Six- to eight-week-old female C57BL/6 [wild-type (WT)] mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). p40^phox knockout mice, which have been backcrossed to the B6 background for 10 generations, were kindly provided by Dr. Ramnik Xavier (Massachusetts General Hospital, Boston, MA, USA) (19). All mice were fed autoclaved food and water and maintained in a specific-pathogen-free facility at Massachusetts General Hospital.

Mice were infected orally with 3 × 10⁸ CFU of the streptomycin-resistant SL1344 strain of S. Typhimurium. After 72 h of infection, mice were euthanized by CO₂ inhalation. Spleens and livers were harvested to determine bacterial translocation and systemic inflammatory response. To study whether p40^phox deficiency affects the early mucosal innate immune and inflammatory response to Salmonella infection, the streptomycin-pretreatment mouse model was used as described previously (4). Briefly, mice were each given 20 mg of streptomycin, followed 24 h later by oral infection with 10⁸ CFU of S. Typhimurium. Ceca were collected for histopathology 24 h after infection. To determine the impact of p40^phox deficiency on the host during Salmonella infection, WT mice and mice lacking p40^phox were infected orally with 3 × 10⁸ CFU of S. Typhimurium. Body weight loss and survival of mice were monitored daily. Mice were humanely euthanized if they exhibited greater than 15% weight loss.

Spleens and livers collected from Salmonella-infected mice (WT and p40^phox−/− mice) (with or without streptomycin treatment), were excised using sterile procedure, weighed, homogenized, and plated on Luria–Bertani plates containing 50 µg/ml streptomycin (3). The CFU were quantified. To examine the appearance and distribution of Salmonella in cecum and spleen, cryostat sections of the intestinal tissues and spleens were incubated with a rabbit antibody against Salmonella (Thermo Fisher Scientific), followed by fluorescein isothiocyanate (FITC) or Cy3-conjugated goat anti-rabbit IgG antibody (Biosource Cat. #554020), and analyzed by immunofluorescence microscopy.

At necropsy, tissue samples of spleen, liver, and cecum were collected, frozen in Tissue-Tek OCT compound (Miles Inc., Elkhart, IN, USA) and then stored at −80°C. Then, 5-µm sections were cut and stained with hematoxylin and eosin (H&E). Pathology was scored using a modified histological scoring systems previously published in the literature (5, 29, 30). The scores were assessed by determination of infiltration of inflammatory cells, with scores ranging from 0 to 4 (0, normal cell pattern; 1, scattered inflammatory cells in the lamina propria; 2, increased numbers of inflammatory cells in the lamina propria; 3, confluence of inflammatory cells extending into the submucosa; and 4, transmural extension of the infiltrative inflammatory cells), together with the evaluation of cecal tissue damage (score range, 0–4; 0, normal tissue pattern; 1, minimal inflammation and crypt hyperplasia; 2, mild crypt hyperplasia with or without focal invasion of epithelium; 3, obvious crypt hyperplasia, invasion of epithelium, and goblet cell depletion; and 4, extensive mucosal damage and extension through deeper structures of the bowel wall). Stained sections were analyzed without prior knowledge of the type of treatment.

Tissue cryosections were fixed in ice-cold acetone, washed, and then blocked with avidin/biotin agent (Vector Laboratories, Burlingame, CA, USA). To analyze the location and abundance of macrophages and neutrophils, cecal slides were stained with FITC-labeled anti-mouse F4/80 (eBioscience Cat. #11-4801-81) and Cy5-labeled anti-mouse Ly6G (Biolegend Cat. #108410). DNA was stained and mounted using the 4′, 6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). To examine the bacterial killing capacity, peritoneal macrophages were collected from uninfected WT and p40^phox KO mice, grown on cover slips, and infected with S. Typhimurium for 1 h (cell/bacteria, 1:10). Cells were then stained with phycoerythrin (PE)-conjugated anti-F4/80 at both 2 and 6 h after infection (24). Internalization and elimination of bacteria by macrophages were visualized by detecting Salmonella with a rabbit antibody against Salmonella, followed by FITC-labeled anti-rabbit IgG antibody.

Total RNA was isolated from spleen, liver, cecum, and peritoneal macrophages using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) or RNeasy Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instruction. All RNA samples were reverse transcribed into cDNA using the Superscript First-Strand Synthesis System (Invitrogen Life Technologies). The cDNA samples were then tested for the expression of TNF-α, IL-1β, KC, MCP1, and MIP2 by real-time RT-PCR performed as previously described (29). Results were normalized to GAPDH expression and relative quantification was calculated using the 2^−ΔΔCT method.

