Therapeutic blockade of CD54 attenuates pulmonary barrier damage in T cell-induced acute lung injury

Abstract

Acute respiratory distress syndrome (ARDS) is a serious, often fatal condition without available pharmacotherapy. Although the role of innate cells in ARDS has been studied extensively, emerging evidence suggests that T cells may be involved in disease etiology. Staphylococcus aureus enterotoxins are potent T-cell mitogens capable of triggering life-threatening shock. We demonstrate that 2 days after inhalation of S. aureus enterotoxin A, mice developed T cell-mediated increases in vascular permeability, as well as expression of injury markers and caspases in the lung. Pulmonary endothelial cells underwent sequential phenotypic changes marked by rapid activation coinciding with inflammatory events secondary to T-cell priming, followed by reductions in endothelial cell number juxtaposing simultaneous T-cell expansion and cytotoxic differentiation. Although initial T-cell activation influenced the extent of lung injury, CD54 (ICAM-1) blocking antibody administered well after enterotoxin exposure substantially attenuated pulmonary barrier damage. Thus CD54-targeted therapy may be a promising candidate for further exploration into its potential utility in treating ARDS patients.

despite decades of research, acute lung injury (ALI)/acute respiratory distress syndrome (ARDS) remains an underdiagnosed and undertreated life-threatening condition and accounts for more than 10% of all intensive care unit admissions (9). ALI/ARDS is a syndrome of acute lung inflammation that presents with bilateral lung infiltrates, pulmonary edema, and hypoxemia (43). The mechanism of ALI/ARDS involves a pulmonary or extrapulmonary insult such as pneumonia, aspiration, sepsis, or major surgery, leading to a recruitment of leukocytes and platelets, release of proinflammatory factors, and injury to the endothelial and epithelial layers. Disruption of the pulmonary endothelial barrier ultimately precipitates the characteristic pathophysiological changes of increased vascular permeability, accumulation of protein-rich fluid, and impaired gas exchange (42, 43).

The two most frequent underlying causes are pneumonia and sepsis, with most patients developing ALI/ARDS secondary to an established bacterial, viral, or fungal infection (43). Both Gram-positive and Gram-negative bacteria can be involved (7, 72), but previous studies have preferentially focused on Gram-negative bacteria and, more specifically, the effects of their bacterial-derived LPS (45). Importantly, however, there are many cases of ALI/ARDS that are likely associated with Gram-positive bacteria, and Staphylococcus aureus, in particular, which is a major contributor to morbidity and mortality, resulting from serious infections (31, 32). One of the critical virulence factors of S. aureus capable of inducing massive inflammation is enterotoxins (20, 60). These superantigens bypass classical antigen presentation processes and, instead, induce oligoclonal expansion of T cells by bridging MHC II with a specific T-cell receptor Vβ chain (20). Superantigens are known for their extreme potency; unlike conventional antigens activating 1 out of 10⁴–10⁶ T cells, superantigens can activate up to 1 out of 4 T cells (26). The resultant T cell-induced inflammatory response and cytokine storm (most notably, IL-2, IFNγ, and TNF) can have disastrous consequences, leading to toxic shock, tissue damage, organ dysfunction, and even death (20, 73). Most S. aureus strains produce superantigen toxins, and recent evidence suggests that they may be involved in a number of serious illnesses, including pneumonia, sepsis, and endocarditis (8, 73). S. aureus enterotoxin A (SEA) has been found in patients with sepsis, and its prevalence correlated with infection severity (6, 19). In animal studies, organ damage and lethality caused by S. Aureus- induced bacteremia or necrotizing pneumonia were shown to be superantigen dependent (69, 74, 83). Furthermore, it was demonstrated that CD4⁺ T-cell activation significantly exacerbated murine lung pathology and impaired bacterial clearance in pneumonia caused by an enterotoxin-producing S. aureus strain (56). Thus S. aureus enterotoxins likely play a crucial role in the severity of sepsis, pneumonia, and the associated ALI/ARDS.

Previous studies showed that administration of S. aureus enterotoxin in animal models resulted in acute pulmonary inflammation (17, 58, 62, 63), and this response appeared to be mediated by T cells (27, 34, 54). In particular, inhalation of S. aureus enterotoxin first induced a systemic inflammatory response characterized by rapid T-cell activation, cytokine and chemokine release, and a T cell-orchestrated recruitment of innate immune cells into the circulation, lymphoid tissues, and lung (34, 63, 76, 77). This early response occurring within several hours of S. aureus enterotoxin exposure was followed by development of considerable lung pathology at 48 h after inhalation, which was marked by a massive T-cell expansion in lymphoid tissues and lung (54, 63). Importantly, no lung pathology was found in the absence of T cells, in particular, CD8⁺ T cells (54). The pulmonary response presented with perivascular and peribronchial inflammation, disruption of terminal vessels, and accumulation of proteins, red blood cells, and leukocytes in the airways (50, 54, 63, 68). These pathological features strongly resemble the histological findings in ALI/ARDS patients (42, 43), suggesting that S. aureus enterotoxin-activated T cells may be capable of inducing ALI/ARDS. Although T cells were previously found to orchestrate both early inflammatory responses and the subsequent lung inflammation (34, 54, 76), the mechanism driving the development of vascular permeability is not fully understood.

The goal of this work was to define how SEA inhalation alters the pulmonary barrier over time and to establish the main molecular players involved in the development of lung injury, to identify clinically translatable therapeutic targets. We show that S. aureus enterotoxin inhalation caused increased vascular permeability, elevated expression of endothelial and epithelial injury markers, increased caspase expression in lung, and a temporal differential cytokine/chemokine profile distinguishing intrapulmonary and systemic responses. Mechanistically, S. aureus enterotoxin triggered rapid activation of pulmonary endothelial cells in the early phase of inflammation, which was followed by significant reductions in endothelial cell number during the late phase of inflammation, marked by massive T-cell expansion and cytotoxic differentiation. The early inflammatory responses due to S. aureus enterotoxin-induced T-cell activation, in part, determined the extent of pulmonary barrier damage. However, remarkably, lung injury could be mitigated by therapeutic blockade of CD54, also known as intercellular adhesion molecule-1 (ICAM-1), when administered 36 h after initiation of inflammation. Thus our data show that ALI/ARDS may be treatable with immunotherapeutic agents given hours, or perhaps even days, beyond the onset of inflammatory response.

METHODS

Mice.

