Introduction

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Silicosis, a chronic diffuse interstitial lung disease, develops in workers exposed to the inhalation of respirable crystalline silica mineral particles. Industrial hygiene measures that control the concentration of airborne dust in the workplace have succeeded in reducing the prevalence and severity of silicosis in many industries, but it remains a common health problem in developing nations throughout the world. Experimental exposure of mice or other animals to silica inhalation offers the means to study the cellular events leading to human silicosis, and may serve to elucidate the more general pathophysiologic mechanisms of lung injury, inflammation, and fibrosis.

Human silicosis is characterized by initial sites of macrophage and lymphocyte aggregation associated with the accumulation of mineral particles. These early silicotic lesions appear to enlarge over time with progressive localized proliferation of fibroblasts and deposition of excess collagen, forming silicotic nodules. As the process continues, these isolated lesions fuse to form the large inflammatory fibrotic conglomerates of progressive massive fibrosis.

In both humans and animals with silicosis, lymphocytes accumulate in lung tissue, in bronchoalveolar lavage (BAL) fluid (BALF), within intrapulmonary lymphoid aggregates and bronchial-associated lymphoid tissue (BALT), and in enlarged draining lymph nodes. These lymphocytes are mostly T cells, with enrichment of CD4⁺ and, to a lesser degree CD8⁺, phenotypes (1). The T cells appear to be activated, express the interleukin (IL)-2 receptor, and show enhanced spontaneous DNA synthesis (6). The nodular lesions of silicosis bear features resembling granulomas, suggesting that a T-helper (Th)-1-like lymphocyte subpopulation might dominate in this disease (7). Further, the interferon (IFN)- produced by such a lymphocyte population would be a candidate cytokine to participate in the macrophage activation characteristic of silicosis (10, 11). IFN- is a cytokine with broad biologic effects; it is produced by lymphocytes, particularly by natural killer (NK) cells and by subsets of T cells (12). IFN- is recognized as the foremost important cytokine in converting macrophages from a resting to an activated state.

Our recent reports have described the development of silicosis in mice exposed by aerosol inhalation to cristobalite silica, with the extent of disease related to the ambient concentration of dust and to the duration of exposure (15). We reported that mice with silicosis overproduce IL-1 and tumor necrosis factor (TNF)-, and that the expression of these cytokines occurs predominantly within the silicotic lung lesions and in BALT (16). We described overproduction of IFN- by lymphocytes in mice with silicosis, with no change or a slight decrease in the abundance of IL-4 in these animals (17). Intracellular IFN- was detected in small mononuclear cells from the lungs and the spleens of air-sham control mice, as well as in the mice with disease. By flow cytometry, the fraction of cells producing IFN- was increased significantly in mice with silicosis from both the lung (19.3 ± 6.0% versus 11.3 ± 1.9%, P < 0.03) and the spleen (6.3 ± 1.0% versus 4.0 ± 0.7%, P < 0.003) (17).

In the current research we sought to determine the cellular mechanisms through which IFN- was produced in the lungs of mice with silicosis. The present studies were undertaken to test two alternative hypotheses: (1) that the increase in the number of small mononuclear cells producing IFN- in silicosis is due to the expansion of a specific subpopulation of lymphocytes that may be unique or distinctive, or (2) that the increase in the number of small mononuclear cells producing IFN- in silicosis is due to parallel expansion of the multiple cell types that produce this cytokine in the normal lung. This research was also designed to address several important questions for which detailed answers were not available in the literature: What is the mixture of lymphocyte phenotypes in the normal lung? How does the proportion of lymphocyte phenotypes in the normal lung compare with those in an extrathoracic peripheral lymphoid organ? How does the proportion of lung lymphocyte phenotypes change in silicosis? Does the mixture of lymphocyte phenotypes in the spleen (a peripheral lymphoid organ remote from direct silica contact) change in pulmonary silicosis?

We focused attention on mice with well-developed silicosis, 16 to 20 wk after a 5-h/d, 12-d inhalation exposure to 70 mg/m³ cristobalite silica (15) to test the hypothesis and address the questions listed earlier. Previous experiments indicated that by 16 wk after mineral inhalation the silica-exposed mice demonstrated lung pathology with focal lesions consisting of mononuclear cell infiltrates and fibroblasts, alveolar debris, perivascular and peribronchial lymphocytic aggregates, and enlargement of BALT structures. BAL cell numbers were increased 3-fold, with significant enrichment in the percentage of lymphocytes. Total lung collagen (hydroxyproline) was increased significantly (15). We have used this model system to reveal novel aspects of the murine lung lymphocyte population and to gain insight into the cellular mechanisms of increased IFN- production in silicosis.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Silica Exposure of Mice

The methods for exposure of mice to silica by inhalation aerosol have been reported in detail elsewhere (15). Mice (Mus musculus) of the strain C3H/HeN were obtained at 5 to 6 wk of age through the National Institutes of Health, supplied by Harlan Sprague-Dawley (Indianapolis, IN). Exposure to cristobalite silica (C&E Minerals, King of Prussia, PA) by aerosol was carried out using a horizontal flow inhalation toxicology system (15, 18- 20). The mice were divided into air-sham control and silica-exposed groups. The two groups were exposed simultaneously to the high efficiency particulate air (HEPA)-filtered, humidified, temperature-controlled carrier airstream or to the mineral dust in two identical chambers for 5 h/d for 12 d. For the experiments described in this report the mice were exposed in two separate procedures to ambient cristobalite concentrations of 70 ± 19 mg/m³ or 70 ± 11 mg/m³ (mean ± standard deviation of the mean [SD]).

