Sepsis lethality via exacerbated tissue infiltration and TLR-induced cytokine production by neutrophils is integrin α3β1-dependent

Abstract

Integrin-mediated migration of neutrophils to infected tissue sites is vital for pathogen clearance and therefore host survival. Although β2 integrins have been shown to mediate neutrophil transendothelial migration during systemic and local inflammation, relatively little information is available regarding neutrophil migration in sepsis beyond the endothelial cell layer. In this study, we report that integrin α3β1 (VLA-3; CD49c/CD29) is dramatically upregulated on neutrophils isolated from both human septic patients and in mouse models of sepsis. Compared with the α3β1 (low) granulocytes, α3β1 (high) cells from septic animals displayed hyperinflammatory phenotypes. Administration of a α3β1 blocking peptide and conditional deletion of α3 in granulocytes significantly reduced the number of extravasating neutrophils and improved survival in septic mice. In addition, expression of α3β1 on neutrophils was associated with Toll-like receptor-induced inflammatory responses and cytokine productions. Thus, our results show that α3β1 is a novel marker of tissue homing and hyperresponsive neutrophil subtypes in sepsis, and blocking of α3β1 may represent a new therapeutic approach in sepsis treatment.

Figure 1

Integrin α₃β₁ surface expression increases on human neutrophils during sepsis. (A) Integrin surface expression on neutrophils from SIRS patients (n = 9), sepsis patients (n = 15), and healthy donors (n = 7). Flow cytometry results expressed as a ratio of the integrin mean fluorescence intensity (MFI) to isotype control MFI of the same donor. *P < .05 (Wilcoxon rank-sum test). (B) Neutrophils isolated from healthy donors were stimulated with PMA (20 ng/mL), TNF-α (20 ng/mL), LPS (100 μg/mL), or fMLP (1 μM) for 1 or 3 hours. Fold changes in Itga3 gene expression, compared with unstimulated cells, were determined by reverse transcription polymerase chain reaction (upper panel), and surface expression of α₃β₁ was measured by flow cytometry (bottom panel). Data are expressed as mean ± SEM of 3 separate donors. *P < .05 (Mann-Whitney test). Integrin α₃β₁ is upregulated on mouse neutrophils in (C) CLP surgery and (D) endotoxemia models of sepsis. Cells isolated from bone marrow, peritoneal lavage, and peripheral blood of naive and septic mice at the indicated time points were gated for neutrophils using forward scatter/side scatter and Gr1^high expression. The results are expressed as % increases in MFI compared with naive controls (naive blood MFI was used to calculate the % change in peritoneal lavage [PL] expression). Data are expressed as mean ± SEM of 4 animals/time point.

Figure 2

Gr1^highCD11b^highα₃β₁^high neutrophils have pro-inflammatory phenotypes. (A) Total bone marrow (BM) cells from C57BL/6 mice (6 hours after CLP) were stained with Ly6G and CD11b antibodies. Cells were sorted into Ly6G^high and Ly6G^low populations, as indicated, and stained with hematoxylin and eosin after cytospin. Ly6G^high cells demonstrate neutrophil morphology with a multilobed nucleus, and Ly6G^low cells have a monocyte-like appearance. (B) The gating strategy for α₃β₁^high and α₃β₁^low neutrophil populations in mice is shown. The pseudo color plots demonstrate α₃β₁ upregulation on mouse BM neutrophils (gated as Ly6G^highCD11b^high single cells) 6 hours after CLP compared with neutrophils from naive mice. (C) Neutrophils from total BM of endotoxemia-treated mice were gated as Ly6G^highCD11b^high cells and fluorescence-activated cell sorted based on their α₃β₁ expression levels into α₃β₁^high and α₃β₁^low populations, as shown in panel B. Fold changes in Il6 and Il10 gene expression 12 hours after sepsis induction, compared with neutrophils from naive mice, were quantified by reverse transcription polymerase chain reaction. Data are expressed as mean ± SD of 3 separate experiments. *P < .05 (Mann-Whitney test). (D) Fluorescence-activated cell -sorted α₃β₁^high and α₃β₁^low neutrophils isolated from endotoxemia mice 12 hours after sepsis induction were in vitro stimulated with LPS (100 ng/mL) to induce cytokine production. Culture supernatants were collected after 24 hours of stimulation and secreted IL-6 and IL-10 levels were measured by enzyme-linked immunosorbent assay (ELISA). Data are expressed as mean ± SD of 3 separate experiments. *P < .05 (Mann-Whitney test). (E) MPO activity of α₃β₁^high and α₃β₁^low BM neutrophils from CLP and endotoxemia mice was measured using bioluminescence. Sorted cells (5 × 10⁴ cells/well) were stimulated with PMA (1 μM) in the presence of luminol (1 mg/well), and the luminescence intensity was imaged. The graphs of MPO activity kinetics (left) in α₃β₁^high and α₃β₁^low cells and the representative luminescence images (right) are shown. Endotox., endotoxin.

