INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The diffuse lung injury process known as the acute respiratory distress syndrome (ARDS) remains a major cause of mortality and morbidity (1). Phagocytes, in particular macrophages and polymorphonuclear neutrophils, are recognized as major components of inflammatory and immunologic reactions in the lung (2). As a result, many candidate molecules involved in the pathogenesis of ARDS have been thought to originate from these inflammatory cells and little is known about the role of infiltrating lymphocytes and their products.

Cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells are known to induce lethal damage on their target cells via two independent death pathways by inducing granule exocytosis or the Fas ligand (FasL)/Fas system (3, 4). The first pathway acts via degranulation of serine proteases, granzymes, together with the pore-forming protein, perforin, to induce rapid death of the target cells. The second pathway is mediated by the cell surface receptor Fas (Apo-1, CD95), expressed on a variety of cells and FasL present on the surface of CTLs. Cells expressing Fas rapidly undergo apoptosis through a rapid intracellular signaling pathway in the presence of FasL. This depends on distinct cytoplasmic regions known as "death domains" (5). Although expression of FasL was originally thought to be limited to CTLs, FasL can be found as a soluble form in the circulation or on the surface of other cells such as activated polymorphonuclear neutrophils (6). Several reports have suggested that apoptosis is necessary for the resolution of fibroproliferative responses after lung injury (7, 8). However, there have been very few reports describing the immunological events in the early phase of ARDS. Interestingly, Hagimoto and coworkers demonstrated that intratracheal administration of bleomycin induces Fas expression in lung epithelial cells and FasL in infiltrating lymphocytes in the early and late phase of lung injury. They speculated that these effector molecules not only play an important role in remodeling of the cell layer lining the alveolar spaces in the late phase of lung injury but also may damage tissues by overexpression in the early phase (9).

Thus, we expected that apoptosis molecules might also participate in the pathogenesis of ARDS. In the present study, we observed gene expressions of perforin, granzyme A, granzyme B, FasL, and Fas in bronchoalveolar lavage fluid (BALF) collected from patients during the acute phase of septic ARDS. We used reverse transcription polymerase chain reaction (RT-PCR) method for their detection. Immunofluorescence flow cytometric analysis of lavaged cells and measurement of soluble FasL in the BALF were also carried out. Expression of messenger RNA (mRNA) of the inflammatory cytokines such as interleukin-1 (IL-1), tumor necrosis factor-alpha (TNF-), IL-6, IL-8, and the inducible nitric oxide synthase (iNOS) were also measured. All data obtained in acute phase of septic ARDS patients were compared with data from patients with normal lung function, from septic patients without lung damage, and from patients in the late phase (7 d after the onset) of septic ARDS.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

This study was conducted between January 1997 and March 1999 at Kyoto Prefectural University of Medicine. Patients with ARDS were prospectively identified. ARDS was defined according to Murray's definition and the American-European consensus conference on ARDS as (1) lung injury scores higher than 2.5; (2) wedge pressure < 18 mm Hg measured by a pulmonary catheter, or no clinical evidence of congestive heart failure; (3) no other explanation for the above findings; and (4) ratio of arterial oxygen tension to fraction of inspired oxygen (Pa_O2/FI_O2) < 200 (1, 10). The patient profile is summarized in Table 1. Of 28 patients enrolled, 14 patients developed ARDS. Seven nonsmoking patients had normal lung functions and were used as the control group. The remainder of the patients enrolled were mechanically ventilated and clinically diagnosed with sepsis syndrome. Their systemic responses were manifested by at least two of the following conditions resulting from an infection: temperature > 38° C, heart rate > 90 beats/min, spontaneous respiratory rate > 20 breaths/min or Pa_CO2 < 32 mm Hg, white blood cell counts (WBC) > 12,000 cells/ mm³, < 4,000 cells/mm³, or > 10% immature forms (11).

There were seven patients who were clinically septic with lung injury score less than 1.0 and Pa_O2/FI_O2 higher than 300 mm Hg (septic non-ARDS group), 10 patients with acute phase (within 24 h after diagnosis) of septic ARDS (acute ARDS group), and 10 patients with late phase (7 d after diagnosis) of septic ARDS (late ARDS group). Six patients (Patients 19 to 24) in the acute ARDS group had a second bronchoalveolar lavage (BAL) 1 wk after the first BAL and were also used as Patients 29 to 34. We could not carry out a second BAL for the other four in the acute ARDS group owing to clinical or ethical considerations. A BAL just after onset of ARDS was not carried out in Patients 25 to 28 because of the later admission to the intensive care unit (ICU). Six patients in the late ARDS group (Patients 25, 26, 27, 30, 33, 34) received high dose of methylprednisolone (1 g/d for 3 d) before the BAL. A history of obstructive pulmonary disease and severe asthma was excluded in all patients. The ethical committee of Kyoto Prefectural University of Medicine approved the study. Informed consent for bronchoscopy was obtained from the patients or their relatives.