Isolation of Lung Cells and Spleen Cells

Lung mononuclear cells were isolated by collagenase enzyme digestion and gentle mechanical disruption using a method we have reported recently (17), a modification of published techniques (3, 21). Mice were killed with sodium pentobarbital, the trachea was cannulated, the thorax was opened, and the lungs were reinflated with a solution of collagenase Type I (324 U/ml; Sigma Chemical, St. Louis, MO), bovine pancreatic DNase I (75 U/ml; Sigma), and porcine heparin (25 U/ml; Elkins-Sinn, Inc., Cherry Hill, NJ) in Dulbecco's phosphate-buffered saline with Mg²⁺ and Ca²⁺ (Life Technologies, Grand Island, NY). The lungs were perfused with the enzyme solution and the lobes were excised, minced, agitated in the enzyme solution for 1 h at 37°C, gently sheared by aspiration through a 14-gauge cannula, and strained through nylon mesh (Falcon Plastics, Franklin Lakes, NJ). For each experiment the lungs from two or three mice were pooled for digestion. The lung cell suspension was centrifuged at 225 × g for 10 min at 5°C, resuspended in RPMI 1640 medium (RPMI) with 5% fetal calf serum (FCS) and with 100 U/ml penicillin and 100 µg/ml streptomycin (P-S) (Life Technologies), and counted by visual hemocytometer.

Spleen cell suspensions were prepared by pressing the spleen between the frosted ends of glass microscope slides to disrupt the tissue by gentle shearing pressure, and were rinsed into RPMI/FCS. For each experiment the spleen cells from two or three mice were mixed, debris was allowed to settle for 5 min at 5°C, the supernatant cell suspension was removed and centrifuged at 225 × g for 10 min at 5°C, the cell pellet was resuspended in 1 to 2 ml of RPMI/FCS, and the leukocyte cell number was counted by hemocytometer.

The lung cells or splenocytes were cultured with brefeldin A to block the secretion of cytokines and allow these peptides to accumulate within the cells for staining (25, 26). For each flow cytometry sample, an aliquot of 1 × 10⁶ lung or spleen cells was placed in a 22-mm-diameter culture plate well (Corning/Costar, Cambridge, MA) with 1.0 ml of RPMI/FCS/P-S. Brefeldin A (Sigma) was prepared as a 100× stock solution in ethanol and added at a final concentration of 10 µg/ml. The cells were cultured for 4 h at 37°C in a 5% CO₂ environment, the plates were struck sharply to dislodge loosely adherent cells, the cell suspension was aspirated into 12 × 75-mm polystyrene tubes, and the plates were rinsed with additional medium. The cell suspension was centrifuged at 225 × g for 5 min at 5°C to pellet the cells for staining.

Propidium Iodide Staining for Cell Viability

Samples of 2 × 10⁶ washed cells were resuspended in 3.0 g/liter polyethylene glycol-8000, 5 mg/liter propidium iodide (PI), 18,000 U/liter ribonuclease (RNase) (Worthington Biochemicals, Lakewood, NJ), Triton-X 100 0.1%, and 0.4 mM sodium citrate in phosphate-buffered saline (pH 7.40) with 0.1% sodium azide (PBS/A), and incubated for 20 min at 37°C. The cells were centrifuged and the pellet was again suspended in the mixture just described, but without RNase or sodium citrate and with 40 mM sodium chloride; the cells were then incubated at 4°C for 30 min or more and analyzed by flow cytometry.

Surface Antigen Phenotype Staining

The cell pellet in each tube was resuspended in 100 µl of RPMI/ FCS with rat polyclonal immunoglobulin (Ig)G (Zymed, San Francisco, CA) at 50 µg/ml to block nonspecific adherence of staining antibodies, incubated for 10 min in the dark at 5°C, and centrifuged, and the supernate was discarded. The pellet was then suspended in 100 µl of RPMI/FCS containing a mixture of the fluorochrome-labeled surface marker antibodies or with appropriate Ig isotype controls. The antibodies used are summarized in Table 1. Cells were stained for 30 min in the dark at 5°C, diluted in 2.0 ml of PBS/A, and centrifuged.

Intracellular Cytokine Staining

Staining for intracellular cytokines was performed using modifications of published methods (26) as we reported recently (17). Fixation, permeabilization, and staining were performed at room temperature. After staining for surface antigen phenotype markers, the cell pellet was again resuspended in 50 µl of PBS/A, 100 µl of fixation medium was added (Fix & Perm, Reagent A; Caltag, Burlingame, CA), and the pellet was incubated for 15 min. The mixture was diluted with 2.0 ml of PBS/A, and the cells were pelleted by centrifugation at 350 × g for 5 min. The fixed cell pellet was suspended in 100 µl of PBS/A with polyclonal rat IgG at 50 µg/ml to block nonspecific adherence of staining antibodies, incubated for 10 min, diluted with 2.0 ml of PBS/A, and centrifuged. The pellet was then mixed with 100 µl of saponin permeabilization solution (Fix & Perm, Reagent B; Caltag), 20 µl of antibody against IFN- labeled with r-phycoerythrin (PE) or fluorescein isothiocyanate (FITC) (PharMingen, San Diego, CA) or the isotype control was added, and the sample was incubated for 15 min in the dark. The suspension was diluted with 2.0 ml of PBS/A, centrifuged, resuspended in 350 µl of 1% paraformaldehyde (EMS, Ft. Washington, PA) in PBS/A, and held at 5°C for up to 24 h until analysis by flow cytometry.

Cytofluorograph Analysis

The cell samples were examined in a dual-laser four-color Coulter EPICS Elite cytofluorograph using Coulter Elite analysis software, evaluating 40,000 or more cells per sample. Each sample population was classified for cell size (forward scatter) and complexity (side scatter), and a gate of interest (Figure 1, Gate A) was drawn around the viable small mononuclear cell population, excluding debris and erythrocytes. Each cell within this gate was categorized for fluorescence intensity in the color channel relevant for each of the surface antigen or cytokine antibody fluorochromes. The results were expressed as the percentage of cells within the size/complexity gate of interest that stained positively for each marker after subtracting the percent positive cells in the isotype control. The results were obtained from two separate inhalation exposure sequences. The cells from two or three mice were mixed, stained, and analyzed, and the values from one to five replicates of each stain combination were averaged to generate the values for each experiment. The mean ± SD (or range) of two to four replicate experiments are reported herein. The significance of differences in cell recovery between groups was assessed by Students' t test (Systat 7.0; SPSS, Inc., Chicago, IL). Differences in the fractions of lymphocyte phenotypes between exposure groups were tested for significance ( = 0.05) by analysis of variance (ANOVA) using the data from two to four replicate experiments.

