Introduction

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Silicosis is a chronic diffuse parenchymal lung disease that results from exposure for prolonged periods to airborne respirable dust containing crystalline silica particles. Although industrial hygiene measures to control ambient dust in the workplace have reduced its prevalence and severity, silicosis remains a prevalent health problem in developing nations throughout the world. In addition to its importance as an occupational lung disease, silicosis may serve as a paradigm for mechanisms of lung injury, inflammation, and fibrosis.

Silicosis in humans is characterized by nodules of mononuclear cell inflammation at sites of mineral particle accumulation, with progression of localized and ultimately diffuse fibrosis. The pulmonary macrophage is believed to be the key initial responder cell that ingests silica and consequently becomes activated to produce a variety of cytokines and other mediators that may augment tissue injury, promote neutrophil and lymphocyte accumulation, and stimulate fibroblast proliferation with altered connective tissue matrix deposition (1, 2). Although direct interactions between silica and pulmonary macrophages are undoubtedly important, most macrophages in the lungs of humans, rats, or mice exposed to silica do not contain particles (3, 4). This observation suggests a reiterative amplification cycle in which macrophages with silica produce cytokines that attract and activate lymphocytes, and suggests that these lymphocytes then produce additional mediators that attract and activate an expanded secondary population of macrophages. The goal of our current research was to identify mediators that might serve as mechanisms for macrophage recruitment and activation.

Lymphocytes accumulate in lung tissue, in bronchoalveolar lavage fluid (BALF), within intrapulmonary lymphoid aggregates and bronchial-associated lymphoid tissue (BALT), and in enlarged draining lymph nodes of both humans and animals with silicosis. These lymphocytes are mostly T cells, with relatively greater enrichment of CD4⁺ than CD8⁺ phenotypes, although both are increased (5). The T cells appear to be activated, express the interleukin (IL)-2 receptor, and show enhanced spontaneous DNA synthesis (10). Preliminary studies with intratracheal silica instillation reported that mice genetically deficient in mature T-cell responses (Balb/c nu/nu) evidenced less macrophage and lymphocyte inflammation than their T cell-sufficient counterparts, but comparable pulmonary fibrosis (11).

A subset of naive CD4⁺ T cells (Th-0) differentiates upon activation into at least two major functional phenotypes, Th-1 cells or Th-2 cells (12). Differentiation is driven predominantly toward one or the other subset by local influences. These two subsets are polarized opposites, and each mutually suppresses the other. They are recognized by their functional characteristics and by the cytokines they produce, rather than by unique surface antigens. The Th-1 subset dominates macrophage activation, granuloma formation, and delayed-type hypersensitivity, and is a primary response to infectious intracellular pathogens such as Listeria. It is induced by interferon- (IFN-) and IL-12, is suppressed by IL-4, and produces primarily IL-2, IL-3, granulocyte macrophage colony-stimulating factor, tumor necrosis factor (TNF), and IFN-. The Th-2 subset is abundant in allergic responses and helminthic infections, promotes immunoglobulin (Ig)E and IgG1 antibody responses, and is induced by IL-4, is suppressed by IFN-, and produces IL-4, IL-5, IL-10, and IL-2. The cytokines IFN- and IL-4 may serve as markers, respectively, for the Th-1 and Th-2 subsets. These functional profiles suggest that a Th-1-like population of lymphocytes might be expected in silicosis.

IFN- is a lymphocyte cytokine with broad biologic effects. It is produced by lymphocytes, particularly by natural killer cells and by subsets of CD4⁺ and CD8⁺ T cells (15). Its effects on target cells are mediated through a dimeric membrane receptor with the assistance of a third protein. IFN- is recognized as the most important cytokine in converting macrophages from a resting to an activated state. IFN- augments expression of major histocompatibility complex Class II molecules both on macrophages and on other cell types. IFN- also appears to augment TNF- production by macrophages subjected to a second stimulus, and to synergize many of the effects of TNF-. In contrast to the proinflammatory effects of IFN- noted previously, administration of the cytokine may diminish allograft rejection, reduce the severity of autoimmune nephritis, downregulate selected macrophage cytokines, and possibly diminish skin fibrosis (18, 19).

IL-4 is a peptide produced by activated T cells and mast cells that modulates immune responses, growth, and differentiation in a variety of target cell types, including B cells, monocytes, and fibroblasts (20). It is a key mediator in promoting asthma and allergic responses in the respiratory tract (23, 24). IL-4 appears to have an important effect in promoting fibrosis. IL-4 added to cell cultures stimulated collagen synthesis in normal and scleroderma human skin fibroblasts, and in fibroblasts from other organs (25). The effect of IL-4 appears to be on gene transcription, mediated by activation of the procollagen gene promotors (28). Human and murine fibroblast subpopulations demonstrate differing collagen synthesis responsiveness to IL-4 based on the abundance or heterogeneity of the IL-4 receptor.

Lung-tissue sections from patients with idiopathic pulmonary fibrosis demonstrated substantially increased numbers of cells expressing IL-4 messenger RNA (mRNA) (31). Skin biopsy specimens and skin fibroblasts from scleroderma patients demonstrated IL-4 mRNA and IL-4 protein, whereas normal skin materials rarely showed evidence for this cytokine (32). In the liver, anti-IL-4 antibody treatment decreased hepatic fibrosis in Schistosoma-infected mice. In experimental radiation pneumonitis in the rat, IL-4 mRNA expression, soluble IL-4 protein, and cells stained in situ for IL-4 were all increased (33). In mice with bleomycin-induced lung injury, lymphocyte IFN- production was increased and IL-4 production decreased, but anti-CD3 antibody reduced IL-4 further and decreased pulmonary fibrosis (34). Intratracheal instillation of silica into mice resulted in a prompt increase of IL-10 protein in lung tissue and BALF cells; IL-10-deficient genetically modified mice demonstrated a greater inflammatory response immediately after silica instillation, but slightly reduced pulmonary fibrosis at 30 d (35). These observations suggest that IFN- and IL-4 have counterbalancing and/or synergistic effects on granulomatous inflammation and on fibrosis.

