Resolution of Toll-like receptor 4-mediated acute lung injury is linked to eicosanoids and suppressor of cytokine signaling 3

Animals

All studies were reviewed and approved by the Harvard Medical Area standing committee on animals. Male 8- to 12-week-old C3H/HeJ (HeJ) and wild-type matched C3H/HeOuJ (OuJ) mice (Jackson Laboratories, Bar Harbor, ME, USA) were maintained in a barrier facility under specific pathogen-free conditions. HeJ mice exhibit resistance to lipopolysaccharide (LPS) due to a spontaneous mutation at the LPS response locus in the intracellular domain of TLR4 (Tlr4^LPS-d), which is not shared by control OuJ mice that express the wild-type TLR4 receptor (18). Given the presence of additional genetic differences between HeJ and OuJ mice, select experiments were performed with congenic C.C3-Tlr4^Lps-d/J mice (C.C3) that have the C3H/HeJ point mutation in Tlr4 in the BALB/cByJ (BALB) background (Jackson Laboratories; ref. 19). In view of the TLR4 defect in HeJ and C.C3 mice, the microbiome was determined in stool samples diluted in PBS (n=3) from all strains using CHROMagar Orientation plates (Becton Dickinson, Franklin Lakes, NJ, USA). Enterococci predominated, with minor variation in less abundant bacteria (Supplemental Fig. S1). In addition, there was no detectable endotoxin in samples of bronchoalveolar lavage fluid (BALF) or serum as measured by the Limulus amebocyte lysate test (Lonza, Walkersville, MD, USA; limit of detection 0.25 endotoxin units/ml).

Acid-initiated ALI

Mice were anesthetized with i.p. injections of ketamine (80 mg/kg body wt; Phoenix Scientific, St. Joseph, MO, USA) and xylazine (10 mg/kg body wt; Phoenix Scientific). Hydrochloric acid (0.1 N HCl, pH 1.5, 50 μl, endotoxin free; Sigma-Aldrich, St. Louis, MO, USA) was instilled selectively into the left lung, followed by a bolus of air (150 μl), as described previously (20). At timed intervals (2, 12, 24, 48 h), anesthetized mice (pentobarbital sodium, 70 mg/kg body wt i.p.; Abbott Laboratories, North Chicago, IL, USA) were mechanically ventilated with a flexiVent small animal ventilator (Scireq, Montreal, QC, Canada) to determine airway mechanics. Whole-lung BALF was obtained at the same time points (1 ml PBS with 0.6 mM EDTA×2). In some animals, lung tissue was snap-frozen for RNA or protein analysis. Total cells in BALFs were counted using a hemocytometer, and differential cell counts were determined after cytospin using Wright-Giemsa staining. Some animals received recombinant murine interleukin-6 (rmIL-6; 2 μg; Biolegend, San Diego, CA, USA) in 1% BSA or vehicle intranasally 2 h after initiation of ALI. Cell-free BALFs were portioned into aliquots and kept at −80°C.

Endothelial permeability

Evans Blue dye was utilized as a marker for endothelial barrier function (21). At 30 min prior to termination of the experiments, Evans Blue dye (40 mg/kg) was injected via tail vein, and extravasation of dye into BALFs was quantified by spectrophotometry (absorbance at 650 nm). In some experiments, HeJ mice were given montelukast (10 μg) or vehicle (0.5% ethanol) in 0.1 ml sterile 0.9% saline (i.v.) 30 min before initiation of ALI. Montelukast sodium was isolated from Singulair tablets by extraction and chromatography at Berlex Biosciences (Richmond, CA, USA). Chemical and NMR analytical data were consistent with the desired structure (as in ref. 22).

Measurement of mediators in BALFs

The amounts of lipoxin A₄ (LXA₄; Neogen, Lexington, KY, USA), cysteinyl leukotrienes (CysLTs) and 8-isoprostane (Cayman Chemicals, Ann Arbor, MI, USA) were measured in BALFs using specific ELISA kits, according to the manufacturer's instructions. Select cytokines and chemokines were measured in BALFs by protein bead array (Aushon Biosystems, Billerica, MA, USA).