View larger version (40K): [in this window] [in a new window]   Figure 1.   Flow cytometry diagrams. Flow cytometry diagrams illustrate typical results for the cell populations and phenotypes examined. Cells were examined in a dual-laser four-color Coulter EPICS Elite cytofluorograph evaluating 40,000 or more cells per sample. The histograms in panels b through f display cell number (ordinate) categorized by fluorescence brightness (abscissa) with a horizontal bar defining positive cells with fluorescence above the isotype control antibody level. (a) Size and granularity characteristics of the lung digest cells. The distribution of cells displayed by size (forward scatter) and complexity (side scatter) is shown with Gate A drawn around the region containing small mononuclear cells with the features of lymphocytes. Approximately 42 to 46% of lung digest cells fell within this gate. Subsequent analyses were performed on the cells within this region. A sample from two air-sham control mouse lungs is shown. (b) Cells containing IFN-. Cell samples were stained with fluorochrome-labeled anti-IFN- antibody after cytokine secretion blockade with brefeldin A and permeabilization with saponin. Negative and positive cells show a bimodal distribution, with cells staining positively for intracellular cytokine demonstrating fluorescence brightness above the isotype control. A sample from two silica-exposed mice is shown. (c) Expression of CD4 on lung lymphocytes. Cells expressing the CD4 adhesion molecule appear as a peak of bright cells beneath the horizontal bar. CD4+ cells within this gated region were then selected for analysis of additional features such as TCR epitopes or IFN- production ( f ). (d) Expression of -TCR on lung lymphocytes. Cells expressing the -TCR phenotype appear as a secondary peak of bright cells beneath the horizontal bar. (e) NK cells among lung lymphocytes. NK cells were identified as cells staining brightly with the DX5 antibody and appear as the peak beneath the horizontal bar. ( f ) IFN- production by CD4+ lung cells. Small mononuclear cells were gated for CD4+ expression (c) and simultaneously examined for the presence of intracellular IFN- (b). The CD4+ IFN-+ cells are shown beneath the horizontal bar. This sample is from a pool of two silica-exposed mice.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We examined mononuclear cells isolated from the lungs and the spleens of C3H/HeN mice 16 to 20 wk after inhalation exposure to cristobalite silica (70 mg/m³, 12 d, 5 h/d) or sham-air without the mineral dust. This interval after exposure was chosen because previous experiments demonstrated well-developed silicosis at this time, with significantly increased production of IFN- (15, 17). Lung mononuclear cells were prepared by enzyme digestion and gentle shearing of tissue to recover lymphocytes from the interstitium, BALT, and parenchymal lymphoid aggregates as well as the air spaces of the lung. Spleen cells were prepared by shearing disruption of the organ. Spleen cells were used to compare lymphocyte populations in the lung with those of a remote reticuloendothelial organ to determine whether changes that occurred in lymphocyte populations in direct proximity to silica were paralleled by systemic alterations. As summarized in this paper's opening paragraphs, our previous research demonstrated in the lungs and spleens of mice with silicosis an increased proportion of lymphocytes producing IFN- (17). Our current research was designed to determine whether the mechanism of increased production of IFN- was attributable to a single lymphocyte phenotype.

The total number of cells recovered from lung tissue was increased significantly in mice exposed to silica (8.4 × 10⁶ ± 2.0 × 10⁶ cells/mouse, 12 mice, 7 experiments), as compared with air-sham control animals (4.9 × 10⁶ ± 1.4 × 10⁶ cells/mouse, P < 0.01). The number of cells recovered from the spleen was similar to that from the silica-exposed (15.8 × 10⁶ ± 9.5 × 10⁶ cells/mouse, 17 mice, 10 experiments) and the air-sham control mice (13.3 × 10⁶ ± 5.0 × 10⁶ cells/mouse, P = NS).

The small mononuclear cells of interest were gated within the flow cytometry diagram according to their size (forward scatter) and complexity (90-degree side scatter), as shown in Figure 1. The gates used for the spleen cells and lung cells were nearly identical in all experiments. The cells falling within each gate represented a slightly smaller fraction of the total cells for the silica-exposed mice than for the control animals (silica 41.4 ± 4.8% versus control 46.1 ± 6.3%, P < 0.04). The fraction of spleen cells encompassed within the gate was not different between air-sham control and silica exposed mice (68.5 ± 4.5% versus 66.5 ± 4.7%, P = NS).

Cell Surface Antigen Phenotypes in Normal Mice

Lung digest cells from control mice were examined to determine the distribution of major phenotypes within the population gated as small mononuclear cells. The phenotypes were examined in six separate experiments in which the lungs of two or three mice were pooled for the experiment. Each phenotype was enumerated in two to four experiments. The data are presented as means ± SD of the results from each experiment. CD4⁺ cells represented 15.5 ± 4.4% of the gated cells, and CD8⁺ cells comprised 6.0 ± 1.5%, a CD4:CD8 ratio of 2.8 ± 1.4. NK cells were stained as 14.5% of the total, whereas T-cell antigen receptor (TCR)⁺ cells comprised 4.2 ± 1.9% of the population. B lymphocytes (CD45R⁺/ B220⁺) formed 5.9 ± 2.4% of the lung digest small mononuclear cells. Macrophages (MAC3⁺) were 12.3% of the gated population. A total of 58.4% of the lung cells within each gated region were identified among these major categories, and 41.6% of the cells remained unidentified. Figure 2 illustrates the phenotypes of cells in the lungs of control mice.