We reported recently on the features of silicosis in mice exposed by inhalation to an aerosol of cristobalite silica (36). These animals developed diffuse pulmonary pathologic changes consistent with silicosis in rodents; demonstrated an expanded airspace population of cells enriched for macrophages, lymphocytes, and neutrophils; and showed increased lung hydroxyproline content as a reflection of increased lung collagen. We observed overexpression of both IL-1 and TNF- genes in the lungs of mice with silicosis, evidenced by increased mRNA extracted from whole-lung tissue and by localization of message by in situ hybridization in macrophages within pathologic lesions (37). These two macrophage-derived cytokines, particularly IL-1, would be candidates for inclusion in the mechanisms by which lymphocytes are mobilized and activated in silicosis (6, 38).

We undertook the present research to determine lymphocyte cytokine production patterns in silicosis. We hypothesized that the lymphocyte response profile in silicosis would follow a Th-1-like pattern, and that IFN- production would be increased. We developed this hypothesis because the pathology of silicosis bears features of a granulomatous response, and thus a Th-1 lymphocyte subset would be a likely participant, and because IFN- would be a likely candidate for augmenting the previously identified macrophage activation state that is characteristic of silicosis. Alternative hypotheses could be developed regarding the production of IL-4. A Th-1-like lymphocyte response with abundant IFN- might be expected to suppress IL-4 production and the Th-2 lymphocyte subset. Conversely, the presence of pulmonary fibrosis as a part of silicosis could be the result, in part, of increased IL-4. To test these hypotheses, we examined the abundance of IFN- and IL-4 mRNA transcripts, cells expressing message in situ, and the relative numbers of isolated lung cells producing cytokines.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Animals and Silica Exposure

The methods for exposure of mice to silica by inhalation of aerosol have been reported in detail (36). Mice (Mus musculus) of the strain C3H/HeN were obtained at 5 to 6 wk of age through the National Institutes of Health (Washington, DC), supplied by Harlan Sprague-Dawley (Indianapolis, IN). Aerosol exposure to cristobalite silica (C&E Minerals, King of Prussia, PA) was carried out using a horizontal-flow inhalation toxicology system (36, 49). The mice were divided into air-sham control or silica-exposed groups, and were exposed to the HEPA-filtered, humidified, temperature-controlled carrier airstream or to the mineral dust in identical chambers for 5 h/d for 12 d. The mice were exposed in two separate experiments to ambient cristobalite concentrations of 70 ± 19 or 70 ± 11 mg/m³ (mean ± SD).

Tissues were collected for extraction of mRNA, and for histology and in situ hybridization, 2 and 16 to 18 wk after the completion of the 12-d silica or air-sham exposure, and cells were isolated for flow cytometry studies 16 to 20 wk after exposure. In addition to the lung tissues, we collected materials from two other lymphoid organs. The mediastinal lymph nodes and thymus were harvested as a source of mRNA likely to contain substantial IL-4 and IFN- transcripts and thus serve as a "positive control" for detecting expression of these genes. Mononuclear cells were isolated from the spleen as a source of lymphocytes that would be expected to include cells producing IFN- and IL-4. The abundance of IFN- and IL-4 in these organs was also compared between air-sham control and silica-exposed mice.

We have reported the results of total lung silica content, histopathology, lung collagen (hydroxyproline), and BALF cell recovery from C3H/HeN mice under identical exposure conditions in a previous publication (36). Our methods for preparation of molecular probes and extraction of nucleic acids have been described (37). Mice were killed by intraperitoneal injection of an overdose of sodium pentobarbital. For extraction of nucleic acids, the thorax was opened, the mediastinal lymph nodes and the adjacent (usually inseparable) thymus gland were removed, and the lobes of the right and left lungs were cut away at the hilum. The tissues were plunged immediately into liquid nitrogen, and were stored frozen at 90°C for future analysis. For histopathology, the neck was incised, the trachea was cannulated with a polyethylene catheter, the abdomen and thorax were opened, and the lungs were exposed. The lungs were reinflated by gentle injection of 4% phosphate-buffered formalin fixative through the airway catheter until they reached approximately normal anatomic volume. Cubes of lung tissue were removed after several minutes fixation in situ, gently agitated in the fixative at room temperature for 24 h, and then washed and stored in 70% ethanol fixative. Tissue was embedded in paraffin, sectioned, and stained for histology with hematoxylin and eosin (H&E) or prepared for in situ hybridization.

Molecular Methods

Nucleic acids were isolated from frozen mouse tissues using Trizol Reagent (GIBCO BRL Life Technologies, Grand Island, NY) (52). Complimentary DNA (cDNA) probes were prepared to detect mRNA encoding the ubiquitous enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and the cytokines IFN- and IL-4. Twenty-base primer pairs were designed from published gene sequences to encompass approximately 450 to 500 exon base pairs (bp), spanning introns when possible. The primer pairs (and calculated strand lengths) were:

GAPDH-5' = CGG ATT TGG CCG TAT TGG GC; 3' = TGA TGG CAT GCA CTG TGG TC (518 bp).

IFN--5' = GCA TCT TGG CTT TGC AGC TC; 3' = CGA CTC CTT TTC CGC TTC CT (445 bp).

