INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Despite tremendous efforts to develop new therapeutic approaches, the acute respiratory distress syndrome (ARDS) remains associated with a major morbidity and a high mortality. The pathogenic mechanisms that caracterize ARDS are complex and still incompletely understood. The inflammatory response involves both cellular and humoral mediators, including bacterial products, cytokines, lipid mediators of inflammation, and reactive oxygen species (ROS) (1). It is generally postulated that tissue injury occurs as a result of activation of inflammatory cells (peripheral blood monocytes [PBM], alveolar macrophages, and neutrophils), resident or recruited to the lungs, where they lead to an increase in protein permeability and the disruption of basal membranes across both endothelium and alveolar epithelium. Recent studies have shown that mechanical ventilation per se could play a role in initiating or propagating a systemic or intrapulmonary proinflammatory response (2, 3).

ROS have been suggested to play a central role in the pathogenesis of ARDS (4). Although ROS are involved in lesional processes, their increased production also leads to the upregulation of temptative protective antioxidant mechanisms, including the classic scavengers (e.g., superoxide dismutases) as well as the stress protein families. Heat shock/stress proteins (HSP) represent an extremely conserved response to many different types of cellular injuries, including ROS (5). In particular, the inducible, cytosolic, 72 kD HSP (hereafter named Hsp70), is transcriptionally regulated by selected ROS and post-transcriptionally by protein kinases (6, 7). Hsp70 has been shown to confer to cells and tissues protection against the deleterious effects of ROS or cytokines, both in vitro and in vivo (8, 9 for review). In several rodent models of ARDS in which tracheal instillation of phospholipase A2 or caecal ligation and perforation-mediated sepsis leads to a mortality of  30%, in vivo induction of Hsp70 in the lung, using either heat shock or sodium arsenite, is associated with a remarkable decrease in both pulmonary inflammation and mortality (10, 11).

In a number of chronic inflammatory diseases such as asthma or acute conditions such as sepsis, an increased expression of Hsp70 has been observed, the significance of which remains elusive, whether marker of injury or protective mechanism (12). In human ARDS, an increased spontaneous expression of Hsp70 was found both in circulating granulocytes and in alveolar macrophages of critically ill patients as compared with normal donors, though expression levels did not correlate with clinical parameters or outcome. However, only basal expression of the inducible Hsp70 was studied at a single time. Furthermore, septic patients with ARDS were excluded from these studies (15, 16).

We addressed the issue as to whether the expression of Hsp70 in PBM, unlike the granulocyte choice mentioned above, of ventilated patients with ARDS, did correlate with the severity of the disease. As an index of inflammation, we characterized the activation state of PBM by their surface expression of FCR III/CD16 (17). As an index of oxidative stress on circulating blood, we determined plasma lipid peroxidation products. Using flow cytometry on permeabilized cells, both basal expression and inducibility of Hsp70 were examined. Basal expression is considered as indicative of the stress associated with the disease itself, and inducibility as indicative of the adaptative ability in the given stress situation. In order then to distinguish between the effects of the disease and those of ventilation, we also analyzed a group of ventilated patients with ARDS.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

Prospective study in a 24-bed tertiary medical intensive care unit (MICU) of a tertiary referral hospital.

Patient Groups

We prospectively studied mechanically ventilated immunocompetent patients admitted for ARDS in our MICU from December 1996 through May 1997 (Tables 1 and 2). The patients with ARDS (Group 1) (n = 13) met the following diagnostic criteria: Pa_O2/FI_O2 < 200, regardless of positive end-expiratory pressure level, bilateral infiltrates on a frontal chest radiograph, pulmonary wedge pressure < 18 mm Hg, or the lack of echographic evidence of left ventricular failure (18). As an approach to distinguish between disease and ventilation per se, we also studied a group of ventilated patients without ARDS (non- ARDS group). These non-ARDS patients (Group 2, n = 9) required mechanical ventilation and oxygen supplementation for respiratory failure distinct from ARDS or acute lung injury (ALI), or for nonpulmonary disease (Table 2). The control group (Group 3, n = 14) consisted of healthy white caucasian volunteer blood donors without current pulmonary or infectious disease. In particular, the healthy control subjects older than 60 yr of age were devoid of any major age-related disease. Exclusion criteria consisted of age younger than 16, pregnancy, and absence of ventilation.