Liver and spleen homogenate was prepared for cytokines measurement through ELISA. Protein concentrations were determined using a bicinchoninic acid protein assay. IL-1β was measured by DuoSet ELISA kit from R&D Systems (Cat. #DY401) according to the manufacturer’s instruction. ELISA capture antibodies and biotinylated secondary antibodies for IFN-γ (BD Cat. #554410; 551216), TNF-α (BD Cat. #554416; 557432), and IL-17A (BD Cat. #555068; 555067) were purchased from DB Bioscience. Standard curves were obtained using recombinant murine IFN-γ (BD Cat. #554587), TNF-α (BD Cat. #554589), and IL-17A (eBioscience Cat. #14-8171-80) from BD Bioscience and eBioscience.

Splenic single-cell suspensions were prepared from uninfected mice and mice infected orally with 3 × 10⁸ CFU S. Typhimurium at 24 h after bacterial infection. Cells were surface stained with fluorescent-conjugated antibodies against CD11b (eBioscience, Cat. #11-0112-85; BD Cat. #553311), Ly6G (BioLegend, Cat. #108410), and F4/80 (eBioscience, Cat. #12-4801-80) before being subjected to flow cytometry analysis with an Accuri C6 FACS machine. Gates set on forward and side angle light scatter were used to exclude dead cells and debris.

Reactive oxygen species generation was measured using lucigenin-dependent chemiluminescence. Peritoneal macrophages were collected by washing the peritoneal cavity of WT or p40^phox−/−mice with ice-cold PBS. Neutrophils were purified using EasySep™ Neutrophil Enrichment Kit (Stemcell Technologies, Catalog #19762) according to the manufacturer’s instructions. Briefly, mouse bone marrow cells were flushed from femurs and tibias, centrifuged and, after hypotonic lysis of erythrocytes, mixed with biotinylated antibodies and magnetic particles. Neutrophils were then separated by immunomagnetic negative selection. Macrophages and neutrophils were then resuspended in 100 µM lucigenin (5 × 10⁷/ml) and incubated for 15 min on ice. Subsequently, cells were seeded in a black 96-well plate (4 × 10⁶/well) and stimulated with 200 nM PMA (Sigma-Aldrich). Macrophages without PMA treatment were used as a control. Light emission was measured on a TopCount NXT scintillation and luminescence reader (PerkinElmer, Waltham, MA, USA) every 5 min for a period of 1 h at 37°C. Data were expressed as relative light units (RLU) normalized to cell number and corrected for a sample blank. Each experiment was performed in triplicate.

In addition, intracellular ROS generation was measured using DCFDA Cellular Reactive Oxygen Species Detection Assay Kit (Abcam, Ab113851) in accordance with the manufacturer’s instructions and published methods (31). DCFDA is a fluorogenic dye that measures ROS activity within the cell. For these experiments, both intestinal and peritoneal macrophages were isolated from C57BL/6 and p40^phox−/− mice. The macrophages from intestine were isolated using published methods (32). Briefly, small and large intestine were collected and Peyer’s patches were removed. The intestinal tissues were cut open longitudinally (to 1.5 cm pieces) and washed in PBS. The epithelial layer was removed by shaking the tissues in EDTA-containing HBSS/FBS buffer (250 rpm for 20 min at 37°C). The intestinal tissue was then digested with CollagenaseVIII/DNAse I solution and filtered through a 100 µm cell strainer. After washing, the cell pellets were collected. Intestinal macrophages were enriched using magnetic beads (anti-F4/80 and CD11b). The cells were then stained in the diluted DCFDA solution at 37°C for 30 min in the dark, then washed. Cells were seeded in a dark side, clear bottom 96-well microplate (2 × 10⁵ stained cells/well). The cells were then stimulated with 20 nM PMA (Sigma-Aldrich) or infected with S. Typhimurium (cell/bacteria, 1:100). Macrophages without PMA treatment or Salmonella infection were used as a control. The microplates were read using a fluorescent plate reader (FluoSTAR OPTIMA-BMG Labtech) with the excitation wavelength set at 485 nm and emission wavelength at 535 nm. Results are expressed as fluorescent intensity of samples minus blank (DCFDA-unstained cells). Each experiment was performed in triplicate.