IL-4-5' = TCG TCA CTG ACG GCA GAG AG; 3' = CGA GTA ATC CAT TTG CAT GA (446 bp).

Primer oligonucleotides were synthesized by the Molecular Diagnostic Laboratory (Department of Pathology, University of Vermont College of Medicine, Burlington, VT) (GAPDH) or by Genosys Biotechnologies, Inc. (The Woodlands, TX) (IFN- and IL-4). Total RNA isolated from C3H/HeN mouse lymphoid tissue was used to generate cDNA first-strand products from mRNA by reverse transcription (RT) using standard techniques (53). The cDNA probes were generated with the first-strand cDNA products and the specific primer pairs for each gene of interest using heat-stable DNA polymerase (TAQ; GIBCO BRL) and repeated thermal cycling. The purified RT- polymerase chain reaction (PCR) cDNA probes were linked with biotin through a photokinetic reaction with ultraviolet (UV) light (Rad-Free; Schleicher & Schuell, Keene, NH), and were used for Northern analysis and for semiquantitative PCR analysis to estimate the abundance of the genes of interest in various tissues. For in situ hybridization, the cDNA probes were tagged for colorimetric detection by internal labeling with digoxigenin-linked uridine triphosphate during PCR cycling (Boehringer Mannheim, Indianapolis, IN).

Semiquantitative RT-PCR was used to measure expression of the genes of interest in mouse lung tissue. Similar amounts of total RNA from air-sham control and silica-exposed lung tissues were subjected simultaneously to RT. The first-strand product from each tissue was then serially diluted and subjected to PCR expansion with each primer pair. Aliquots (1 µL) of the PCR products from each serial dilution were dotted onto a nylon membrane (Nytran; Schleicher & Schuell), cross-linked with UV light exposure, and subjected to hybridization with the biotinylated cDNA probes using the Rad-Free system (Schleicher & Schuell). The membranes were then blocked with casein reagent, reacted with streptavidin-alkaline phosphatase, and washed. The membranes were placed against chemiluminescent phosphate substrate sheets with Biomax MR film (Kodak, Rochester, NY), and the optical density of each reaction dot was determined with a flat-bed scanner (ES-1200C; Epson America Inc., Torrance, CA) using a computer analysis system (Macintosh Power PC 7500, Apple Computer, Cupertino, CA; Photoshop 3.0, Adobe Systems, Inc., Mountain View, CA; and ScanAnalysis 2.55, Biosoft, Cambridge, UK). The optical density measurements for IFN- and IL-4 samples were adjusted for the ratio of GAPDH in control and silica samples: (Sample gene optical density [OD]/Average OD of all samples) divided by (Sample GAPDH OD/Average GAPDH OD of all samples). The adjustment provided by the GAPDH comparison was less than 40% in all instances. The primer pairs generated PCR products of the appropriate sizes (see Figure 1) for the designated genes. The sensitivity and specificity of the cDNA probes were demonstrated by Northern blot analysis, digestion of hybridization targets with ribonuclease (RNase) but not deoxyribonuclease (DNase), and failure to interact with irrelevant targets.

View larger version (53K): [in this window] [in a new window]   Figure 1.   Agarose gel electrophoresis of PCR products. Comparable amounts of total RNA extracted from air-sham control and silica-exposed mouse lungs 16 wk after exposure were subjected to RT and PCR expansion with 20-mer primer pairs specific for murine GAPDH or IFN- (IFN-g), and then to agarose gel electrophoresis. The oligonucleotide size-marker mixture displays bands at 100-bp intervals, indicating appropriate sizes as designed for GAPDH (518 bp) and for IFN- (445 bp). Comparable amounts of PCR products for the abundant ubiquitous gene GAPDH indicates similar loading of control and silica total RNA. Substantial gene product for IFN- is seen in silica-treated, but not control, lung tissue.

In Situ Hybridization

Formalin-fixed, paraffin-embedded tissues were subjected to in situ hybridization using modifications of techniques described by Davis and associates (37) and Fleming and colleagues (54). The tissues were subjected to gentle digestion with protease type VIII (Sigma Chemical Co., St. Louis, MO) to improve penetration of probes, denatured with formamide, hybridized with the digoxigenin-labeled probes, reacted with alkaline-phosphatase-linked anti-digoxigenin antibody (Boehringer Mannheim), and reacted with a substrate solution of nitro blue tetrazolium and bromo-chloro-indoyl phosphate. The sections were examined microscopically for localization of the blue-black reaction product. Appropriate controls for each cDNA probe were created by omitting probe or protease digestion, by using irrelevant human tissues with the murine probes, or by digesting the tissues with RNase A (Sigma) or DNase (Sigma) prior to hybridization.

Lung and Spleen Cell Isolation

Lung mononuclear cells were isolated by collagenase enzyme digestion and gentle mechanical disruption using modifications of published techniques (7, 57). Mice were killed with an overdose of sodium pentobarbital, the trachea was cannulated, the thorax opened, and the lungs reinflated with a solution of collagenase Type I (324 U/ml; Sigma), 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 (PBS) with Mg²⁺ and Ca²⁺ (GIBCO BRL, Life Technologies). The aorta and the inferior vena cava were cut, and the lungs were perfused with this enzyme solution via the right ventricle until white. The lobes of the lungs were excised, sliced into small cubes, and agitated in 30 ml of the enzyme solution for 1 h at 37°C. Lungs from two mice were pooled for digestion. At 30 and 60 min the tissue suspension was gently aspirated and expelled from a syringe through a 14-g blunt cannula to assist tissue disruption, and finally passed through a nylon mesh strainer (Falcon Plastics, Franklin Lakes, NJ) to exclude large fragments. The isolated cell suspension was centrifuged at 225 × g for 10 min at 5°C, resuspended in RPMI 1640 medium with 5% fetal calf serum and with penicillin 100 U/ml and streptomycin 100 µg/ ml (RPMI/FCS/P-S) (Life Technologies), counted by visual hemocytometer, and aliquoted for flow cytometry staining.