View larger version (45K): [in this window] [in a new window]   Figure 2.   Mononuclear cell phenotypes in air-sham control mice. The major categories of cell types based on surface antigen phenotypic staining in control lung and spleen cell preparations are shown as stacked bar diagrams. The fractions of CD4/CD8 -TCR+ cells are shown, while CD4+ or CD8+ -TCR+ are included with those cell types. Macros, monocytes and macrophages; unknown, calculated difference from 100%.

Spleen cell samples from air-sham control mice served as a reference for mononuclear cell phenotypes in a peripheral lymphoid organ. The small mononuclear spleen cells within the gate of interest included 18.8 ± 6.8% CD4⁺ cells and 8.7 ± 2.2% CD8⁺ cells, an average CD4⁺: CD8⁺ ratio of 2.3 ± 0.3. Macrophages, NK cells, and -TCR⁺ cells were identified respectively as 8.8%, 5.9%, and 4.6% of the total population. In contrast to the lung digest samples, B lymphocytes formed 60.8 ± 3.5% of the spleen cells. The fraction of B cells in the spleen was significantly greater than in the lung (P < 0.01); the other phenotypes were not statistically different. For the control spleen samples, essentially all of the cells within the region of interest were identified within the categories of the cell types listed (sum of 107.6%). Figure 2 summarizes the phenotypes of cells in the spleens of control mice.

Cells expressing the -TCR comprised 20.0 ± 7.4% of the total lung cells, whereas cells expressing the -TCR were 4.2 ± 1.9% of the total. Essentially all -TCR⁺ cells also expressed CD4 or CD8 surface antigens. In contrast, most of the cells expressing the -TCR did not exhibit CD4 or CD8 surface antigen (double-negative cells, CD4/ CD8); the -TCR⁺ cells included 19.2% that were CD4⁺ and 1.8% that were CD8⁺. In the air-sham control spleen samples, the distribution of the TCR phenotypes among the gated cells was 4.6 ± 3.0% -TCR and 25.4 ± 2.6% -TCR. As in the lung, most of the cells expressing the -TCR did not exhibit CD4 or CD8 (76% double-negative). The fractions of CD4/CD8 -TCR⁺ cells in the lungs and spleens of control mice are included in Figure 2.

Changes in the Frequencies of Cell Phenotypes in Silicosis

The abundance of the various phenotypes of T lymphocytes in whole-lung digest preparations did not change extensively in silicosis. As shown in Figure 3, the percentages of lung cells expressing CD4⁺ were not different in the lungs of air sham control and silica-exposed mice (15.5 ± 4.4% versus 15.9 ± 5.8%). The percentage of CD8⁺ lung cells decreased slightly (4.6 ± 0.9% versus 6.0 ± 1.5%) and the CD4:CD8 ratio increased slightly (3.4 ± 1.0 versus 2.8 ± 1.2) in silica-exposed compared with control samples, but these differences did not reach statistical significance. In the spleen, air- and silica-exposed mice showed similar proportions of CD4⁺ (18.8 ± 6.8% versus 18.6 ± 4.9%) and CD8⁺ cells (8.7 ± 2.2% versus 7.2 ± 1.5%), and the CD4:CD8 ratios were not different (2.3 ± 0.2 versus 2.7 ± 0.3). The proportions of -TCR⁺ cells, NK cells, B cells, or macrophages, as shown for air-sham control mice in Figure 2, did not change significantly in either the lung or the spleen as a result of silica exposure.

View larger version (25K): [in this window] [in a new window]   Figure 3.   Distribution of CD4+ and CD8+ phenotypes in lung cells. The percent of total gated cells in the lung digest population exhibiting CD4 or CD8 are shown for air-sham control and silica-exposed mice. The CD4:CD8 ratio is indicated for each exposure group. The differences in percentages or ratios between control and silica-exposed mice were not significant.

The percentage of cells expressing the -TCR phenotype did not change significantly in the lung (4.2 ± 1.9% versus 4.2 ± 2.8%) or the spleen (4.6 ± 3.0% versus 5.8 ± 4.0%) in silica-exposed animals. In contrast to other cell types, the percentage of -TCR cells also expressing CD4 increased significantly in the lung samples (control 5.3 ± 1.0% versus silica 8.7 ± 0.3%, P < 0.03), as shown in Figure 4.

View larger version (54K): [in this window] [in a new window]   Figure 4.   Fraction of -TCR+ cells expressing CD4+. The percentages of cells that express CD4 within the population of -TCR+ cells from the lung and the spleen are shown for air-sham control and silica-exposed mice (*P < 0.03).

Effects of Preparation Methods on Surface Antigen Staining

The lung cells were prepared by enzyme digestion, physical shearing of tissue, and culture with brefeldin A. We wished to determine whether these procedures might have affected the viability of the treated cells or the abundance of surface antigen epitopes. By comparison, spleen cells were prepared by gentle disruption of the tissue without enzyme digestion. We subjected spleen cells to collagenase/deoxyribonuclease enzyme digestion and shearing by passage through a 14-gauge cannula to determine whether this treatment affected surface antigen phenotype staining. Samples of standard and enzyme-digested spleen cells were stained in duplicate for surface antigen expression. Slightly decreased or increased absolute percentages of cell types were observed as a result of the enzyme digestion and shearing. For example, the fraction of CD4⁺ splenocytes decreased from an average of 10.5 to 8.5%, while the fraction of CD8⁺ cells increased from 3.9 to 6.1%. We believe that these differences are within the variation of the method, and conclude that enzyme digestion did not greatly influence the results we observed for lung cells. Control or silica-exposed lung or spleen cells expressed similar frequencies of CD4 or CD8 when cultured in medium with or without brefeldine A, indicating that this reagent did not substantially change the recognition of lymphocyte phenotypes.