Sample Timing

Blood samples was taken with concomitant usual blood tests. For each sample, the total duration of mechanical ventilation (MV) and ARDS were precise. The variation of first blood sample realization was explained by several factors as a different duration of previous hospitalization before admission in our institution (tertiary referral center), availability of flow cytometer facilities, and necessity to get fresh blood samples before cell preparation. The time spent in another hospital before being referred to Hopital Cochin greatly varied from case to case and was thus not detailed. Subsequently, first blood samples were taken on average 8 d after ARDS onset in Group 1 (Table 2). Repeated sampling was performed on surviving patients still undergoing mechanical ventilation, with a minimal interval of 1 wk.

Reagents

Ficoll-Paque used for gradient centrifugation was from Pharmacia (Uppsala, Sweden). Saponin, paraformaldehyde, bovine serum albumine fraction V (BSA) were from Sigma Chemical (St. Louis, MO). Phosphate-buffered saline (PBS) without calcium or magnesium, RPMI 1640 medium, fetal calf serum (FCS), L-glutamine, and Hepes buffer were obtained from ICN Biomedicals (Orsay, France). The anti-CD14- and CD16-FITC-conjugated antibody and the secondary FITC-conjugated rabbit antimouse antibody were purchased from DakoA/S (Glostrup, Denmark). The primary antibody used to detect Hsp72, SPA-810, was from Stress Gen (Victoria, BC, Canada).

Measurement of Plasma Lipid Peroxidation Products Level

Lipid peroxidation represents a downstream effect of damage to cell membranes induced by ROS and diffusible lipid hydroperoxides. The latter can decompose into gases such as ethane or pentane and aldehydes such as malondialdehyde (MDA) as by-products. Lipid peroxides were determined as MDA-thiobarbituric on HPLC, as previously reported (19). To adjust the results to the nutritional status of the subjects, the results were expressed as MDA (mmol/L):cholesterol (g/L) ratio. Plasma cholesterol was measured on a Roche centrifugal analyzer (Roche Diagnostics, Welwyn Garden City, UK) using enzymatic reagents (Merck-Clevenot, Nogents/Marne, France).

Cell Preparation and Heat Shock

Fifteen milliliters of peripheral blood were collected in heparinized glass tubes and kept at room temperature until preparation of the cells. Blood was diluted at 1:1 in PBS, and peripheral blood mononuclear cells were isolated by Ficoll-Paque gradient centrifugation. After centrifugation, the pellets were resuspended in 10 ml RPMI 1640 culture medium supplemented with 10% FCS, 2 mM L-glutamine, and 25 mM Hepes. Cell viability was estimated by Trypan blue exclusion and was always above 95%. Cells were resuspended at a final concentration of 5 · 10⁶ cells/ml in supplemented RPMI, and submitted to heat shock (44° C for 20 min by immersion of the cell suspension in a thermostatically regulated water bath), or maintained at 37° C (control cells). Cells were then incubated overnight for 14 h at 37° C in a humidified atmosphere containing 5% CO₂ in air.

Determination of FcR III (CD16) Membrane Expression on PBM

The CD14+/CD16+ cells usually account for about 10% of total PBM, and exhibit features common with mature tissue macrophages (20). These PBM subpopulations expand during in vitro culture and in various clinical situations such as sepsis (17). Cells were isolated as described above and PBM stained on ice with 3 µl antibody CD16-FITC conjugated for 15 min, then washed with PBS containing 1% BSA (PBS-BSA). Cells were analyzed using an EPICS ELITE fluorescence-activated cell sorter (Coulter, Miami, FL). The background level of staining was determined using a FITC-isotype IgG₁ control antibody, and both mean fluorescence (MF) and the percentage of CD16-positive cells were determined.