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. A pool of spleen cells from two mice was mixed, the larger 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 in the cells for staining. Brefeldin A is a nontoxic inhibitor of intracellular transport that serves to disaggregate the Golgi apparatus (61, 62). For each flow cytometry sample, an aliquot of 1 × 10⁶ 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.

Intracellular Cytokine Staining

Staining for intracellular cytokines was performed using modifications of published methods (62). The cell pellet in each tube was resuspended in 100 µl of RPMI/FCS with rat polyclonal IgG (Zymed, San Francisco, CA) at 50 µg/ml, incubated for 10 min in the dark at 5°C, and centrifuged, and the supernate was discarded. Fixation, permeabilization, and staining were performed at room temperature. The cell pellet was resuspended in 50 µl of PBS, 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, 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, incubated for 10 min, diluted with 2.0 ml of PBS, and centrifuged. The pellet was mixed with 100 µl of saponin permeabilization solution (Fix & Perm, Reagent B; Caltag); 20 µl of antibody against IFN- or IL-4 (both rat-IgG1- labeled with r-phycoerithrin or fluorescein isothiocyanate; 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, centrifuged, resuspended in 350 µl of 1% paraformaldehyde (EMS, Ft. Washington, PA) in PBS, 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 (Miami, FL) 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 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 the cytokine antibody fluorochrome. 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 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). A cultured Th-2 CD4⁺ T-cell clone (courtesy of Karen Roessner, Department of Medicine, University of Vermont) in which approximately 30% of the cells produced intracellular IL-4 staining was used as a positive control for this cytokine. Differences between exposure groups were tested for significance ( = 0.05) by Student's t test (Systat 7.0; SPSS Inc., Chicago, IL).

	Results

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Silicosis in Mice

We reported previously the features of mice exposed to aerosolized cristobalite at 70 mg/m³ for 5 h/d for 12 d (36). We selected an early time point with few or no histopathologic changes (2 wk) and a late time point with obvious silicosis (16 to 20 wk) at which to examine IFN- and IL-4 production. Two weeks after silica exposure, the lung tissues appeared normal or showed slight accumulation of neutrophils and macrophages in the alveolar spaces; increased connective tissue matrix or widening of the alveolar septae was not apparent. At the same time, BALF cell samples showed increased macrophages, lymphocytes, and polymorphonuclear leukocytes with a twofold increase in the total number of cells recovered, and lung collagen (hydroxyproline) was increased slightly but significantly in silica-exposed mice. At 16 wk after cristobalite aerosol exposure, the mice showed the pathology of silicosis in rodents: focal parenchymal lesions, air spaces filled with cells and exudate, increased cellularity, and interstitial thickening. Increased lymphoid tissue was a prominent feature in the lungs from silica-exposed mice, with enlarged and more numerous BALT, and with parenchymal and subpleural lymphoid aggregates. At 16 to 20 wk, lung collagen was increased to 155% of control levels, and BALF cell recovery was increased more than threefold. The number of lymphocytes recovered by BAL from air-sham control mice was 9.21 ± 6.83 × 10⁴ (mean ± SD; 6.1% of total) compared with 33.48 ± 17.15 × 10⁴ (7.5% of total) from silica-exposed mice. Thus, the inflammatory-cell response and the accumulation of excess connective tissue matrix began within 2 wk and continued through 16 to 20 wk after silica exposure.

Expression of IFN- and IL-4

We assessed the production of IFN- and IL-4 in mouse lungs and in the mediastinal lymph nodes/thymus indirectly by measuring the abundance of steady-state mRNA transcripts for these two genes of interest. The purpose of these experiments was to determine whether the steady-state expression of these genes was altered in the lung or in nearby lymphoid organs in silicosis. Semiquantitative RT-PCR products for GAPDH, IFN-, and IL-4 were detected in mediastinal lymph node/thymic tissue and in lung tissue from both air-sham and silica-exposed mice. GAPDH was abundant in all tissues, whereas the cytokines were detected only when relatively high numbers of thermal cycles were employed. Estimation of the abundance of mRNA for IFN- and IL-4 was based on selective enrichment in comparable amounts of total RNA.

The total RNA recovered per lung from control mice was greater at 16 wk (240 µg) than at 2 wk (183 µg). We hypothesize that this change reflects an increase in size due to normal lung growth, and parallels the increases in body weight (23.6 to 41.9 g) and wet lung weight (150.2 to 168.5 mg) over the same time period. Total RNA recovery averaged 294 and 309 µg per lung at 2 and 16 wk in the silica-exposed mice, respectively, with similar increases in lung wet weight (182.4 and 212.5 mg, respectively). The body weights of air- and silica-exposed mice were not significantly different at 2 or at 16 wk. The lung wet weights of the silica-exposed mice were significantly greater than control mice at both 2 and 16 wk (P < 0.005). The increases in total RNA and wet lung weight in the mice with silicosis are believed to be due to the greater numbers of inflammatory cells, other lung cells, connective tissue matrix, and edema in these tissues.