We examined the viability of the isolated cell populations to determine whether dying cells might have contributed to the results. The flow cytometry size/complexity gate we selected excluded small dead cells with picnotic nuclei, but would have included cells of normal size and nuclear morphology that were early in the process of cell death. We evaluated the DNA content of the lung and spleen cells within the size/complexity gate of interest to determine whether a substantial portion of the cells were dying as a result of enzyme digestion, cell isolation, storage before analysis, or incubation in culture with brefeldine A. The DNA content of the cells was assessed by staining with PI; cells containing less than diploid DNA were defined as dying (presumed apoptosis). Immediately after isolation 10 to 15% of lung cells but less than 1% of spleen cells registered subdiploid DNA. After storage and culture with brefeldin A, 34% of spleen cells and 38% of lung cells contained less than diploid DNA. The frequency of cells expressing CD4⁺, CD8⁺, or NK-cell surface antigens was not different between the cells with diploid or less than diploid DNA. We infer that cell death did not selectively strike a particular phenotype to alter our results, but that all phenotypes in the population appeared to be affected similarly by the isolation and culture process. Color interference between the PI nuclear DNA stain and the PE or FITC intracellular cytokine stains did not permit simultaneous determination of DNA content and cytokine production.

Production of IFN- by Lung and Spleen Cells

The small mononuclear cells within the region of interest were enumerated for the percentages of cells expressing specific surface antigen phenotypes that stained for intracellular IFN- above the background level of the isotype control. The results in each category are shown as the average of two to four determinations on pooled specimens from two or three mice per experiment in two replicate inhalation exposure experiments. The specificity of intracellular cytokine staining was demonstrated by negative results or by decreased percentages of stained cells when brefeldin A was omitted, the saponin permeabilization step was omitted, or the cells were stained with unlabeled antibody before staining with fluorochrome-linked antibody (data not shown). The results were analyzed in two ways. First, the staining characteristics were gated to selected cells expressing a specific surface antigen phenotype, and the fraction of cells within that phenotype that produced IFN- was determined. Alternatively, the population was gated to select the cells staining positively for intracellular IFN- and the fraction of cells expressing each surface antigen phenotype was assessed.

In the lungs of control mice 11.3 ± 1.9% of the gated cells contained IFN-, whereas silica-exposed mouse lungs demonstrated 19.3 ± 6.0% of cells with IFN- (P < 0.03) as previously reported (17). In the spleen, control mice demonstrated 4.0 ± 0.7% cells with IFN-, and silica-exposed mice showed 6.3 ± 1.0% cells with the cytokine (P < 0.03) (17). We now report that the production of IFN- was attributable to several specific phenotypes in both organs.

In air-sham control mice, 4.7 ± 1.9% of the lung cells expressing the CD4⁺ surface antigen contained IFN-. This fraction was significantly increased in silica-exposed mice, where 9.3 ± 2.4% of CD4⁺ lung cells produced IFN- (P < 0.05). This comparison is shown graphically in Figure 5. Small fractions of CD8⁺ cells produced IFN- in the lungs of control or silica-exposed mice (1.9 ± 1.1% versus 3.0 ± 1.0%, P = NS). Cells identified as NK cells by staining with DX5 antibody also produced IFN-. In control lung specimens, 2.3% of the NK cells contained IFN- and the fraction was increased more than 3-fold after silica exposure (8.2%), as shown in Figure 5. Cells containing IFN- represented 4.8 ± 2.4% of the cells expressing the -TCR phenotype in the air-sham control mouse lung population. This fraction was increased to 15.9 ± 5.1% among -TCR⁺ lung cells from silica-exposed mice, as shown in Figure 5. In preliminary experiments with lung digest cells, minimally detectable fractions of B cells or macrophages appeared to produce IFN-, and these cell types were not examined for cytokine production in subsequent studies. ANOVA demonstrated that the effect of silica exposure on the percentages of cells producing IFN- was significant (P = 0.006), and that differences among specific cell types were significant (P = 0.018). The differences between air-sham and silica-exposed lung cells reached significance (P < 0.05) for CD4⁺ cells and -TCR⁺ cells when post hoc analysis of differences for specific cell types were performed.

View larger version (39K): [in this window] [in a new window]   Figure 5.   Fraction of specific phenotype lung cells that produce IFN-. The fractions of lung digest cells producing IFN- within the subpopulations of CD4+, CD8+, NK+, or -TCR+ cells are shown for air-sham control and silica-exposed mice. The population of small mononuclear lung cells was gated for each surface antigen phenotype, and the percentage of cells staining positively for intracellular cytokine was measured.

In the spleen, similar percentages of air-sham control and silica-exposed cells produced IFN- when CD4⁺ cells (control 2.3 ± 0.3% versus silica 2.2 ± 1.7%), CD8⁺ cells (0.5 versus 0.9%), or -TCR⁺ cells (6.1 ± 3.5% versus 5.7 ± 0.1%) were examined. The fraction of spleen NK cells producing IFN- was increased slightly in silicosis (8.9% versus 11.9%). The effect of exposure on the percentage of cells producing IFN- was not significant by ANOVA for any of the cell types tested.

The phenotypes of cells producing IFN- were analyzed by an alternative approach in which IFN-⁺ cells were gated in the cytofluorograph display, and the fraction of these cells that also exhibited each of the surface antigen phenotypes was calculated. Figure 6 illustrates this comparison for lung cells from air-sham and silica-exposed mice. Production of IFN- was greatest among CD4⁺ cells (7.5 ± 3.1%), and the fraction increased with silica exposure (10.3 ± 1.2%). -TCR⁺ cells provided 5.9 and 8.1% of the IFN- lung cells in control and silica-exposed mice, respectively, presumably representing most of the CD4⁺ cells as well. CD8⁺ cells represented 1.3 ± 0.5% and 1.6 ± 0.4% of control and silica-exposed IFN--producing lung cells, respectively. Cells marked as -TCR⁺ contributed small fractions to the production of IFN- among control lung cells (3.1 ± 1.0%), but increased significantly in mice with silicosis (8.4 ± 0.3%). Similarly, NK cells provided a smaller fraction of the IFN- cells in the lungs of control mice compared with those with silicosis (2.7 ± 1.4% versus 9.5 ± 2.8%). For the proportion of IFN--producing cells representing each phenotype, the effects of both silica exposure and cell phenotype were significant by ANOVA, except for the fraction of cells staining as CD8⁺. The distribution of phenotypes in the cells producing IFN- in the spleen samples from control mice was similar to that in the control lung population, and did not change substantially after silica exposure.