Figure 3

Blocking and conditional deletion of α₃β₁ inhibits neutrophil migration and improves survival. (A) Total cells were isolated from PL and lungs of septic mice (8 hours after CLP) injected either with LXY2 or control peptide (88 μg or 100 nmole/dose, IV) 2 hours after CLP. Bar graphs show total number of neutrophils (Gr1/Ly6G^highCD11b^high cells), 5 mice/group. (B) CLP surgery was performed and either LXY2 or control peptide (100 nmole/dose) was injected post-CLP, as shown in supplemental Figure 3C. Mouse survival was analyzed using Kaplan-Meier log-rank test (9/group). (C) Serum concentrations of IL-6 (8 hours after CLP) from LXY2- and control peptide–treated mice were measured by sandwich ELISA. Values are expressed in ng/mL (5 mice/group). (D) Bacterial loads from PL of LXY2- and control peptide–treated mice (5/group). Diluted samples were streaked on tryptone soya agar blood agar and colonies counted after 24 hours’ incubation at 37°C. (E) Total cells were isolated from PL and lungs of α₃ cKO mice (Itga3 ^flox/flox; Ela-Cre) and littermate Ela-Cre controls (Itga3 ^wt/wt; Ela-Cre) 8 hours after CLP. Neutrophils were gated as described earlier (4 mice/group). (F) CLP surgery was performed and survival analyzed using Kaplan-Meier log-rank test to compare α₃ cKO and Ela-Cre control mice (12/group). (G) Serum levels of IL-6 (8 hours after CLP, 4 mice/group). (H) Bacterial loads in PL of α₃ cKO and Ela-Cre mice 8 hours after CLP (4 mice/group). (A,C,D,E,G,H) Data are expressed as mean ± SEM. *P < .05 (Mann-Whitney test).

Figure 4

Integrin α₃β₁ blockade or targeted genetic depletion impedes neutrophil extravasation at the basement membrane. (A) The effect of α₃β₁ blockade on neutrophil migration in vivo was assessed by intravital multiphoton microscopy in the 1 μM fMLP-superfused cremaster vessels of LysM-GFP mice after IV injections of either control (Ctrl) peptide or LXY2 (88 μg or 100 nmole). (B) Transmission electron microscopy imaging of fMLP-superfused cremaster microvasculature of mice administered LXY2. Endo, endothelium; L, lumen; peri, pericyte; P, polymorphonuclear leukocytes. (C) The effect of α₃β₁ conditional deletion in α₃ cKO (Itga3^flox/flox; Ela-Cre) mice compared with control Ela-Cre (Itga3^wt/wt; Ela-Cre) mice on neutrophil transmigration in response to fMLP (1 μM), assessed by intravital multiphoton microscopy. (A-C) Results are representative of 3 independent experiments. (A,C) Results were quantified and expressed as mean ± SEM. *P < .05 (Mann-Whitney test).