IFN- mRNA was more abundant in the lung tissues of mice exposed to silica than in air-sham control animals at both 2 and 16 wk after exposure. Figure 1 shows agarose gel electrophoresis of RT-PCR products from air-sham and silica-exposed mice 16 wk after exposure, demonstrating comparable abundance of GAPDH mRNA, but detection of IFN- mRNA in silica-exposed lung tissue at a dilution at which little or no message was detected in control tissues. Samples from two or three individual mice or from pools of two to four mice at each time point were examined by semiquantitative limiting-dilution RT-PCR analysis. The amounts of cytokine gene products were adjusted for the proportion of GAPDH products in control versus silica-exposed tissues. As shown in Figure 2A, IFN- mRNA transcripts were detected in silica-exposed lung tissue at three times the abundance of control levels 2 wk after exposure, and at twice control levels at 16 wk after exposure. The expression of IFN- was increased in the intrathoracic lymphoid tissue at 2 wk (202%) and at 16 wk (119%) in silica-exposed compared with air-sham tissue, as shown in Figure 2B.

View larger version (38K): [in this window] [in a new window]   Figure 2.   Semiquantitative RT-PCR for expression of IFN-. Total RNA samples extracted from lung (A) and mediastinal lymph node/thymus (B) tissues from air-sham control and silica-exposed mouse lungs were analyzed by semiquantitative limiting-dilution RT-PCR for the expression of IFN-. The data represent the average of two to four specimens pooled and are adjusted for the abundance of GAPDH message in each specimen.

In contrast to the increase in IFN- gene expression, IL-4 mRNA transcripts were less abundant in the lung tissue from silica-exposed mice compared with the air-sham controls. Figure 3A illustrates the results of semiquantitative RT-PCR analysis, showing that IL-4 mRNA was slightly reduced at 2 wk (90% control) and substantially reduced at 16 wk (52% control) in the lungs from silica- exposed mice. In the control lung tissue, the mRNA transcripts for both IFN- and IL-4 were apparently somewhat more abundant at 16 wk than at 2 wk when adjusted for comparable expression of GAPDH. The abundance of IL-4 message was also decreased in the mediastinal lymph node/ thymus tissue of silica-exposed mice, with a slight reduction at 2 wk (81% control) and a very large reduction (3% control) by 16 wk after exposure, as shown in Figure 3B.

View larger version (33K): [in this window] [in a new window]   Figure 3.   Semiquantitative RT-PCR for expression of IL-4. Total RNA samples extracted from lung (A) and mediastinal lymph node/thymus (B) tissues from air-sham control and silica-exposed mouse lungs were analyzed by semiquantitative limiting-dilution RT-PCR for the expression of IL-4. The data represent the average of two to four specimens pooled and are adjusted for the abundance of GAPDH message in each specimen.

Localization of Cytokine Production in Tissues

Sites of IFN- and IL-4 production in the lung were determined by in situ hybridization of the digoxigenin-labeled cDNA probes with cytokine mRNA transcripts. The purpose of these experiments was to determine the anatomic sites in the lung where production of these cytokines occurred in normal mice, and whether the same or different sites were involved in silicosis. Cells containing substantial blue-black reaction product in the cytoplasm and nucleus stood out against pale gray background tissues, as shown in Figure 4. The staining was eliminated when the tissues were treated with RNase before in situ hybridization, whereas staining was not altered by pretreatment with DNase. As previously reported (37), the digoxigenin-labeled GAPDH cDNA probe generated abundant reaction product in almost all nucleated cells in the alveolar walls, bronchial mucosa, or lymphoid tissues.

View larger version (151K): [in this window] [in a new window]   Figure 4.   Detection of IFN- and IL-4 gene expression by in situ hybridization. (a) Silicosis in the mouse. Conventional histology of the lung of a mouse 16 wk after a 70-mg/m3, 5-h/d, 12-d inhalation exposure to cristobalite silica demonstrates focal lesions with mononuclear cell infiltrates, increased connective tissue matrix, and accumulations of lymphocytes around small blood vessels (V) and airways (A) (H&E stain). (b) IFN- mRNA in BALT. IFN- mRNA transcripts were detected by in situ hybridization with digoxigenin-labeled cDNA probes in cells in the BALT of an air-sham control mouse. Cells with dense black reaction product stand out against the background of pale gray surrounding cells and tissue (A = airway). (c) IFN- mRNA in silicotic lung lesions. Small mononuclear cells with black reaction product for IFN- mRNA transcripts are found in the parenchymal lung lesions of a mouse 16 wk after silica exposure. (d) IL-4 mRNA in BALT. IL-4 mRNA transcripts were detected in occasional cells in the BALT structures of both air-sham (illustrated) and silica-exposed mice (A = airway).

IFN- mRNA transcripts were detected in the BALT structures in normal mouse lung tissue by in situ hybridization, as shown in Figure 4b. The cells with reaction product appeared to be small mononuclear cells consistent with the morphology of lymphocytes. IFN- mRNA-stained cells were very rarely found among the alveolar structures and bronchial walls of the control mice.

Small mononuclear cells with IFN- mRNA reaction product were found in the silicotic lesions of the lungs from silica-exposed mice, as shown in Figure 4c. The positive cells appeared to be mixed among larger macrophages, other cells, and debris. IFN--stained cells were also found in the enlarged BALT structures and in the parenchymal lymphoid aggregates of the silicotic mice. IFN--stained cells appeared to be much more numerous in the lung tissues from the exposed mice. In part, this may have been because the BALT were larger and more frequently found, and the silicotic lesions and parenchymal lymphoid aggregates were not seen in the normal tissue.

Cells staining with reaction product for IL-4 mRNA were found rarely in either the air-sham or the silica-exposed mouse lungs. Almost all of the IL-4-stained cells appeared to be small mononuclear cells within BALT structures, as illustrated in Figure 4d. Very rare, isolated cells with apparent staining were found in the lung parenchyma. Cells with staining for IL-4 were notably less frequent than cells staining for IFN- in all of the tissues examined.