View larger version (39K): [in this window] [in a new window]   Figure 6.   Fraction of IFN--producing lung cells that express specific phenotypes. The fractions of lung digest cells expressing specific surface antigen phenotypes within the subpopulation of cells producing IFN- are shown for air-sham control and silica-exposed mice. The population of small mononuclear lung cells was gated for cells staining positively for intracellular IFN-, and the percentage of these cells expressing the various phenotypic surface antigens was determined.

The production of IFN- within the lung by specific phenotypes of lymphocytes can be considered as the total burden of each cell type producing the cytokine. We estimated the lung burden for several phenotypes of interest by multiplying the total number of cells recovered from the organ by the percentage of cells framed within the cytofluorograph gate of interest, the percentage of cells within that gate expressing the phenotype, and finally the percentage of that phenotype producing IFN-. For example, the total number of CD4⁺ IFN--producer cells in an air-sham control mouse lung was calculated as (4.9 × 10⁶ total cells) × (46.1% of cells in gate A) × (15.5% CD4⁺ cells) × (4.7% CD4⁺ IFN-⁺ cells) = 16.5 × 10³ cells per lung. These results are illustrated in Figure 7, and demonstrate that IFN- production was dominated by CD4⁺ cells with a major contribution from NK cells. The burden of CD4⁺IFN-⁺ lung cells increased 3.1-fold, -TCR⁺ IFN-⁺ cells increased 5.1-fold, and NK cells producing IFN- increased 5.5-fold in mice with silicosis. The increased numbers of IFN--producer cells within each phenotype was not caused by enrichment of the phenotype within the population, but rather by the increased total number of cells recovered in the lung digest preparations (1.7-fold) and the increased fraction of cells (1.9- to 3.6-fold) producing IFN- within each phenotype.

View larger version (37K): [in this window] [in a new window]   Figure 7.   Total numbers of lung cells producing IFN-. The total number of cells per lung producing IFN- of the phenotypes that express CD4 (CD4+), carry the -TCR (gd-TCR+), or express NK cell (NK-cells) antigens were estimated by multiplying the total number of cells recovered from the lung by the percentage of cells framed within the cytofluorograph gate of interest, the percentage of cells within that gate expressing the phenotype, and finally the percentage of that phenotype producing IFN-.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Several types of lymphocytes, not a single phenotype alone, produced IFN- in the normal mouse lung. CD4⁺ T cells were the most numerous producers of IFN- (Figure 7), with lesser numbers of NK cells and of CD4CD8 -TCR⁺ cells. In the normal lung, only 2 to 5% of the cells within each of these surface antigen phenotypes produced IFN- (Figure 5). Both the absolute number and the percentage of IFN--producer cells within each of these three phenotypes increased substantially in fully developed silicosis. The absolute increase was greatest for CD4⁺ T-cells, but the relative 3- to 5-fold increase in producer cell number was similar for all three phenotypes. Thus, our initial hypothesis that one cell type might be enlarged selectively in silicosis appears to be unsupported, and the production of IFN- appears to be derived from at least three distinct lymphocyte phenotypes.

Relatively little information has been published regarding the abundance of different functional types of lymphocytes identified in situ (29) or by tissue digestion (30, 31) in the lungs of normal mice, with some variation in results attributable to mouse strain, age, and preparation methods. We observed in normal adult C3H/HeN mice a predominance of T lymphocytes and NK cells. The low percentage of B cells (6% B220+) we noted is similar to the results reported in BALF (32, 33) and tissue specimens (31) from normal mice, and in BALF from normal human subjects (34). We identified the proportions of CD4⁺ and CD8⁺ cells, and a CD4:CD8 ratio that were similar to those reported previously in normal mice (31, 35). Changes in the numbers and proportions of BAL leukocyte types have been described in silica-exposed human workers (36) and in experimental silicosis in mice (15, 32, 40) and other species (5, 41, 42).

Within the population of cells extracted from the whole lung by enzyme digestion, we found that NK cells were almost as numerous as CD4⁺ T cells, and were relatively much more abundant than in the spleen. CD4CD8 -TCR⁺ cells represented a small proportion of the lymphocytes in both the lung and the spleen. We did observe an increase in the CD4⁺:CD8⁺ ratio in murine silicosis, as had been reported previously (4, 32), but the change did not reach statistical significance (Figure 3). The percentages of NK cells and -TCR⁺ cells did not change substantially after silica exposure. Because the total number of lymphocytes recovered from the lung nearly doubled in silicosis, all phenotypes were increased substantially in absolute numbers of cells.

Limitations of the Methods

The methods chosen for these experiments carry some limitations that may affect interpretation of the results. We prepared lung mononuclear cells for analysis by enzyme digestion of the whole lung, including the large airways and blood vessels. BAL has been used to sample air-space cells and secretions in many human diseases and animal disease models, including silicosis in mice. Other investigators performed BAL before lung digestion in an attempt to distinguish parenchymal from air-space cells (3, 35, 43). Rather than BAL, we selected the enzyme digestion method to obtain samples from all anatomic locations, not just the free cells of the airspace. Our previous observations of silicosis in the mouse lung showed numerous lymphocytes within enlarged BALT structures and parenchymal lymphoid nodules, not only within the alveolar spaces. We reasoned that lymphocytes at all of these locations might produce IFN- and could be important biologically in the pathogenesis of silicosis, and thus should be included for analysis of phenotype.

