INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hypersensitivity pneumonitis (HP) is an inflammatory lung disease caused by an allergic reaction to a variety of inhaled antigens (1). Sensitization to offending antigens probably occurs only after repeated exposures. The disease is characterized by a large accumulation of T cells in the lungs (2).

As many as 50% of individuals exposed to environmental antigens that can cause HP develop a lymphocytic alveolitis but remain asymptomatic (3), whereas only a few develop clinical symptoms of the disease (estimated at three in 1,000). Cigarette smokers have a very low incidence of HP as compared with nonsmokers (4), whereas cofactors, such as a viral infection or endotoxin exposure, might enhance the risks of developing the disease. We have previously shown that after an infection with Sendai virus, mice become more responsive to antigens of Saccharopolyspora rectivirgula, the bacterium responsible for farmer's lung. This enhanced response persisted for up to 6 mo after the viral infection (5). Similarly, Fogelmark and colleagues have shown that the reaction to -glucans was increased by the addition of endotoxins producing a lung response similar to the one observed in HP; when given separately, -glucans and endotoxins produce a nonspecific inflammation (6).

T lymphocytes play an important role in the pathogenesis of HP. The number and percentage of T cells is increased in the bronchoalveolar lavage fluid (BALF) of patients with HP, and the latter usually constitutes 60 to 80% of the recovered cells (7). These cells are activated (8, 9), react specifically to causative antigens (10) and, in animal models, are able to transfer sensitization to naive animals (11). Since most T cells in HP are phenotypically and functionally activated, we investigated the mechanism by which T-cell activation preferentially occurs in HP.

The differentiation and activation of T cells requires the recognition by the T cell receptor of a foreign antigen on antigen-presenting cells (APC), and in addition the ligation of costimulatory molecules CD80 (B7-1) and CD86 (B7-2) on APC to two homologous receptors on T lymphocytes: CD28 and CTLA4 (12). CD28 is expressed on resting and activated T cells, whereas CTLA4 is induced after T-cell activation. Signaling through CD28 is thought to promote T-cell proliferation (13), whereas signaling through CTLA4 may induce a negative regulation that is presumed to control inappropriate T-cell expansion (14).

CD80 and CD86 are expressed by potent APC including dendritic cells, monocytes, macrophages, and activated B cells (15). Alveolar macrophages (AM) of the normal lung have a low level of expression of B7 molecules and a poor capacity to function as APC (16). In diseases with lymphocytic alveolitis, such as HP and sarcoidosis, AM are activated and show an increased antigen-presenting capacity as compared with AM from normal subjects (17). We have also previously shown that AM of patients with HP have a potentiating effect on lymphocyte proliferation, contrasting with the normal suppressive activity of AM from healthy subjects (18). In the present study, we hypothesized that an upregulation of the expression of B7 molecules on AM may account for the activation and proliferation of T lymphocytes in HP, and that this expression could be modulated by viral infection and cigarette smoking. The aims of the study were to compare the expression of CD80 and CD86 on the AM cell surface of patients with active HP, asymptomatic antigen-exposed subjects, normal nonsmoking subjects, and smokers, and to assess the effect of an in vitro viral infection on the expression of these molecules on AM from normal nonsmokers and from smokers.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

The study population included 18 patients (11 male and seven female) with active HP (nine with farmer's lung, four bird fanciers, two peat moss workers, and three with humidifiers lung (mean age: 45 yr; range: 16 to 77 yr); 10 asymptomatic, antigen-exposed male subjects (seven dairy farmers and three peat moss factory workers; mean age: 35 yr; range: 19 to 64 yr), 18 normal nonsmokers (12 male and six female; mean age: 26 yr; range: 20 to 35 yr); and 12 healthy smokers (eight male and four female; mean age: 29 yr; range: 20 to 51 yr). Smokers smoked from 10 to 35 cigarettes per day and had smoked from 5 to 15 yr. The diagnosis of HP was based on previously described criteria (19). The asymptomatic farmers recruited to participate in the study had normal lung function and chest radiographs. Both groups were from family dairy farms of similar size. As a group, the asymptomatic subjects (AS) were younger and therefore had been exposed to this environment for a fewer number of years. However, age is not a risk factor for HP, and two of the HP patients were 16 yr old. All HP and AS subjects were nonsmokers. None of the subjects was taking any steroids or other antiinflammatory medication at the time of the study.

