Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

We have previously shown that an F₂ hybrid of a cross between C57BL/6 and 129 mouse strains develops asbestos- induced fibroproliferative lung disease (1), as previously described in rats (2). When the genes that code for the 55- and 75-kD receptors for tumor necrosis factor (TNF)- are knocked out (3), the hybrid mice are protected from the fibrogenic effects of asbestos (1), as well as from silica and bleomycin (4). We demonstrated that after asbestos exposure these TNF- receptor knockout (TNF-RKO) mice exhibit reduced expression of several peptide growth factors thought to be key to the development of interstitial pulmonary fibrosis (1). These factors are platelet-derived growth factor (PDGF)-A and -B, and transforming growth factor (TGF)- and -. TNF- production remained high in the exposed animals (1). We have postulated that protection from the fibrogenic agents is due to this reduced expression of peptide growth factors such as PDGF and TGF-₁ that are downstream from the control of TNF- (5, 6). Even though TNF- production remains high in the asbestos-exposed KO mice, the lack of TNF- receptors results in a loss of the signaling mechanisms through which TNF- influences the expression of other growth factors. Inasmuch as PDGF is the major mitogen for mesenchymal cells and TGF-₁ is a potent inducer of extracellular matrix production by these cells, it is reasonable to postulate that the mice fail to develop a fibroproliferative process because of the lack of TNF- signaling as indicated earlier. To test this postulate, it will be necessary to manipulate the individual growth factors and establish their roles in development of disease. Thus, we have asked whether overexpression of TGF-₁ by means of an adenovirus vector in the lungs of the fibrogenic-resistant TNF-RKO mice is sufficient to mediate the development of fibroproliferative lung disease.

Adenoviral vectors (ADVs) used to transduce gene expression in the lung have provided a powerful tool to approach the question framed above (7). Indeed, Sime and colleagues used a replication-deficient adenovirus to overexpress TGF-₁ (8) and TNF- (6) in the lungs of normal rats. Using this very ADV, we show here that transduction of TGF-₁ in the lungs of the TNF-RKO mice is sufficient to produce inflammation and fibrogenesis, thus adding strong support for the popular concept that TGF-₁ plays a significant role in the early development of interstitial fibrogenesis. Before this report, there was no information available that would allow an investigator to predict whether the overexpression of any growth factor would induce inflammation and fibrogenesis in the lungs of the fibrosis-resistant TNF-RKO mice, animals that provide increasingly popular models of human disease (1, 3, 4, 9).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Recombinant ADV

The recombinant adenovirus constructed with the complementary DNA (cDNA) of the coding region of full-length porcine TGF-₁ was kindly provided by Dr. Jack Gauldie (Department of Pathology, McMaster University, Hamilton, ON, Canada). This cDNA sequence contains a mutation from cysteine to serine at positions of 223 and 225, which results in the expression of biologically active TGF-₁ (AdTGF-₁^223/225). Control vector MG3 was constructed in the laboratory of Dr. David Curiel (University of Alabama at Birmingham, Birmingham, AL). The methods for construction, amplification, purification, and calculation of the recombinant adenovirus were previously described (6, 8).