Determination of Intracellular Hsp70 Expression in PBM

We have recently established that flow cytometry is superior to Western blotting for Hsp70 detection in terms of both sensitivity and quantification: first, flow cytometry allowed us to distinguish between different population expressing variable amounts of Hsp70, and, second, flow cytometry is a more suitable method as compared with Western blotting, in particular when it is necessary to determine more than one parameter at a given time point (21). After a 14-h recovery period at 37° C (22), cell suspensions were centrifuged (5 min at 470 × g) and supernatants were discarded. Briefly, the cell pellets were fixed with paraformaldehyde (3% in PBS) for 10 min and then incubated for 10 min at room temperature with saturating amounts of the monoclonal anti-Hsp70 antibody (diluted to 1/100 in PBS-BSA), in the presence of the permeabilizing agent saponin (0.3%). After washing in PBS-BSA, cells were labeled with the secondary antibody, FITC-conjugated rabbit antimouse IgG₁, diluted to 1/30 in PBS-BSA for 10 min at room temperature. Cells were washed in PBS-BSA and stored at 4° C until flow cytometry was performed. Viable cells were gated on the basis of forward angle and 90-degree light scatter and fluorescence measured with EPICS ELITE fluorescence-activated cell sorter. PBM were identified with anti-CD14, and fluorescence histograms of at least 5,000 counts were obtained for each sample. The background level of staining was determined using a FITC-isotype IgG₁ control antibody. The difference between MF for heat shocked and control cells, i.e., delta MF, stands for the Hsp70 inducibility. An Hsp70 inducibility example in a PBM population of patients with ARDS is shown in Figure 1.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Example of flow cytometric measurement for Hsp70 inducibility over time in one patient with ARDS patient (long-term survivor) (Patient 7). PBM were stained as described, and for each sample, 5,000 cells were analyzed. The histogram dark fluorescence curves represent basal or spontaneous Hsp70 level expression, and the shift clear histogram represents the inducible expression after in vitro heat shock (1, 2, 3). Differences between control and heated cells represent mean fluorescence for Hsp70, a reflection of the cell stress inducibility. In the dot plot, gates are set around the viable CD14+ cell-monocytes (4).

Statistical Analysis

All data were expressed as means ± SD. We compared the results of Hsp70 basal and inducible expression in patients with ARDS versus control subjects, patients with ARDS versus non-ARDS, surviving patients with ARDS versus nonsurvivors with ARDS. Moreover we compared results of repeated sample in a subset of six patients with ARDS. Intergroup comparisons were performed using the non parametric Mann-Whitney test and Kruskall-Wallis analysis of variance for quantitative data as appropriate. Spearman's rank correlation was used for nonparametric correlations and chi-square analysis was used to compare categorical variables. A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical and Usual Blood Parameters

The main clinical and biologic characteristics of the patients are listed in Tables 1 and 2. No significant difference was noted between the three groups regarding age, sex, and smoking history. The simplified acute physiology score (SAPS II) and the mean number of organ failures (OSF) were 41 ± 13 and 1.6 ± 0.9 in Group 1, respectively, 49 ± 14 and 1.8 ± 0.8 in Group 2. There were no statistical differences between Groups 1 and 2 regarding blood leukocytosis, duration of mechanical ventilation before first blood sampling, mortality or disease severity score (SAPS II, OSF) (Table 1). ARDS nonsurvivors (n = 4) exhibited higher mean age and mean plasma lactate level than survivors (n = 9), but no significant differences regarding Murray score and duration of mechanical ventilation on the first blood sampling day. The severity of the disease was also assessed by the values of the mean initial Pa_O2-to-FI_O2 ratio: 172 ± 91 (range, 57 to 350) and the mean Murray score: 2.69 ± 0.7 (range, 1.75 to 4) on the day of first blood sampling. All patients with ARDS initially met the gazometric criteria mentioned above (Pa_O2-to-FI_O2 ratio < 200), either before or at first blood sampling or both.

Plasma Lipid Peroxidation Products Level

In order to evaluate the oxidative status of our subjects, we determined the MDA:cholesterol ratio. This ratio was significantly higher in the ARDS group (Group 1) (2.49 ± 0.86) than in the healthy volunteers group (1.29 ± 0.07, p < 0.02). These results indicate that, as previously described (23), the patients with ARDS included in this study were submitted to an increased oxidative burden (Table 1). The non-ARDS group displayed intermediary levels (1.92 ± 0.61) that were not different from either the ARDS or the control groups.

Membrane Expression of CD16 Receptor

In order to evaluate the activation state of PBM, FcIII R/ CD16 membrane expression was determined as an index of inflammation. The results are shown in Table 1 and indicate that PBM from the ARDS group (Group 1) expressed significantly higher CD16 (79 ± 16%) than did cells from the control group (57 ± 14%, p < 0.005). These results suggest that in patients with ARDS, PBM acquire at least some characteristics of more mature macrophages. Cells from the non-ARDS group expressed intermediary levels (68 ± 10%) (Table 1), that were not different from either the ARDS or the control groups.

Basal Levels of Hsp70 Expression

Basal levels of Hsp70 in PBM from the ARDS group were not different from those from the control group (Figure 2). Basal levels of Hsp70 in PBM from the non-ARDS group were not different from those from the ARDS group, but slightly higher than those from the control group, with marginal statistical significance (p = 0.048). These differences remain unclear, although there was no correlation between individual basal Hsp70 levels in the ARDS group and the Murray score, Pa_O2-to-FI_O2 ratio, or the MDA:cholesterol ratio. No difference was found between survivors and nonsurvivors in the ARDS group.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Basal levels of Hsp70 in PBM from the three groups studied (Group 1: ARDS, Group 2: non-ARDS, Group 3: control). *p = 0.048.