Samples

BALF was obtained at the time of induction of general anesthesia in the control group, 3 to 7 d after artificial ventilation in the septic non-ARDS group, within 24 h after ARDS was diagnosed in the acute ARDS group, and 7 d after diagnosis of ARDS in the late ARDS group. All patients were sedated and preoxygenated (FI_O2 = 1.0). A flexible bronchoscope (FB-18x; Pentax, Tokyo, Japan) was inserted and wedged into the right middle lobe bronchus; five aliquots of 20 ml of 0.9% saline were instilled and then gently aspirated. The lavage fluid was strained through sterile gauze to remove mucus and debris. After gauze filtration, the BALF was centrifuged immediately at 500 × g for 10 min at +4° C, and the cell-free supernatant was stored at 80° C for further analysis. The cells in the pellet were used for total RNA extraction. In some patients, cells were washed and resuspended to prepare 1 × 10⁶ cells/ml in PBS solution for flow cytometric measurements. A small aliquot of pooled, well-mixed lavage fluid was placed on a grid hemocytometer for measurement of total cell counts. The percentage of different cell types was determined by counting 400 cells stained with Diff-Quik (American Scientific, McGraw Park, IL) on cytocentrifuge preparations (Cytospin 3; Shandon Scientific, Cheshire, UK). Total levels of protein, and albumin were determined in the cell-free supernatant by a nephelometric and a modified colorimetric method, respectively.

RT-PCR Procedure

Total RNA was extracted from the BALF by the acid guanidium phenol chloroform (AGPC) method. After purification, RNA was dissolved in diethylpyrocarbonate (DEPC)-treated water and quantified. Five micrograms of total RNA were used for reverse transcription to make complementary DNA (cDNA) with a SuperScript preamplification kit for first-strand cDNA synthesis (Life Technologies, Gaithersburg, MD). After the reverse transcription reaction, 250 ng of cDNA were used as a template in the polymerase chain reaction (PCR) procedure. The PCR mixture (total volume 50 µl) consisted of cDNA, 2 µl of 25 mM primers, 10 µl of 10 mM deoxyribonucleoside triphosphate (dNTP), and 2 units of Taq polymerase. The mixture was subjected to PCR amplification. Oligonucleotide primers for each molecule were: GAPDH: (sense 5' ACC ACC ATG GAG AAG GCT GG 3', antisense 5' CTC AGT GTA GCC CAG GAT GC 3'); perforin (PCR product: 561 bp, sense 5' AAG TCA GCT CCA CTG AAG CT 3', antisense 5' GGT AGG TTT GGT GGA AGG AG 3'); granzyme A (PCR product: 560 bp, sense 5' ACC CTA CAT GGT CCT ACT TAG 3', antisense 5' AAG TGA CCC CTC GGA AAA CA 3'); granzyme B (PCR product: 425 bp, sense 5' GCT TAT CTT ATG ATC TGG GAT C 3', antisense 5' AAG TCA GAT TCG CAC TTT CGA 3'); FasL (PCR product: 447 bp, sense 5' CAG CTC TTC CAC CTA CAG AAG G 3', antisense 5' AAG ATT GAA CAC TGC CCC CAG G 3'); Fas (PCR product: 324 bp, sense 5' CAA TGG GGA TGA ACC AGA CTG C 3', antisense 5' GGC AAA AGA AGA AGA CAA AGC C 3'); IL-1 (PCR product: 802 bp, sense 5' ATG GCA GAA GTA CCT AAG CTC GC 3'; antisense 5' ACA CAA ATT GCA TGG TGA AGT CAG TT 3'); TNF (PCR product: 444 bp, sense 5' GAG TGA CAA GCC TGT AGC CCA TGT TGT AGC A 3', antisense 5' GCA ATG ATC CCA AAG TAG ACC TGC CCA GAC T 3'); IL-6 (PCR product: 628 bp, sense 5' ATG AAC TCC TTC TCC ACA AGC GC 3', antisense 5' GAA GAG CCC TCA GGC TGG ACT G 3'); IL-8 (PCR product: 289 bp, sense 5' ATG ACT TCC AAG CTG GCC GTG GCT 3', antisense 5' TCT CAG CCC TCT TCA AAA ACT TCT C 3'); iNOS (PCR product: 231 bp, sense 5' CTT CAA CCC CAA GGT TGT CTG CAT 3', antisense 5' ATG TCA TGA GCA AAG GCG CAG AAC 3'). Primers for human/mouse iNOS were purchased from Maxim Biotech (South San Francisco, CA) and IL-1, IL-6, IL-8, and TNF- were purchased from Clontech (Palo Alto, CA). Primers for perforin, granzyme A, granzyme B, FasL, and Fas were designed for this study. After the PCR, a 9-µl aliquot from the PCR reaction was electrophoresed on a 2% agarose gel stained with ethidium bromide. The gel was then photographed under ultraviolet transillumination. For quantification, the gel was scanned under ultraviolet transillumination using a densitometer (ATTO, Osaka, Japan). Each signal was normalized relative to its corresponding reduced glyceraldehyde-phosphate dehydrogenase (GAPDH) signal from the same RNA, and expressed as the mRNA/GAPDH ratio.

