Modulation of infection-mediated migration of neutrophils and CXCR2 trafficking by osteopontin

Abstract

Osteopontin (OPN) is a pro-inflammatory protein that paradoxically protects against inflammation and bone destruction in a mouse model of endodontic infection. Here we have tested the hypothesis that this effect of OPN is mediated by effects on migration of innate immune cells to the site of infection. Using the air pouch as a model of endodontic infection in mice, we showed that neutrophil accumulation at the site of infection with a mixture of endodontic pathogens is significantly reduced in OPN-deficient mice. Reduced neutrophil accumulation in the absence of OPN was accompanied by an increase in bacterial load. OPN-deficiency did not affect neutrophil survival, CXCR2 ligand expression, or the production of inflammatory cytokines in the air pouch. In vitro, OPN enhanced neutrophil migration to CXCL1, whereas in vivo, inhibition of CXCR2 suppressed cellular infiltration in air pouches of infected wild-type mice by > 50%, but had no effect in OPN-deficient mice. OPN increased cell surface expression of CXCR2 on bone marrow neutrophils in an integrin-α_v -dependent manner, and suppressed the internalization of CXCR2 in the absence of ligand. Together, these results support a model where the protective effect of OPN results from enhanced initial neutrophil accumulation at sites of infection resulting in optimal bacterial killing. We describe a novel mechanism for this effect of OPN: integrin-α_v -dependent suppression of CXCR2 internalization in neutrophils, which increases the ability of these cells to migrate to sites of infection in response to CXCR2 ligands.

Keywords: CXCR2; endodontic infection; integrin-αv; osteopontin; receptor recycling.

Figure 1

In the air pouch model of infection, reduced neutrophil accumulation in osteopontin knockout (OPN KO) mice corresponds to higher bacterial number. Endodontic pathogen (EP) bacteria mix was injected into previously formed air pouches at 1 × 10⁹ CFU/ mouse. (a–c) At each time‐point, the cells were harvested and total cells were counted (a). The percentage (b) and total number of neutrophils (c) were determined by FACS using FITC anti‐mouse Ly6G antibody. (d) Viable bacteria (CFU) were enumerated using the drop plate method. n = 9, *P < 0·05, **P < 0·01, ***P < 0·001.

Figure 2

Neutrophil survival is not affected in the absence of osteopontin (OPN). (a) Air pouch neutrophils were stained with Ly6G and propidium iodide (PI), and the proportions of apoptotic (R1, high PI staining); necrotic (R2, intermediate PI staining) and live (R3, negative PI staining) cells were determined by FACS. The sum of the percentage cells in the R1 and R2 gates is shown. (b) Cytotoxic effect of endodontic pathogen (EP) bacteria is only detected 48 hr after infection. The hallmark cytotoxic effect of the EP on mouse neutrophils is the generation of a substantial R2 necrotic population. The proportion of neutrophils in the three gates is shown for wild‐type (WT) mice. *P < 0·05 compared with all other groups by one‐way analysis of variance.

Figure 3

Cytokine and chemokine accumulation in infected air pouches. KC (A), interleukin‐6 (IL‐6) (B), macrophage inflammatory protein 2 (MIP‐2) (C), IL‐1α (D), and osteopontin (OPN) (E) levels were determined in the lavage fluid from air pouches at different times after infection using ELISA. With the exception of OPN, none of the differences between wild‐type (WT) and OPN knockout (KO) mice were significant. For both WT and OPN KO, KC and IL‐6 showed significantly higher levels at 6 hr after infection than at other times. *P < 0·05; (a) P < 0·05 compared with all other time‐points of the same genotype; (b) P < 0·05 compared with 0, 16 and 48 hr, OPN KO mice. Significance determined by one‐way analysis of variance, n = 4. (F) Immunoprecipitation/Western blot for OPN in air pouch fluid. Lanes 1–4: samples were immunoprecipitated with mouse monoclonal antibody 2A1 then subjected to Western blot with goat polyclonal antibody AF808 (R&D). Lane 1: WT air pouch fluid; Lane 2: OPN –/– air pouch fluid; Lane 3: PBS; Lane 4; PBS + 25 ng recombinant OPN. Lane 5: 100 ng recombinant OPN (R&D) only. Arrows indicate full‐length OPN, and a possible polymeric form; * indicates likely polymeric OPN in the recombinant preparation.