The Fraction of Cells Producing Cytokines

Cells containing intracellular cytokines were identified by staining with flourescent-labeled antibodies against IFN- or IL-4 and enumerated by flow cytometry. The purpose of these experiments was to determine the proportion of cells producing the cytokines among the mononuclear cell population of the normal lung, and to assess whether the proportions changed in silicosis. The flow cytometry approach was selected to provide information about the production of immunoreactive protein to complement the mRNA expression data offered by our other methods, and to provide a quantitative analysis enumerating thousands of events. Cell suspensions from the lung were prepared by enzyme digestion and gentle disruption, and from the spleen by mechanical disruption. Small mononuclear cells with the size (forward scatter) and complexity (side scatter) characteristics of lymphocytes were gated as an area of interest in the cytofluorograph diagram, as shown in Figure 5. Intracellular staining for IFN- or IL-4 was defined by cell fluorescence above the background level generated by a similarly labeled irrelevant immunoglobulin of the same species and isotype as the antibody. Staining for IFN- was readily identified as a second peak of positive cells brighter than the majority of cells within the gate of interest; a larger peak was evident in samples from silica-exposed mice compared with the air-sham controls. These features are illustrated in Figure 5. Cells staining for IL-4 represented a smaller proportion of the total population, but were evident as a bright subset (Figure 5). The small mononuclear cells within the defined gate of interest produced both the greatest proportion of positive cells and the clearest definition of bimodal peaks compared with regions outside of this gate. A very small proportion of cells (0.2 to 0.5%) stained positively for both IFN- and IL-4 when dual antibody staining with two fluorochromes was used on the same specimen. The flow cytometry diagrams shown in Figure 5 are representative of a typical experiment in which the lung digest cells and the spleen cells from two air-sham control mice or two silica-exposed mice 20 wk after exposure were pooled for analysis. Similar results were obtained from the multiple repetitions with mice 16 to 20 wk after exposure that were used to develop the methods. The numerical data and statistical analysis (Figure 6) were generated by merging the results of two to five replicates from each pooled specimen and from two separate inhalation-exposure experiments.

View larger version (35K): [in this window] [in a new window]   Figure 5.   Flow cytometry diagrams of intracellular cytokine staining. Cytofluorograms from a representative experiment show the distribution of lung digest cells (upper row) and spleen cells (lower row). The entire cell populations are shown in the left-hand panels, displayed according to their size (forward scatter, y-axis) and complexity (90° scatter, x-axis) characteristics. A gate defining the cells of interest has been drawn around a subset of small mononuclear cells with the features of lymphocytes. Intracellular staining for IFN- (IFN-g) is shown in the middle panels, displayed as the number of cells (y-axis) at each level of brightness (x-axis). IFN-- positive cells appear as a second, brighter peak compared with the negative majority, and are enumerated within the region defined by the horizontal marker line. The distributions are shown for the immunoglobulin isotype (dotted line), the air-sham control cells (thin line), and the silica-exposed cells (thick line). Intracellular staining for IL-4 is shown in the right-hand panels, arranged similarly.

View larger version (26K): [in this window] [in a new window]   Figure 6.   Small mononuclear cells with intracellular cytokines detected by flow cytometry. (A) Percentage of cells producing IFN-. The fractions of cells staining positively for IFN- by flow cytometry from the lung digest and spleen cell populations of air-sham control and silica-exposed mice 16 to 20 wk after exposure are shown as the means ± SD of data from pooled samples in two replicate aerosol exposures. The percentages of positive cells were increased in silica-exposed mice at both anatomic sites (*P < 0.03). (B) Percentage of cells producing IL-4. The fractions of cells producing IL-4 were not different in air-sham control and silica-exposed mice in both the lungs and the spleen. The conditions were as described for A.

The percentages of small mononuclear cells producing IFN- were significantly increased in both the lungs and the spleens of mice with silicosis compared with the air-sham control animals, as shown in Figure 6A. In the lung digest samples from control mice, an average of 11.3 ± 1.9% (mean ± SD) of lymphocytes stained positively for intracellular IFN-; whereas 19.3 ± 6.0% of cells were positive in silica-exposed mice (P < 0.03). Splenocytes were positive at 4.0 ± 0.7% and 6.3 ± 1.0%, respectively (P < 0.01). Approximately 3% of cells stained positively for IL-4 in both the lung and spleen cell samples from control and silica-exposed mice, with no significant differences between them (Figure 6B).

The number of cells recovered from the lungs of mice exposed to silica was significantly increased compared with the lungs of control mice (84 ± 20 × 10⁵ cells/mouse, versus 49 ± 14 × 10⁵ cells/mouse, P < 0.01). The cytofluorograph gate of interest encompassed approximately 65% of the total cell population, with no consistent differences between exposure groups. Thus, the total number of IFN-- producing cells would have averaged about 10 × 10⁵ in a silica-exposed mouse lung, and 3.6 × 10⁵ in an air-sham control mouse, an approximately 3-fold increase in this cell population. The absolute number of cells in the lung producing IL-4 would also have increased from approximately 0.38 × 10⁵ in control mice to 1.63 × 10⁵ in silica-exposed mice. The number of cells recovered from the spleens of silica-exposed mice was increased slightly but not significantly (158 ± 95 versus 133 ± 50 × 10⁵ cells/mouse).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We have found evidence for increased expression of IFN- mRNA, increased numbers of cells expressing this gene, and increased numbers of cells with intracellular immunoreactive IFN- protein in the lungs of mice with silicosis as compared with air-sham control mice. The cells responsible for the IFN- production had the morphologic features of lymphocytes; were located mainly in the BALT structures, lymphoid aggregates, and infiltrates in the lung; and were intermingled with other cell types within the silicotic lesions of the lung. The detailed phenotypic features of the cells producing IFN- are under investigation but were not the subject of the present research.