The enzyme digestion method could carry the disadvantage that cells might be uniformly or selectively damaged by the extraction process. We focused our cytofluorograph analysis on cells with the size and complexity characteristics of normal lymphocytes, and excluded picnotic cells and debris. Nonetheless, a substantial portion of the lung cells analyzed showed subdiploid DNA content, suggesting apoptosis. Because the fraction of subdiploid cells was greater in normal mouse lung digest preparations (10%) than in spleen cell preparations (1%), we inferred that the enzyme digestion and the physical shearing procedure may have damaged some of the lung cells. In the spleen, virtually all of the cells examined in the flow cytometry gate of interest could be accounted for within the categories of surface antigen phenotypes that we examined (Figure 2). A substantial proportion of the lung cells within the gate of interest could not be identified among the phenotypes examined. These unidentified cells may have been structural lung cells that truly did not bear any of the surface antigens tested, or lymphocytes with surface antigens masked or damaged by the extraction process. However, our experiment exposing spleen cells to enzyme digestion did not show a substantial decrease in the recognition of the phenotypes examined.

The method of staining immunoreactive cytokines within cells provides a tool for simultaneously identifying producer cells and defining their other features. Cytokines are often secreted from cells immediately after synthesis, and are usually not stored in large quantities within the cell. Thus, brefeldin A was used to block secretion and allow cytokine accumulation before staining. In addition, the cells were permeabilized with saponin to allow penetration of the staining antibody into the cytoplasm. This approach might underestimate the true proportion of cells producing IFN- if the synthetic rate were below the level of detection sensitivity, if brefeldin blockade were incomplete, or if permeabilization did not permit entry of adequate antibody for effective staining. Conversely, this method could overestimate the fraction of IFN- producer cells if a substantial proportion of cells synthesized the protein in precursor form but did not normally secrete it. These limitations might alter the accuracy of the absolute number of cells counted as producing IFN-, but should not substantially change the relative fractions found in normal mice compared with those with silicosis. Analysis of cells by flow cytometry permits the examination of each cell individually and large numbers of events can be recorded to permit statistical strength. This approach is limited by the availability of reagents, and by the combinations of reagents possible with four colors or fluorochromes for simultaneous labeling.

Limitations of the Results

For the experiments in this report we chose a time point 16 wk after inhalation exposure at which both the cellular inflammatory and fibrotic aspects of silicosis were fully developed. Only one time point was analyzed because of the large number of experiments and combinations of reagents required. The data represent a single time point in the course of an evolving disease. Our previous results demonstrated increased steady-state abundance of IFN- messenger RNA (mRNA) at an early time point, 2 wk after silica exposure, as well as at the 16-wk time point (17). It is not yet known whether the same proportion of lymphocyte phenotypes produce IFN- at the earlier time, or whether there might be a single dominant phenotype that initiates the cytokine response soon after silica inhalation.

Lung digestion and cytofluorograph analysis did not permit the localization of different phenotypes of IFN-- producer cells to specific anatomic sites within the lung. Our previous research showed IFN- mRNA, detected by in situ hybridization, expressed in cells at several locations within the mouse lung with silicosis. IFN--producer cells were found within the parenchymal silicotic lesions, within parenchymal lymphoid nodules, and within enlarged BALT structures (17). We do not know from the present research whether all three predominant IFN--producer lymphocyte phenotypes (CD4⁺ T cells, NK cells, and -TCR⁺ cells) are distributed uniformly throughout these locations, or whether there is anatomic compartmentalization for each phenotype.

Functions of Lymphocyte Phenotypes

The Th1-Th2 paradigm. Lymphocytes expressing the surface adhesion molecule CD4 are the single most abundant phenotype of T cells in peripheral blood, secondary lymphoid organs, and most nonlymphoid tissues (44, 45). With activation, CD4⁺ Th cells may be driven from a naive or neutral status (Th0) toward differentiation into one of two largely exclusive categories with different functional attributes that are recognized by the profile of cytokines they produce: Th1 or Th2 cells (9, 46). Th1 cells produce IFN- and IL-2. Th1 cells appear to be induced by the action of IL-12 acting in the absence of IL-4, and are dominant in cell-mediated immune responses, infections caused by intracellular pathogens, macrophage activation, and granuloma-like tissue reactions. Th2 cells produce IL-4, -5, and -10, and are associated with IgE production, mast cell activation, allergic reactions, and helminthic infections. Recent data suggest that a spectrum of responses is possible, and that cytokines attributed to both categories may be produced simultaneously by parallel cell populations (48).

-TCR T cells. T cells that express the -TCR phenotype are mostly CD4CD8, but a minority proportion may be CD8⁺ or CD4⁺ (44, 51). -TCR⁺ cells are rare in lymphoid organs or blood but abundant in epithelial organs such as the respiratory, gastrointestinal, and genitourinary systems, and in the skin. The exact functions of this cell population have not been fully defined, but include production of IFN- after activation and direct cell-killing capabilities (45). Augustin and associates estimated that 8 to 20% of resident pulmonary lymphocytes in the mouse were CD3⁺ -TCR cells, inferring that these were T cells (54). This estimate is similar to our results, showing 3.2% of all lung digest mononuclear cells or 6.9% of identified lymphocytes marking as -TCR⁺ cells in the normal lung (see Figure 2).