The resident peritoneal macrophages were collected from uninfected WT and p40^phox knockout mice. After 2 h of incubation in complete DMEM, non-adherent cells were removed by washing. The adherent cells were cultured at 37°C overnight and then exposed to 10⁷ S. Typhimurium (cell/bacteria, 1:10) for 1 h in antibiotic-free medium. After completion of the infection period, cells were washed with PBS and then incubated with gentamicin-containing medium (100 µg/ml) for 2 h to eliminate extracellular bacteria. Cells were then lysed immediately in sterile 1% Triton X-100 in water or, after a further 4 h, in medium containing 10 µg/ml gentamicin. Serial dilutions of the cell lysates were plated on LB agar containing streptomycin overnight. Colonies were counted after incubation and bacterial numbers present inside the cells at each time point were calculated (24, 33).

Results are expressed as the mean ± SEM. Survival curves of infected mice (WT and p40^phox−/−) were compared using Kaplan–Meier analysis followed by log-rank test. Statistical differences were determined using one-way ANOVA with Tukey’s post hoc test, the two-tailed Student’s t-test or paired t-test (body weight loss) with GraphPad Prism (GraphPad Software, San Diego, CA, USA). Significance was defined as a P-value <0.05.

NADPH oxidase is critical in host defense against invading pathogens (34). To investigate the role of p40^phox, one of the subunits of the NADPH oxidase, in host protective and inflammatory responses against Salmonella infection and in the regulation of the severity of bacterial-induced disease, p40^phox−/− mice and C57BL/6 WT control mice were infected orally with 3 × 10⁸ CFU of S. Typhimurium (SL1344 strain). As expected, mice infected with S. Typhimurium exhibited significant body weight loss and high mortality (100%) by day 5 of infection in WT mice. By contrast, p40^phox−/− mice developed more severe disease, resulting in a more significant body weight loss and early mortality reaching 100% mortality by 4 days after inoculation (Figures 1A,B). To examine whether the increased susceptibility of p40^phox-deficient mice to Salmonella infection was related to poor control of bacterial replication, bacterial translocation to spleens and livers was determined 72 h after infection. Significantly, greater bacterial numbers were present in the spleens (Figure 1D) and livers (Figure 1C) of p40^phox−/− mice compared with C57BL/6 controls, suggesting an impaired mucosal barrier function and inability to effectively clear microbes in p40^phox-deficient mice. This is further supported by the results from our immunofluorescence microscopic analysis revealing enhanced bacterial dissemination in the spleen tissues of p40^phox-deficient mice (Figure 1E).

The increased tissue bacterial loads were accompanied by exacerbated inflammation and tissue damage, which was evidenced by the observations from analysis of histopathology and pro-inflammatory cytokine profiles of these organs from WT and p40^phox−/− mice with and without S. Typhimurium infection. Histological analysis of infected mice revealed extensive pathological changes in the spleen, evidenced by a neutrophilic infiltrate and congestion within the red pulp in Salmonella-infected WT mice (Figure 2A). In contrast to WT mice, p40^phox knockout mice developed more severe tissue damage, characterized by red pulp congestion, thrombosis, and splenic architecture disruption (Figure 2A). Histological analysis of the liver sections from Salmonella-infected mice showed massive necrosis accompanied by a mixed inflammatory cell infiltration in both WT and p40^phox-deficient mice. However, more pronounced and extensive hepatic lesions were noted in Salmonella-infected p40^phox−/− mice compared to WT mice (Figure 2B). These results suggest that Salmonella infection of p40^phox KO mice results in an exacerbated inflammation and tissue injury in the spleen and livers due to the higher numbers of bacteria in the tissues.

Consistent with more severe inflammatory changes in the spleen and liver, we found there was a significant upregulation of gene expression of inflammatory cytokine IL-1β (Figure 2C) and TNF-α (Figures 2D,E) in infected p40^phox KO mice. Furthermore, ELISA analysis of cytokine concentration in spleen and liver homogenates confirmed that Salmonella infection induced a marked increase in production of TNF-α in spleen from mice lacking p40^phox (Figure 2H). Moreover, a trend of increase was found for IL-1β production in spleen from p40^phox−/− mice (Figure 2I) while there was no statistical significance. However, the levels of IFN-γ (Figure 2F) and IL-17A (Figure 2G) present in the livers and spleens of infected p40^phox−/− mice were comparable to WT levels. These findings suggest that p40^phox deficiency results in increased pro-inflammatory cytokine production (including TNF-α and IL-1β) during bacterial infection, which may be responsible for the observed enhanced tissue damage. Collectively, the data demonstrate that p40^phox is critical for host resistance and survival during Salmonella infection.

