Using a score, the degree of lung inflammation was assessed. Briefly, a score of 0 meant there were no leukocytes around bronchioles or alveoli; a score of 1 meant there were a few leukocytes there; a score of 2 meant there were 1–3 layers of leukocytes there; and a score of 3 meant there were 5 layers of leukocytes there.

Immunofluorescence stain

The paraffin-embedded lung tissue was stained to identify neutrophils and macrophages. After deparaffinization and rehydration using xylene at decreasing concentrations (100% to 50%), the slides underwent heat-induced Ag retrieval in 10 mM sodium citrate buffer (pH 6.0) for 20 min at 95°C. The slides were then blocked with PBS containing 2% BSA and 2% donkey serum for 2 h at room temperature (RT). They were stained with DAKO Ab diluent (Agilent, Carpinteria, CA, USA), including neutrophil Abs (Abcam, Cambridge, UK) (1:100), and anti-F4/80 Abs (Novus Biologicals, LLC, Centennial, CO, USA) (1:100) overnight at 4°C. After the washing step, the slides were stained with secondary donkey anti-rat IgG Abs conjugated with Alexa Fluor 488 (1:1,000) and donkey anti-rabbit IgG Abs conjugated with Alexa Fluor 546 (1:1,000) for 1 h in the dark at RT. They were mounted with Fluoroshield containing DAPI (Sigma-Aldrich, St. Louis, MO, USA). The stained slides were analyzed using a Carl-Zeiss confocal microscope LSM 700. For quantification, the mean fluorescence intensity (MFI) of Ly6G or F4/80 relative to DAPI was processed using ZEN 2.3 black software (Carl Zeiss, Dresden, Germany).

BALF protein and cytokine measurement

To measure BALF protein, BALF supernatant was collected after centrifugation. The supernatant was diluted by one-fifth and quantified by Pierce™ BCA Protein Assay Kit (ThermoFisher Scientific, Waltham, MA, USA). Cytokines were measured in the supernatant by using Duoset ELISA kits for mouse TNFα and VEGF (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s protocol.

AM depletion and in vivo inhibitor treatment

To deplete AMs, mice were anesthetized and then intranasally pre-treated with clophosome A clodronate liposome (f70101c; FormuMax Scientific, Sunnyvale, CA, USA) 48 h before bacterial infection. Control mice were intranasally pre-treated with control liposome for clophosome (f70101; FormuMax Scientific). Mice were anesthetized and then intranasally treated with IgG control or anti-VEGF Ab (5 µg per mouse; R&D Systems) to block local alveolar VEGF at the same time as infection.

Flow cytometry and FACS sorting

Single cells from the dissociated lung or BALF pellet were lysed with RBC lysis buffer for 5 min at 4°C and centrifuged. After removing the supernatant, the cell pellet was washed twice with PBS containing 2% FBS. The washed pellet was counted using a hemocytometer and stained with Abs for 30 min at 4°C. The Abs were PE-Cy7-anti-CD45 (BD Bioscience, San Jose, CA, USA), eFlour-450-anti-Ly6G (eBioscience, San Diego, CA, USA), APC-anti-CD11c (BioLegend, San Diego, CA, USA), PE-anti-CD170 (SiglecF; BioLegend), FITC-anti-CD11b (BD Bioscience), Percp-cy5.5-anti-CD64 (BioLegend), and Brilliant-Violet711- anti-F4/80 (BioLegend). To examine in vivo AM cell death, BALF cells were stained with PE-cy7-anti-CD45, APC-anti-CD11c, PE-anti-CD170 (SiglecF), DAPI, and Annexin V-FITC Apoptosis Kit (BioVision Inc., Milpitas, CA, USA). The stained cells were analyzed using a FACS Fortessa BD flow cytometer. To isolate AMs, stained cells were isolated with CD45⁺CD11c⁺SiglecF⁺ using BD FACS Aria II. The data was analyzed using FlowJo software (BD Biosciences).