Male and female C57BL/6J and TCR βδ^−/− mice were obtained from Jackson Laboratory (Bar Harbor, ME) and used between 1.5 and 4 mo of age. All mice were kept in the Central Animal Facility at UConn Health in accordance with federal guidelines. All experimental procedures were approved by the Institutional Animal Care and Use Committee of UConn Health.

Immunization and therapy.

A highly purified form of SEA was purchased from Toxin Technology (Sarasota, FL). SEA (1 μg) and vehicle control were administered intranasally, as described previously (34). For neutralization therapy experiments that blocked early responses, mice received 250 μg cytotoxic T lymphocyte-associated protein 4-Ig (CTLA4-Ig; Sigma-Aldrich, St. Louis, MO), 500 μg anti-TNF (clone XT3.11), or 500 μg of rat IgG (both from BioXCell, Lebanon, NH) intraperitoneally 2 h before SEA inhalation. For neutralization therapy experiments blocking late T-cell responses, mice received 250 μg anti-CD178 (clone 101626; R&D Systems, Minneapolis, MN), 500 μg anti-CD54 (clone YN1/1.7.4; BioXCell), or 500 μg rat IgG intraperitoneally 36 h after SEA inhalation.

In vivo permeability assays.

The in vivo permeability assays were performed using FITC-dextran, similarly as described before (13, 22). To assess lung permeability from lung to blood, mice received 50 μl of FITC-dextran (3,000 mW; 5 mg/ml; Life Technologies, Grand Island, NY) intranasally, and serum from tail blood was obtained 1 h later. For blood into lung permeability assay, mice received 100 μl of FITC-dextran intravenously through the retro-orbital route, and bronchoalveolar lavage (BAL) fluid was collected 1 h later. FITC fluorescence in the serum (1:70) or BAL fluid (undiluted) was determined by a plate reader (483-14/530-30; CLARIOstar, BMG LABTECH) and was normalized either to the vehicle only-treated mice (Fig. 1) or the IgG-treated mice (Fig. 7).

Fig. 1.

SEA increases pulmonary permeability, cellular recruitment, and expression of epithelial and endothelial markers of injury 2 days after inhalation. A: permeability in the lung was measured by FITC-dextran fluorescence assay. Left: FITC-dextran was administered intranasally, and serum was obtained 1 h later (24 and 48 h after Staphylococcus aureus enterotoxin (SEA) or vehicle inhalation). FITC fluorescence was measured in serum and normalized to the vehicle control. At least three independent experiments, with n = 8–11 per group. Right: FITC-dextran was administered intravenously, and BAL fluid was obtained 1 h later (48 h after SEA or vehicle). FITC fluorescence was measured in bronchoalveolar lavage (BAL) fluid and normalized to vehicle control. Two independent experiments with n = 6–10 per group. B: concentration of albumin in BAL fluid 24 and 48 h after SEA or vehicle inhalation. Three independent experiments with n = 5 per group. C: representative confocal microscopy images of lung tissue showing clusters of recruited myeloid cells (CD11b⁺ 7/4⁺) and SEA-specific T cells (Vβ3⁺) in proximity to macrophages (CD169⁺) 48 h after SEA inhalation. B, bronchiole; V, blood vessel. Scale bar = 80 μm. Images are representative of three independent experiments with n = 4 for vehicle and n = 6 for SEA. D: concentration of the lung injury markers CD54, angiopoietin-2 (Ang-2), and RAGE in BAL fluid 24 and 48 h after SEA or vehicle inhalation. Three experiments with n = 5–8 per group. Data are shown as means ± SE. Two-way ANOVA with Sidak’s test or unpaired t-test (for Fig. 1A, right); **P < 0.01, ***P < 0.001.

Fig. 7.

Delayed CD54 blockade substantially minimizes lung permeability, cell injury, and inflammation after SEA inhalation. A: diagram of experimental setup. Mice were administered SEA intranasally, and then they were treated with IgG control, anti-CD178, or anti-CD54 intraperitoneally 36 h later. BAL fluid and tissue were obtained 48 h after SEA inhalation. B: total number of BAL fluid cells from mice treated with IgG control, anti-CD178, or anti-CD54, as described in “A”. C: composition of cells within the BAL fluid following IgG control, anti-CD178, or anti-CD54 treatment. Average percentages of CD45⁺CD3⁺Vβ3⁻ T cells, CD45⁺CD3⁺Vβ3⁺ T cells, neutrophils (CD45⁺CD11b⁺Ly6C⁺Ly6G⁺), and monocytes (CD45⁺CD11b⁺Ly6C⁺Ly6G⁻) in BAL fluid are represented by stacked bars. D: concentrations of injury markers Ang-2, RAGE, and CD54, inflammatory markers Gzm B, IFNγ, IL-6, and TNF, and albumin in BAL fluid from mice treated with IgG control, anti-CD178, or anti-CD54. E: total number of Vβ3⁺ T cells, monocytes, and neutrophils in lung in mice treated with IgG control, anti-CD178, or anti-CD54. Three independent experiments with n = 9 per group. F: lung permeability measured by FITC-dextran permeability assay. Mice were given SEA or vehicle intranasally, and the SEA-exposed mice were treated with IgG control or anti-CD54 36 h later. FITC-dextran was administered intravenously 1 h before euthanasia. FITC fluorescence was measured in BAL fluid and normalized to IgG-treated mice. Data were combined from three independent experiments with n = 7–11. G: expression (median fluorescent intensity, MFI) of Gzm B and IFNγ in CD8⁺Vβ3⁺ T cells (CD45⁺CD3⁺CD8⁺Vβ3⁺). The lung cells were cultured with brefeldin A only (Gzm B) or brefeldin A + PMA/ionomycin (IFNγ) for 5 h and stained with antibodies. Two independent experiments with total n = 6–8 per group. Data are shown as means ± SE. One-way ANOVA with Dunnett’s test (for Fig. 7, B–E) or one-way ANOVA with Tukey’s test (for Fig. 7, F and G); *P < 0.05, **P < 0.01, ***P < 0.001.

Pan-caspase imaging.

The amount of apoptosis in lung and spleen was detected by CAS-MAP near-infrared in vivo fluorescent-imaging probe (Vergent Bioscience, Minneapolis, MN), according to the manufacturer’s instructions. Briefly, the CAS-MAP probe was intravenously injected 60–90 min before euthanasia. Lung and spleen were then imaged by Odyssey CLx Imager (LI-COR, Lincoln, NE).

Tissue preparation, BAL fluid collection, flow cytometry, and cell sorting.