Flow Cytometry and Soluble FasL Measurement

The immunofluorescence flow cytometric analysis was carried out using cells obtained from six patients in the acute ARDS group, three patients in the septic non-ARDS group, and six patients in the late ARDS group. Cells (1 × 10⁶) in a 1-ml tube were incubated for 30 min with a fluorescein isothiocyanate (FITC)-conjugated monoclonal antibody against anti-human Fas (UB; Medical and Biological Laboratories, Nagoya, Japan) or against FasL (NOK-2; Pharmingen, San Diego, CA). In another setting, the cells were first prepared with permeabilization solution to degenerate cell membrane, and incubated with anti-human perforin antibodies (G9; Ancell Corp. Bayport, MN), or anti-human granzyme B antibodies (Parmacell, Paris, France), and then directly incubated with streptavidin-phycoerythrin (Pharmingen, San Diego, CA). Expressions of these molecules were detected by an EPICS profile analyzer (Coulter Corp., Hialeah, FL).

Soluble FasL in the supernatant of BALF was measured by an enzyme-linked immunosorbent assay (ELISA) kit (Medical and Biological Laboratories, Nagoya, Japan). The lower limit of detection levels of soluble FasL was 50 pg/ml.

Data Analysis

Data were expressed as mean ± SEM. Kruskal-Wallis nonparametric analysis of variance for factorial experiments was used. Dunn's procedure was used for post hoc multicomparison analysis among the groups. We regarded the data as statistically significant at a p value of < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Seven of 14 ARDS patients enrolled died during their stay in the ICU. The patients died primarily due to multiple organ failure within 6 to 14 d after entry in the ICU. No patients in the septic non-ARDS group developed lung injury.

Cell Counts

The cell counts of the BALF from each group and protein and albumin levels are summarized in Table 2. Percentages of alveolar macrophages in BALF from the acute ARDS and the late ARDS groups were significantly lower and the percentages of neutrophils were higher than those in the control group. No significant differences were observed in the percentages of alveolar macrophages or neutrophils between the septic non-ARDS group and the control group. No statistical differences were observed in the percentage of lymphocytes among the groups. Concentraions of total protein and albumin were also increased in the ARDS groups.

Expression of Apoptotic Molecules, Cytokines, and iNOS

The PCR products, amplified using specific primers for the apoptotic molecules, inflammatory cytokines, and iNOS, demonstrated clear bands as shown in Figure 1. These bands were absent in PCR-amplified products lacking a cDNA template. Thus, there was no evidence of DNA contamination in any of the samples. This implied that the amplified cDNA originated from mRNA and not from genomic DNA or other contamination. In the control group, there were no mRNA expressions of apoptotic molecules, inflammatory cytokines, or iNOS. All patients in the acute ARDS group demonstrated high mRNA expression of all the effector molecules for apoptosis except for two patients (Patients 15 and 20) who did not show upregulated FasL expression (Figure 1 and Figure 2A). Interestingly, these expressions disappeared in the late phase of septic ARDS. In some cases of septic non-ARDS, moderate expression of Fas (Patients 11, 12, and 14) was observed. All inflammatory cytokines and iNOS mRNA were highly expressed in the acute ARDS group, whereas the reduction of these expressions was arbitrary in the late ARDS group. In the non-ARDS group, high expressions of IL-1 and IL-8 mRNA were observed. However, the expression ratios of these mRNAs were significantly higher in the acute ARDS group (Figure 2B). There was no distinct relationship between the mortality of the patients and the intensities of the expressions.