MACS for CD31⁺ endothelium and BMDN

The right lobes of the lung were chopped into small pieces and dissociated using a collagenase type II solution for 1 h at 37°C. The samples were filtered twice through a 70 µm strainer and washed. The single cells were washed after being processed in RBC lysis buffer for 5 min at 4°C. The single cells were stained with CD31 microbeads (130-097-418; Miltenyi Biotec, Bergisch Gladbach, Germany) for 20 min at 4°C. After washing, the stained samples were isolated by using positive selection in an autoMACS Pro separator (Miltenyi Biotec).

For BMDN sorting, the femurs and tibias of mice were removed and cut at both ends with scissors. A 5 ml of PBS was added, and the cells were lysed with RBC lysis buffer. The single cells were stained with a neutrophil isolation kit (130-097-658; Miltenyi Biotec) for 20 min at 4°C. After washing, the stained samples were isolated by using positive selection in an autoMACS Pro separator (Miltenyi Biotec).

Immunoblotting

Cells were lysed in RIPA buffer (Sigma-Aldrich) containing 1% Protease and Phosphatase Inhibitor Cocktail (Sigma-Aldrich) for 10 min on ice. The cells were sonicated and centrifuged for 15 min at 4 °C at 13,000 rpm. The supernatant was collected and quantified by Pierce™ BCA Protein Assay Kit (ThermoFisher Scientific). A total of 10 µg protein was loaded onto an 8% or 12% SDS-PAGE gel and then transferred to a polyvinylidene difluoride membrane by using a wet/tank transfer system (Bio-Rad, Hercules, CA, USA). The membrane was blocked by Tris-buffered saline with Tween 20 (TTBS) containing 5% Difco skim milk (BD Biosciences) for 1 h at room temperature. After blocking, the membrane was incubated with Purified anti-CD54 (ICAM1) Ab (BioLegend) or anti-VCAM1/CD106 Ab (R&D Systems) or anti-β-actin Ab (Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4°C with rotation. The membrane was washed with TTBS three times and incubated with secondary Abs linked with HRP for 2 h at RT.

Real time PCR

Following MACS or FACS sorting, total mRNA was extracted from the isolated cells using GeneAll Hybrid-R (305-101; Geneall, Seoul, Korea). The extracted mRNA was quantified using a NANO DROP 1000 spectrophotometer (ThermoFisher Scientific). cDNA was obtained with PrimeScript RT Master Mix (RR036A; Takara, Kusatsu, Japan). Real-time PCR was performed with KAPA SYBR FAST (KK4605; Roche, Basel, Switzerland) and QuantStudio 3 Real-Time PCR System (Applied Biosystems, Waltham, MA, USA). The sequences were as follows: mouse NOX4 sense primer 5′-TTG CCT GGA AGA ACC CAA GT-3′, NOX4 antisense primer 5′-TCC GCA CAA TAA AGG CAC AA -3′, mouse ICAM1 sense primer 5′- AGG TGG TTC TTC TGA GCG GC-3′, ICAM1 antisense primer 5′-AAA CAG GAA CTT TCC CGC CA -3′, mouse GAPDH sense primer 5′-TAG GGC CTC TCT TGC TCA GT-3′, GAPDH antisense primer 5′-GGA CCT CAT GGC CTA CAT GG-3′. Gene specific expression was normalized against the GAPDH housekeeping gene. The relative expression was determined by the comparative 2^−ΔΔCT method.

Ea.hy926 cell culture

Ea.hy926 cells were obtained from Yonsei University (Seoul, Korea) and maintained in Endothelial Cell Growth Medium-2 Bulletkit (EGM2, CC-3162; Lonza, Basel, Switzerland) at 37°C in a humidified 5% CO₂ incubator. Ea.hy926 cells in 6-well plates were pre-treated with 5 µM GLX353122 (HY-100111; MedChemExpress, Monmouth Junction, NJ, USA), a NOX4 inhibitor. The cells were treated for 6 h with 25 ng/ml recombinant human VEGF 165 (rhVEGF₁₆₅) at 37°C in a humidified 5% CO₂.