Bronchoalveolar Lavage

All patients and volunteers underwent fiberoptic bronchoscopy and bronchoalveolar lavage (BAL) with 300 ml of lavage solution. With the subject under local anesthesia, a 5.5-mm O.D. fiberoptic bronchoscope was advanced through the mouth and into the trachea, and was wedged into a segmental or subsegmental bronchus. The wedged lung segment was lavaged with five aliquots of 60 ml each of normal sterile saline prewarmed to 37° C; the fluid was gently aspirated after each aliquot. From 52 to 65% of the lavage fluid was recovered and centrifuged, and the resulting BALF cells were washed three times, after which cell counts were made in a hemocytometer. A significant difference in the amount of recovered BALF was found for patients with HP versus normal subjects (p = 0.008). No differences between the other groups were found. To standardize the results, the number of cells is expressed per milliliter of recovered BALF. Cell differential counts were performed after Diff-Quik (Dade Diagnostics, Aguada, PR) staining on glass coverslips (20).

Flow-Cytometric Analysis

BALF cells were washed in phosphate-buffered saline with 1% bovine serum albumin. Fc receptors were blocked by incubation with an excess of human IgG (10 µg/10⁶ cells) for 20 min at 4° C. Cells were then washed and further incubated with phycoerythrin (PE)-labeled anti-CD80 or anti-CD86 (Pharmingen, Mississauga, ON, Canada) for 45 min at 4° C. PE-labeled isotype mouse immunoglobulins were used as negative controls. BALF lymphocytes were double stained with peridinin chlorophyll protein (PerCP)-labeled anti-CD3 and PE-labeled anti-CD28 or anti-CTLA4 antibody.

Cells were analyzed in an EPICS Elite ESP flow cytometer equipped with a 488-nm argon laser (Coulter, Hialeah, FL). As shown in Figure 1, AM and lymphocytes were gated on a logarithmic scale according to their different light-scattering properties: forward scatter, which reflects cell size, and side scatter, which depends on intensity and diffraction capacity of the cells (complexity/granularity). In some preliminary experiments, sorting of gated cells allowed assessment of the purity of the AM population by Diff-Quik staining; the purity was greater than 95%. From 15,000 to 20,000 gated cells were analyzed. The gates chosen to define the percentage of positive cells were established in order to have less than 2% positive cells with the isotypic IgG control.

View larger version (37K): [in this window] [in a new window]   Figure 1.   Flow cytometric distribution of BALF cells from an asymptomatic antigen-exposed subject. Patients with HP had more lymphocytes, normal nonsmokers and smokers had more macrophages. Cells are distributed according to their size (forward scatter [FS]) and their complexity/granularity (SS). AM and lymphocyte populations are delineated with the rectangle used to analyze them selectively.

In Vitro Viral Infection

BALF cells from normal nonsmokers and smokers were exposed for 18 to 20 h to influenza virus (H1N1) obtained from the American Type Culture Collection (Rockville, MD) at a multiplicity of infection of 1, and were incubated at 37° C in RPMI medium +10% fetal bovine serum. Infection was confirmed by detection of influenza virus RNA with a reverse transcription-polymerase chain reaction (RT- PCR) (data not shown). The expression of CD80 and CD86 molecules was measured by flow cytometry as described earlier. Exclusion of nonviable cells was assessed by incorporation of propidium iodide.

Statistical Analysis

Means and SD were determined for continuous variables. Two analyses were performed. The first was a comparison of patients with active HP, AS, normal nonsmokers, and smokers. The statistical approach used to perform this analysis was a one-way analysis of variance (ANOVA) with a factor representing the group effect (the comparison between the four groups). Normality and variance assumptions were tested. The second analysis was a comparison of the percent of CD80⁺ and CD86⁺ cells with and without viral infection. This comparison was done separately for nonsmokers and smokers, using Student's paired t test or Wilcoxon's signed ranks test to evaluate the virus effect. Relationships between percent lymphocytes and CD86 were expressed with Spearman's correlation coefficients, since linearity of the relationships between these parameters was not observed. The results were considered significant at p  0.05. Data were analyzed with the SAS statistical package program (SAS Institute Inc., Cary, NC).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BAL

The total number of cells in the recovered BALF in each of the different groups is given in Figure 2. The total number of cells in patients with active HP (788 × 10³ cells/ml) was greater than in normal nonsmoking subjects (85 × 10³ cells/ml) (p = 0.0001). AS had intermediate values (350 × 10³ cells/ml) that were significantly different from those of patients with acute HP (p = 0.015) and normal subjects (p < 0.0001). The number of cells recovered from the BALF was also higher in smokers than in normal nonsmokers (329 × 10³ cells/ml, p < 0.0001). The cell characteristics of the different groups are also given in Figure 2. In accord with previous reports, the percentage of lymphocytes was markedly increased in patients with HP (64 ± 0.3%), was intermediate in AS (43 ± 4.6%) as compared with nonsmokers (13 ± 2%), and was much lower in smokers (4.3 ± 1.1%). AM were the predominant BALF cell population in normal volunteers and smokers, with values of 84 ± 2.3% and 95 ± 1.3%, respectively.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Number of total cells and subpopulations expressed per milliliter of recovered BALF (mean ± SEM) for the various groups. HP = hypersensitivity pneumonitis, AS = asymptomatic antigen-exposed subjects, NS = normal nonsmokers. Hatched columns = total cells, black columns = AM, white columns = lymphocytes, gray columns = neutrophils.