BAL fluid was collected into 2 ml of PBS by lung lavage in situ and then centrifuged. The lung tissue was perfused with PBS by inserting a butterfly needle through the heart. Alternatively, in experiments in which BAL fluid was collected, lung tissue was harvested without perfusion. Lung tissue was digested with 150 U/ml collagenase (Sigma-Aldrich), 60 U/ml DNase I (Sigma-Aldrich), 2% FBS, 1 mM Ca₂Cl, and 1 mM Mg₂Cl in balanced salt solution for 30 min at 37°C. RBCs in lung were lysed with NH₄Cl solution, and cells were counted using a Z1 particle counter (Beckman Coulter, Brea, CA). For flow cytometry, nonspecific binding was blocked with Fc receptor blocking solution (76) and was stained with antibodies and LIVE/DEAD Fixable Blue Dead Cell Stain (Life Technologies) for 30–45 min on ice. For intracellular staining of granzyme B and IFNγ, cells were cultured with brefeldin A (1 μg/ml; Calbiochem, San Diego, CA) and with or without PMA (50 ng/ml, Calbiochem)/ionomycin (1 μg/ml, Fisher Scientific, Waltham, MA) for 5 h. The cells were stained with surface antibodies, fixed in 2% formaldehyde, permeabilized with 0.25% saponin, and stained with intracellular antibodies overnight at 4°C. Flow cytometry was performed using a FACS-LSRII (BD Biosciences, San Jose, CA), and data were analyzed using FlowJo software (Tree Star, Ashland, OR). For cell sorting, lung tissue was prepared, as described above, and live endothelial cells (CD45⁻CD31⁺CD54⁺) were sorted with FACS Aria (BD Biosciences). Fluorochrome-conjugated antibodies were obtained from BD Biosciences, eBioscience (San Diego, CA) or BioLegend (San Diego, CA). The following antibodies were used: anti-CD45.2 (shown as CD45; 104), anti-CD31 (390), anti-CD54 (YN1/1.7.4), anti-Vβ3 (KJ25), anti-CD3 (17A2), anti-granzyme B (NGZB), anti-CD95 (Jo2), anti-CD11a (H155–78), anti-IFNγ (XMG1.2), anti-CD8 (53–6.7), anti-CD11b (M1/70), anti-Ly6C (HK1.4), and anti-Ly6G (18A).

Cytometry by time-of-flight.

Lungs were obtained from mice 14 or 40 h after SEA or vehicle inhalation, digested with collagenase/DNase solution and purified using Lympholyte M (Cedarlane, Burlington, NC). Each one of the four samples was labeled with Cell-ID cisplatin to identify live and dead cells and subsequently barcoded with Cell-ID Pd barcoding kit [all cytometry by time-of-flight (CyTOF) reagents are from Fluidigm, San Francisco, CA]. The four samples were pooled together, permeabilized with methanol, and incubated with Fc receptor blocking solution and 24 heavy metal-conjugated antibodies, including seven signaling markers. DNA was labeled using Cell-ID Intercalator-Ir. The pooled sample was spiked with normalization beads and analyzed by a mass cytometer (Helios, Fluidigm). The data were debarcoded using the Fluidigm Debarcoder v1.04. A single-cell and live-cell gate were obtained from each sample. The samples were further subgated to include the same number of cells per sample and merged to create ViSNE maps (MatLab, MathWorks, Natick, MA) using only the nonsignaling markers. Histograms comparing the expression of signaling and activation markers were generated using MatLab. Median intensity values were obtained by FlowJo software.

ELISA and multiplex assays.

BAL fluid and serum (from tail blood) were collected and cytokine and chemokine multiplex assays (Bio-Rad, Hercules, CA) and ELISAs for albumin (Bethyl, Montgomery, TX), CD54, angiopoietin-2 (Ang 2), receptor for advanced glycation end products (RAGE), granzyme B (R&D Systems), and cytokine ELISAs for IL-6, TNF, and IFN-γ (BD Biosciences) were performed according to the manufacturer’s instructions.

Quantitative real-time PCR.

For quantitative real-time PCR analysis of sorted endothelial cells (CD45⁻CD31⁺CD54⁺), the total RNA was isolated with RNeasy mini kit (Qiagen, Valencia, CA), and cDNA was obtained using iScript cDNA synthesis kit (Bio-Rad). The relative gene expression levels of target genes and the endogenous control gene Gapdh were determined by real-time PCR using TaqMan primers (Thermo Fisher Scientific) and a CFX96 real-time PCR instrument (Bio-Rad). The gene expression was normalized to Gapdh using a standard curve method.

Confocal microscopy.

Frozen tissue sections were prepared as described before using the periodate-lysine-paraformaldehyde (PLP) fixation method (76). The antibody-stained sections were examined with Zeiss LSM780 confocal microscope mounted on an inverted Axio Observer Z1 (Zeiss, Oberkochen, Germany). The images were acquired as confocal z-stack projections using ×0.8 numerical aperture objectives. The following antibodies were used: PE-conjugated anti-Vβ3 clone KJ25, V450-conjugated anti-CD11b clone M1/70, allophycocyanin-conjugated anti-CD11c clone HL3, PE-conjugated anti-active caspase-3 clone C92-605 (BD Biosciences), eFlour 660-conjugated anti-CD169 clone SER-4 (eBioscience), and DyLight 488-conjugated anti-Ly6B.2 clone 7/4 (labeled as 7/4; Novus Biologicals, Littleton, CO). Image analysis, including adjustments for contrast and brightness and cellular quantification, was performed with Imaris software (Bitplane, Zurich, Switzerland). The number of active caspase-3⁺ cells in the SEA-exposed mice was determined in loci of increased inflammation (areas with an increased number of inflammatory cells). No such areas were found in the vehicle control.

Statistical analysis.

The following tests were used to determine statistical significance (P < 0.05): one-way ANOVA (with Dunnett’s or Tukey’s multiple-comparisons tests), two-way ANOVA (with Sidak’s multiple-comparisons test), two-tailed Student’s unpaired t-test, or paired t-test. All statistical tests were performed using Prism-GraphPad (La Jolla, CA).

RESULTS

S. aureus enterotoxin inhalation induces lung injury marked by cell death.

We have previously shown that upon local inhalation, SEA was detectable in serum within minutes and triggered systemic activation of T cells. This systemic inflammatory response was marked by cytokine and chemokine release, as well as innate cell migration into the circulation and lymphatic tissue (76). Furthermore, activation of SEA-specific αβ T cells induced IL-17 release by γδ T cells, which further aided recruitment of neutrophils into the lung (34). Despite these rapid inflammatory responses occurring within the first 8 h, lung alveolitis manifested only 2 days after SEA inhalation (54). These data suggested that in the time period between the first minutes of systemic T-cell activation and the appearance of extensive lung histopathology at 48 h, a communication web was formed between the systemic response and the lung mucosa, ultimately resulting in deterioration of the pulmonary barrier function.