Figure 4

Osteopontin (OPN) enhances migration of bone marrow neutrophils (BMN) to KC. BMN were labelled with calcein AM and added to the upper chambers of a 48‐well chemotaxis system (5‐μm filter). Various concentrations of OPN or KC (in ng/ml) were added in triplicate into the upper and/or lower wells of the chamber as indicated. After 90 min of incubation at 37°, non‐migrated cells were removed from the top of the filter, and fluorescence on the polycarbonate membrane was measured using an Alpha Innotec Imager (a). A similar set of experiments was performed without calcein AM labelling: in this case cells were fixed with methanol, stained with Giemsa (b), and counted (40 × objective) (c). *P < 0·05; **P < 0·01, ***P < 0·001.

Figure 5

CXCR2 inhibition suppresses neutrophil accumulation in response to infection in wild‐type (WT) but not in osteopontin (OPN) knockout (KO) mice. Air pouches were raised in mice as described in Materials and methods. One hour before and again 1 hr after infection with the bacterial species, CXCR2 inhibitor (SB225002, 50 μg/mouse) or vehicle (0·9% NaCl with 0·33% Tween‐80) was injected intraperitoneally. Mice were killed and cells were harvested 6 hr after infection. Total cells (a) and neutrophil numbers (b) were determined as in Fig. 1; cell numbers from uninfected pouches (0 hr) are shown for comparison. n = 3 to n = 6 *P < 0·05. (c) Migration of isolated bone marrow neutrophils to KC (20 ng/ml,lower chamber) or in the presence of OPN (1000 ng/ml, upper chamber) was performed in the presence or absence of CXCR2 inhibitor (SB225002, 300 nm). n = 3, representative of three independent experiments. *P < 0·05.

Figure 6

Osteopontin (OPN) enhances the surface expression of CXCR2 on neutrophils in vitro. Bone marrow neutrophils from wild‐type (WT) (a, b) and integrin‐α _V ^{fl/–Tie2Cre} [α _V knockout (KO)] (c) mice were incubated with varying concentrations of KC (a, ng/ml), or OPN (b, c, ng/ml) and CXCR2 cell surface receptor expression was analysed by flow cytometry. CXCR2 median fluorescent intensity (MFI) was measured on Ly6G positive cells. *P < 0·05, **P < 0·01, compared with cells treated with medium only (Media). (d) Integrin‐α _v expression on Ly6G⁺ neutrophils was determined using anti‐mouse α _v antibody HMa5‐1. WT and α _v KO cells were collected from blood, bone marrow and peritoneal exudate as indicated. (e) WT bone marrow neutrophils (BMN) were incubated with or without OPN (1000 ng/ml) in the presence of blocking antibody to integrin‐α _v (HMa5.1, 10 μg/ml) for 2 hr, then CXCR2 cell surface receptor expression analysed by flow cytometry. This experiment was performed using a different flow cytometer, resulting in increased MFI. *P < 0·05; **P < 0·01.

Figure 7

Osteopontin (OPN) suppresses the internalization of CXCR2 in bone marrow neutrophils (BMN). (a) Experimental procedure: cell surface proteins of BMN were biotin labelled at 4°. Labelled neutrophils were then incubated with or without OPN for the times as indicated at 37° to allow internalization. Remaining surface biotinylation was removed, and biotinylated internalized proteins were collected with streptavidin agarose and analysed by Western blot (b), followed by densitometry analysis (c). (d) Neutrophils were surface biotinylated, then incubated for 2 hr with or without OPN or blocking antibody to the integrin‐α _v (HMa5·1, 10 μg/ml). Two independent experiments are shown in (b); normalized average data are presented in (c) and (d) (n = 2). Unst., unstimulated.