Phenotypic Analysis of AM

The percentages of CD80⁺ and CD86⁺ AM for the different groups of subjects are presented in Figure 3. For all groups, CD86 expression was higher than that of CD80. The percentage of CD86⁺ AM was higher in HP patients than in nonsmokers (62.3 ± 5.9% versus 24.2 ± 2.8%, p = 0.0001). AM from AS had intermediate values (38.1 ± 9.2%). The percent of AM for smokers expressing CD86 was significantly lower than for nonsmokers' AM (p < 0.001). No relationship between age and expression of these molecules was found within the different groups.

View larger version (20K): [in this window] [in a new window]   Figure 3.   Percentage of AM from the various groups expressing CD80 and CD86. The data are presented as percentage of CD80+ and CD86+ cells with a gate with which less than 2% of control mouse isotype antibody was detected. Columns with different letters are significantly different. For CD80, no significant difference was detected between the HP and AS groups (p = 0.33). A significant difference was found between HP patients and nonsmokers (p < 0.001) as well as smokers (p = 0.002).

The proportion of CD86⁺ cells correlated with the percentage of lymphocytes recovered in the BALF for AS (r = 0.87, p = 0.0009; Figure 4). No correlation was found in the HP group (r = 0.192, p = 0.548).

View larger version (11K): [in this window] [in a new window]   Figure 4.   Correlation between the percent of CD86+ cells and the percent of lymphocytes in lavage fluid (Spearman's regression analysis). No correlation was found in the group with active HP (n = 12), but a good correlation was observed in the AS group (n = 10).

Phenotypic Analysis of BALF Lymphocytes

CD28 and CTLA4 expression were measured on the surfaces of lymphocytes from patients with active HP, asymptomatic subjects, and normal nonsmoking subjects. This analysis could not be performed in smokers because very few lymphocytes were detected in the BALF from these subjects (Figure 5). Equal levels of CD28 expression were found on BALF lymphocytes from the three groups of subjects. A very strong expression of CTLA4 was detected on a high percentage of BALF lymphocytes from normal subjects (98.83 ± 0.49% positivity), AS (88.84 ± 10.68%), and HP patients (93.91 ± 4.65% positivity).

View larger version (67K): [in this window] [in a new window]   Figure 5.   Percent of CD28+ and CTLA4+ lymphocytes in the BALF of HP patients, AS, and normal subjects. CTLA4 (hatched columns) was expressed on most BALF lymphocytes from the three studied groups. Similarly, no significant difference was detected for CD28 expression (gray columns).

In Vitro Viral Infection

As illustrated in Figure 6, the proportion of CD86⁺ cells was significantly increased after in vitro infection of AM from normal nonsmokers with influenza virus (p = 0.0064). A slight but not significant increase in CD80⁺ cells was observed after this infection (p = 0.07). No changes in the proportion of CD80⁺ or CD86⁺ cells were seen after infection of AM from smokers with influenza virus. These experiments were not performed with AM from patients or AS, since a highly upregulated expression of the costimulatory molecules was already observed.

View larger version (21K): [in this window] [in a new window]   Figure 6.   Effect of influenza virus infection on CD80 and CD86 expression on AM from normal nonsmokers (n = 9) and smokers (n = 9). White columns = before influenza virus infection, gray columns = 24 h of viral infection. Virus-infected AM from normal subjects had a small but not significantly increased expression of CD80 and a significantly enhanced expression of CD86. The viral infection produced a small but not significant effect on costimulatory molecule expression on AM from smokers (p = 1.00 and p = 0.16 for CD80 and CD86, respectively). Within groups, columns with different letters are significantly different.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study clearly show a marked difference in the expression of costimulatory molecules of the B7 family (CD80 and CD86) by AM of patients with active HP and AS with a lymphocytic alveolitis, as compared with normal nonsmokers and smokers. These findings are not surprising, given the lymphocyte activation associated with HP. Similar results were obtained in other interstitial lung diseases such as sarcoidosis (17). For all of the groups in the present study, the percentage of CD86⁺ cells was consistently higher than that of CD80⁺ cells. This is consistent with the rapid upregulation of CD86 expression in response to several stimuli, whereas CD80 is induced later and at lower levels on APC (13). An almost complete lack of expression of costimulatory molecules was observed in smokers, suggesting an active suppression of AM activity. Another possibility is that smoking induces the recruitment to the airways of a subpopulation of AM that fail to express costimulatory molecules.