To further study the influence of p40^phox deficiency on host innate immunity against Salmonella infection, we next analyzed the frequency of macrophages and neutrophils in the splenic tissue of p40^phox knockout mice with Salmonella infection. Our FACS analysis of spleen cells revealed that the frequency of macrophages expressing CD11b⁺ F4/80⁺ was significantly enhanced in infected p40^phox−/− mice compared to WT mice (Figures 3A,C). A significant increase in splenic Ly6G⁺CD11b⁺ neutrophil numbers was also observed between WT mice and p40^phox-deficient mice with Salmonella infection (Figures 3B,D). These results are further supported by the results from our immunofluorescence microscopic analysis of spleen revealing enhanced macrophages and neutrophil in the spleen tissues of p40^phox-deficient mice (Figure 3E). These data suggest that loss of p40^phox results in increased transmigration of macrophages and neutrophils to the site of infection. We next determined the role for p40^phox in the regulation of bacteria-induced chemoattractant expression, which regulates macrophage and neutrophil recruitment. Accordingly, we assessed the expression of the three major chemoattractants: KC (CXCL1), MCP1 (CCL2), and MIP2 (CXCL2) in the spleen tissues. Our RT-PCR results showed that Salmonella infection induced an upregulation of KC, MCP1, and MIP2 gene expression in the spleens. This bacteria-induced upregulation of chemokines expression was more pronounced in the spleen of p40^phox-deficient mice than that detected in WT mice (Figures 3F–H). These observations show that p40^phox deficiency in mice may enhance the recruitment of phagocytes to the infection site by regulating the expression of KC, MCP1, and MIP2 during Salmonella infection and suggest that the excessive accumulation of granulocytes in the spleens of infected p40^phox KO mice may thereby enhance the inflammatory response.

Our data mentioned above suggest that Salmonella induced more severe systemic infection (i.e., spleen and liver infection) in p40^phox-deficient mice. To study whether and how p40^phox deficiency affects the early mucosal innate immune and inflammatory responses, we utilized the streptomycin-pretreatment model of Salmonella infection, which has been widely used in the field to analyze Salmonella-induced intestinal inflammation (2, 4, 5, 29). In brief, mice were treated with streptomycin prior to infection with S. Typhimurium (10⁸ CFU) and sacrificed 24 h post infection. Our results showed that mice infected with Salmonella developed acute disease, as indicated by a hunched posture, disturbed body hair, soft stool, and body weight loss soon after bacterial inoculation. A more severe disease status was observed in p40^phox−/− mice after Salmonella infection (data not shown). Histological examination of cecum showed extensive pathological changes that were typical in this model, including pronounced submucosal edema, dramatic disruption of epithelial architecture, and marked inflammatory cell infiltrates (Figure 4A). The ceca of p40^phox−/− mice showed a more pronounced edema of the intestinal tissue and increased inflammatory cellular infiltration (Figure 4A) than that detected in WT mice. These cecal inflammatory features were absent from uninfected mice (Figure 4A). Using the pathological scoring system (29), we found that the cecal inflammation was notably more severe in p40^phox KO mice than in WT mice (Figure 4B). Immunofluorescence microscopic analysis of infected intestine (cecum) revealed increased bacterial loads in the tissue of the p40^phox KO mice (Figure 4D) compared to WT B6 mice (Figure 4C). Further analysis of tissue bacterial loads showed that the enhanced severity of intestinal injury in the p40^phox KO mice was associated with significantly higher bacterial numbers in livers (Figure 4E) and spleens (Figure 4F) compared to WT mice, again an indication of more severely impaired bacterial killing activity in p40^phox-deficient mice. Moreover, and consistent with the severe intestinal injury, the mRNA expression levels of IL-1β (Figure 4G) and TNF-α (Figure 4H) were upregulated markedly in the cecum of infected mice compared with uninfected mice at 24 h post exposure to Salmonella.

Next, we examined whether p40^phox deficiency affected the infiltrating cell population in cecal lamina propria of mice in the streptomycin-pretreated Salmonella-colitis model using the immunofluorescence approach. Ceca sections were stained with anti-F4/80 (35) for macrophages and anti-Ly6G for neutrophils (29). As shown in Figures 5C,D, the immunofluorescence examination revealed a clear infiltration of F4/80⁺ macrophages and Ly6G⁺ neutrophils in the cecal submucosa in the infected mice, but was absent in the ceca of the uninfected mice in both groups (Figures 5A,B). A significantly higher frequency of both F4/80⁺ macrophages and Ly6G⁺ neutrophils was noted in the cecal lamina propria of p40^phox knockout mice with Salmonella infection (Figures 5E,F) compared to WT mice. The observed increased macrophage and neutrophil infiltration in the intestinal tissue was found to be correlated with upregulation of KC (Figure 5G) and MCP1 (Figure 5H) in the cecum of infected mice compared with uninfected mice at 24 h post exposure to Salmonella. Notably, a much higher expression level of KC was observed in the inflamed cecum of p40^phox−/− mice than in WT mice (Figure 5G). Taken together, these results suggest that p40^phox−/− mice develop exacerbated Salmonella-induced enterocolitis, characterized by severe tissue injury and enhanced macrophages and neutrophil infiltration in the intestinal tissue.