Quantitative PCR

Total RNA was extracted from snap-frozen left lung using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA). Total RNA (2 μg) was reverse transcribed using TaqMan qPCR gene expression assays (Applied Biosystems, Foster City, CA, USA), according to the manufacturer's instructions. TaqMan probes specific for murine SOCS1 and SOCS3 were used as target genes and murine cyclophilin A served as the reference gene. Cycle threshold (C_T) values were determined in triplicate using an Applied Biosystems 7300 real-time PCR system, and relative expression was analyzed using the comparative 2^−ΔΔCT method.

Western blot analysis

Analysis of SOCS1, SOCS3, STAT3, and phosphoSTAT3 protein was performed by Western blot analysis. β-Actin was used as a control. Antibodies were used in the manufacturer's recommended concentrations (1 μg/ml; Abcam, Cambridge, MA, USA and Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Lung histopathology

After dissection, lungs were perfusion fixed at 20 cmH₂O in IHC zinc fixative buffer (BD Pharmingen, San Diego, CA, USA). Tissue blocks were obtained from midsagittal slices of lungs embedded in paraffin. After deparaffinization, tissue sections were incubated for 1 h at room temperature with LY-6G (1:100 dilution, BD Pharmingen), or SOCS3 (1:200 dilution, Abcam), followed by horseradish peroxidase-conjugated secondary antibody. H&E stains were also obtained.

Macrophage transfection

The murine alveolar cell line AMJ2-C8 was obtained from American Type Culture Collection (ATCC; Manassas, VA, USA). Macrophages were plated 24 h (37°C, 5% CO₂) at 1 × 10⁶ cells/ml in culture plates containing serum-free modified Eagle's medium (MEM) supplemented with 10% FBS (ATCC) and penicillin-streptomycin. Chariot (Active Motif, Carlsbad, CA, USA), a noncytotoxic protein transfection reagent (23–25), was used to transfer SOCS3 antibody and rabbit IgG into the cells, according to manufacturer's directions. This approach was chosen because of the potential for posttranscriptional regulation of SOCS3. Macrophages were transfected with anti-SOCS3 or polyclonal IgG antibody (1 μg/10⁵ cells) for 2 h. After transfection, cells were exposed to 1 μg/ml LPS (Escherichia coli 055:B5; Sigma-Aldrich) or vehicle (H₂O) for 4–24 h (37°C, 5% CO₂).

Statistical analysis

Numerical results are expressed as means ± se. ANOVA was used to determine significance for differences between >2 groups. For analyses between 2 groups, cohorts were compared by Student's t test. Values of P < 0.05 were considered significant. Statistics were performed using GraphPad Prism 5 for Windows (GraphPad, San Diego, CA, USA).

TLR4 influences the time course for acid-initiated ALI and its resolution

To determine the time course for host responses to ALI, a self-limited experimental model of gastric acid aspiration was used. Sterile hydrochloric acid (0.1 N HCl, pH 1.5, 50 μl) was instilled selectively into the left main stem bronchus. This nonlethal model of ALI did not require mechanical ventilation for survival and provided an experimental window into cellular and molecular mechanisms of resolution. After acid instillation, lung elastance (E) and dynamic resistance (R) were measured (flexiVent) at baseline and 2, 12, 24, and 48 h after initiation of ALI (Fig. 1A, B). E and R values significantly increased and peaked 24 h (Φ_max) after ALI in wild-type OuJ mice, returning to near-baseline levels at 48 h. HeJ mice have a TLR4-receptor defect and were protected after ALI with significant relative decrements in both peak E and R at 24 h. The time to maximal response Φ_max was 12 h for both E and R in HeJ mice, and the resolution interval (R_i, defined as the time from Φ_max to half-maximal response) for these parameters was ∼6 h. These measures of organ injury, Φ_max, and resolution, R_i, in the TLR4-receptor defective HeJ mice were decreased by ∼50% compared to wild-type OuJ mice, indicating that TLR4 signaling both increases the extent of organ injury and delays its resolution.

Given these transient changes in lung mechanics, the effect of acid injury on airway barrier integrity was next determined. Leakage permeability changes were measured by extravasation of intravenous Evans Blue dye into BALFs. Wild-type OuJ mice had a significant increase in BALF Evans Blue dye 24 h after ALI (Fig. 1C). In contrast, HeJ mice did not display significant changes in endothelial permeability in response to lung injury at this time point, and the amounts of BALF Evans Blue dye were significantly less than in the OuJ mice (Fig. 1C), suggesting a more rapid restitution of barrier integrity in the HeJ mice.