Limitations Imposed by the Methods

The apparent increase in IFN- in the lungs of mice with silicosis was found using three independent methods, suggesting that the central observation is robust. Nonetheless, each approach has limitations that must be considered. Increased abundance of mRNA encoding for IFN- was observed by semiquantitative PCR analysis of tissue extracts. This technique is very sensitive but does not indicate in which cells, or where within the organ, the message originated. Furthermore, post-transcriptional regulation of gene function could result in upregulation of mRNA without ultimate increases in mature secreted cytokine peptides. The quantitation of cytokine gene expression was based on the assumption that expression of the constitutive enzyme GAPDH did not increase, and the results could be modified if GAPDH was upregulated in parallel with IFN-. This did not appear to be the case because similar amounts of total RNA from control and silica-exposed tissues yielded comparable expression of GAPDH. Changes in the rates of mRNA transcription or degradation could not be estimated because only steady-state mRNA levels were measured. This approach could not distinguish between a large number of cells producing small amounts of relevant mRNA as opposed to a small number of cells producing abundant copies.

The technique of in situ hybridization was used to localize cytokine message expression in tissue sections. This method is limited by the relative insensitivity of the technique, and it is likely that only cells expressing the gene of interest at a high level are detected. The localization of reaction product depends on the conditions of the method. Excessive development in the color substrate solution could lead to diffuse nonspecific staining, whereas inadequate tissue digestion, hybridization, or color development could miss genuine reaction sites. The intensity of staining varies somewhat across a large tissue section, and regions with excessive or absent reaction product must be avoided. As a strength, in situ hybridization permits confirmation that cells expressing the gene of interest appear in locations where it would be expected. The cells expressing IFN- or IL-4 appeared to be small mononuclear cells predominantly located in lymphoid organs, findings consistent with the expectation that lymphocytes are primarily responsible for producing these cytokines. Conversely, GAPDH expression was detected in virtually all nucleated cells but not in erythrocytes, consistent with the ubiquitous distribution of this enzyme. It is likely that in situ hybridization for cytokine gene expression mainly detects cells with abundant mRNA copies, and thus might miss cells producing the cytokine from small numbers of transcripts.

Flow cytometry to identify cells producing cytokines had the advantages that an antibody was used to detect immunoreactive peptide rather than mRNA, and that automation permitted counting thousands of events with a high level of precision. The sensitivity of the technique for identifying cells producing cytokines is not known, and cells with low levels might be missed. The disadvantages of this method include the preparation of lung cells by digestion and the need to culture the cells before staining them. It is possible that the digestion process selected a subset of cells enriched with or depleted of the cytokine of interest; for example, cells producing IFN- might be liberated or might survive preferentially in comparison with cells producing IL-4. Similarly, digestion might alter gene expression in a manner that would change cytokine production. The culturing of digested cells for 4 h with brefeldin A could alter cytokine production so that it did not reflect accurately the behavior of the same cell in the lung or spleen. Our methods of analysis categorized cells as either positive or negative for the production of cytokines, and may have been insensitive to slight changes in the production of these peptides by the population as a whole. This approach seemed justified because of the bimodal peaks seen in the flow cytometry histograms (Figure 5).

Our previous findings demonstrated substantial enlargement of BALT and the appearance of parenchymal lymphoid aggregates. Because these lymphocyte-rich foci would be likely sources for the cytokines of interest, we chose methods we believed would be successful in sampling these sites: isolation of lung cells from all sites by enzyme digestion of the whole lung, in situ hybridization in whole lung sections, and extraction of mRNA from the whole lung. BAL samples the free cells and fluids of the bronchial and alveolar air spaces of the lung. We chose not to examine lavage specimens in detail, because the BAL results could represent a biased sample. Preliminary studies with BALF cell populations did show increased percentages of cells producing IFN- in mice with silicosis (data not shown).

Cytokines in the Normal Lung

The production of both IFN- and IL-4 was detected by all three methods in the lungs of normal air-sham control mice. On the basis of the abundance of mRNA, the number of cells found by in situ hybridization, and the percentage of cells identified by flow cytometry, it appeared that IFN- was produced in somewhat greater amounts than IL-4. It is possible that the steady-state levels of both message and peptide could be a false reflection of biologic activity or rapid turnover, and that the detection methods might differ in sensitivity. In control lung tissues the abundance of mRNA transcripts for IFN- and IL-4 both appeared to increase relative to GAPDH at 16 wk compared with 2 wk. This difference may be due to more intense expression of these genes as the mice mature. It is possible as well that cytokine mRNA and GAPDH mRNA are differentially susceptible to degradation during extraction, and that unintended slight differences in handling of tissues at 2 and 16 wk led to the differences in control cytokine/ GAPDH mRNA ratios at these time points. Because control and exposed tissues were handled simultaneously and identically at each time point, the possibility of differential degradation would not explain the differences observed between the air-sham control and the silica-exposed mice.

The cells producing both IFN- and IL-4 appeared to be lymphocytes on the basis of their size and complexity characteristics identified by flow cytometry, and on the basis of their morphology by in situ hybridization. The mRNA-positive cells in the normal lung appeared to be almost exclusively located within defined lymphoid aggregates and BALT. These findings suggest that lymphocytes with Th-1- and with Th-2-like features are part of the lymphoid constituency of the normal lung. Production of IFN- and IL-4 was also identified in the lymphoid organs of the thorax (mediastinal lymph nodes and thymus) and in the spleens of the air-sham control mice, as would be expected. These tissues were included in our experiments to serve as positive sources of cells producing the cytokines of interest.