In line with this hypothesis, we tested for increased lung permeability and the appearance of cellular injury markers and cell death following SEA inhalation. Lung permeability was assessed by administering FITC-dextran intranasally or intravenously, and then measuring fluorescence in the serum (lung to blood) or in the BAL fluid (blood to lung), respectively. While there was no difference in permeability at 24 h in the SEA-treated mice compared with vehicle, permeability was significantly increased by 48 h after SEA (Fig. 1A). Importantly, albumin was also increased in BAL fluid at 48 h (Fig. 1B).

It was previously shown that after S. aureus enterotoxin inhalation, acute alveolitis correlated with increased T-cell and innate immune cell presence in the BAL fluid and lung (34, 54, 62, 63, 68). Here, it is shown that 48 h after SEA inhalation, clusters of SEA-specific Vβ3⁺ T cells were concentrated at the interstitial areas surrounding bronchioles and perivascular regions (Fig. 1C). Additionally, these specific Vβ3⁺ T cells colocalized with CD11b⁺7/4⁺ inflammatory innate cells, and this area of inflammation was further characterized by the presence of enlarged CD169⁺ lung macrophages (Fig. 1C). The dramatic changes in the inflammatory cell number and lung permeability suggested that SEA inhalation could induce injury to pulmonary epithelial and endothelial cells. Indeed, CD54, RAGE, and Ang-2, which are established markers of epithelial and endothelial cell injury in ALI/ARDS (81), were significantly increased in BAL fluid 48 h after SEA inhalation but not at 24 h post-SEA (Fig. 1D).

In a previous report, it was shown that mice lacking T cells (TCR βδ^−/− mice) did not develop any lung pathology following SEA inhalation. Furthermore, CD4-depleted WT mice showed a partial abrogation of the inflammatory response, whereas CD8-depleted mice had an almost complete abrogation of pulmonary inflammation (54). These findings suggested that the inflammatory response to SEA could not occur in the absence of T cells. Nevertheless, T cell-independent responses to a S. aureus enterotoxin have been reported (18). Therefore, to further establish the role of T cells in SEA-induced pulmonary injury, WT and TCR βδ^−/− mice were exposed to SEA or vehicle, and BAL fluid was collected 48 h after inhalation. Unlike WT mice, SEA-exposed TCR βδ^−/− mice did not show any differences in the total number of BAL fluid cells, albumin concentration, as well as the endothelial injury marker Ang-2 (21), when compared with the vehicle control (Fig. 2A), confirming the critical role of T cells in inducing lung injury after SEA inhalation.

Fig. 2.

SEA induces apoptosis in the lung in a T-cell-dependent manner. A: number of BAL fluid cells and the concentration of albumin and Ang-2 in BAL fluid 48 h after SEA or vehicle inhalation in WT and TCR βδ^−/− mice. Two independent experiments were performed with a total of n = 6 per group. B: apoptosis in the lung and spleen of WT and TCR βδ^−/− mice 54 h after SEA or vehicle inhalation, as measured by a pan-caspase-binding fluorescent probe. After injecting the probe intravenously, tissue was removed and scanned by a fluorescence imager. C: quantification of the pan-caspase-binding fluorescent probe in the lung and spleen of WT and TCR βδ^−/− mice 54 h after SEA or vehicle inhalation. Three independent experiments with n = 8–18 mice per group. D: representative confocal microscopy images of lung tissue showing the presence of apoptotic cells (active caspase-3⁺) 48 h after SEA inhalation. Scale bar = 80 μm. Right panels depict magnified images of the fields indicated in left panels. E: number of active caspase-3⁺ cells per field. The number of positive cells 48 h after SEA or vehicle inhalation was determined from three images per mouse using Imaris (Bitplane). Counts in the SEA mice were performed in areas of inflammation. Data are representative of three independent experiments with n = 4 for vehicle and n = 6 for SEA. Data are shown as means ± SE. Two-way ANOVA with Sidak’s test or unpaired t-test (for Fig. 2E). ***P < 0.001.

It was next hypothesized that the increased number of inflammatory cells and the presence of injury markers in BAL fluid were associated with apoptosis, particularly occurring in tissues with significant T-cell expansion, i.e., lung and spleen (54). To test this, SEA- or vehicle-exposed mice received a pan-caspase-binding fluorescent probe before euthanasia, and the relative fluorescence levels in lung and spleen were measured. Both lung and spleen from SEA-treated mice showed significantly greater fluorescence than vehicle control, indicating an increase in cell death in tissues with expanded T-cell population (Fig. 2, B and C). To confirm that T cells were mediating the observed increase in caspase expression in the lung and spleen, the study was repeated using TCR βδ^−/− mice. No differences in relative fluorescence between the tissues were found, demonstrating that T cells are required for the induction of caspases in the lung (Fig. 2, B and C). To further validate these findings, lung tissue sections were examined for the presence of active caspase-3, an irreversible marker of cell death (55). Lungs from SEA-treated mice had a greater number of active caspase-3⁺ cells relative to vehicle control (Fig. 2, D and E). Thus SEA inhalation caused clustering of inflammatory cells to the interstitial bronchioles and perivascular regions, expression of epithelial and endothelial injury markers, and an increase in cell death, ultimately giving rise to increased pulmonary permeability.

S. aureus enterotoxin triggers early and delayed inflammation, which activates and subsequently reduces the number of lung endothelial cells.

Previous studies showed that development of pulmonary inflammation at 48 h was preceded by a rapid systemic inflammatory response in the first several hours after S. aureus enterotoxin inhalation. This response involved activation of enterotoxin-specific T cells, robust cytokine and chemokine release, and a recruitment of innate immune cells into circulation, lymphoid tissues, and lung (34, 63, 76). Our findings in Fig. 1, however, suggested that there was no pulmonary injury or change in lung permeability at 24 h. To better understand the relationship between the early inflammatory events and development of lung injury at 48 h, we performed a multiplex assay on mouse serum and BAL fluid obtained 4, 14, 24, or 48 h after SEA inhalation. As predicted, SEA induced massive increases in serum cytokines and chemokines 4 h after inhalation (Fig. 3). These levels dropped off by 14 h and were not different from the vehicle control by 24 h. In sharp contrast, the proinflammatory cytokines/chemokines in BAL fluid were increased only at 48 h, with the exceptions of CXCL1 and IL-12p40 (Fig. 3). The early presence of CXCL1 and IL-12p40 in BAL fluid could not be due to leakage from systemic circulation because, like CXCL1 and IL-12p40, IL-6, G-CSF, and CCL2 were markedly elevated in blood at 4 h but were not found in BAL fluid at the early time points. Therefore, we concluded that CXCL1 and IL-12p40 were likely produced locally rather than permeated from blood to BAL fluid. This finding was consistent with our previous report showing that IL-17 release by pulmonary γδ T cells promotes recruitment of neutrophils to the lung by 8 h after SEA inhalation (34).