In addition to TLR4, HeJ and OuJ mice display other genetic differences that are not well characterized. To provide further evidence for TLR4 as a critical locus for regulation of host responses to sterile ALI, select experiments were performed with congenic C.C3-Tlr4^Lps-d/J and control BALB/cByJ mice. Similar to the HeJ mice, the C.C3 mice were protected from acid-initiated ALI (at 24 h), with E and R significantly increased in only wild-type BALB mice (Fig. 2A, B). The effect of acid injury on endothelial permeability was also determined in both strains. While both wild-type BALB and TLR4-defective C.C3 mice showed a significant increase in BALF Evans Blue dye 24 h after ALI, the changes in permeability in the C.C3 strain were significantly less than in wild-type BALB mice (Fig. 2C).

CysLTs lead to endothelial leak after ALI

CysLTs are potent vasoactive mediators (26), so their levels were determined during this self-limited model of ALI. Compared to HeJ mice, BALF levels of CysLTs were significantly increased in OuJ mice 24 h after acid instillation (OuJ vs. HeJ: 1008±321.6 vs. 206.8±253.7 pg/ml; P<0.05, n=4; Fig. 3A). To examine the connection between CysLTs and pulmonary endothelial permeability, the CysLT1 receptor antagonist montelukast was administered intravenously (10 μg in 100 μl) to OuJ mice 30 min prior to initiation of lung injury. There was significantly less extravasation of Evans Blue dye into BALFs in OuJ mice treated with montelukast relative to control animals (vehicle, 0.5% ethanol in 0.9% NaCl, 100 μl; Fig. 3B). Because LXA₄ is an endogenous counterregulatory eicosanoid that blocks cysLT-mediated actions in the lung (22) and vasculature (27), BALF LXA₄ levels were determined after ALI. Similar to CysLTs, LXA₄ was significantly increased in OuJ mice compared to HeJ mice 24 h after ALI (OuJ vs. HeJ: 227.6±131.2 vs. 62±78.3 pg/ml; P<0.05, n=4; Fig. 3C). Of interest, the relative conversion of arachidonic acid to CysLTs rather than LXA₄ was increased by ∼30% in the OuJ mice (mean CysLT/LXA₄ ratio was 4.4 in OuJ vs. 3.3 in HeJ), favoring increased vascular permeability changes with intact TLR4 signaling.

TLR4 influences the time course for acid-initiated lung inflammation and its resolution

In contrast to the decreased perturbations in lung mechanics and permeability in the HeJ mice, acid-initiated ALI in these animals was not similarly associated with decreased inflammation (Fig. 4). After ALI, leukocyte counts transiently increased with Φ_max for BALF PMNs and macrophages at 24 h (Fig. 4A, B), which was concomitant with peak increases in E and R (Fig. 1A, B). Of interest, the modest increases in BALF PMNs and macrophages in HeJ mice 24 h after ALI were ∼2-fold greater than in wild-type OuJ mice 24 h after ALI (Fig. 4A, B). While Φ_max for BALF PMNs and macrophages was increased in the HeJ mice, R_i was not significantly changed for these parameters of lung inflammation, indicating that TLR4 signaling in ALI leads to both proinflammatory and anti-inflammatory responses without significantly influencing the pace of leukocyte resolution.

Lung histopathology 24 h after ALI revealed pulmonary edema as the predominant consequence of acid injury, including thickening of the alveolar septa and interstitium that was concomitant with peak changes in airway mechanics (Fig. 4C). The alveolar and interstitial changes were less marked in HeJ mice (Fig. 4D). Relative to uninjured lung (data not shown), Ly-6G immunostaining identified an increase in leukocyte trafficking in OuJ mice 24 h after ALI, and, consistent with the BALF analyses (Fig. 4), HeJ mice had modestly increased lung PMNs at this time point (Fig. 4E, F).