Ex vivo AM apoptosis assay

AMs (CD45⁺CD11c⁺CD11b⁻SiglecF⁺), isolated by FACS, were seeded and cultured ex vivo in DMEM (Lonza) in 12-well plates. The cells were infected with S. aureus (multiplicity of infection [MOI] 10) and incubated for 300 min at 37°C with 5% CO₂. They were then collected and stained using the Annexin V-FITC/PI Apoptosis Kit (BioVision Inc.). The stained cells were analyzed using a FACS Fortessa BD flow cytometer.

S. aureus-GFP uptake assay

BMDN isolated by MACS and AMs isolated by FACS were seeded in 12-well plates. The cells were infected with GFP-tagged S. aureus (MOI 10 or 20) and incubated for 1 h at 37°C with 5% CO₂. Then, they were collected in a tube and centrifuged at 4,000 rpm, 5 min, RT. The supernatant was cultivated on a BHI agar plate for measuring CFU. The pellet was washed twice and analyzed for GFP-positive populations using FACS as described previously.

Statistical analysis

Comparisons of 2 samples were processed by an unpaired Student’s t-test, whereas analysis of variance was followed by one-way ANOVA using Bonferroni’s multiple comparisons test in all statistical analyses. Comparisons of multiple samples, including time variance, were analyzed by two-way ANOVA test. Data were expressed as mean ± SEM for all graphs. p-values <0.05 were considered statistically significant. Statistical analyses were performed using the GraphPad PRISM 8 software (GraphPad, San Diego, CA, USA).

NOX4 is responsible for attenuation of lung inflammation at 24 h after S. aureus infection

To identify the immunologic role of NOX4 in bacterial lung infection, we examined the concentration of BALF proteins as a lung injury index in WT or NOX4 KO mice after S. aureus lung infection. In WT mice, BALF protein levels significantly increased at 12 h and then subsequently decreased until 120 h after infection. Conversely, in NOX4 KO mice, BALF protein levels exhibited a significant increase at 12 h, remained elevated at 24 h, and then decreased until 120 h after infection (Fig. 1A). Notably, although the levels in NOX4 KO mice were similar to those in WT mice at 12 h, they were three times higher than those in WT mice at 24 h (Fig. 1A). Furthermore, BALF protein levels showed a decline in WT mice after 12 h and in NOX4 KO mice after 24 h. Importantly, these levels tended to be higher in NOX4 KO mice than in WT mice from 48 h to 120 h after infection, although these differences did not reach statistical significance (Fig. 1A). To determine whether the lung injury in NOX4 KO mice was accompanied by an increased bacterial burden after infection, we quantified the number of remaining bacteria in BALF. The bacterial burden in BALF increased at 12 h and subsequently decreased until 120 h in WT mice (Fig. 1B). However, in contrast to the pattern of BALF protein levels, there was no discernible difference in bacterial burden between WT and NOX4 KO mice up to 120 h after infection (Fig. 1B). The levels of TNFα, a pro-inflammatory cytokine, increased at 12 h and subsequently decreased until 120 h in WT mice (Fig. 1C). Notably, at 24 h, TNFα levels were significantly higher in NOX4 KO mice than in WT mice (Fig. 1C). Therefore, these results showed that NOX4 KO mice had more inflammation than WT mice at 24 h after infection, and the increased inflammation was not derived from the difference in bacterial load.

Figure 1
NOX4 is responsible for attenuation of lung inflammation at 24 h after S. aureus lung infection (n=2–9).
WT and NOX4 KO mice were intranasally infected with S. aureus (3×10⁷ CFU) until 120 h. (A) Protein concentration in BALF after infection. (B) S. aureus CFU in the BALF. (C) Concentration of TNFα in the BALF. (D) Representative H&E images and (E) inflammatory scores of lung tissue (scale bar=500 µm). (F) Representative immunofluorescence staining images of Ly6G (green), F4/80 (red), and DAPI (blue). MFI of (G) Ly6G or (H) F4/80 relative to DAPI (scale bar=100 µm). All data were presented as mean ± SEM. The p-value using a two-way ANOVA (A-C, E, G, H).