Inducible Levels of Hsp70

As an approach to evaluate the ability of our study population's PBM to respond to an additional stress, we then exposed monocytes to heat shock in vitro. The results of all determinations are expressed as delta MF for Hsp70, i.e., the inducibility of Hsp70. Hsp70 inducibility was significantly lower in the ARDS group (9.09 ± 10.1) than in the healthy volunteers group (19.2 ± 12.3) (p = 0.02) (Figure 3A). As in patients with ARDS, when compared with control subjects, Hsp70 inducibility was significantly lower in the non-ARDS group (9.6 ± 7.5) (p = 0.02).

View larger version (26K): [in this window] [in a new window]   Figure 3.   Inducibility of Hsp70 in patients with ARDS and control subjects. (A) Mean fluorescence, which represents the difference between control and heated cells (inducibility), was significantly higher in the control group (19.2 ± 12.3) than in the ARDS group (pooled samples) (9.09 ± 10.1). (B) The difference in Hsp70 inducibility was no longer found when blood sampling was performed more than 15 d after the initial ARDS diagnosis. *p = 0.02.

Seven patients with ARDS underwent repeated blood collection and determinations of Hsp70 inducibility over time (Patients 1, 7, 8, 9, 11, 12, 13). The difference in Hsp70 inducibility between the subset of the ARDS group with repeated samples and the control groups was not significant (p = 0.2) when blood sampling was performed late in the course of the disease, i.e., more than 15 d after the initial ARDS diagnosis was made (Figure 3B).

Inducible levels of Hsp70 strongly correlated with duration of both ARDS and mechanical ventilation (rho = 0.62, p < 0.003 for the duration of ARDS and rho = 0.604, p < 0.003 for the duration of ventilatory support) (Figures 4A and 4B). Intraindividual correlations between the duration of mechanical ventilation and Hsp70 inducibility in the seven patients with ARDS in whom the latter could be determined over time, are shown in Figure 5. For example, results for Hsp70 inducibility over time are shown in detail for Patient 7, in whom blood sampling could be done six times (Figure 1). In contrast, no correlation was found between Hsp70 inducibility and Murray score, Pa_O2-to-FI_O2 ratio, or MDA:cholesterol ratio. No correlation was found between Hsp70 inducibility and age in the ARDS group, and there was no difference in Hsp70 inducibility in patients with ARDS who died as compared with survivors (data not shown). As an approach to distinguish between disease and ventilation per se, we also studied a non-ARDS group of ventilated patients (n = 9). The strong correlation found in the ARDS group between the duration of ventilatory support and Hsp70 inducibility was lacking in the non-ARDS group (rho = 0.2, p = 0.5) (Figure 4C).

View larger version (17K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   View larger version (16K): [in this window] [in a new window]   Figure 4.   Correlation between duration of ARDS (A) or mechanical ventilatory support (B) and Hsp70 inducibility in PBM. No correlation was obtained in the non-ARDS group (C ).

View larger version (16K): [in this window] [in a new window]   Figure 5.   Individual plots for Hsp70 inducibility (mean fluorescence) during the course of mechanical ventilation for the seven patients with ARDS who underwent two or more Hsp70 determinations. Patient 12 (open circles) showed a decreased level of Hsp70 expression before death. Patient 1 (open squares) dead likewise inverse relationship and Patient 8 (open triangles) are survival.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Here we report that Hsp70 inducibility was significantly lower in PBM from patients with ARDS as compared with healthy volunteers despite higher oxidative stress in ARDS cells, whereas basal expression of Hsp70 was not different. Although no differences were found regarding basal or inducible Hsp70 expression between ARDS and non-ARDS groups, in patients with ARDS, there was a positive correlation between Hsp70 inducibility in PBM and the duration of mechanical ventilation (Figures 4 and 5).

In this study, we used PBM because of their availability as peripheral blood cells, which makes them interesting candidates for future large-scale clinical studies, and of their higher constitutive and inducible Hsp70 expression as compared with either neutrophils or lymphocytes (24). We have previously shown that PBM are activated in subacute endotoxemia in sheep and increasing evidence places circulating monocytes in direct line with resident lung macrophages (25, 26), as is also suggested by the higher percentage of CD16+ PBM in our patients with ARDS versus patients without ARDS and control subjects.