Fig. 3.

There are differential patterns of cytokine and chemokine release in BAL fluid vs. serum after SEA inhalation. BAL fluid and serum were collected 4, 14, 24, and 48 h after SEA or vehicle inhalation, and protein levels of various analytes were measured by multiplex immunoassay. Data were combined from three independent experiments with n = 4–7 per group. Data are shown as means ± SE. Two-way ANOVA with Sidak’s test; **P < 0.01, ***P < 0.001.

To further study pulmonary response diversity before the onset of injury, lung cells were obtained from mice 14 or 40 h after SEA or vehicle inhalation, barcoded, stained with heavy metal-conjugated antibodies, and analyzed by CyTOF, also known as mass cytometry. The high-dimensional cytometry data were visualized as colorimetric two-dimensional ViSNE maps (4), which cluster single-cell data based on similarity in marker expression patterns. Furthermore, because the individual samples were barcoded, pooled, and acquired simultaneously by CyTOF, each map represents all four samples (14 h vehicle, 14 h SEA, 40 h vehicle, and 40 h SEA). Using combinations of lineage markers, we identified various lung cell populations to discern which cells were stimulated after SEA inhalation. We first looked at SEA-specific Vβ3⁺ T cells; however, insufficient numbers of specific T cells could be collected at 14 h, likely due to activation-induced increased adherence to the tissue (48). Nevertheless, CD8⁺Vβ3⁺ and CD4⁺Vβ3⁺ T cells (Fig. 4A) isolated 40 h after SEA inhalation showed upregulation of pSTAT5 and DNA content (Fig. 4B), consistent with activation and proliferation. Interestingly, even nonresponding bystander CD4⁺Vβ14⁺ T cells showed some increase in STAT phosphorylation (Fig. 4, A and B) but not to the extent of SEA-specific T cells. These findings confirmed that CyTOF could be used to identify changes in cell signaling. CD11c⁺MHC II⁺ dendritic cells (Fig. 4C) were slightly activated at 14 h but more robustly at 40 h, as indicated by increased pSTAT3 (Fig. 4D). Notably, CD45⁻CD31⁺CD54⁺ endothelial cells (Fig. 4E) had upregulated pSTAT1 and MHC II and some downregulation of IκB, as early as 14 h after SEA inhalation (Fig. 4F). These results were replicated in three additional independent experiments (Fig. 4, G–I). Finally, STAT phosphorylation in other lung cell populations trended with their lineages (Fig. 4, J and K). Overall, these data suggest that endothelial cells became activated early but were able to maintain their function, as there were no changes in pulmonary permeability or injury markers at 24 h (Fig. 1).

Fig. 4.

Endothelial cells become activated before the appearance of lung pathology. Lung tissue was obtained from mice 14 or 40 h after SEA or vehicle inhalation. Individual samples were barcoded, pooled, and then stained with heavy metal-conjugated antibodies. The pooled sample was analyzed by mass cytometry (Helios, Fluidigm). ViSNE maps generated using MatLab were used to represent clusters of cells identified according to expression of markers used in the experiment. Expression of each marker can be visualized on the basis of the color gradient indicated on the left. A: ViSNE maps to identify CD3⁺ cells and subpopulations of CD3⁺ cells (SEA-specific CD45⁺CD3⁺CD8⁺Vβ3⁺ and CD45⁺CD3⁺CD4⁺Vβ3⁺ T cells and bystander CD45⁺CD3⁺CD4⁺Vβ14⁺ T cells). B: histograms showing the expression of signaling and activation markers in T cells 40 h after vehicle (blue) vs. SEA (red) inhalation. C: ViSNE maps identifying dendritic cells (CD45⁺CD11c⁺MHC II⁺). D: histograms showing the expression of signaling and activation markers in dendritic cells 14 (top) or 40 h (bottom) after vehicle (blue) vs. SEA (red) inhalation. E: ViSNE maps identifying endothelial cells (CD45⁻CD31⁺CD54⁺). F: Histograms showing the expression of signaling and activation markers in endothelial cells 14 (top) or 40 h (bottom) after vehicle (blue) vs. SEA (red) inhalation. G: pSTAT median intensity (MI) in T-cell populations 40 h after SEA or vehicle inhalation. H: pSTAT3 median intensity in dendritic cells 14 or 40 h after SEA or vehicle inhalation. I: MHC II median intensity in endothelial cells 14 or 40 h after SEA or vehicle inhalation. J: pSTAT5 median intensity in CD45⁺NK1.1⁺CD3⁻ NK cells 40 h after SEA or vehicle inhalation. K: pSTAT3 median intensity in CD45⁺CD11c⁺SIGLEC F⁺ alveolar macrophages 40 h after SEA or vehicle inhalation. Each experiment is represented as a pair of connected symbols (blue square denotes vehicle, while red square denotes SEA) with a mean shown as a horizontal line. Data are representative of four independent experiments with n = 4 per group. Paired t-test; *P < 0.05, **P < 0.01.

To study endothelial cell activation, lung cells were collected 14, 24, or 40 h after SEA or vehicle inhalation, sorted for CD45⁻CD31⁺CD54⁺ endothelial cells, and then analyzed for gene expression. Sorted cells displayed a number of dynamic changes in response to SEA (Fig. 5A). Expression of integrin Vcam1 were increased at 14 h (Fig. 5A), likely signifying early activation. Interestingly, the expression of Nos3 (endothelial nitric oxide synthase), an enzyme essential to endothelial cell homeostasis (25, 59), was first upregulated at 24 h and then downregulated at 40 h (Fig. 5A). Finally, Icam1 (CD54) expression was increased twofold over vehicle at 40 h post-SEA inhalation (Fig. 5A). In addition to these sequential phenotype changes in pulmonary endothelial cells, there was a significant reduction in their numbers at 48 h after SEA (Fig. 5B), and endothelial cell percentage inversely correlated with percentage of SEA-specific Vβ3⁺ T cells in the lung (Fig. 5C). Endothelial cells from SEA-exposed mice also showed increased expression of CD54 and CD95 (Fas) (Fig. 5D), and similarly, SEA-specific Vβ3⁺ T cells had greater levels of CD11a expression (receptor for CD54) and CD95 (Fig. 5E). Granzyme B was also measured, since it was found to increase in cases of septic ALI/ARDS patients (24). Indeed, SEA-specific Vβ3⁺ T cells were found to have upregulated granzyme B expression (Fig. 5E), but, most importantly, robust levels of granzyme B were detected in BAL fluid at 48 h (Fig. 5F). Thus, as lung permeability manifested at time points beyond 24 h (Fig. 1A), there were coincident increases in markers of cytotoxicity and cell death, particularly granzyme B and CD95.

