Abstract
Myeloid leukocyte recruitment into the lung in response to environmental cues represents a key factor for the induction of lung damage. We report that Hck- and Fgr-deficient mice show a profound impairment in early recruitment of neutrophils and monocytes in response to bacterial LPS. The reduction in interstitial and airway neutrophil recruitment was not due to a cell-intrinsic migratory defect, because Hck- and Fgr-deficient neutrophils were attracted to the airways by the chemokine CXCL2 as wild type cells. However, early accumulation of chemokines and TNF-α in the airways was reduced in hck−/−fgr−/− mice. Considering that chemokine and TNF-α release into the airways was neutrophil independent, as suggested by a comparison between control and neutrophil-depleted mice, we examined LPS-induced chemokine secretion by neutrophils and macrophages in wild type and mutant cells. Notably, mutant neutrophils displayed a marked deficit in their capability to release the chemokines CXCL1, CXCL2, CCL3, and CCL4 and TNF-α in response to LPS. However, intracellular accumulation of these chemokines and TNF-α, as well as secretion of a wide array of cytokines, including IL-1α, IL-1β, IL-6, and IL-10, by hck−/−fgr−/− neutrophils was normal. Intriguingly, secretion of CXCL1, CXCL2, CCL2, CCL3, CCL4, RANTES, and TNF-α, but not IL-1α, IL-1β, IL-6, IL-10, and GM-CSF, was also markedly reduced in bone marrow–derived macrophages. Consistently, the Src kinase inhibitors PP2 and dasatinib reduced chemokine secretion by neutrophils and bone marrow–derived macrophages. These findings identify Src kinases as a critical regulator of chemokine secretion in myeloid leukocytes during lung inflammation.
Introduction
Polymorphonuclear neutrophil (PMN) recruitment into the lung represents a key feature of host defense against infection. However, accumulating evidence points to an important role of PMN in driving lung pathology in several diseases, including acute lung injury (ALI) (1), cystic fibrosis (2, 3), and tuberculosis (4). The aim to identify effective ways to reduce lung inflammation has prompted intense investigation on mechanisms that regulate PMN recruitment into the lung. These investigations have led to the view that stimulation of lung epithelial and innate immune cells by bacterial components or other inflammatory mediators triggers NF-κB–dependent synthesis and secretion of a wide array of chemokines and cytokines that promote PMN recruitment. Consistent with this view, blocking either NF-κB activation or chemokine–receptor interactions results in a marked decrease of PMN recruitment in different models of lung inflammation (4–12).
Several studies have identified Src-family kinases among the possible target molecules regulating inflammatory cell recruitment into the lung. Mice expressing a constitutively active form of Hck or with the selective granulocyte inactivation of the Src-family kinase inhibitor C-terminal Src kinase develop an exaggerated pulmonary inflammation spontaneously and are hyperresponsive to systemic or intranasal instillation of LPS (13, 14). Excessive inflammation in motheaten mice, a phenotype resulting from a mutation in the gene Ptpn6 that encodes for the nonreceptor protein-tyrosine phosphatase Shp1, is caused by enhanced signaling via Src kinases, and the Src downstream target Syk, in neutrophils (15). Consistent with the evidence that excessive Src kinase activity results in innate immune cell–mediated inflammatory responses, either genetic deficiency of Src kinases or their inhibition by drugs results in a marked reduction in granulocyte recruitment into the lung and other tissues (16–20).
In this report, we address whether Hck and Fgr regulate PMN and monocyte recruitment and development of lung inflammation in an LPS-induced model of ALI and found that deficiency of these kinases results in a markedly reduced susceptibility to ALI induction. Experiments performed with the aim to identify mechanisms by which Src-family kinases regulate myeloid cell recruitment excluded a role for Hck and Fgr in regulation of intrinsic neutrophil migratory ability. In fact, PMN recruitment into the airways of hck−/−fgr−/− mice in response to PMN-attractive chemokines was comparable with that detected in wild type (WT) mice. However, we found that secretion of four different chemokines, as well as TNF-α, was markedly defective in hck−/−fgr−/− PMNs and macrophages challenged with LPS. As a result, these chemokines accumulate to a lower extent in the airways of hck−/−fgr−/− mice. These findings concur with previous evidence that deficiency of Src kinases does not impair the chemotactic responses of PMNs in transwell assays in vitro or in chemical peritonitis in vivo and actually enhances the response of PMNs and dendritic cells to chemokines recognizing CXCR2 and CCR1 receptors (21, 22). In addition, they extend to lung inflammation the recently established concept that Src kinases are indispensable for autoantibody-induced inflammation in the joint and the skin due to their role in triggering PMN activation, but not in regulating their intrinsic migratory ability (23).
Materials and Methods
Mice and bone marrow cells
Generation and maintenance of hck−/−fgr−/− double-knockout mice in the C57BL/6J background were as described by Lowell and Berton (16). WT and knockout animals used in the experiments were 8–10 wk of age. Animals were housed at a pathogen-free facility at the University of Verona and treated according to protocols approved by the Minister of Health of Italy and the university animal care committee. Bone marrow neutrophils (PMNs) were isolated by centrifugation of bone marrow cells flushed from femurs and tibias over a Percoll discontinuous density gradient (Amersham, Arlington Heights, IL) as described by Fumagalli et al. (24). Bone marrow–derived macrophages (BMDMs) were isolated from femurs and tibias as previously described (25). Cells were resuspended in DMEM supplemented with Glutamax (BioWhittaker, Walkersville, MD) 15% FCS, 10% L cell conditioned medium (LCM) as a source of CSF-1, 100 U/ml penicillin, and 100 mg/ml streptomycin (BMDMs complete medium), and cultured at 37°C/5% CO2 in 75-cm2 flasks. After 24 h, the nonadherent cells were removed, counted, plated on multiwell plates, and incubated for 6–7 d in the above medium to allow differentiation to BMDMs.
Lung cells and fluid
Mice were anesthetized and were given LPS (5 μg) or PBS by intranasal instillation. At 4 or 24 h from challenge, mice were euthanized and cannulated through the trachea for the recovery of bronchoalveolar lavage fluid (BALF) cells. Airways were washed four times with 0.5 ml ice-cold PBS, and after centrifugation, the supernatant was collected and stored at −80°C. Total cells in the pellet were resuspended in PBS and counted. For cytospin preparations, 5 × 104 cells were centrifuged onto glass slides at 400 rpm. Cytospins were stained with May-Grünwald-Giemsa or nonspecific esterase, coverslipped, and examined by light microscope, and the differential cell counts performed on 300 cells. For calgranulin B staining, BALF cells recovered after 2 h from intranasal instillation of PBS or LPS were centrifuged onto glass slides at 400 rpm. Cytospins were kept in absolute ethanol for 30 min, then immunohistochemistry staining was done using a goat HRP-polymer kit (Biocare Medical, Concorde, CA) and the primary anti-mouse S100A9 Ab (R&D Systems, Minneapolis, MN). The assay was developed according to the manufacturer’s instructions.