^*p<0.05, ^**p<0.01, ^***p<0.001, ^****p<0.0001.

• Click for larger image
• Download as PowerPoint slide

Next, we investigated H&E staining images to identify the invasion of cells into the lung after S. aureus infection. The inflammatory score, based on H&E staining, indicated that immune cell infiltration in lung tissues peaked at 24 h after infection and then gradually decreased until 120 h in both WT and NOX4 KO mice (Fig. 1D and E). Moreover, NOX4 KO mice exhibited more extensive immune cell infiltration than WT mice at 24 h after infection (Fig. 1D and E). Furthermore, we performed immunofluorescence staining of neutrophils (Ly6G) and macrophages (F4/80) to examine the infiltrated cell types in the lung after infection (Fig. 1F-H). The immunofluorescence staining of neutrophils displayed a pattern similar to that observed in the H&E staining (Fig. 1E and G). At 24 h after infection, NOX4 KO mice exhibited significantly higher neutrophil infiltration into the lung compared to WT mice (Fig. 1F and G). However, the infiltration of macrophages was not as extensive as that of neutrophils, and there were no discernible differences between WT and NOX4 KO mice after infection (Fig. 1F and H). These results suggest that NOX4 is responsible for attenuation of neutrophilic inflammation at 24 h after S. aureus lung infection. Furthermore, the inflammation induced by S. aureus was resolved more rapidly in WT mice than in NOX4 KO mice from 24 to 120 h, implying that the absence of NOX4 delayed the resolution of inflammation.

NOX4 is responsible for AM recruitment at 24 h after S. aureus lung infection

Next, we examined the number of total cells in BALF in WT or NOX4 KO mice at 12 h and 24 h after S. aureus lung infection because NOX4 KO mice showed higher inflammation than WT mice at 24 h after infection. The number of cells in BALF increased in both WT and NOX4 KO mice at 12 h (Fig. 2A). Even though infiltrated cells in BALF were similar between WT and NOX4 KO mice at 12 h, the recruited cells in NOX4 KO mice were two times higher than those in WT mice at 24 h (Fig. 2A). Next, we performed flow cytometry analysis to examine the infiltrated cell types in the BALF after infection (Supplementary Fig. 1). In BALF from NOX4 KO mice, the number of neutrophils is almost the same as that of total cells (Fig. 2A and B), and the percentage of neutrophils was almost 80% of total cells at 24 h after infection (Supplementary Fig. 2A). Moreover, the number of neutrophils in NOX4 KO mice was similar to that in WT mice at 12 h, whereas the cells in NOX4 KO mice were three times higher than those in WT mice at 24 h (Fig. 2B). Although the number of monocytes and monocyte-derived macrophages was too small compared to that of neutrophils, the number of these cells increased in WT mice at 12 and 24 h after infection (Fig. 2C). In contrast, in NOX4 KO mice, the number of these cells increased at 12 h but decreased at 24 h (Fig. 2C). Furthermore, NOX4 KO mice had a lower count of these cells than WT mice at 24 h after infection (Fig. 2C). However, the percentages of monocytes and monocyte-derived macrophages did not exhibit significantly increases at 12 h in either WT or NOX4 KO mice (Supplementary Fig. 2B and C). Nevertheless, at 24 h after infection, the percentages of these cells were lower in NOX4 KO mice than in WT mice (Supplementary Fig. 2B and C). These findings suggest that the neutrophil population is the main cellular type that is infiltrated in NOX4 KO mice’s lungs, as well as that the neutrophil population is sustained in NOX4 KO mice at 24 h following infection.