Figure 5

Integrin α₃β₁ regulates TLR2-, but not TLR4-induced cytokine responses. (A) ELISA of IL-6, TNF-α, and IL-10 in supernatants of neutrophils, isolated by negative selection, from BM of α₃ cKO (Itga3^flox/flox; Ela-Cre) mice (white bars) and control Ela-Cre (Itga3^wt/wt; Ela-Cre) mice (black bars). Cells (10⁶) were stimulated for 24 hours with Pam₃CSK₄ (100 ng/mL) and supernatants tested in duplicate. The results from 3 independent experiments are expressed as mean ± SD. *P < .05 (Mann-Whitney test). (B) ELISA of IL-6, TNF-α, and IL-10 in supernatants of neutrophils, isolated by negative selection, from BM of α₃ cKO (white bars) and control (black bars) mice. Cells (10⁶ cells/300 μL) were stimulated for 24 hours with ultrapure LPS (100 ng/mL) and supernatants tested in duplicate. The results from 3 independent experiments are expressed as mean ± SD. *P < .05 (Mann-Whitney test).

Figure 6

Integrin α₃β₁-binding peptide LXY2 differentially regulates cytokine secretion due to TLR2 and TLR4 stimulation. (A) ELISA of IL-6, TNF-α, and IL-10 in supernatants of neutrophils, isolated by negative selection, from BM of wild-type C57Bl/6 mice. Cells (10⁶ cells/300 μL) were incubated for 24 hours with Pam₃CSK₄ (100 ng/mL) with and without α₃β₁-binding peptide LXY2 (1 μM). Supernatants were tested in duplicate or triplicate. The results from 3 independent experiments are expressed as mean ± SD. *P < .05 (Mann-Whitney test). (B) Left: Representative western blot images of FAK phosphorylation kinetics in neutrophils (600 000 cells/100 μL) from C57BL/6 mice after Pam₃CSK₄ (10 μg/mL) stimulation ± LXY2 (100 μM) over indicated periods. Total cell lysates were analyzed by western blotting to determine the extent of FAK phosphorylation at Tyr³⁹⁷. Right: Densitometric analysis of FAK phosphorylation in WT neutrophils from C57BL/6 mice. Levels of phospho-FAK were normalized to total FAK at each time point and expressed as mean fold increase compared with control (time = 0 minutes). Data are expressed as mean ± SEM from 3 independent experiments. *P < .05 (2-way repeated measures ANOVA with Bonferroni posttest).

Figure 7

Compensatory CD11b upregulation in α₃β₁-depleted mice negatively regulates TLR2 responses. (A) Surface expression of CD11b levels on BM neutrophils (gated as Ly6G^highCD115^low cells) from α₃ cKO and control Ela-Cre mice. Total BM cells (10⁶) were stimulated with Pam₃CSK₄ (1 μg/mL) over time, as indicated (n = 4/time point). Data are expressed as mean MFI ± SD (top) or mean fold change compared with unstimulated neutrophils ± SD (bottom). *P < .05 (2-way repeated measures ANOVA with Bonferroni posttest). (B) Left: Representative western blot images of Syk phosphorylation kinetics in neutrophils from α₃ cKO and Ela-Cre mice after Pam₃CSK₄ (1 μg/mL) stimulation over indicated periods. Total cell lysates were analyzed by western blotting to determine the extent of Syk phosphorylation at Tyr³⁴⁶ (equivalent to Tyr³⁵² for humans). Right: Densitometric analysis of Syk phosphorylation. phospho-Syk was normalized to total Syk at each time point and expressed as mean fold increase compared with time = 0 minutes. Data are expressed as mean ± SEM from 3 independent experiments. *P < .05 (2-way repeated measures ANOVA with Bonferroni posttest). (C) Left: Representative western blot images of total MyD88 in neutrophils from α₃ cKO and Ela-Cre mice after Pam₃CSK₄ (1 μg/mL) stimulation over time. Right: Densitometric analysis of bands expressed as % of control group (time = 0 minutes). MyD88 levels were normalized to actin at each time point. Data are expressed as mean % change ± SEM from 4 independent experiments.