Histological analysis and immunostaining
After euthanasia and isolation of BALF, the left lobe of the lung was formalin fixed, paraffin embedded, sectioned at 3–4 μm, and then stained with H&E for histological analysis or with anti-mouse F4/80 (Ly71) Ag (Cl:A3-1; Serotec, Oxford, U.K.) or anti-mouse Gr-1 (Clone RB6-8C5; R&D Systems) Abs for immunohistochemistry.
Neutrophil depletion
C57BL/6J and hck−/−fgr−/− mice were depleted using an anti-Ly6G mAb (BioXCell, West Lebanon, NH), as described by Daley et al. (26). In brief, anti-Ly6G mAb was diluted into sterile endotoxin-free 0.9% NaCl saline solution at a concentration of 1 mg/ml. The Ab was injected i.p. at a dose of 0.5 mg/mouse, 17 h before intranasal instillation of LPS (5 μg) or PBS. Control mice were injected i.p. with saline. Two hours after intranasal treatments of depleted and control mice, peripheral blood was collected from all experimental animals to confirm neutrophil depletion by cytofluorimetric analysis. This was performed using a panel of five fluorochrome-conjugated Abs to CD11b, GR-1 (clone RB6-8C5; Biolegend), CD11c, Ly6C, and Ly6G.
Cytokine detection by Multiplex
Cytokines and chemokines were measured in BALF and in supernatants of cultured cells. PMN or BMDM supernatants were collected after 4 or 24 h of culture in DMEM, 10% FCS, penicillin, and streptomycin. Cells were then lysed using the Milliplex Map Lysis Buffer (Merck Millipore, Billerica, MA) containing 1 μM DFP, 10 μM PAO, and the complete Protease Inhibitor Cocktail tablets (Roche Diagnostic GmbH, Mannheim, Germany). BALF, cell culture supernatants, and lysates were analyzed in triplicate for CXCL1/KC, CXCL2/MIP-2, CCL2/MCP-1, CCL3/MIP-1α, CCL4/MIP-1β, CCL5/RANTES, IL-1α, IL-1β, IL-6, IL-10, and TNF-α by multiplex bead array assay (Milliplex magnetic mouse cytokines panels; Merck Millipore) and acquired on a MagPix instrument (Luminex, 's-Hertogenbosch, The Netherlands). All reagent dilutions (beads, cytokine standards, cytokine controls, biotinylated detection Ab, etc.) were prepared and assay developed according to the manufacturer’s instructions.
Other assays
For cytofluorimetric analysis, after 4 h from the LPS challenge, BALF cells were recovered as described earlier, counted, and incubated with purified rat anti-mouse CD16/CD32 (clone 2.4G2; BD Biosciences, San Jose, CA) and mouse IgG (Sigma-Aldrich, St. Louis, MO) to block Fc receptors. Then cells were stained with a mixture of four fluorochrome-conjugated Abs to CD11b (clone M1/70; eBioscience, San Diego, CA), CD11c (clone HL3; BD Biosciences), Ly6C (clone AL-21; BD Biosciences), and Ly6G (clone 1A8; Biolegend, San Diego, CA). Data were acquired on a MACSQuant Analyzer (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany), and data analyses were performed using FlowJo software (Tree Star, Ashland, OR).
Statistical analysis
Data are expressed as mean values ± SD. Statistical significance between cell accumulation in the lung of different groups of mice was calculated by unpaired Student t test. Statistical significance of differences between cytokine/chemokine release by PMNs and BMDMs was evaluated by ANOVA with Bonferroni posttests, which was performed using GraphPad Prism software.
Results
Hck and Fgr regulate LPS-induced PMN recruitment into the lung
In mice, LPS inhalation induces a rapid recruitment of PMNs and monocytes into the bronchoalveolar space. In WT C57BL/6J mice, we detected a marked increase in the number of cells present in the BALF as early as 4 h after intranasal instillation of 5 μg LPS (Fig. 1A). At this early time point, myeloid leukocytes recruited into the airways were mostly PMNs (70.6 ± 19% of total cells; Fig. 1B). However, nonspecific esterase staining of BALF cells allowed us to detect a significant, albeit small, increase in the absolute number of esterase-positive mononuclear phagocytes (Mo; Fig. 1C). In hck−/−fgr−/− mice, PMN recruitment into the bronchoalveolar space at 4 h was nearly abolished after LPS inhalation (Fig. 1B). In addition, numbers of esterase-positive Mos was similar in hck−/−fgr−/− mice treated with vehicle (PBS) or LPS, thus making Mo recruitment in response to LPS in hck−/−fgr−/− mice virtually undetectable (Fig. 1C). The profound defect of the early myeloid leukocyte recruitment into the bronchoalveolar space of hck−/−fgr−/− mice (Fig. 1) is secondary to impaired cell migration from the blood to the lung interstitium. In fact, staining of paraffin-embedded lung sections from hck−/−fgr−/− mice with the granulocyte-specific Ab Gr-1 (Fig. 1D) or the Mo-specific Ab F4/80 (Fig. 1E) revealed a marked reduction of interstitial PMNs and Mos after LPS inhalation. The reduced recruitment of myeloid leukocytes into the airways was not a consequence of an impairment of the general response to LPS. In fact, 4 h after intranasal instillation of LPS, neutrophil counts in the blood were increased to the same extent in WT and hck−/−fgr−/− mice [control WT and hck−/−fgr−/− mice: 650 ± 80 and 900 ± 50/μl (n = 3), respectively; LPS-treated WT and hck−/−fgr−/− mice: 3000 ± 850 and 3200 ± 550/μl (n = 4), respectively]. Consistent with this finding, in previous studies with a model of systemic endotoxemia, hck−/−fgr−/− mice demonstrated many systemic signs characteristic of endotoxic shock (16).
PMN recruitment into the airways in response to LPS is markedly defective in hck−/−fgr−/− mice. After 4 h following intranasal instillation of 5 μg LPS or vehicle (PBS), BALF cells were enumerated as described in 2Materials and Methods. Total cells (A), PMNs (B), and Mo (C) numbers, as determined by cytospin staining and counting, in the BALF of WT (PBS, n = 12; LPS, n = 20) or hck−/−fgr−/− mice (PBS, n = 12; LPS, n = 18) are shown. *p < 0.05, ***p < 0.001. (D) After 4 h following intranasal instillation of LPS, lungs were removed, formalin-fixed, and paraffin-embedded. Sections of 3–4 μm of WT or hck−/−fgr−/− lung tissue were stained with the anti-granulocyte Ab Gr-1, followed by HRP-labeled secondary Ab. Insets show higher magnification of peribronchial area. A few dark brown PMNs are indicated by arrows. Note the very low number of PMNs in the lung of hck−/−fgr−/− compared with WT mice. (E) Lung sections prepared as described earlier were stained with the anti-macrophage Ab F4/80, followed by HRP-labeled secondary Ab and photographed with a 10× objective. Insets show higher magnification (images taken with a 40× objective) of peribronchial area. Note the very low number of F4/80+ mononuclear cells in the lung of hck−/−fgr−/− compared with WT mice. A few dark brown peribronchial monocytes are indicated by arrows.