Figure 2
NOX4 is responsible for alveolar macrophage recruitment at 24 h after S. aureus lung infection (n=3–9).
WT and NOX4 KO mice were intranasally infected with S. aureus (3×10⁷ CFU). (A) Total cell and (B) neutrophil counts (CD45⁺CD11b⁺CD11c⁻Ly6G⁺ SiglecF⁻) in BALF. (C) Monocyte (left; CD45⁺CD11b⁺CD11c⁻Ly6G⁻SiglecF⁻CD64⁺F4/80⁻) and monocyte-derived macrophage (right, CD45⁺CD11b⁺CD11c⁻Ly6G⁻SiglecF⁻CD64⁺F4/80⁺) counts in BALF. (D) AM counts (CD45⁺CD11c⁺SiglecF⁺). (E) Gate and percentages of AM in BALF. (F) Gate and Percentage of Annexin V⁺DAPI⁺ in AM gate (CD45⁺CD11c⁺SiglecF⁺). The data showed three or more experiments, as in the symbol indicating individual mice. All data were presented as mean ± SEM. The p-value using a two-way ANOVA (A-F).

^*p<0.05, ^**p<0.01, ^***p<0.001, ^****p<0.0001.

• Click for larger image
• Download as PowerPoint slide

AMs have been reported to be differentiated from monocytes (7) and to play an important role in the early host response to bacterial lung infections by regulating the inflammatory response, removing neutrophils, and promoting tissue repair (8, 26). Therefore, we checked the number of AMs in WT or NOX4 KO mice at 12 and 24 h after S. aureus infection (Fig. 2D). The number of AMs in BALF was similar between WT and NOX4 KO mice in the absence of infection. At 12 h after infection, the number of AMs in both WT and NOX4 KO mice was almost the same. Interestingly, the number of AMs decreased in NOX4 KO mice compared to that in WT mice at 24 h after infection (Fig. 2D). Following infection, the total cell count in BALF increased due to inflammation, mainly attributed to the recruitment of neutrophils (Fig. 2A and B, Supplementary Fig. 2A). Consequently, the percentage of AMs in BALF appeared to decrease at 12 and 24 h after infection (Fig. 2E). Although it may appear that the percentage of AMs decreases at 12 and 24 h, it was significantly lower in NOX4 KO mice than in WT mice at 24 h after infection (Fig. 2E). To investigate the cause of the reduction of AMs in NOX4 KO mice after infection, we examined in vivo flow cytometry-based apoptosis analysis (Annexin V & DAPI) in AM population. The percentage of necrotic cells (Annexin V⁺ DAPI⁺) in AMs increased in WT mice at 12 and 24 h after infection (Fig. 2F). The percentage of necrotic cells in AMs was comparable between WT and NOX4 KO mice at 12 and 24 h after infection (Fig. 2F). Therefore, the reduction of AM in NOX4 KO mice at 24 h after infection was not due to increasing cell death. These results indicated that NOX4 is responsible for AM recruitment into the lung at 24 h after S. aureus lung infection. Moreover, the increased level of lung injury and neutrophil number in NOX4 KO mice at 24 h after S. aureus infection may be attributed to the decreased level of AMs in NOX4 KO mice at 24 h after infection.

NOX4 is not involved in the antibacterial activity of neutrophils and AMs upon S. aureus infection

We questioned whether intrinsic NOX4 affected the function of neutrophils after S. aureus infection. To investigate the phagocytic capacity of neutrophils, we isolated bone-marrow derived neutrophils (BMDN) from WT or NOX4 KO mice using MACS. In BMDN from WT and NOX4 KO mice, there were no differences in the uptake of GFP-tagged S. aureus at different time points within 1 h (Fig. 3A). The percentages of apoptotic and necrotic cell death were also similar in both WT and NOX4 KO BMDN after GFP-tagged S. aureus infection (Fig. 3B and C). Moreover, BMDN isolated from both WT and NOX4 KO mice exhibited similar bacterial killing ability with MOI 5 or 20 at 30 min (Fig. 3D).