Fig. 5.

SEA inhalation causes both early and late changes in endothelial cells. A: endothelial cell gene expression 14, 24, or 40 h after vehicle or SEA inhalation. Endothelial cells (CD45⁻CD31⁺CD54⁺) were then sorted from total lung cells and lysed to obtain RNA. Gene expression levels are normalized to Gapdh and are shown relative to vehicle control (set to 1). Three independent experiments with n = 6 per group. B: number of lung endothelial cells (ECs; CD45⁻CD31⁺) 48 h after SEA or vehicle inhalation, as measured by flow cytometry. C: correlation between endothelial cell and CD3⁺Vβ3⁺ T-cell percentages in the lung 48 h after SEA inhalation. Linear regression curve fit P value was calculated via F test. Two independent experiments with n = 8 per group. D: expression of surface CD54 and CD95 on endothelial cells (CD45⁻CD31⁺) 48 h after SEA or vehicle inhalation. E: surface expression of CD11a and CD95 and intracellular expression of granzyme B (Gzm B) in CD3⁺Vβ3⁺ T cells 48 h after SEA or vehicle inhalation. Histograms are representative of two independent experiments with n = 8 per group. Median fluorescent intensities in the upper corner of each histogram are shown as an average ± SD of one out of the two experiments. Isotype control (gray), vehicle (V; black dashed line), and SEA (S; black thick line). F: Gzm B concentration in BAL fluid of SEA or vehicle-treated mice 24 and 48 h after inhalation. Three independent experiments with n = 8 per group. Data are represented as means ± SE. Two-way ANOVA with Sidak’s test or unpaired t test (for Fig. 5B). *P < 0.05, **P < 0.01, ***P < 0.001.

S. aureus enterotoxin-induced lung injury is dependent on T-cell costimulation, but therapy is possible through CD54 interference.

It was previously demonstrated that systemic recruitment of neutrophils and monocytes after S. aureus enterotoxin-induced T-cell activation was uniquely regulated by TNF and CD28 costimulation (76). Thus it was hypothesized that preventing early inflammation with CD28 or TNF blockade would minimize the increase in pulmonary permeability. Mice were pretreated with the CD28 signaling inhibitor CTLA4-Ig, anti-TNF antibody, or IgG control 2 h before SEA inhalation, and at 48 h post-SEA, BAL fluid and lung were harvested (Fig. 6A). CTLA4-Ig and anti-TNF pretreatment significantly reduced BAL fluid cell numbers (Fig. 6B), whereas CTLA4-Ig, but not anti-TNF, decreased the overall proportion of T cells and inflammatory innate cells in the BAL fluid (Fig. 6C; P = 0.043 for CTLA4-Ig and P = 0.12 for anti-TNF). Importantly, anti-TNF and CTLA4-Ig significantly reduced the presence of albumin and some of the injury and inflammation markers, but CTLA4-Ig was more potent overall at attenuating the lung injury (Fig. 6D). Additionally, CTLA4-Ig decreased the total number of monocytes in lung (Fig. 6E). Thus early inflammatory responses define, in part, the severity of S. aureus enterotoxin-mediated lung injury.

Fig. 6.

Blocking T-cell activation with CTLA4-Ig or anti-TNF, in part, attenuates measures of lung permeability, cell injury, and inflammation after SEA inhalation. A: diagram of experimental setup. Mice were treated with CTLA4-Ig, anti-TNF, or IgG control 2 h before SEA inhalation. BAL fluid and lung tissue were obtained 48 h after SEA inhalation. B: total number of BAL fluid cells pretreated with IgG control, CTLA4-Ig, or anti-TNF, as described in A. C: composition of cells within the BAL fluid following IgG control, CTLA4-Ig, or anti-TNF pretreatment. Average percentages of CD45⁺CD3⁺Vβ3⁻ T cells, CD45⁺CD3⁺Vβ3⁺ T cells, neutrophils (CD45⁺CD11b⁺Ly6C⁺Ly6G⁺), and monocytes (CD45⁺CD11b⁺Ly6C⁺Ly6G⁻) in BAL fluid are represented by stacked bars. D: concentrations of injury markers (Ang-2, RAGE, and CD54), inflammatory markers Gzm B, IFNγ, IL-6, and TNF, and albumin in BAL fluid from mice pretreated with IgG control, CTLA4-Ig, or anti-TNF. E: total number of Vβ3⁺ T cells, monocytes, and neutrophils in lung in mice pretreated with IgG control, CTLA4-Ig, or anti-TNF. Data are represented as means ± SE. Three independent experiments with n = 9 per group. One-way ANOVA with Dunnett’s test. *P < 0.05, **P < 0.01.

Despite the role played by early inflammatory responses in influencing the extent of lung injury, therapeutic strategies to block them have limited translational relevance when taking into consideration the typical disease progression and time scales involved, as well as the logistics of diagnosing and treating patients in the hospital. Alternatively, we investigated later events in the SEA-induced response, during which time, entry and accumulation of lung T cells are evident. Pulmonary tissue and endothelial cells, in particular, can be injured in a number of ways, including mechanisms involving cytokines and direct death pathways such as CD95 signaling (38, 59, 80). Additionally, CD54 was previously shown to enhance leukocyte transmigration and promote cytotoxic T-cell function (5, 23, 61). Thus CD95 and CD54 (Fig. 5D) could be essential mediators of pulmonary barrier damage occurring later in the SEA response progression.