Despite the observed increase in macrophages and neutrophils, our results also showed that p40^phox−/− mice have elevated numbers of bacteria present in infected tissue, suggesting that these phagocytes are unable to effectively control bacterial multiplication in p40^phox−/− mice. Previous study showed that the NADPH oxidase deficiency of p40^phox−/− neutrophils were severely deficient in bacterial killing (20). In this study, we evaluated the bacterial killing capacity of p40^phox−/− macrophages, as macrophages are the major population of tissue-resident mononuclear phagocytes, contributing significantly to the elimination of bacteria, and are the major source of critical mediators of inflammation. We isolated peritoneal macrophages from WT and p40^phox−/− mice, infected the cells with Salmonella in vitro, and then used gentamicin protection assay as a measure of macrophage microbicidal function (24). Our results showed that the macrophages isolated from p40^phox−/− mice internalized similar numbers of bacteria at 2 h after infection compared with cells isolated from WT mice (Figure 6A). However, at 6 h after initial bacterial exposure, the number of viable bacteria recovered from p40^phox-deficient macrophages was significantly higher than that in macrophages isolated from WT mice (Figure 6A), indicating an impaired bacterial killing capacity of p40^phox−/− macrophages.

To further determine the functional alterations of p40^phox−/− macrophages, peritoneal macrophages from WT and p40^phox−/− mice were collected, grown on cover slips, and infected with Salmonella. After incubation in gentamicin-containing medium, we visualized the influence of p40^phox deficiency on the uptake and subsequent elimination of Salmonella by macrophages in vitro using immunofluorescent staining approach at 2 and 6 h after bacterial infection. Our results showed that at the early time point (2 h), F4/80⁺ macrophages isolated from WT and p40^phox−/− mice internalized bacteria equally well, evidenced by the detection of similar frequency and intensity of Salmonella in macrophages (Figure 6B). At the 6 h time point, the number of cells that contained bacteria was markedly reduced in WT macrophages and compared to that detected in p40^phox−/− macrophages (Figure 6B), confirming that macrophages from p40^phox−/− mice are impaired in killing internalized bacteria.

Our cytokine production analysis of Salmonella-infected tissues revealed a more pronounced increase in TNF-α and IL-1β response in p40^phox-deficient mice (Figures 2 and 4). To define the impact of p40^phox deficiency on bacterial antigen-induced macrophage cytokine response, we collected peritoneal macrophages from WT and p40^phox−/− mice and exposed the cells to Salmonella LPS in vitro (36). Our real-time PCR analysis revealed that LPS treatment significantly upregulated the gene expression of TNF-α and IL-1β in both WT and p40^phox−/− cells, whereas the expression levels of IL-1β and TNF-α in p40^phox−/− macrophages were significantly higher than those isolated from normal mice (Figure 6C). These results suggest that p40^phox deficiency induces dysregulation of pro-inflammatory responses of macrophages.

Using a lucigenin-dependent chemiluminescence assay, we measured PMA-induced ROS generation by peritoneal macrophages as well as neutrophils isolated from WT and p40^phox−/− mice and confirmed that the production of ROS was markedly reduced in p40^phox−/− cells (Figures 7A,B). However, PMA-induced ROS production was not completely abolished in p40^phox−/− macrophages and neutrophils. Upon phagocytosis of the pathogens, the ROS are produced. In the next set of experiments, using the DCFDA Cellular ROS Detection Assay Kit, we determined the impact of p40^phox deficiency on the intracellular ROS production of both intestinal (Figures 7C,D) and peritoneal (Figures 7E,F) macrophages in response to S. Typhimurium infection. Intestinal and peritoneal macrophages were isolated from both WT and p40^phox-deficient mice and stained in the diluted DCFDA solution before being infected with S. Typhimurium or treated with PMA. Our results show that Salmonella-induced macrophage ROS production was significantly higher in cells from WT mice compared to p40^phox−/− mice (Figures 7C,E). Moreover, PMA-induced ROS production of macrophages was found to be markedly lower in p40^phox−/− macrophages compared to WT cells (Figures 7D,F), confirming the results from lucigenin-dependent chemiluminescence assay. These results demonstrate that intracellular ROS generation by macrophages from p40^phox−/− mice was significantly impaired.