PMN recruitment into the airways in response to LPS is markedly defective in hck−/−fgr−/− mice. After 4 h following intranasal instillation of 5 μg LPS or vehicle (PBS), BALF cells were enumerated as described in 2Materials and Methods. Total cells (A), PMNs (B), and Mo (C) numbers, as determined by cytospin staining and counting, in the BALF of WT (PBS, n = 12; LPS, n = 20) or hck−/−fgr−/− mice (PBS, n = 12; LPS, n = 18) are shown. *p < 0.05, ***p < 0.001. (D) After 4 h following intranasal instillation of LPS, lungs were removed, formalin-fixed, and paraffin-embedded. Sections of 3–4 μm of WT or hck−/−fgr−/− lung tissue were stained with the anti-granulocyte Ab Gr-1, followed by HRP-labeled secondary Ab. Insets show higher magnification of peribronchial area. A few dark brown PMNs are indicated by arrows. Note the very low number of PMNs in the lung of hck−/−fgr−/− compared with WT mice. (E) Lung sections prepared as described earlier were stained with the anti-macrophage Ab F4/80, followed by HRP-labeled secondary Ab and photographed with a 10× objective. Insets show higher magnification (images taken with a 40× objective) of peribronchial area. Note the very low number of F4/80+ mononuclear cells in the lung of hck−/−fgr−/− compared with WT mice. A few dark brown peribronchial monocytes are indicated by arrows.
Impairment of myeloid cell recruitment into the airways in response to LPS required the double inactivation of hck or fgr. In fact, we did not find a significant reduction of PMN recruitment into the bronchoalveolar space after LPS inhalation in single-knockout hck−/− or fgr−/− mice (data not shown). These findings indicate that these kinases play a redundant role in regulation of inflammatory myeloid cell recruitment into the lung, as we previously found for PMN recruitment into the liver (16).
Myeloid cell recruitment into the airways varies considerably based on the LPS dose used and the time after the LPS challenge. For example, as shown in Fig. 2A, after 24 h from challenge with concentrations of LPS as low as 5 ng, the number of PMNs recruited into the airways was even higher than after 4 h in response to 5 μg LPS (see Fig. 1B for comparison). To determine whether Hck and Fgr deficiency affected the kinetics of PMN recruitment, we examined mice at 24 h after high LPS dose (5 μg) instillation. As shown in Fig. 2B, after 24 h, neutrophil recruitment into the airways was comparable in WT and hck−/−fgr−/− mice. In contrast with high-dose LPS, recruitment of hck−/−fgr−/− PMNs into the airways after low-dose LPS (2.5 ng) instillation remained low after 24 h (Fig. 2C). From the findings reported in Figs. 1 and 2, we conclude that early myeloid leukocyte airway recruitment in response to LPS is strictly regulated by Hck and Fgr. However, at very high doses of LPS, other inflammatory mechanisms compensate for loss of Hck and Fgr, allowing neutrophil recruitment to occur with delayed kinetics.
Regulation of LPS-induced PMN recruitment into the lung by Hck and Fgr. (A) After 4 or 24 h following intranasal instillation of different doses of LPS or vehicle (PBS), BALF PMNs were enumerated as described in 2Materials and Methods. ND = not determined. Mean results ± SD of three independent experiments are shown. (B) PMNs were enumerated in the BALF after 24 h from the intranasal instillation of 5 μg LPS. (C) Mice were given 2.5 ng LPS intranasally and after 24 h BALF cells were enumerated as described in 2Materials and Methods. **p < 0.01.
Regulation of LPS-induced PMN recruitment into the lung by Hck and Fgr. (A) After 4 or 24 h following intranasal instillation of different doses of LPS or vehicle (PBS), BALF PMNs were enumerated as described in 2Materials and Methods. ND = not determined. Mean results ± SD of three independent experiments are shown. (B) PMNs were enumerated in the BALF after 24 h from the intranasal instillation of 5 μg LPS. (C) Mice were given 2.5 ng LPS intranasally and after 24 h BALF cells were enumerated as described in 2Materials and Methods. **p < 0.01.
Hck and Fgr regulate LPS-induced monocyte recruitment into the lung
As shown in Fig. 1, within 4 h from the intranasal instillation of LPS, the number of esterase-positive mononuclear cells, also stained by the anti-macrophage Ab F4/80, in the airways increased, thus pointing for an early recruitment of monocytes from the blood. However, considering that both resident alveolar macrophages and recruited blood monocytes are esterase-positive, nonspecific esterase staining did not allow us to appreciate a strong difference in the recruitment of monocytes in WT or hck−/−fgr−/− mice (see Fig. 1C). To better identify monocytes recruited into the airways, we used two different approaches.
In preliminary studies, we exploited the knowledge that blood monocytes and monocytes recruited early from the blood, but not resident macrophages, are positive for the expression of calgranulin (27). However, because the high number of PMNs, which are also calgranulin-positive, could have hampered enumeration of calgranulin-positive monocytes after 4 h from the LPS challenge, we examined recruitment of this cell population at 2 h after the LPS challenge, when the accumulation of PMNs into the airways was very low (see below for quantitative data). In cytocentrifuge preparations from the BALF of WT mice, we clearly detected a calgranulin-positive mononuclear cell population at 2 h after LPS, which was markedly reduced in the BALF of hck−/−fgr−/− mice (Fig. 3A, 3B). Notably, the nuclear/cytoplasmic ratio of calgranulin-positive cells was higher compared with that of resident, calgranulin-negative macrophages, that is, more reminiscent of a monocyte morphology.
Hck and Fgr regulate LPS-induced monocyte recruitment into the lung. WT or hck−/−fgr−/− mice were given 5 μg LPS or vehicle (PBS) intranasally. After 2 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods. (A) Cytocentrifuge preparations were stained for calgranulin and counterstained with hematoxylin as described in 2Materials and Methods. Arrows point to calgranulin-negative cells displaying the classical morphology of resident AMs. Arrowheads point to smaller, strongly calgranulin-positive cells with a higher nuclear/cytoplasmic ratio than resident AMs and that are visible only in LPS-treated WT mice airways. Images were taken with a 40× objective. (B) Percent of calgranulin-positive cells in the airways of WT and hck−/−fgr−/− mice 2 h after LPS challenge was quantified as described in 2Materials and Methods. **p < 0.01. (C and D) WT or hck−/−fgr−/− mice were given 5 μg LPS or vehicle (PBS) intranasally. After 4 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods. BALF cells were pelleted, washed twice with PBS, 2% FBS, 2 mM EDTA, and immunostained for CD11b, CD11c, Ly6G, and Ly6C. After exclusion of doublets and debris, a sequential gating strategy was used to identify populations expressing specific markers: neutrophils (CD11b+, Ly6G+, Ly6C+) and monocytes (CD11b+, Ly6G−, Ly6Clo/−). (E and F) Number of CD11b+/Ly6G+ (granulocytes) and CD11b+/Ly6G− (monocytes) cells in the BALF of vehicle (PBS) or LPS-treated WT and hck−/−fgr−/− mice. Mean results ± SD of three independent experiments are shown. **p < 0.01; ***p > 0.001.