View larger version (54K): [in this window] [in a new window]   Figure 1.   mRNA levels of GAPDH, perforin, granzyme A, granzyme B, FasL, Fas, IL-1, TNF-, IL-6, IL-8, and iNOS in cells obtained from BALF as assessed by RT-PCR. Data from control group is not shown, as only the housekeeping GAPDH gene was visibly expressed. See also Figure 2 for densitometric ratio of each expression.

View larger version (24K): [in this window] [in a new window]   View larger version (41K): [in this window] [in a new window]   Figure 2.   (A) Densitometric ratio of mRNA expressions for apoptotic molecules against GAPDH mRNA expression in each group. Almost no expression was noted in the control, septic non-ARDS, and late phase of ARDS group. Data shown as mean + SEM. *p < 0.05 versus control group. (B) Densitometric ratio of inflammatory cytokines and iNOS mRNA expressions against GAPDH mRNA expression in each group. Significant increases in the expressions of IL-1 and IL-8 were noted in all septic patients groups. Furthermore, TNF-, IL-6, and iNOS mRNA were upregulated in the septic acute ARDS group. Data shown as mean + SEM. *p < 0.05 versus control, septic non-ARDS and late ARDS group.

Flow Cytometry and Soluble FasL

To clarify which cells were responsible for the expression of these apoptotic molecules, the immunofluorescence flow cytometric analysis was carried out. A scattergram of BALF cells is depicted in the upper column of Figure 3. Expressions of these molecules were detected only in the population containing mainly the lymphocytes. In the acute ARDS group (n = 6), expressions of granzyme B (Figure 3) and perforin were enhanced (22 ± 5%, ranging from 8 to 38% of cells in the population that were positive for granzyme B and 15 ± 3%, ranging from 6 to 23% for perforin), whereas almost no positive cells were observed in other cell populations. In the late ARDS group and the non-ARDS group, there was no or very weak expression of these molecules in all cell populations. Expressions of Fas and FasL in the cells were weak in all groups. High levels of soluble FasL in the BALF supernatant were detected only in the acute ARDS group (n = 10) (Figure 4). No soluble FasL was detected in other groups (n = 6 in the control and the non-ARDS group, and n = 10 in the late ARDS group).

View larger version (47K): [in this window] [in a new window]   Figure 3.   The representative immunofluorescence flow cytometry for granzyme B. Upper panel shows the scattergram of lavaged cells, clearly separating three populations, lymphocytes, polymorphonuclear neutrophils, and macrophages. The immunofluorescence flow cytometric analysis carried out using each population demonstrated that only the lymphocyte population in the acute phase of septic ARDS expressed granzyme B or perforin. The lower panel depicts one representative flow cytometry; 38% of cells were positive against the anti-granzyme B antibody.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Soluble FasL in the BALF supernatant in each patient group. Only the acute phase of septic ARDS patients expressed high concentration of soluble FasL in the BALF. *p < 0.05 versus control, septic non-ARDS, and late ARDS group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The studies described here suggest that two major apoptosis pathways, perforin/granzyme and FasL/Fas systems, are highly upregulated in the acute phase of lung injury in septic ARDS patients. In contrast, sepsis itself seems to have little effect on these cytotoxic pathways in the lung. An interesting finding was that mRNA expressions of these effector molecules were observed concurrently with expression of inflammatory cytokines and iNOS. Expression of the apoptotic molecules was clearly reduced as the ARDS progressed toward a nonacute state as compared with the expression of the inflammatory cytokines or iNOS.