Animals, Delivery of Recombinant Adenovirus, and Tissue Preparation

Mice with both the p55 and p75 TNF- receptor genes knocked out (TNF-RKO) were purchased from Jackson Laboratories (Bar Harbor, ME). These mice were produced on a mixed genetic background of the C57BL/6 and 129 inbred strains (B6129). B6129 F2 mice were used as wild-type controls for the TNF-RKO mice. The mice were 8 to 10 wk old, weighing 25 to 35 g. All mice were maintained on the basis of NIH guidelines under specific pathogen-free conditions. The mice were instilled intratracheally, 50 µl/each, with AdTGF-₁^223/225 or control virus MG3 diluted in phosphate-buffered saline (PBS) to a concentration of 5 × 10⁸ plaque-forming units (pfu); or with PBS control, 50 µl/each mouse. Five B129 hybrid controls and 10 TNF-RKO mice constituted the individual groups, which were treated either with the viral vector plus active TGF-, with the viral vector plus saline, or with saline alone, as just described. The animals were killed by intraperitoneal injection of 0.5 ml of 100 mg/ml ketamine hydrochloride at periods of 7 d after the instillation of ADV. The animals were exanguinated by cutting the renal artery and the lungs were perfused with fresh 4% paraformaldehyde in phosphate buffer, pH 7.4, through the trachea, at a pressure of 25 cm H₂O for 20 min. After perfusion, the trachea was clamped and the lung was removed and placed in fresh fixative overnight at 4°C. The fixed lung samples were embedded in paraffin, and 4-µm-thick sections were prepared on positively charged slides (Superfrost; CMS, Houston, TX) for immunohistochemistry and hematoxylin and eosin (H&E) staining. The AdTGF-₁^223/225 adenovirus instillation and tissue preparation protocols were performed for two separate experiments several months apart, with no apparent differences in any of the parameters studied.

Bromodeoxyuridine Labeling and Trichrome Staining

All mice were injected intraperitoneally with bromodeoxyuridine (BrdU) at the concentration of 50 mg/kg 4 h before the animals' death. Immunohistochemical staining for BrdU in mouse lung tissues was performed at 7 d after treatment using the immunoperoxidase technique described previously (1, 2). Briefly, the deparaffinized tissue sections were pretreated with 2 N HCl for 20 min, followed by incubation with 0.01% trypsin in 0.05 mol/liter Tris-HCl (pH 7.8), containing 0.1% CaCl₂ for 6 min at 37°C. The sections were incubated with 0.3% hydrogen peroxide and 5% normal goat serum for 30 min, respectively. The sections then were incubated with rat monoclonal antibody against BrdU (Harlan Sera Lab., Ltd., Loughbrough, UK) at room temperature for 1 h. The slides were then incubated with biotin-conjugated goat antirat (1:4,000; Jackson Immunoresearch, West Grove, PA) and streptavidin-horseradish peroxidase (1;2,000, Jackson Immunoresearch), at room temperature for 1 h, respectively. After washing, the peroxidase activity was visualized with a 10-min incubation in 0.05 M Tris-HCl at pH 7.6 containing 0.02% diaminobenzidine (Sigma, St. Louis, MO) and 0.006% hydrogen peroxide. The slides were counterstained with Lerner-3 hematoxylin (Lerner, Inc., Pittsburgh, PA). An equivalent dilution of normal mouse immunoglobulin G was used in place of the primary antibody as a specificity control. Mouse small intestine was used as a positive control each time.

Gomori trichrome staining was used routinely for demonstrating collagen distribution.

Quantitative Analysis of BrdU Labeling

To quantify BrdU labeling, positive-stained cells were counted in three separate anatomic locations in the lung: (1) terminal bronchiolar airway epithelial cells, (2) terminal bronchiolar airway interstitial cells, and (3) parenchymal alveolar cells of both epithelial and mesenchymal origin.

In the airways, epithelial cells were identified as lining the bronchiolar lumen and clearly above the basement membrane. All of the epithelial cells were counted in cross-sectioned airways and longitudinally sectioned airways; counting of 100 cells started with the last epithelial cell before the alveolar duct and proceeded proximally up the airway. Interstitial cells were determined as clearly beneath the basement membrane, and all of the interstitial cells underneath the same 100-epithelial-cell area were counted, as were all of the interstitial cells in the cross-sectioned airways. A total of 10 bronchioles was analyzed per section, typically five longitudinal and five cross-sectioned airways.