Hck and Fgr regulate LPS-induced monocyte recruitment into the lung. WT or hck−/−fgr−/− mice were given 5 μg LPS or vehicle (PBS) intranasally. After 2 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods. (A) Cytocentrifuge preparations were stained for calgranulin and counterstained with hematoxylin as described in 2Materials and Methods. Arrows point to calgranulin-negative cells displaying the classical morphology of resident AMs. Arrowheads point to smaller, strongly calgranulin-positive cells with a higher nuclear/cytoplasmic ratio than resident AMs and that are visible only in LPS-treated WT mice airways. Images were taken with a 40× objective. (B) Percent of calgranulin-positive cells in the airways of WT and hck−/−fgr−/− mice 2 h after LPS challenge was quantified as described in 2Materials and Methods. **p < 0.01. (C and D) WT or hck−/−fgr−/− mice were given 5 μg LPS or vehicle (PBS) intranasally. After 4 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods. BALF cells were pelleted, washed twice with PBS, 2% FBS, 2 mM EDTA, and immunostained for CD11b, CD11c, Ly6G, and Ly6C. After exclusion of doublets and debris, a sequential gating strategy was used to identify populations expressing specific markers: neutrophils (CD11b+, Ly6G+, Ly6C+) and monocytes (CD11b+, Ly6G−, Ly6Clo/−). (E and F) Number of CD11b+/Ly6G+ (granulocytes) and CD11b+/Ly6G− (monocytes) cells in the BALF of vehicle (PBS) or LPS-treated WT and hck−/−fgr−/− mice. Mean results ± SD of three independent experiments are shown. **p < 0.01; ***p > 0.001.
As a second approach to demonstrate that deficiency of Hck and Fgr results in reduced recruitment of monocytes to the airways, we examined changes in the presence of CD11b+/Ly6G−/Ly6C+ or CD11b+/Ly6G−/Ly6Clo/− cells, which are known to represent two distinct populations of blood monocytes named inflammatory or patrolling, respectively (28). As shown in Fig. 3C, in WT mice, after 4 h after LPS stimulation, both a CD11b+Ly6G+Ly6C+ population (granulocytes) and a small CD11b+Ly6G−Ly6Clo/− population (patrolling monocytes) could be clearly detected. Deficiency of Hck and Fgr resulted in a reduction in the recruitment of the total CD11b+ cells, including the Ly6G+/Ly6C+ granulocytes and the Ly6G−Ly6Clo/− patrolling monocytes (Fig. 3D). Quantification of these defects across a number of mice confirms that both monocyte and PMN recruitment are significantly reduced in the hck−/−fgr−/− mice (Fig. 3D, 3F).
Hck and Fgr do not regulate PMN chemotactic response but do regulate accumulation of chemokines in the airways
At least two explanations may account for the earlier findings. The first one is that deficiency of Hck and Fgr results in an intrinsic defect in the ability of myeloid leukocytes to migrate toward an inflammatory site. We therefore asked whether PMN migration into the bronchoalveolar space in response to CXCL2/MIP-2 was defective in hck−/−fgr−/− mice. As shown in Fig. 4A and 4B, direct instillation of CXCL2/MIP-2 into the airways induced a dose-dependent and marked increase in the total number of cells recruited into the lung and most of these cells were PMNs (WT: 87.4 ± 38.2%; hck−/−fgr−/−: 90.8 ± 53.3%). Notably, no difference was found in the number of airway PMNs between WT and hck−/−fgr−/− mice. These findings confirm previous results obtained with a thioglycollate-induced peritonitis model of PMN recruitment in vivo (21) and recent studies examining PMN migration to inflamed joints and skin (23).
Hck and Fgr deficiency does not result in any alteration of CXCL2-induced PMN recruitment into the lung. (A and B) WT or hck−/−fgr−/− mice were given the indicated doses of CXCL1/MIP-2 or vehicle (PBS) intranasally (n = 4 for each condition with both mouse strains). After 4 h, mice were sacrificed and cell accumulation into the airways quantified as described in 2Materials and Methods. (C–F) WT or hck−/−fgr−/− mice were given LPS or vehicle (PBS) intranasally. After 4 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods for multiplex bead array analysis of CXCL1 (C), CCL3 (D), and TNF-α (E) in cell-free supernatants. The number of neutrophils (F) in the BALF used for detection of CXCL1, CCL3, and TNF-α is shown for comparison. **p < 0.01, ***p < 0.001.
Hck and Fgr deficiency does not result in any alteration of CXCL2-induced PMN recruitment into the lung. (A and B) WT or hck−/−fgr−/− mice were given the indicated doses of CXCL1/MIP-2 or vehicle (PBS) intranasally (n = 4 for each condition with both mouse strains). After 4 h, mice were sacrificed and cell accumulation into the airways quantified as described in 2Materials and Methods. (C–F) WT or hck−/−fgr−/− mice were given LPS or vehicle (PBS) intranasally. After 4 h, mice were sacrificed and BALF prepared as described in 2Materials and Methods for multiplex bead array analysis of CXCL1 (C), CCL3 (D), and TNF-α (E) in cell-free supernatants. The number of neutrophils (F) in the BALF used for detection of CXCL1, CCL3, and TNF-α is shown for comparison. **p < 0.01, ***p < 0.001.
A second possible explanation is that reduced amounts of PMN-attracting chemokines accumulate in the airways of hck−/−fgr−/− mice. Notably, we found that the increase in CXCL1/KC (Fig. 4C) and CCL3/MIP-1α (Fig. 4D), as well as the proinflammatory cytokine TNF-α (Fig. 4E), in the BALF of hck−/−fgr−/− mice was reduced after 4 h from the LPS challenge, that is, at a time when a much lower number of PMNs was recruited into the airways of hck−/−fgr−/− mice compared with WT ones (Fig. 4F).
The experiments reported in Fig. 4 raised the issue of the origin of chemokines/cytokines released into the airways. Epithelial or innate immune cells resident in the lung parenchyma release PMN-attracting chemokines (4–12). Thus, reduced chemokine/TNF-α release in the airways of hck−/−fgr−/− mice could reflect a role of these kinases in regulating lung parenchymal cell activation. However, to our knowledge, these kinases were not reported to be expressed in lung cells and RT-PCR analysis did not allow us to detect Hck or Fgr expression in type II airway epithelial cell (C.A. Lowell, unpublished observations). Notably, at early time points after lung injury, PMNs were reported to be indispensable for the induction of lung inflammation (29). This raises the possibility that PMNs are required for the early release of chemoattractants, and Hck and Fgr regulate this function.