Figure 3
NOX4 is not involved in the antibacterial activity of neutrophils and AMs upon S. aureus infection.
(A-D) Ex vivo BMDNs from WT or NOX4 KO mice were infected with S. aureus at the indicated time point. (A) Percentages of GFP⁺ BMDNs with MOI 10. (B) Percentages of apoptotic cells (Annexin V⁺PI⁻) in BMDNs with MOI 10. (C) Percentage of necrotic cells (Annexin V⁺PI⁺) in BMDNs with MOI 10. (D) S. aureus CFU in media with MOI 5 or 20 at 30 min after infection. (E-G) Ex vivo AMs from WT or NOX4 KO mice were infected with S. aureus. (E) Percentages of GFP⁺ AMs after infection with MOI 1. (F) S. aureus CFU in media at 60 min after infection with MOI 1 or 10. (G) Percentage of necrotic cells (Annexin V⁺PI⁺) in AMs with MOI 10. All data were presented as mean ± SEM. Two-way ANOVA (A-C, G).

• Click for larger image
• Download as PowerPoint slide

Next, we investigated whether intrinsic NOX4 directly influenced the function of AMs after infection. To examine the phagocytic capacity and killing ability of AMs, we isolated them from WT or NOX4 KO mice using FACS. AMs from WT or NOX4 KO mice showed similar patterns in bacterial uptake and bacterial killing ability at different time points and doses (Fig. 3E and F). Additionally, we conducted ex vivo flow cytometry-based apoptosis analysis (Annexin V & PI) in AMs with S. aureus (MOI 10). AMs isolated from WT or NOX4 KO mice exhibited that no significant differences in necrotic cell death in response to S. aureus infection (Fig. 3G). These findings suggest that intrinsic NOX4 has no discernible effect on neutrophil and AM function, including antibacterial activity and cell death, in response to S. aureus infection.

NOX4 is necessary to induce ICAM1 expression on endothelial cells after S. aureus infection

Given that NOX4 is not intrinsically involved in the antibacterial activity of neutrophils and AMs, we hypothesized that NOX4 could operate in cells other than neutrophils and AMs to control their numbers upon S. aureus infection. Since NOX4 is reported to be expressed in lung endothelial cells (19, 22), and lung endothelium is responsible for monocyte recruitment from blood to alveoli (5, 27), we focused on the role of NOX4 in lung endothelial cells upon S. aureus infection. We isolated lung CD31⁺ endothelial cells from WT or NOX4 KO mice using MACS and examined the expression of NOX4 after S. aureus infection. In the WT endothelial cell, we observed a significant increase in NOX4 gene expression at 12 h, followed by a decrease at 24 h after infection (Fig. 4A). To investigate whether NOX4 in endothelial cells could be involved in the recruitment of AMs into the lung after infection, we examined the expression level of endothelial adhesion molecules such as ICAM1 which are known to play an important role in monocyte recruitment on endothelial cells (5, 27). The level of ICAM1 mRNA expression in WT mice endothelial cells increased at 12 h and decreased at 24 h after S. aureus infection (Fig. 4B), which were similar patterns to NOX4 expression (Fig. 4A). Interestingly, the mRNA expression level of ICAM1 in endothelial cells from NOX4 KO mice was lower than that in endothelial cells from WT mice at 12 h after infection (Fig. 4B). Similar to the alteration pattern of mRNA expression, the protein level of ICAM1 in NOX4 KO mice endothelial cells was lower than that in WT mice endothelial cells at 12 h after infection (Fig. 4C). The protein level of vascular cell adhesion molecule-1 was also known to be associated with monocyte recruitment (6, 28). However, there was no difference in the protein level of VCAM1 in the endothelial cells from WT and NOX4 KO mice after infection (Fig. 4C). Taken together, we showed that NOX4 is necessary to induce the expression of ICAM1, but not VACM1, in endothelial cells upon S. aureus infection, suggesting that NOX4 might induce recruitment of AMs by controlling ICAM1 expression in lung endothelial cells.