Therefore, we hypothesized that CD95 and/or CD54 blockade would prevent pulmonary barrier leakiness at clinically relevant time points beyond the initiation of inflammation. Thus, 36 h after SEA inhalation, mice were treated with anti-CD178 (ligand for CD95; FasL), anti-CD54, or IgG control, and BAL fluid and lung were collected at 48 h (Fig. 7A). While anti-CD178 showed no change in total cell number, anti-CD54 significantly reduced the total number of cells in BAL fluid (Fig. 7B) and also decreased the overall proportion of T cells and inflammatory innate cells present (P = 0.0012, Fig. 7C). Importantly, anti-CD54 therapy potently decreased the measured injury markers (CD54 and RAGE), inflammatory markers (granzyme B, IFNγ, and IL-6), and albumin in BAL fluid (Fig. 7D). Paradoxically, BAL fluid TNF levels were increased with anti-CD54 treatment (Fig. 7D). This overall reduction in inflammation was not limited to soluble factors, since CD54 blockade also reduced the total numbers of SEA-specific Vβ3⁺ T cells in the lung (Fig. 7E). Interestingly, anti-CD178 treatment only increased the number of neutrophils in the lung, while having no effect on measures of lung injury (Fig. 7, D and E). To validate the effect of anti-CD54 on preserving lung barrier integrity, mice were treated with anti-CD54 or IgG control 36 h after SEA inhalation, and lung permeability was measured with FITC-dextran (as described previously) at 48 h. Indeed, the anti-CD54-treated mice showed significantly lower levels of FITC fluorescence in BAL fluid relative to IgG-treated mice, but not statistically different from vehicle control (P = 0.27; Fig. 7F). While CD54 blockade reduced total numbers of SEA-specific T cells in the lung by 45% (Fig. 7E), we also observed substantial decreases in BAL fluid IFNγ and granzyme B levels (73% and 76%, respectively) in the treated mice (Fig. 7D). These findings suggest that anti-CD54 therapy might directly diminish T-cell cytotoxic potential, in addition to reducing pulmonary T-cell infiltration. However, there was no difference in IFNγ and granzyme B expression by SEA-specific CD8⁺ T cells (Fig. 7G), suggesting that the synthesis of cytotoxic molecules was not markedly impacted by the therapy. In summary, these findings demonstrate that anti-CD54 therapy significantly attenuates S. aureus enterotoxin-mediated lung injury well after the initiation of inflammation.

DISCUSSION

ALI/ARDS is a life-threatening condition that affects ~200,000 people in the United States annually, but despite recent advances in supportive care, the mortality rates remain unacceptably high (9, 42, 43). Although many clinical studies have investigated various pharmacotherapies for ALI/ARDS, results have been mostly disappointing (10, 42). Thus there is an urgent need to better understand the mechanisms of tissue injury to design novel treatments for ALI/ARDS, and we propose CD54 as a promising immunotherapeutic target.

While the role of innate cells, particularly, neutrophils and macrophages, has been studied extensively in ALI/ARDS (2, 82), the involvement of adaptive immunity, T cells, in particular, is only now beginning to emerge (57). Prior studies showed that ALI/ARDS patients present with greater numbers of lymphocytes in circulation (86) and in BAL fluid (39). Furthermore, the T-cell growth factor IL-2 was found to be elevated in BAL fluid of ARDS patients and was strongly associated with increased mortality (37, 78). ARDS alveolar T cells also showed significant increases in activation, proliferation, and cytokine secretion (67). In mouse models of ALI, Th17 cells were shown to be pathogenic (39), while Tregs were essential for recovery (14). However, the mechanism of T cell-based pulmonary injury is unclear. Similar to LPS and bleomycin inhalation in mice (1), SEA models acute lung injury, but with greater emphasis on the role of T cells. It is important to note that enterotoxins, including SEA, have been detected in the circulation of intensive care unit patients (6) and also in S. aureus isolated from individuals with complications from pneumonia (71, 83). This only suggests an association of S. aureus enterotoxins with ARDS, but nonhuman primates exposed to aerosolized enterotoxin exhibited severe pulmonary lesions, with lethality in some subjects (44). Thus there is a critical need to attempt to identify S. aureus enterotoxin protein in BAL fluid of ALI/ARDS patients.

To explore the mechanisms governing the development of ALI following S. aureus enterotoxin inhalation, we first defined the inflammatory changes that take place in the lung in terms of two distinct phases. The initial response is caused by rapid, oligoclonal T-cell activation leading to robust cytokine and chemokine release (Fig. 3), and recruitment of neutrophils and monocytes into circulation followed by their appearance in the lung (34, 76). This phase of the response occurs within hours after SEA inhalation. Here, we found that the early inflammatory responses were accompanied by changes in cell signaling and the appearance of activation markers of endothelial cells. Upregulation of pSTAT1 and MHC II [potentially in response to IFNγ stimulation (12, 16)] and changes in mRNA expression levels of integrin Vcam1 were evident (Figs. 4F and 5A). However, there were no changes in lung permeability or signs of damage noted (Fig. 1), and the barrier was also impermeable, even to the high levels of circulating cytokines found in the serum at 4 h (Fig. 3).

Lung permeability was clearly increased at 48 h (Fig. 1A), which coincides with the second phase of the inflammatory response marked by robust T-cell expansion in the lung (54). This occurs just before peripheral T-cell deletion and anergy, which is a noted outcome of enterotoxin T-cell response (30, 49). The change in lung permeability was further accompanied by increased levels of soluble CD54, Ang-2, and RAGE (Fig. 1D), increased expression of caspases (Fig. 2), and elevated cytokines and chemokines in BAL fluid (Fig. 3). At 40 h after SEA inhalation, the onset of the second phase was marked by changes in signaling events and activation of different pulmonary cell populations (Fig. 4). In particular, pSTAT5 expression was increased in T cells and NK cells (Fig. 4, G and J), whereas pSTAT3 expression was increased in dendritic cells and alveolar macrophages (Fig. 4, H and K). We posit that STAT5 phosphorylation is in response to T-cell release of IL-2, while IL-6 release from innate cells promotes STAT3 phosphorylation (16). On the other hand, endothelial cells upregulated Icam-1 and downregulated Nos3 transcription (Fig. 5A), suggesting a loss of homeostatic mechanisms and polarization toward a more activated phenotype (25, 52). Finally, this later effect likely portended the reduction of endothelial cells, which also coincided with increased CD54 and CD95 expression (Fig. 5, B and D). Thus lung endothelial cells became activated and later began to disappear, probably by a death process, as S. aureus enterotoxin induced T-cell activation and expansion.