To address this issue in our experimental setting, we examined chemokine/TNF-α accumulation into the airways at an early time point (2 h) after intranasal instillation of LPS, that is, when the number of PMNs accumulated into the airways is still low, and after depletion of PMNs by injection of the anti-PMN Ab LY6G. As shown in Fig. 5A, and consistent with the results reported in Fig. 1, at this early time point LPS-induced PMN recruitment into the airways was defective in hck−/−fgr−/− mice compared with WT animals. Anti-Ly6G Ab treatment dramatically reduced PMN numbers in the blood of WT and hck−/−fgr−/− mice, as well as PMN recruitment into the airways (Fig. 5A, see legend for absolute PMN numbers). Similar to data concerning PMN recruitment, accumulation of chemokines and TNF-α into the airways was reduced in hck−/−fgr−/− mice (Fig. 5B–E). However, airway chemokine/TNF-α accumulation was totally blood and airway PMN independent. Experiments performed with WT mice showed that chemokine/TNF-α accumulation in the airways was comparable in PMN-depleted and control mice also after 4 h from intranasal instillation of LPS (data not shown).
Chemokine and TNF accumulation into the airways in response to LPS is PMN independent, but defective in hck−/−fgr−/− mice. Mice were injected i.p. with vehicle (0.9% NaCl) or an anti-Ly6G Ab as described in 2Materials and Methods. After 17 h from the injection mice were intranasally injected with vehicle (PBS) or 5 μg LPS. After 2 h, a sample of blood was withdrawn for cytofluorimetric analysis of C11b+/Ly6G+ cells, and BALF cells were enumerated as described in 2Materials and Methods. (A) Percent PMNs in the blood and the lung of control and PMN-depleted mice. Absolute numbers of PMNs in the BALF of LPS-challenged mice were: WT mice injected i.p. with vehicle, 8.133 ± 5.693/mouse (n = 4); WT mice injected i.p. with anti-Ly6G Abs, 350 ± 300; hck−/−fgr−/− mice injected i.p. with vehicle, 1.235 ± 460; hck−/−fgr−/− mice injected i.p. with anti-Ly6G Abs, 482 ± 480. PMNs were undetectable in the lung of mice intranasally injected with PBS (not shown). CXCL1 (B), CXCL2 (C), CCL3 (D), CCL4 (E), and TNF-α (F) were assayed in cell-free supernatants by multiplex bead array analysis as described in 2Materials and Methods. **p < 0.01, #p < 0.001. n.s., not significant.
Chemokine and TNF accumulation into the airways in response to LPS is PMN independent, but defective in hck−/−fgr−/− mice. Mice were injected i.p. with vehicle (0.9% NaCl) or an anti-Ly6G Ab as described in 2Materials and Methods. After 17 h from the injection mice were intranasally injected with vehicle (PBS) or 5 μg LPS. After 2 h, a sample of blood was withdrawn for cytofluorimetric analysis of C11b+/Ly6G+ cells, and BALF cells were enumerated as described in 2Materials and Methods. (A) Percent PMNs in the blood and the lung of control and PMN-depleted mice. Absolute numbers of PMNs in the BALF of LPS-challenged mice were: WT mice injected i.p. with vehicle, 8.133 ± 5.693/mouse (n = 4); WT mice injected i.p. with anti-Ly6G Abs, 350 ± 300; hck−/−fgr−/− mice injected i.p. with vehicle, 1.235 ± 460; hck−/−fgr−/− mice injected i.p. with anti-Ly6G Abs, 482 ± 480. PMNs were undetectable in the lung of mice intranasally injected with PBS (not shown). CXCL1 (B), CXCL2 (C), CCL3 (D), CCL4 (E), and TNF-α (F) were assayed in cell-free supernatants by multiplex bead array analysis as described in 2Materials and Methods. **p < 0.01, #p < 0.001. n.s., not significant.
The data reported in Figs. 4 and 5 suggest a complex scenario in mechanisms of regulation of PMN recruitment by the Src-kinases Hck and Fgr. These kinases seem to be dispensable for chemokine-induced PMN recruitment (Fig. 4A, 4B). However, their deficiency results in reduced chemokine/TNF-α accumulation into the airways independently of the blood and airway PMN number (Figs. 4C–E, 5). Elucidating which cells display an Hck/Fgr-dependent pathway of chemokine/TNF-α secretion in a complex multicellular organ is a worthy object of future investigation. To start to address this issue, we adopted a reductionist approach starting with the characterization of the role of Hck and Fgr in regulation of myeloid cell chemokine secretion.
hck−/−fgr−/− PMNs release a lower amount of chemokines in response to LPS
To address the issue of PMN chemoattractant release, we examined secretion of different chemokines acting on granulocytes (CXCL1/KC, CXCL2/MIP-2, CCL3/MIP-1α, CCL4/MIP-1β) by WT and mutant PMNs in response to different doses of LPS (Fig. 6). As shown in Fig. 6A, hck−/−fgr−/− PMNs released much lower amounts of all of the four chemokines after 4 h following stimulation with LPS and independently of the stimulus dose used. The defect in chemokine release by hck−/−fgr−/− compared with WT PMNs may explain differences we observed in PMN polarization between the two mouse strains (Fig. 6B). In fact, whereas WT PMNs maintained for 4 h in the presence of LPS displayed a clearly polarized morphology typically occurring in chemoattractant-stimulated cells, hck−/−fgr−/− PMNs remained rounded.
Hck−/−fgr−/− PMNs are defective in the ability to secrete neutrophil-attractive chemokines in response to LPS. (A) Bone marrow PMNs were maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of the indicated doses of LPS. After 4 h, the medium was collected and chemokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from three different mice are reported. *p < 0.05, **p < 0.01, ***p < 0.001. If not highlighted by an asterisk, differences between WT and hck−/−fgr−/− PMNs were not statistically significant. (B) Bone marrow PMNs treated as in (A) were photographed with a 40× phase-contrast objective. Arrows point to elongated, polarized PMNs.
Hck−/−fgr−/− PMNs are defective in the ability to secrete neutrophil-attractive chemokines in response to LPS. (A) Bone marrow PMNs were maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of the indicated doses of LPS. After 4 h, the medium was collected and chemokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from three different mice are reported. *p < 0.05, **p < 0.01, ***p < 0.001. If not highlighted by an asterisk, differences between WT and hck−/−fgr−/− PMNs were not statistically significant. (B) Bone marrow PMNs treated as in (A) were photographed with a 40× phase-contrast objective. Arrows point to elongated, polarized PMNs.