Alveolar cells within the lung parenchyma were counted as all of the cells within a randomly selected field. Five random fields per section were counted. Each field was selected by moving the micrometer stage one-half micron. Labeled cells were identified by light microscopy and counted at ×400 magnification. Counts were reported as a percentage of labeled cells among all the cells counted for each category. One-way analysis of variance (ANOVA) was performed to determine differences between groups and exposures.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Mice used were the F2 generation of a cross between C57BL/6 and 129 strains. The normal untreated animals and those exposed to saline as a control (Figure 1A) always had normal lung histopathology. The ADV used to transduce TGF-₁ expression in the lungs of rats typically produces mild perivascular and peribronchiolar inflammation (6, 8). This proved to be the case in mice as well, at a dose of 5 × 10⁸ pfu of ADV alone (Figure 1B). This immune response was confined to the small vessels and bronchioles and did not extend into the alveolar walls. At 7 d after treatment with 5 × 10⁸ pfu of ADV that transduced the expression of active TGF-₁, the animals had developed diffuse interstitial inflammation and fibrogenesis (Figures 1C and 1D). The process was largely interstitial with dramatic increases in alveolar wall thickening and numbers of chronic inflammatory cells (Figures 1C and 1D). These cells have not yet been characterized, but appear to be predominantly small lymphocytes, monocytes, and macrophages.

View larger version (134K): [in this window] [in a new window]   Figure 1.   H&E-stained sections of lung tissue from control and viral vector-exposed mice. Terminal bronchiole (TB), blood vessel (V). Bars = 20 µm. (A) Untreated and saline-treated mice always had normal well- expanded airways and parenchyma. (B) Section from the lung of a C57BL/6-129 mouse treated with the viral vector alone. Perivascular accumulation of chronic inflammatory cells was caused by vector treatment, but there were no changes in the surrounding alveolar walls. (C and D) C5BL/ 6-129 mice treated with the ADV transducing active TGF-1 developed diffuse interstitial inflammation and fibrogenesis. The alveolar walls were thickend with an expanded interstitial matrix (inset in D). (E and F ) TNF-RKO mice treated with the ADV transducing active TGF-1 developed diffuse interstitial inflammation and fibrogenesis. There were no apparent differences in histopathologic pattern of disease between the C57BL/6-129 and TNF-RKO mice.

The TNF-RKO mice exposed to ADV transducing TGF-₁ expression exhibited the same histopathologic patterns as did the treated hybrid animals described earlier (Figures 1E and 1F).

The tissues were stained with trichrome as a relatively crude assay for the presence of mature collagen. Trichrome staining was not apparent in any control tissue except around vessels and airways as expected. Lung tissue from the mice with inflammation exhibited some early evidence of fibrogenesis, with blue staining in the background and interstitial patches of clearly stained tissue (Figure 2A).

View larger version (100K): [in this window] [in a new window]   Figure 2.   (A) Histopathologic section stained with Gomori's trichome to identify collagen. Both the normal hybrid mice and TNF-RKO mice exhibited diffuse trichrome staining along the alveolar walls. (B) BrdU staining showed proliferating cells, primarily in the interstitial compartments of the airways (arrowheads) and alveolar walls (arrow) (see Table 1). Few airway epithelial cells were labeled. The inset shows proliferating intestinal epithelium as a positive control for BrdU incorporation.

BrdU staining was carried out at 7 d after treatment to determine whether cells were dividing at an increased rate. This turned out to be the case in all the mice exhibiting diffuse inflammation (Figure 2B), but not in any of the control animals, as demonstrated by quantitative analysis of the stained cells (Table 1). The airway epithelium was not significantly increased, but the interstitial compartment of the airways was clearly significantly different (P < .05). In the areas of lung parenchyma exhibiting inflammation, a variety of cell types incorporated BrdU at significantly increased rates (Table 1).