Taken together, these findings suggested a two-hit mechanism of endothelial cell injury: first, pulmonary endothelial cells became activated (Figs. 4F and 5A) due to early inflammatory responses occurring systemically or locally, but the endothelial barrier remained intact throughout this phase (Fig. 1). More specifically, the massive levels of serum cytokines detected at 4 h (Fig. 3) or the transiently recruited neutrophils in the lung (34) could activate the endothelial cells without causing their injury or death. A second hit, coinciding with the systemic and intrapulmonary expansion of SEA-specific T cells, induced the endothelial cells to downregulate homeostatic factors and acquire a more activated and death-prone phenotype, eventually leading to cellular dysfunction and death. Recently, it was proposed that sequential hits to the lung from various insults, such as pneumonia, sepsis, or mechanical ventilation, may cause the severe pathology of ARDS. This proposed model for disease etiology and progression is known as the “multiple hit theory” (10). Similarly, animal models of ALI that consist of two inflammatory insults to the lung may better mimic the clinical scenario (36).

To unravel the mechanism responsible for the observed increase in pulmonary permeability, we designed our experimental strategy based on recent data showing that the early inflammatory responses were differentially affected by blocking TNF or CD28 costimulation (76). In the current report, we demonstrate that TNF or CD28 blockade elicited a profound reduction of albumin in BAL fluid, as well as diminished inflammation and expression of injury markers in the lung (Fig. 6). Thus, although there were no changes in pulmonary permeability initially (Fig. 1), early blockade of T-cell activation reduced the subsequent lung injury. Consistent with the two-hit injury model, these findings suggested that the activation of endothelial cells due to an initial inflammatory stimulus may not necessarily induce their injury and death but may increase their sensitivity to secondary injury.

Understanding that the initial T-cell activation influences lung injury at later time points is undoubtedly important; however, it may be difficult to translate these findings into clinical care, since delivering early therapy during the initial insult is often not possible. Previous studies attempting to reduce the inflammatory response after S. aureus enterotoxin inhalation administered therapy prophylactically or immediately after toxin exposure (64, 65, 75, 79). Alternatively, we sought a therapeutic approach to reduce S. aureus enterotoxin-induced lung injury in a scenario mimicking a clinical setting, wherein enterotoxin inhalation preceded initiation of therapy by many hours. Because CD95 and CD54 were upregulated on endothelial cells after SEA inhalation (Fig. 5D), we administered blocking antibodies targeting CD178 and CD54 at 36 h post-SEA. Previous studies found that both CD95 and CD178 were elevated in BAL fluid of ALI/ARDS, and the CD95/CD178 pathway may be critical for inducing epithelial cell injury (3, 24, 46, 47). However, with the exception of increasing lung neutrophil numbers (Fig. 7E), blocking CD178 had no effect on lung injury measures in our model (Fig. 7, B–D). In sharp contrast, CD54 blockade profoundly reduced pulmonary injury, by means of decreasing BAL fluid cell number, preventing increases in proinflammatory cytokine release, minimizing expression of markers of endothelial and epithelial cell damage, and reducing lung permeability (Fig. 7). Being a costimulatory molecule, CD54 is expected to inhibit T-cell activation at the time of antigen presentation (15, 33, 35, 85), but our results show that many hours after SEA-induced activation, CD54 plays an active role in shaping T-cell responses. In particular, specific T-cell numbers were significantly reduced after CD54 therapy (Fig. 7E). In addition to reducing T-cell expansion in the lung, another proposed mechanism by which CD54 blockade might maintain the lung barrier is based on data showing that CD54 was important for enhancing cytotoxic function of lymphocytes (11, 66, 80). Similarly, we observed substantial reductions in granzyme B and IFNγ in the BAL fluid of anti-CD54-treated mice (Fig. 7D). Intriguingly, there was no difference in the cytotoxic potential of SEA-specific CD8⁺ T cells after CD54 therapy (Fig. 7G), suggesting effector T-cell differentiation was not markedly impaired. Rather than the synthesis, anti-CD54 therapy could be involved in regulating the release of cytotoxic and inflammatory molecules (5). Finally, although SEA-induced lung injury is T cell-dependent (Fig. 2), other cell populations, including inflammatory innate cells are recruited to the lung (Fig. 1C). Therefore, anti-CD54 could also ameliorate the SEA-induced lung injury by affecting other cell types, such as neutrophils and monocytes. Although recruitment of neutrophils and monocytes to the lung was not affected by anti-CD54 (Fig. 2E), the treatment could affect the function of these cells, such as the production of reactive oxygen species (84). CD54-targeted therapy was previously shown to attenuate pulmonary damage in two models of neutrophil-dependent lung injury (40, 53), and anti-CD54 antibody (enlimomab) has already been tested in several clinical trials (e.g., transplantation, burn injuries, and refractory rheumatoid arthritis) (29, 51, 70). Therefore, anti-CD54 may be a potential therapeutic target for ALI/ARDS, worthy of further investigation. Finally, taking into account the role of CD54 in leukocyte migration during injury or infection (61), it will be important to accurately define therapeutic windows for CD54-blocking therapies given to ALI/ARDS patients. Furthermore, determining the molecular players downstream of CD54 engagement may also enable development of more targeted treatment options.

The utility of CD54 therapy is well matched to a typical ALI clinical scenario. A patient suffering from a severe pneumonia or trauma may experience overactivation of the immune system and/or direct damage to the tissue, resulting in impaired lung barrier defenses. This deficit may predispose the patient to additional insults, such as toxins and danger- and pathogen-associated molecular patterns that can give rise to systemic inflammatory response syndrome and/or ALI with T cells, likely playing a crucial role (41). An excellent example is the appreciation of S. aureus enterotoxins in septic patients (6, 19, 28). Here, S. aureus enterotoxin inhalation was found to induce sequential changes in the lung, defined by initial rapid endothelial cell activation coinciding with inflammatory responses due to T-cell priming followed by pulmonary injury and cell death coinciding with T-cell expansion in the lung. Importantly, even when administered therapeutically at time points well beyond the inciting inflammatory trigger, anti-CD54 effectively reduced all measures of lung inflammation and injury and restored pulmonary barrier integrity. Thus an immunotherapeutic agent targeting the CD54 pathway may be a suitable treatment option for ALI/ARDS patients, especially in suspected cases of T cell-induced inflammation.

GRANTS

The study was supported by National Institute of Allergy and Infectious Diseases Grant P01AI056172.

DISCLOSURES

No conflicts of interest, financial or otherwise are declared by the authors.

AUTHOR CONTRIBUTIONS

J.S. and A.T.V. conceived and designed research; J.S., A.M., P.M., and J.M.R. performed experiments; J.S., A.M., P.M., J.M.R., and J.A.B. analyzed data; J.S. and A.T.V. interpreted results of experiments; J.S. prepared figures; J.S. drafted manuscript; J.S., A.M., P.M., J.M.R., and A.T.V. edited and revised manuscript; J.S. and A.T.V. approved final version of manuscript.