To know whether the mutant PMN defect is selective for chemokines active on granulocytes, we examined secretion of a few other cytokines (IL-1α, IL-1β, IL-6, IL-10, and TNF-α) by WT and hck−/−fgr−/− PMNs (Fig. 7). Notably, secretion of IL-1α, IL-1β, IL-6, and IL-10 was robust after 24 h, but almost undetectable after 4 h of stimulation with LPS. Intriguingly, the secretion of these four cytokines after 24 h from the LPS challenge was comparable in WT and hck−/−fgr−/− PMNs. Differently from IL-1α, IL-1β, IL-6, and IL-10, TNF-α was detectable in the incubation medium after 4 and 24 h from LPS stimulation, and hck−/−fgr−/− PMNs displayed a reduced ability to secrete this cytokine at both time points.
Hck−/−fgr−/− PMNs are not defective in the ability to secrete IL-1α, IL-1β, IL-6, and IL-10 in response to LPS. Bone marrow PMNs were maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of the indicated doses of LPS. After 4 or 24 h, the medium was collected and cytokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from two to three different mice are reported. *p < 0.05, ***p < 0.001. If not highlighted by an asterisk, differences between WT and hck−/−fgr−/− PMNs were not statistically significant.
Hck−/−fgr−/− PMNs are not defective in the ability to secrete IL-1α, IL-1β, IL-6, and IL-10 in response to LPS. Bone marrow PMNs were maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of the indicated doses of LPS. After 4 or 24 h, the medium was collected and cytokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from two to three different mice are reported. *p < 0.05, ***p < 0.001. If not highlighted by an asterisk, differences between WT and hck−/−fgr−/− PMNs were not statistically significant.
It is important to note that findings similar to those described earlier were obtained examining chemokine/cytokine secretion by syk−/− PMNs. In fact, Syk-deficient PMNs secrete lower amounts of CXCL1, CXCL2, CCL3, and TNF-α, but normal amounts of IL-1β and IL-6 in response to Escherichia coli and Staphylococcus aureus (30). Notably, this last study demonstrated that Syk regulates secretion, but not expression and intracellular storage, of a few chemokines and TNF-α. To address whether the Src-family kinases Hck and Fgr are also implicated in the regulation of chemokines and TNF-α secretion selectively, we compared the secreted and intracellular pool of CXCL1, CXCL2, CCL3, CCL4, and TNF-α in WT and hck−/−fgr−/− PMNs (Fig. 8). Notably, although secretion of these four cytokines and TNF-α was suppressed or strongly reduced in hck−/−fgr−/− PMNs, their intracellular content was comparable with that of WT PMNs. We conclude that an Src/Syk signaling pathway regulates PMN chemokine secretion in response to LPS.
Hck−/−fgr−/− PMNs are defective in the ability to secrete, but not to synthetize and store intracellularly, neutrophil-attractive chemokines in response to LPS. Bone marrow PMNs were maintained in 48-well plates in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of 100 ng/ml LPS without or with 1 μM dasatinib or 10 μM PP2. After 4 h, PMNs were resuspended by pipetting and cells were pelleted in a microfuge. After aspiration of the supernatant, pelleted cells were lysed in an equal volume of lysis buffer (see 2Materials and Methods). Chemokines and TNF-α released in the supernatant or present in the cell lysate were assayed by multiplex bead array. Mean results of three experiments in each of which cells were pooled from three different mice are reported. *p < 0.05, **p < 0.01, #p < 0.001. n.s., not significant.
Hck−/−fgr−/− PMNs are defective in the ability to secrete, but not to synthetize and store intracellularly, neutrophil-attractive chemokines in response to LPS. Bone marrow PMNs were maintained in 48-well plates in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of 100 ng/ml LPS without or with 1 μM dasatinib or 10 μM PP2. After 4 h, PMNs were resuspended by pipetting and cells were pelleted in a microfuge. After aspiration of the supernatant, pelleted cells were lysed in an equal volume of lysis buffer (see 2Materials and Methods). Chemokines and TNF-α released in the supernatant or present in the cell lysate were assayed by multiplex bead array. Mean results of three experiments in each of which cells were pooled from three different mice are reported. *p < 0.05, **p < 0.01, #p < 0.001. n.s., not significant.
To strengthen the notion that Src-family kinases are implicated in regulation of LPS-induced chemokine release, we addressed whether compounds targeting Src kinases inhibit this response. As shown in Fig. 8, both dasatinib, a dual-specificity Abl/Src inhibitor, and PP2, a selective Src kinase inhibitor, suppressed secretion of CXCL1, CXCL2, CCL3, CCL4, and TNF-α by WT PMNs. However, they also had a strong inhibitory effect on CXCL2, CCL3, CCL4, and TNF-α intracellular accumulation, suggesting either that, by targeting other Src-family members, they have a broader inhibitory effect on Src kinase activities than that deriving from Hck/Fgr deficiency or that off-target effects of these compounds result also in inhibition of gene transcription.
Hck and Fgr regulate cytokine secretion by murine macrophages
The issue of the role of Src-family kinases in regulation of chemokine and cytokine secretion by macrophages is somehow controversial. Early studies implicated Src-family kinases in regulation of the response to LPS (reviewed in Ref. 31), but triple-deficient hck−/−fgr−/−lyn−/− peritoneal and BMDMs display no alteration in LPS/IFN-γ–induced release of IL-1, IL-6, or TNF-α (25). The defective capability of hck−/−fgr−/− PMNs to release CXCL1, CXCL2, CCL3, CCL4, and TNF-α prompted studies to examine the role of Src-family kinases in chemokine/cytokine secretion by macrophages. To this purpose, we stimulated WT or hck−/−fgr−/− BMDMs with LPS, a TLR4 ligand, PAM3, a TLR2 ligand, or R848, a TLR7/8 ligand, at doses found to be optimal in preliminary experiments and assayed the release of a wide array of chemokines and cytokines (Fig. 9, Supplemental Fig. 1). Similar to results obtained with PMNs (Fig. 7) and BMDMs stimulated with LPS/IFN-γ (25), we did not find major alterations in the release of IL-1α, IL-1β, IL-6, IL-10, and GM-CSF by hck−/−fgr−/− BMDMs in response to LPS, with the only exception of TNF-α, whose release by mutant BMDMs was ∼3-fold lower compared with WT BMDMs (580.13 ± 81.1 versus 1832.17 ± 417.4 pg/ml; Supplemental Fig. 1). However, LPS-induced secretion of the chemokines CXCL1, CXCL2, CCL2/MCP-1, CCL3, CCL4, and CCL5/RANTES was significantly reduced in hck−/−fgr−/− BMDMs (Fig. 9). Notably, hck−/−fgr−/− BMDMs also displayed a reduced ability to secrete CXCL1, CXCL2, and CCL2 in response to PAM3 and R848 (Fig. 9), but similar to data obtained using LPS as a stimulus, they were not defective in PAM3- and R848-induced secretion of IL-1α, IL-1β, IL-6, IL-10, and GM-CSF (Supplemental Fig. 1). With the exception of CCL4, whose secretion was already maximal at 4 h, the release of CXCL1, CXCL2, CCL2, CCL3, and CCL5 was higher after 24 h from stimulation, but hck−/−fgr−/− BMDMs displayed a markedly reduced ability to secrete these chemokines in response to LPS also after this longer time point (Fig. 9). In addition, after 24 h, hck−/−fgr−/− BMDMs secreted lower amounts of CXCL1, CXCL2, and CCL2 also in response to PAM3 and R848 (Fig. 9). Despite some variability in the response to different agonists (CCL3 and CCL4 release in response to PAM3 and R848 after 4 and 24 h were comparable in WT versus hck−/−fgr−/− BMDMs), these findings implicate Hck and Fgr in regulation of several chemokines release in murine macrophages.