TLR4 signaling in ALI regulates cytokine and SOCS3 expression

BALF levels of interleukin-6 (IL-6) were significantly increased in OuJ mice at 2 and 12 h after ALI compared to baseline and HeJ mice (Supplemental Fig. S2A). To determine whether this cytokine would increase lung injury (i.e., lung elastance), recombinant murine IL-6 (rmIL-6, 2 μg, i.n.) was administered to HeJ mice 2 h after ALI. Animals receiving rmIL-6 displayed significantly increased airway PMNs 24 h after ALI, but no significant changes in lung elastance were elicited (Supplemental Fig. S2B, C). BALF levels of additional mediators linked to ALI were also determined by protein bead array, and strain-specific differences were present for keratinocyte chemoattractant (KC) and monocyte chemotactic protein-1 (MCP-1) (Supplemental Fig. S3). Because the levels of KC were discordant with BALF PMN numbers, the kinetics for KC production in this model were not further examined. BALF LTB₄ was also detected 24 h after ALI, but only in low picogram amounts in both cohorts (OuJ vs. HeJ: 5.7±2.7 vs. 1.8±0.2 pg/ml; P<0.05, n=4).

Without significant differences in BALF cytokine levels, the SOCS family proteins, which regulate cytokine signaling, were next examined. Acid-initiated ALI induced a marked increase in lung SOCS3 expression in both mouse strains as early as 2 h after injury (Fig. 5A). Of interest, SOCS3 expression in HeJ mice remained markedly increased 12 h after ALI, while SOCS3 expression in the OuJ mice decreased substantially at this time point (ΔC_T, OuJ vs. HeJ: 6.6±0.26 vs. 11.03±1.96; P<0.05). In addition, 24 h after acid-initiated ALI, SOCS3 expression was still 4-fold increased over baseline in HeJ mice. For purposes of comparison, expression of the related gene SOCS1 was also determined and found to be increased early (2 h) after ALI and decreased with time. Relative to SOCS3, these changes in SOCS1 expression were of much smaller amplitude, and no significant differences between genotypes were present (Fig. 5B). Western blot analyses confirmed the increased SOCS3 expression in HeJ mice 12 h after ALI (Fig. 5C). SOCS1, as well as STAT3 protein, a pivotal intracellular signaling checkpoint for SOCS proteins (reviewed in ref. 28), was not significantly increased in HeJ mouse lung compared to OuJ mice (Fig. 5C). Immunohistochemistry 24 h after ALI revealed increased lung SOCS3 expression in HeJ mice in the cytoplasm of both airway lining epithelial cells and alveolar macrophages (Fig. 5D, E).

SOCS3 regulates alveolar macrophage responses to soluble stimuli

Because of the reported relationship between TLR4 signaling in ALI and oxidative stress (8) and now in view of the change in SOCS3 expression in alveolar macrophages (Fig. 5), we investigated a potential link between TLR4, SOCS3, and oxidative stress by introducing an anti-SOCS3 antibody into murine alveolar macrophages (AMJ2-C8 cell line) to regulate SOCS3 function in these cells. First, to determine the effect of the introduction of this anti-SOCS3 antibody on cell signaling, cells were exposed to the TLR4-agonist LPS (E. coli 055:B5, 1 μg/ml) for 4 h (37°C, 5% CO₂) and phosphorylation of STAT3 on tyrosine 705 (pYSTAT3) was quantitated by densitometry on immunoblot. Cell activation by LPS increased the ratio of pYSTAT3 to STAT3, and introduction of anti-SOCS3 led to further increases in pYSTAT3 that were significantly greater than the changes in cells receiving control IgG (0.35 ± 0.08 vs. 0.14 ± 0.05 densitometric units, respectively; P<0.03, n=6), indicating that this anti-SOCS3 antibody displayed blocking properties for SOCS3 signaling. Next, in the presence of anti-SOCS3 or control antibody, the AMJ2-C8 cells were exposed to LPS (E. coli 055:B5, 1 μg/ml) for 24 h (37°C, 5% CO₂), and levels of 8-isoprostane, a sensitive marker of oxidative stress (29), were determined in cell supernatants. After transfection of anti-SOCS3 antibody, LPS initiated a significant increase in 8-isoprostane relative to cells transfected with polyclonal IgG or unstimulated control cells (Fig. 6A). In addition, 8-isoprostane levels were significantly increased in BALFs 24 h after ALI in the OuJ mice compared to baseline and relative to HeJ mice (Fig. 6B).