NK cells. NK cells are abundant mature lymphocytes that do not express T-cell (CD3) or B-cell (human CD19, mouse CD45/B220) surface epitopes, and are found in lymphoid organs, blood, and peripheral tissues (44, 64- 66). NK cells possess the ability to kill tumor cells, virus-infected cells, and other targets on direct contact without prior immunization, in part through Fas/Fas ligand-triggered apoptosis (67). The lung NK-cell population expanded substantially, while that in spleen and blood did not, after influenza inoculation (68). Thus, these cells may be critical in the host defenses against cancer or virus infections (69, 70). Activation and proliferation of NK cells is driven by a variety of cytokines produced by macrophages (IL-12, IL-15) and T cells (IL-2, IFN-) (71, 72). NK cells carry receptors for MHC class I molecules, and if this receptor is engaged upon cell-cell contact, then NK-cell cytotoxicity is inhibited. NK cells produce IFN- or IL-4 upon activation (67, 73). Newman and associates found no increase in the percentage of NK cells in the peripheral blood of silicosis patients or workers exposed to silica dust (74). Our observation that the percentage of NK cells in the lungs of mice with silicosis is not increased matches their finding in humans, and extends it by reporting that the absolute number of NK cells was increased substantially and that a greater proportion of the NK cells produced IFN-. We postulate that NK cells in silicosis participate in driving the Th1-like response, and may also serve killer functions in terminating damaged cells.

Cell Types Producing IFN-

Interferons are multifunctional glycoproteins with a broad range of antiviral, antiproliferative, and immunoregulatory effects on the target cell. The antiviral type I IFNs (IFN-/) are produced by a large variety of cells in response to virus infection, and confer resistance to the virus on neighboring target cells. The activating (and antiviral) type II IFN (IFN-) is produced almost exclusively by activated lymphocytes, and acts locally or systemically to stimulate macrophages and to modify the function of other target cell types. Although distinct in DNA sequence, peptide structure, and receptors, both the / and  IFNs share similar rapid intracellular transcription pathways involving signal transducer and activator of transcription-1 (75).

The present report indicates that lymphocytes are an important source of IFN- in the lung under normal conditions and in response to silica exposure. Other lung-cell types also might be considered as possible sources for this cytokine. Hahon and Castranova (76) studied rat type II pneumocytes and alveolar macrophages in culture after infection with influenza virus, and demonstrated the secretion by both cell types of materials that inhibited the growth of Sendai virus in target RFL-6 fibroblast cell cultures. The materials had physicochemical properties consistent with IFN(s), but the distinction between IFN-/ and IFN- was not examined.

We do not believe that alveolar type II cells would contribute to IFN- production, and are not aware of any reports describing synthesis of IFN- by epithelial cells of any type. Cultured mouse L-929 fibroblasts and mouse embryo fibroblasts have been reported to produce IFN- along with IFN-/ after stimulation with poly-ICLC or virus infection (77). Alveolar macrophages might theoretically contribute small amounts of IFN- to the total produced. Most reports identify lymphocytes (particularly T and NK cells) as the source of IFN-, but recent publications describe IFN- production by macrophages in response to autocrine (78) or IL-12 plus IL-18 stimulation (79).

We examined production of IFN- by cells other than lymphocytes in two ways. First, cells staining for surface markers associated with macrophages (MAC3) or B lymphocytes (CD45R/B220) were examined simultaneously for intracellular IFN-, and less than 1% of these phenotypes stained positively for the cytokine. Second, the proportion of cells containing IFN- was compared in cells inside the "small mononuclear cell" gate of interest with larger and/or more complex cells outside the gate, and production was always 4- to 5-fold higher among the cells within the gate. We believe these findings support the conclusion that the identified lymphocyte phenotypes account for most of the IFN- production by lung cells, both under normal conditions and in silicosis.

IFN- in Silicosis

IFN-, produced by subsets of T and NK cells, possesses diverse, potent biologic activities, including macrophage activation and the augmentation of MHC class II molecule expression, TNF- production, and TNF- effects (12). IFN- may act in an autoregulatory fashion to drive T cells toward a Th-1 phenotype and to suppress IL-4 and companion Th-2 cytokines. As well as these proinflammatory effects, IFN- administered therapeutically may diminish allograft rejection, reduce the severity of autoimmune nephritis, downregulate selected macrophage cytokines, and possibly decrease skin fibrosis (80, 81).

The exact role of IFN- in the pathogenesis of silicosis is not yet known. We postulate that the increased production of IFN- by multiple phenotypes of lymphocytes in the lungs of mice with silicosis could be driven initially by lung macrophages that have ingested silica particles and become stimulated by them, and consequently secrete cytokines that drive lymphocyte activation and the Th-1-like phenotype. IL-12 or -18 (82) would be leading macrophage-derived cytokines for this role. The resulting expanded population of activated lymphocytes that is producing IFN- and other cytokines could in turn amplify the cellular inflammatory response in silicosis. By this means, additional monocyte-macrophages without silica particles could be recruited and activated. Thus, IFN- could also promote fibrosis through secondary stimulation of TNF- effects and of other macrophage-derived cytokines that stimulate fibroblast proliferation and collagen synthesis. Conversely, IFN- might serve as a modulating influence in diminishing the abundance of IL-4 and decreasing the extent of fibrosis. Future research must interdict specific steps in these pathways to test these hypotheses.

	Conclusions

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We conclude that IFN- is produced in excess in the lungs of mice with silicosis. The IFN- is produced by several phenotypes of lymphocytes, including CD4⁺ T cells, NK cells, and -TCR⁺ T cells. The increase in IFN- cannot be attributed primarily to a change in the proportion of producer phenotypes within the lymphocyte population because the percentages of CD4⁺ cells, NK cells, and -TCR⁺ T cells were not changed. The increase in IFN- is due to an increase in the total number of lymphocytes in the lung, and to an increase in the percentage of cells within each phenotype that produces IFN-.

Future research should focus on determining the anatomic locations where each phenotype is concentrated, defining the initial events after the silica inhalation that triggers expansion of IFN- production, determining whether all phenotypes are expanded simultaneously or sequentially, and assessing the importance of IFN- in the inflammation and fibrosis of silicosis through experiments to interrupt this pathway.