Hck−/−fgr−/− BMDMs are defective in the ability to secrete chemokines in response to LPS and other TLR ligands. BMDMs were isolated and maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of LPS (25 ng/ml), PAM3 (200 ng/ml), or R848 (10 μM). After 4 or 24 h, the medium was collected and chemokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from two to three different mice are reported. *p < 0.05, **p < 0.01, #p < 0.001. ns, not significant.
Hck−/−fgr−/− BMDMs are defective in the ability to secrete chemokines in response to LPS and other TLR ligands. BMDMs were isolated and maintained in RPMI 1640 medium supplemented with 10% FBS in the absence or presence of LPS (25 ng/ml), PAM3 (200 ng/ml), or R848 (10 μM). After 4 or 24 h, the medium was collected and chemokines released in the supernatant assayed by multiplex bead array as described in 2Materials and Methods. Mean results of three experiments in each of which cells were pooled from two to three different mice are reported. *p < 0.05, **p < 0.01, #p < 0.001. ns, not significant.
To strengthen these findings, we addressed whether dasatinib, which suppressed chemokine secretion by PMNs challenged with LPS (Fig. 8), also affected BMDM responses (Supplemental Fig. 2). Notably, dasatinib had a strong inhibitory effect on the capability of BMDMs to release CXCL1, CXCL2, CCL2, CCL3, CCL4, and TNF-α in response to a wide range of LPS doses.
Discussion
The lung is protected from pathogens entering the body through the airways by various mechanisms that include both constitutive, innate defenses, such as the antimicrobial airway surface liquid that is constantly cleared through the ciliary activity and the cough reflex (30, 32), and pathogen-induced lung tissue cell responses (33–36). These last are mainly based on the release of proinflammatory cytokines and chemokines orchestrating the recruitment of leukocytes endowed with effector and regulatory functions. The response of the lung to injury must be finely tuned to avoid an excessive inflammatory reaction and the consequent and irreversible damage of the lung parenchyma and its gas-exchange function; indeed, an excessive inflammatory response contributes to a severe impairment of the lung function.
In this report, we identify Hck and Fgr as essential components of a signaling pathway regulating early recruitment of neutrophils and monocytes into the lung in response to LPS. Our findings that early steps in ALI development are defective in hck−/−fgr−/− mice complement previous findings with mice genetically modified to express activated Src kinases in myeloid leukocytes. Indeed, either mice expressing a constitutively active mutant of Hck carrying a substitution of the C-terminal regulatory tyrosine (Y499) with phenylalanine or mice displaying a higher Src-family kinase activity in their granulocytes, due to the conditional inactivation of the Src inhibitory kinase C-terminal Src kinase, develop a lung pathology characterized by myeloid cell infiltration within the lung interstitium and airways and are, as a result, hyperresponsive to LPS-induced shock (13, 14). Excessive skin and lung inflammation is also a feature of motheaten mice that carry a loss-of-function mutation in the Ptpn6 gene, which encodes for the nonreceptor protein-tyrosine phosphatase Shp1. Notably, Shp1 is believed to counteract Src kinase–mediated signals, and its conditional deletion in granulocytes results in an enhancement of Src kinase activity (15, 37). All these findings point to a critical role of Src kinases in regulation of myeloid leukocyte recruitment into inflammatory sites; this conclusion was also supported by the evidence that recruitment of eosinophils into the lung in a model of allergic inflammation and neutrophil recruitment into the liver in an LPS-induced shock model are defective in mice with the deficiency of Hck and Fgr (16, 17). Importantly, a very recent report showed that the genetic deficiency of all three Src kinases expressed at the highest level in myeloid leukocytes, that is, Hck, Fgr, and Lyn, results in protection from autoantibody-induced arthritis, skin blistering disease, and reverse passive Arthus reaction and a failure of myeloid leukocytes to accumulate at sites of inflammation (23).
Intriguingly, the role of Src kinases in promoting granulocyte recruitment into inflammatory sites does not seem to depend on their capability to regulate their intrinsic migratory ability (see Ref. 38 for a review). Indeed, standard transwell assays failed to reveal any defect in the in vitro migratory ability of murine Src kinase–deficient PMNs or human PMNs treated with an Src inhibitor (21, 22, 24, 39). More importantly, the use of mixed chimeric mice, containing both WT and Src-deficient PMNs, demonstrated that they are recruited to the same extent into the inflamed peritoneum, the joints, and the skin (21, 23). Notably, studies with mixed chimeric mice containing both WT PMNs and PMNs deficient of Syk, the tyrosine kinase acting in concert with Src kinases in granulocyte responses to stimuli acting through integrin and immune receptors (40, 41), also failed to reveal any intrinsic migratory defect of Syk-deficient PMNs (21, 40). Consistent with these findings, we found that airway recruitment of hck−/−fgr−/− PMNs was normal after intranasal instillation of the PMN-attracting chemokine CXCL1 (Fig. 4). Collectively, these results suggest that the Src/Syk signaling pathway does not regulate the intrinsic capability of PMNs to migrate into inflammatory sites in response to chemoattractants.
PMN recruitment into the lung in response to harmful stimuli present in the airways is due to the release of a wide array of cytokines and chemokines (33–35, 42). Cells that have been implicated in the release of proinflammatory and chemotactic factors include epithelial cells lining the bronchial and the alveolar lumen, alveolar and interstitial macrophages, and even monocytes that are recruited early from the blood (4, 33, 43–47). Although not specifically addressed in the context of lung inflammation, recent findings add further complexity to mechanisms that regulate PMN recruitment into inflammatory sites. In fact, full-blown PMN recruitment into the skin has been shown to depend on the release of LTB4 and CXCR2 chemokine ligands by a limited number of scouting PMNs in the extravascular compartment (48, 49). In this report, we show that, at least as far as the bronchoalveolar space is concerned, early accumulation of chemokines and the proinflammatory cytokine TNF-α in response to LPS occurs independently of PMNs (Fig. 5), suggesting that myeloid cells resident in the lung interstitium or the alveoli, together with lung epithelial cells, play a major role in inducing a first wave of chemoattractant generation. Intriguingly, this wave depends on the Src-family kinases Hck and Fgr, which are believed to be restricted in their expression to cells of the myeloid lineage or B cells. Although we cannot exclude that these kinases may also regulate epithelial cell responses, the available knowledge suggests that it is the response of local macrophages, or early recruited monocytes, that play a major role in the LPS-induced chemoattractant release. In addition, although we did not address this issue, activation of PMNs to release PMN-attracting chemokines may represent an important step in the amplification of PMN recruitment into the interstitium before their final migration into the airways.