Figure 4
NOX4 is necessary to induce ICAM1 expression on endothelial cells after S. aureus infection (n=3–5).
(A-C) CD31⁺ endothelial cells were isolated using MACS from WT or NOX4 KO mice lungs at 0, 12, and 24 h after S. aureus infection. (A) Expression of NOX4 mRNA in CD31⁺ endothelial cells. (B) Expression of ICAM1 mRNA in endothelial cells by real-time PCR. (C) Immunoblot of ICAM1, VCAM-1, β-actin, and Coomassie staining. Representative image (left) and relative density quantified from immunoblot (right). (D) Ea.hy926 cells were pretreated with 10 µM GLX351322 to inhibit NOX4 prior to rhVEGF₁₆₅ treatment. Immunoblot of ICAM1. Representative image (left) and relative density quantified from immunoblot (right). All data were presented as mean ± SEM. The p-value using a one-way ANOVA (A) and two-way ANOVA (B-D).

^*p<0.05, ^**p<0.01, ^****p<0.0001.

• Click for larger image
• Download as PowerPoint slide

Previous studies have reported that VEGF can induce ICAM1 expression in various endothelial cells (29, 30, 31), and that NOX4 colocalizes with and interacts with a VEGF receptor 2 (VEGFR2) in the endothelium (32). Furthermore, NOX4 can stabilize VEGFR2 and promote endothelial cell function (33). These reports suggest that VEGF may regulate NOX4-mediated ICAM1 expression in endothelial cells after S. aureus infection. Thus, we investigate whether VEGF regulates the ICAM1 expression by NOX4. We first inactivated NOX4 activity using GLX351322, NOX4 inhibitor, in a stable human endothelial cell line such as Ea.hy926 cells. We next measured ICAM1 expression at 3 and 6 h after treatment with rhVEGF₁₆₅ in the presence or absence of GLX35122. At 6 h after rhVEGF₁₆₅ treatment, ICAM1 expression increased, and increased ICAM1 expression was reduced by GLX35122 treatment (Fig. 4D), indicating that VEGF can induce endothelial ICAM1 expression in a NOX4-dependent manner.

VEGF attenuates lung inflammation via inducing AM recruitment after S. aureus infection

We wondered whether VEGF is induced in WT mice upon S. aureus infection. Therefore, we examined the concentration of VEGF in BALF after S. aureus infection. We found that the level of VEGF was increased at 12 h and decreased at 24 h after lung infection in WT mice (Fig. 5A). In addition, the increased level of VEGF in NOX4 KO mice was almost the same as that in WT mice (Fig. 5A), indicating that VEGF is increased upon S. aureus infection in a NOX4-independent manner.

Figure 5
VEGF attenuates lung inflammation via inducing AM recruitment at 24 h after S. aureus lung infection (n=3–6).
(A) VEGF concentration in BALF from WT and NOX4 KO mice. (B-G) WT and NOX4 KO mice intranasally treated with IgG or anti-VEGF Abs, and they were infected with 3×10⁷ CFU of S. aureus at the same time. The inflammatory responses were assessed at 24 h after the infection. (B) Representative flow cytometry plots (left), and percentages of AM in total BALF cell (right). (C) AM counts in BALF. (D) Protein concentration in BALF. (E) Total cell counts in BALF. (F), neutrophil counts in BALF. (G) S. aureus CFU in BALF. The data showed 3 experiments, with the symbol representing individual mice. All data were presented as mean ± SEM. The p-value using two-way ANOVA (A) and one-way ANOVA (B-G).

^*p<0.05, ^**p<0.01, ^***p<0.001, ^****p<0.0001.