Because the normal and KO mice responded essentially the same to the ADV-TGF-₁ treatment, we wanted to be sure that TNF- receptor expression was indeed deleted for both the 55- and 75-kD receptors as originally described (3). This was the case as demonstrated by Northern analysis of whole-lung messenger RNA (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown that overexpression of active TGF-₁ in the lungs of normal mice resulted in inflammation and fibrogenesis 1 wk after treatment. This was expected on the basis of previous studies using this same ADV to transduce TGF-₁ or TNF- expression in rats (6, 8). However, previous to the findings presented here it was not at all clear whether the TGF-₁ expression would cause disease in TNF-RKO mice. The question was open because these mice have proven to be resistant to the fibrogenic and inflammatory events caused by inhaled asbestos (1) and by instillation of silica or bleomycin (4). Further, the expression of other peptide growth factors such as TGF-, (1) PDGF-A and -B (1), and TGF-₁ (10) was clearly reduced in these KO animals after lung injury. If the expression of a single growth factor would induce interstitial inflammation and fibrosis, it might then be possible to better define the role of this factor in disease development. In this initial phase of our ongoing studies, it is possible to draw at least the conclusion that TGF-₁ overexpression transduced by an ADV in TNF-RKO mice causes interstitial inflammation and fibrogenesis, a result that could not be predicted a priori. In addition, TGF-₁ overexpression appears to trigger a cascade of additional cytokines on the basis of the initial histopathologic picture of increased numbers of interstitial inflammatory cells and an expanded interstitial matrix. Among our goals for studying this useful model system are an analysis of the factors activated by TGF-₁ and an in-depth study of the mechanisms through which TGF-₁ controls interstitial fibrogenesis.

TGF-₁ is well known as a fibrogenic cytokine (11). We (12) and others (13) have demonstrated TGF-₁ expression during fibrogenesis, but its precise role in the pathogenesis of disease remains unclear. This problem is largely due to the multiple biologic activities of TGF-₁, which is a chemoattractant for inflammatory cells, an inducer of extracellular matrix components by fibroblasts, and a downregulator of epithelial cell proliferation (11). These points are consistent with the histopathologic picture presented here, i.e., cellular infiltrates, expanded interstitium, and few epithelial cells that had incorporated BrdU (Figure 2 and Table 1). The presence of this well-known effect of TGF-₁ downregulating epithelial proliferation is intriguing and will have to be confirmed through further studies on the individual cell populations, but incorporation of BrdU in airway and alveolar epithelium typically ranges from 4 to 15% in models of lung injury (14, 15). In that regard, the findings presented here clearly are different than expected.

When TNF- signal transduction is blocked in the TNF-RKO mice, expression of other growth factors is diminished, as discussed earlier (1, 4, 10). This and other published work (5, 6) suggest that TNF- plays a role in activating these more "downstream" factors and is, therefore, essential for the development of the fibroproliferative process that results in interstitial pulmonary fibrosis. Because a cascade of many growth factors is elaborated simultaneously as this disease develops, it has been impossible to define the role of each one. We are asking whether or not the replication-deficient ADV can be used to replace the expression of growth-factor genes that apparently are reduced in the TNF-RKO mice. Here we report some success in this regard because TGF-₁ alone clearly induced a fibroproliferative process. Whether the disease will progress similarly in the normal and KO mice is not yet known. In addition, it is not at all clear how the initial expression of TGF-₁, without the influence of TNF-, mediates the fibrogenic response. Perhaps TGF- expression is dependent on signaling through TNF- under normal conditions, but the TGF- protein obviously functions as expected even in the absence of TNF- signaling. Finally, we reported recently that titering the viral concentration down to 10⁷ pfu provides a more subtle disease process (16). It will be important to determine whether the KO mice respond differently to varying vector concentrations compared with the normal background controls.

In conclusion, we postulate that TNF- is essential for activating a natural cascade of factors that mediate fibroproliferative lung disease. We have shown here that it is possible to start a component of the cascade with overexpression of active TGF-₁ by means of an ADV. This could be an extremely useful model system for determining the roles of each of the growth factors that is expressed during disease development.