Increased Production of IFN- in Silicosis

The production of IFN- was increased in the lungs of mice with silicosis compared with air-sham controls when assessed by all three of the methods used. The abundance of IFN- mRNA transcripts was increased both at an early time point after exposure (2 wk) and at a later time point after obvious pathologic silicosis had developed (16 wk). Because the lung tissue was not distinctly abnormal and the number of lymphocytes was only slightly increased at 2 wk, we focused attention at 16 to 20 wk for flow cytometry and in situ hybridization. By flow, the percentage of cells producing IFN- was increased by 70% and the approximate number of cells in the lung was increased threefold. Many of these cells were localized by in situ hybridization to the enlarged BALT structures and parenchymal lymphoid aggregates that characterized silicosis in these mice. In addition, cells staining positively for IFN- message were found mingled with macrophages, mesenchymal cells, and debris within the silicotic alveolar lesions. These findings appear to support our hypothesis that silicosis involves a Th-1-like lymphocyte-dominated process. The results suggest that IFN- is produced in direct proximity to the developing disease process, and could influence the functions of macrophages and other cell types at those sites.

Changes in IL-4 Production

The production of IL-4 in the lungs of mice with silicosis appeared to be unchanged or slightly downregulated compared with air-sham control animals. When assessed by mRNA abundance, there was no change 2 wk after exposure and there was a reduction to approximately 50% at 16 wk. By flow cytometry, the percentage of cells with IL-4 peptide was not changed significantly, and the approximate absolute number would have been increased. Cells expressing IL-4 mRNA were found in the BALT tissues of both control and silica-exposed mice, but were not observed in the silicotic lung parenchymal lesions. Thus, the relative abundance of IL-4 compared with IFN- appears to be decreased in silicosis, although the absolute abundance of cells producing IL-4 may be increased compared with control mice.

IL-4 potentiates fibroblast collagen production (25, 28- 30, 65), and may be a mechanism to promote fibrosis (31- 34). Hypothetically, IL-4 could also promote pulmonary fibrosis in silicosis. The more prominent production of IFN- and relatively less abundant production of IL-4 that we observed in mice following silica exposure suggest that IL-4 is not a major influence on fibrosis in this disease. Measurements of cytokines are always complex, and demand reservations in their interpretation. Most cytokines interact intensely with others to modulate the balance of the entire cytokine network. This cross-influence is important with IFN- and IL-4, as Th-1 and Th-2 lymphocyte subtypes and their cytokine profiles appear to be opposite and competitive (12). The profile of cytokines observed at any moment in a disease process probably represents both a pathway to the disease and an attempt to restore normal cytokine balance. Experiments that observe but do not perturb the cytokine network cannot distinguish these interrelated processes with confidence. Studies that intervene in the disease process to augment or reduce IFN- and IL-4 will be needed to understand more completely the role and importance of these cytokines in silicosis.

Cytokine Responses in Lymphoid Organs Outside of the Lung

Increased abundance of IFN- mRNA was observed in the intrathoracic lymphoid organs of silicotic mice. We removed the mediastinal lymph nodes and the thymus as a single block of tissue because we were not able to separate these structures reliably in mice. We examined these tissues to serve as an expected source of mRNA transcripts for both IFN- and IL-4 in case either cytokine appeared to be scarce or absent from lung samples, and to determine whether any alterations in cytokine expression found within the lung extended to nearby and/or draining lymphoid organs as well. The observation of increased steady-state IFN- mRNA is consistent with our previous report showing activation of T cells in the enlarged thoracic nodes of rats with silicosis (36).

An increased percentage of cells with intracellular IFN- was detected by flow cytometry in the spleens of silicotic mice. We examined spleen cells as a reliable and abundant source of easily isolated lymphocytes for flow cytometry, and to determine whether any alterations in cytokine regulation in silicosis extended to systemic sites beyond the lung. The finding of upregulation of IFN- at a lymphoid site remote from direct silica contact and within the abdomen could help explain our previous observation that peritoneal macrophages from rats with silicosis are "primed" for increased production of TNF- upon secondary stimulation (46).

The percentage of cells producing IL-4 was the same in the spleens of air-sham control and silica-exposed mice after exposure. IL-4 mRNA was similar in the mediastinal nodes/thymus tissue of both groups at 2 wk, but appeared to be decreased substantially by 16 wk in the silica-exposed mice. These observations suggest that the IL-4 biologic signal appears to be decreased relative to the IFN- signal in both the proximate and remote lymphoid tissues of mice with pulmonary silicosis.

	Conclusions

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We have found evidence for increased production of IFN- in the lymphoid tissue and parenchymal lesions of the lungs of mice with silicosis. The augmented IFN- response appears to develop early, and to persist through at least 20 wk of developing disease. The increased production of IFN- extends to other lymphoid organs within the thorax and at remote sites. At the same time, the production of IL-4 appears to be unchanged or decreased. If these two cytokines are viewed as markers and effectors of Th-1- or Th-2-like lymphocyte-mediated immune-inflammatory responses, then silicosis would follow a Th-1-like pattern within this paradigm. This concept is consistent with the many observations indicating both local and indirect macrophage activation in silicosis, and with the histologic features that resemble granuloma formation. Additional research will be required to identify further the functional phenotypes of the cells producing IFN-, and to understand more fully the exact role of this potent cytokine in the pathogenesis of silicosis.