• Click for larger image
• Download as PowerPoint slide

Next, we investigated the functional role of VEGF in AM recruitment upon S. aureus infection. To test this, we intranasally treated WT and NOX4 KO mice with anti-VEGF Abs together with S. aureus to block local VEGF in alveoli. We then examined the ratio and number of AMs in BALF at 24 h post-infection in the presence of control IgG or anti-VEGF Abs. Anti-VEGF treatment significantly reduced the ratio and number of AMs in WT mice, which were similar to those in IgG-treated NOX4 KO mice (Fig. 5B and C). Furthermore, the ratio and number of AMs in IgG-treated NOX4 KO mice did not differ from those in anti-VEGF-treated NOX4 KO mice (Fig. 5B and C). These results suggest that VEGF, increased by S. aureus infection, can recruit AMs in NOX4-dependent manner. In this context, we wondered whether increased VEGF is responsible for the attenuation of neutrophilic inflammation and lung injury upon S. aureus infection. Anti-VEGF treatment increased the levels of BALF protein (Fig. 5D), BALF cells (Fig. 5E), and neutrophils (Fig. 5F) in WT mice, which were similar to those in IgG- or anti-VEGF-treated NOX4 KO mice (Fig. 5D-F). However, bacterial loads were not affected in all groups (fig. 5G). Taken together, our results demonstrate that VEGF in the alveoli is essential for protecting against S. aureus infection by recruiting AMs in NOX4-dependent manner.

NOX4 attenuates lung inflammation via inducing AM recruitment after S. aureus infection

To investigate the functional role of AMs in neutrophilic inflammation and lung injury upon S. aureus infection, we intranasally treated mice with clodronate liposome to deplete AMs at 48 h prior to S. aureus infection (34). Clodronate liposome treatment significantly reduced the ratio and number of AMs in WT mice, similar to those in isotype liposome-treated NOX4 KO mice (Fig. 6A and B). Moreover, the ratio and number of AMs in isotype liposome-treated NOX4 KO mice were not different from those in clodronate liposome-treated NOX4 KO mice (Fig. 6A and B). Clodronate liposome treatment increased the levels of BALF protein (Fig. 6C), BALF cells (Fig. 6D), and neutrophils (Fig. 6E) in WT mice, similar to those in isotype liposome- or clodronate liposome-treated NOX4 KO mice (Fig. 6C-E). However, bacterial loads were not affected in all groups (Fig. 6F). These results demonstrate that AM recruitment is required for the resolution of neutrophilic inflammation in a NOX4-dependent manner after S. aureus infection.

Figure 6
NOX4 attenuates lung inflammation via inducing AM recruitment at 24 h after S. aureus lung infection (n=4–6).
WT and NOX4 KO mice were intranasally administrated with clodronate liposome or isotype liposome 48 h before S. aureus infection. They were infected with S. aureus (3×10⁷ CFU) and sacrificed after 24 h. (A) Representative flow-cytometry plots (left), and percentages of AM in BALF cells (right). (B) AM counts in BALF. (C) Protein concentration in BALF. (D) Total cell counts and (E) neutrophil counts in BALF. (F) S. aureus CFU in BALF. The data showed two or more experiments, as in the symbol indicating individual mice. Bars represented means ± SEM. The p-value using one-way ANOVA was used for (A-F).

^*p<0.05, ^**p<0.01, ^***p<0.001, ^****p<0.0001.

• Click for larger image
• Download as PowerPoint slide

In this study, S. aureus lung infection simulates the production of VEGF in alveoli. When NOX4 is present, the elevated VEGF levels induce endothelial ICAM1 expression in a NOX4-dependent manner, facilitating the recruitment of AMs to the alveoli. These recruited AMs play a crucial role in reducing neutrophilic inflammation. However, in the absence of NOX4, increased VEGF is unable to significantly enhance endothelial ICAM1 expression, leading to insufficient recruitment of AMs to the alveoli. This deficiency in AMs results in excessive neutrophilic inflammation (Fig. 7).

Figure 7
Schematic of the study.
S. aureus lung infection triggers the production of VEGF in the alveoli. In WT mice, elevated levels of VEGF can stimulate the expression of endothelial ICAM1 in a NOX4-dependent manner. Consequently, the AMs were recruited to the alveoli through the increased endothelial ICAM1, and they reduced neutrophilic inflammation. However, in NOX4 KO mice, the increased VEGF is not able to induce appropriate endothelial ICAM1 expression, resulting in insufficient AM recruitment in the alveoli. As a result, the insufficient AMs cause excessive neutrophilic inflammation.

• Click for larger image
• Download as PowerPoint slide