In the context of lung epithelial and endothelial injury in ARDS, the pathogenesis has been suggested to consist of two pathways; one is from direct effects of an insult on the lung cells and the other is from indirect effects of an acute systemic inflammatory response (1). The latter involves activation of alveolar macrophages and sequestered neutrophils in the lung (1, 2). In fact, the finding that mRNA expression of IL-1, TNF-, IL-6, IL-8, and iNOS was highly upregulated in the ARDS group is consistent with previous reports (12). However, the role of infiltrating lymphocytes or their major products in the pathogenesis of ARDS has never been reported. Our finding that perforin, granzyme A, granzyme B, and FasL were exclusively activated in the early ARDS group strongly suggests that activated lymphocytes, especially CTLs and NK cells, might be involved in the pathogenesis, because these molecules are produced mainly by these cells (15, 16). Perforin and granzyme A and B are the main effector molecules of these cell types and released by the granule-exocytosis pathway (17, 18). Perforin molecules from the tubular transmembrane activate pore formation on the target cells. Then, granzyme A together with granzyme B are introduced into the target cell through the pore and activate the caspase family of proteases leading to apoptosis (19, 20). Thus, in the early phase of ARDS, the perforin/granzyme pathway may be activated to cause lung epithelial and endothelial cell death. Because FasL is preferentially expressed in these lymphocytes and transduces an alternative death-signal to target cells which express Fas, both perforin/granzyme and FasL/Fas pathways might operate to cause destructive cell damage to both the epithelium and endothelium in ARDS. As several reports suggested the potential role of FasL/Fas and perforin/granzyme systems in the pathogenesis of a wide variety of inflammation targeting endothelial or epithelial cells (21), our findings here could be another example in support of cell injury by these effector molecules. Although we observed protein expression of perforin and granzyme in the population of lymphocyte by flow cytometry, we could not detect the definite expression of FasL or Fas in lavaged cells. However, this is more difficult to understand; the detection of soluble FasL might partially explain this outcome. As we only dealt with lavaged cells, it is possible that lung epithelial cells express Fas and reacted with soluble FasL. Relatively weak mRNA expression of FasL in some of the acute ARDS patients might explain the rapid turnover of FasL molecule. Of course, there is an inevitable limitation as we dealt with clinical samples from patients and further basic investigation should be performed.

The presence of these apoptosis molecules in the lung may be very detrimental to lung function if they were excessively induced and expressed in the alveolar epithelial cells. It has been convincingly demonstrated that an intact alveolar epithelial barrier is necessary to prevent alveolar flooding and to recover from ARDS. In a clinical study on ARDS patients, it was demonstrated that patients who were able to concentrate the alveolar edema protein and thus absorb fluid had a much higher chance of recovery and survival than patients who were not able to concentrate the protein (24). Thus, a link between the findings of increased expression of the apoptosis markers and the inflammatory molecules measured in this report and the integrity and function of the alveolar epithelium may exist that could affect the clinical outcome of the patient.

The fact that we could not detect the expression of apoptotic molecules both as mRNA and proteins in the late phase of ARDS suggests that the absence of these molecules might cause the stasis of infiltrating neutrophils which are normally eliminated by apoptosis (25). Furthermore, as an alternative role for the FasL/Fas system, FasL may cause apoptosis in proliferating fibroblasts and endothelial cells that are shed in the process of remodeling of the alveolar structure (7, 8). This is supported by a study using bleomycin-induced lung injury in which FasL and Fas were highly expressed not only in the acute phase of lung injury but also in the resolution phase of pulmonary fibrosis (9). In this scenario, these apoptotic molecules should be essential to protect epithelial cell damage. We speculate that an overexpression of apoptotic molecules in the acute phase of ARDS and the later absence in the subsequent course exacerbate the lung damage. Thus, our study may also validate further studies focusing on apoptosis in the late phase of ARDS.

The clinical benefit of corticosteroids is still controversial and in fact the multicenter trial showed no preventive effect of steroid therapy for ARDS (26). In our study, several patients received large-dose methylprednisolone in the course of ARDS. Nevertheless, as the number of patients studied here was small, it was difficult to determine any beneficial or detrimental effect of this therapy on the mortality in ARDS. However, it is worth noting that the patients in the late ARDS group who received large-dose methylprednisolone (Patients 25-27, 30) showed less expression of inflammatory cytokines mRNA. Because anti-inflammatory genes may also be suppressed by corticosteroids, selective suppression of specific molecules as in anti-IL-8 therapy (27) might be more important for a successful treatment of ARDS. Also, it seemed that the expression of effector molecules was subdued in the time course without regard to steroid therapy.

In conclusion, we observed significant mRNA and protein expression for apoptotic molecules together with proinflammatory cytokines and iNOS in BALF obtained from patients in the early phase of ARDS. These observations warrant further study that apoptosis pathways might be involved in the pathogenesis of acute lung injury.