Of particular interest was the ability of influenza virus to enhance the expression of these CD80 and CD86 on AM from nonsmokers. Viral infections in other systems upregulate B7 molecules in vitro (21, 22), and viral infections in vivo enhance expression of these molecules in mice (23). In the present study, a fixed number of viral particles was added to the same number of AM from normal nonsmokers and smokers. We did not quantify the viral loads present after the incubation period, but infection was assessed by the detection of influenza virus RNA in AM, using the RT-PCR method. It therefore remains possible that AM from smokers were less infected than those from normal subjects; however, previous studies have reported that AM from smokers are as susceptible to viral infection as normal macrophages, and produce even more viral particles than normal AM (24). We therefore do not believe that the lack of enhancement of the expression of costimulatory molecules in AM of smokers in our study was caused by lack of infection of these cells.

Influenza virus infection of monocytes and macrophages induces a full range of cytokine gene transcription and subsequent release of cytokines such as tumor necrosis factor- (TNF-), interleukin (IL)-1, IL-6, interferons, and C-C chemokines (25). Because cytokines can regulate the expression of costimulatory molecules (15), one could speculate that the defective regulation of B7 expression after viral infection in AM from smokers was due to defective cytokine regulation. Indeed, altered cytokine regulation in the lungs of smokers has been reported (26). Human immunodeficiency virus (HIV)- infected AM have an increased accessory function as compared with normal AM, as well as increased IL-1 and IL-6 release, whereas HIV-infected patients who smoke show impaired AM accessory function and IL-1 and IL-6 secretion (27). Similarly to what is observed in HP, lymphocytic alveolitis is also less common in HIV-infected smokers. These data also fit well with our recent observation that AM from HP patients have lost their normal lymphosuppressive function, and actually increase the proliferation of mononuclear cells in response to PHA stimulation (18).

The close correlation between the percent of CD86⁺ cells and percent of BALF lymphocytes observed in AS but not in HP patients could indicate that the induction of these receptors was still under immunoregulatory control in the AS group. The induction of CD86 receptors may depend on the duration or intensity of antigen exposure. Alternatively, a viral infection may increase the accessory function of AM and contribute to the appearance of active HP. No correlation between CD86 expression and percent lymphocytes was seen in active HP, perhaps because a maximal B7 molecule induction is already present.

Given the differences in the structures and kinetics of ligation of CD80 and CD86, different functions have been attributed to these costimulatory molecules. In many experimental systems it was reported that CD80 might preferentially influence the induction of a type 1-like lymphocyte response (characterized by production of IL-2 and interferon- [IFN-]), whereas CD86 promotes a type 2-like response (characterized by production of IL-4, IL-5, and IL-10) (28). However, other studies, both experimental and human, failed to demonstrate such a dichotomy (29). In HP, no clear pattern of cytokine profile has been found. In the experimental murine model, a Th1 pattern seems to be predominant, since CD4⁺ Th1 cells are able to adoptively transfer experimental HP (30); moreover, other studies have shown that IFN- and IL-12, both of which are Th1 cytokines, are essential for the expression of HP in mice (31, 32). However, in human HP, a cellular BALF profile with a predominance of Th2 lymphocytes was described (33).

Analysis of CD28 on BALF T lymphocytes from HP patients, AS, and normal subjects revealed that this receptor is equally expressed in all three groups. This is not surprising, since these molecules are present constitutively on most T cells. This analysis could not be done in smokers because too few lymphocytes were detected in this group. More surprisingly, we found a very high level of expression of CTLA4 on BALF T lymphocytes from all of our subjects including normals. This could be explained by the negative regulatory function of CTLA4, and may provide a protective mechanism against damaging immune reactions in the normal lung. CTLA4 was also highly expressed on BALF T lymphocytes from patients and AS, indicating that a negative signal is still present, but since B7 molecule expression is also upregulated in these groups, one could speculate that the amplification signal is greater than the inhibitory signal.

B7-CD28 ligation increases T-cell survival by upregulating the expression of the intrinsic cell survival factor Bcl-x_L, which confers resistance to apoptosis (34). Further studies need to be done to confirm that this occurs in HP.

The results reported here, taken with the fact that cigarette smoking clinically decreases the risk of HP and that a viral infection upregulates the response to HP antigens (4, 5), further support the important role of antigen presentation by AM in the pathogenesis of HP. More functional studies are needed in which the costimulatory pathways are blocked in order to document the effect of these interventions on the response of the lung to antigenic challenges in an experimental model of HP.