With the aim to elucidate the role of Src-family kinases in lung inflammation, we examined in some detail the capability of PMNs and macrophages to respond to LPS using either cells deficient of Hck and Fgr or WT cells treated with Src kinase inhibitors of broader specificity. Intriguingly, we found that, after an LPS challenge, the intracellular accumulation of CXCL1, CXCL2, CCL3, CCL4, and TNF-α was comparable in WT and hck−/−fgr−/− PMNs; however, their secretion in the external medium was markedly reduced (Fig. 8). Thus, Hck and Fgr do not seem to regulate LPS signaling leading to NF-κB activation, chemokine gene transcription, and chemokine intracellular accumulation before their secretion. In line with this conclusion, other LPS-induced cytokines such as IL-1α, IL-1β, IL-6, and IL-10 were secreted to comparable levels by WT and hck−/−fgr−/− PMNs (Fig. 6). These findings are consistent with the notion that one of the roles played by Src kinases in granulocytes is the regulation of secretory pathways (39, 50–53). Importantly, Syk was reported to regulate secretion, but not intracellular accumulation, of PMN-attracting chemokines and TNF-α in response to pathogenic bacteria (30), suggesting that an Src/Syk signaling pathway targets trafficking of intracellularly stored PMN-attracting chemokines and TNF-α to the external milieu. Notably, the Src/Syk signaling pathway plays a central role in regulation of cytoskeleton dynamics (reviewed in Ref. 40) and this, in turn, is implicated in the series of events underlying granule–plasma membrane fusion (reviewed in Ref. 54). Hence it is conceivable that a possible defect in plasma membrane targeting and docking of chemokine-containing vesicles in Src- and Syk-deficient PMNs is secondary to alterations in rearrangement of the cytoskeleton. Notably, hck−/−fgr−/− PMNs displayed alterations in their capability to polarize in response to LPS (Fig. 6). The role of Src kinases in regulation of chemokine secretion is reinforced by the findings that the selective Src kinase inhibitor PP2 and the dual-specificity Abl/Src inhibitor dasatinib hamper PMN-attracting chemokines and TNF-α secretion (Fig. 8). However, the effect of these drugs goes beyond inhibition of chemokine/TNF-α secretion because also intracellular accumulation of these molecules was markedly inhibited. To establish whether this is due to inhibition of Lyn or, in the case of dasatinib, of Abl or to off-target effects of these drugs is an issue worthy of further work. Independently of its mechanism of action, the capability of dasatinib to inhibit chemokine and TNF-α secretion, coupled with its inhibitory role on several PMN responses (55), make it a good candidate to reduce tissue damage in chronic inflammatory and autoimmune diseases.
In the context of lung inflammation, different populations of resident Mos, or early recruited monocytes, as well as lung epithelial cells can release chemokines and cytokines that augment PMN recruitment. Very recent studies identified in the skin a population of perivascular macrophages that seems to play a crucial role in secretion of the PMN-attracting chemokines CXCL1, CXCL2, and CCL3, and the consequent PMN recruitment in response to S. aureus infection (56). Although it is not known whether perivascular macrophages are implicated in PMN accumulation into the lung, this finding certainly adds further complexity to the repertoire of cells implicated in PMN recruitment. We report that monocyte recruitment into the lung parenchyma and the airways in response to LPS is markedly defective in hck−/−fgr−/− mice (Figs. 1, 3). Considering that Src kinases do indeed regulate macrophage intrinsic migratory ability (reviewed in Ref. 38; see also Refs. 57–60), the inability of monocytes to migrate into the lung interstitium of hck−/−fgr−/− mice may represent another mechanism responsible for a decreased load of PMN-attracting chemokines. Indeed, we detected calgranulin-positive, monocyte-like cells in the airways already after 2 h following the LPS challenge, that is, at a time point at which a very few number of PMNs had migrated into the lung (Figs. 3, 5). Hence we cannot exclude that, because of an intrinsic defect in their migratory ability, hck−/−fgr−/− monocytes cannot contribute to the establishment of a monocyte-dependent neutrophil extravasation circuit (47). To start to address whether Src kinases regulate chemokine secretion also in Mos, we examined chemokine and cytokine secretion by BMDMs (Fig. 9, Supplemental Fig. 1). Similar to PMNs (Figs. 6, 8), BMDMs from hck−/−fgr−/− mice secreted lower amounts of CXCL1, CXCL2, CCL3, and CCL4, but also of CCL2 and CCL5, in response to LPS (Fig. 9). Secretion of CXCL1, CXCL2, and CCL2 by hck−/−fgr−/− BMDMs was also reduced in response to the TLR2 ligand PAM3 and the TLR7/8 ligand R848 (Fig. 9). The finding that Src kinase–deficient macrophages secrete lower amounts of chemokines in response to LPS is in agreement with both early and more recent reports implicating Src-family kinase in LPS-initiated signal transduction in Mos (31, 61–66); however, it contrasts with the evidence that hck−/−fgr−/−lyn−/− macrophages have no major defect in LPS/IFN-γ–induced cytokine secretion (25). However, interpretation of results with triple-mutant hck−/−fgr−/−lyn−/− macrophages is complicated by the fact that Lyn kinase has negative regulatory function in TLR signaling (67), so triple-mutant mice lose both activating and inhibitory signaling at the same time. This nuance was not appreciated when the first studies of triple-mutant cells were reported. Notably, we did not find any difference in LPS-, PAM3-, and R848-induced secretion by WT or hck−/−fgr−/− BMDMs of a number of cytokines including IL-1α, IL-1β, IL-6, IL-10, and GM-CSF, the only exception being, in analogy with mutant PMNs, TNF-α, whose secretion by hck-/-fgr−/− BMDMs was significantly reduced (Supplemental Fig. 1). The Src kinase dependence of chemokine secretion by BMDMs was confirmed by studies with dasatinib (Supplemental Fig. 2) that markedly inhibited chemokine secretion in response to different doses of LPS. Hence macrophages also release a defined spectrum of chemokines in a manner dependent on Src kinase expression or activity.
In conclusion, we report that Src kinases regulate myeloid cell recruitment into the lung likely through their capability to regulate chemokine secretion. Coupled with previous findings establishing the notion that PMN recruitment requires chemoattractant release by early recruited PMNs (23, 48) and that both Src kinases (23) and Syk (68) regulate PMN accumulation into the joints and the skin, our report provides an important piece of information to highlight the Src/Syk signaling pathway as a target to control inflammatory diseases.
Footnotes
This work was supported by the European Community’s Seventh Framework Program (FP7 2007-2013) under Grant 282095-TARKINAID (to G.B.) and National Institutes of Health Grants RO1 AI068150 and RO1 AI065495 (to C.A.L.).
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.