Hsp70 inducibility in PBM was used as a reflection of the ability to mount a stress response in a given situation. We used flow cytometry because this technique requires fewer cells than Western blotting and allows us to determine, on the same blood sample, both basal and inducible HSP expression. Although there is a strong correlation between Western blotting and quantitative flow cytometric analysis for Hsp70 (22), we found that cytometry was superior to Western blotting for Hsp70 detection in terms of both sensitivity and quantification (21). Hsp70 inducibility was tested using a standard in vitro HS procedure, similar for all blood samples. The period of recovery after HS was based on previous studies which demonstrated that Hsp70 expression levels in PBM were highest at this recovery time (21, 22). In animal studies, similar triggers for the stress response, when applied 18 to 24 h prior to sepsis or chemical lung injury, lead to a significantly higher survival rate (10, 11).

We examined the hypothesis that the levels of Hsp70 in PBM could represent a new prognostic factor or a diagnostic biomarker in ARDS. With respect to basal expression of Hsp70 in PBM, our results, however, are not in favor of this hypothesis. If we are postulating that Hsp70 is protective, we might have expected higher levels of Hsp70 at both baseline and after induction in Group 1 as compared with Group 2. These data are in contrast to previous reports which showed that peripheal blood granulocytes of patients with ARDS spontaneously expressed more Hsp72 as detected by Western blotting than did healthy humans (15). These differences might relate to differences in the patient populations (our study included septic ARDS), or in the cells analyzed (PBM versus neutrophils). The results further obtained by Kindas-Mügge and colleagues (16) in alveolar macrophages, showing higher basal mRNA levels for Hsp70 in alveolar macrophages from patients with ARDS, could also relate to cell differentiation. One should also mention that in this study, we tested only for the stress-inducible Hsp70, and thus cannot preclude a difference in basal expression of the constitutive cognate Hsc70 or other forms of Hsp70. We tend, however, to consider that our negative data are an adequate reflection of the interindividual variability in basal Hsp70 expression in circulating cells, for which there is increasing evidence from different laboratories using various techniques (R. I. Morimoto, personal communication; M. Günther, unpublished results). The differences regarding Hsp70 inducibility in PBM might reflect a generally reduced stress response capability during severe disease and could relate to in vivo exposure to LPS and subsequent endotoxin tolerance in our patients. Furthermore, we found that subacute endotoxemia in an ovine model for ARDS (25), also leads to subsequent reduction of Hsp70 inducibility (B. S. Polla and D. R. Morel, unpublished results). These results were recently demonstrated in human PBM from patients with severe sepsis. PBMC from those patients exhibited significantly lower inducible Hsp70 as compared with cells from healthy donors. Interestingly they observed an LPS dose- and time-dependant inhibition of Hsp70 expression in PBMC both from patients and from healthy control subjects (27). Our data in the ARDS group indicate that the ability to produce Hsp70 in response to a stress additive to the disease might be more relevant than basal Hsp70 levels in the follow-up of these patients. The kinetics of increase of the inducible Hsp70 along with the course of the disease may also suggest a repair effect against lung injury. Inducible Hsp70 contributes to protection from and/or repair of ROS-mediated lesions, and there is increasing evidence for the cytoprotective effects of Hsp70 against oxidative stress or hypoxic-ischemia injury in human cell lines or in animal models (14, 28, 29).

As we did not find any correlation between duration of ventilatory support and Hsp70 inducibility in non-ARDS patients, we cannot propose Hsp70 inducibility as a biomarker of mechanical or endothelial-shear stress despite recent convincing datas because our study was initially not designed to analyze this possibilty (30, 31). Our results suggest, however, that the recovery of the ability to mount a stress response relates to an ARDS-specific repair process, although the specificity of Hsp70 value in ARDS remains questionnable as the majority of patients succumb to multisystem organ failure rather than to pulmonary insufficiency itself. Determination of Hsp70 levels and inducibility both in PBM and alveolar macrophages obtained with bronchoalveolar lavage should be performed in further studies to better define the respective role of lung damage as compared with systemic inflammatory response in the results obtained here.

In conclusion, we found that Hsp70 inducibility was significantly reduced in the ventilated patients, whereas basal expression of Hsp70 was not different between patients and control subjects in contrast to previous reports. We propose that when studying HSP in lung diseases, both basal and inducible HSP expression should be considered, as they reflect distinct, complementary variables of the stress response. Among all correlations we considered between Hsp70 inducibility and clinical or laboratory biomakers for disease severity and outcome, the only factors that appeared significant was the duration of the disease and the duration of ventilatory support.