THE ANATOMICAL RECORD 254:61–73 (1999) Differential Immunolocalization of VEGF in Rat and Human Adult Lung, and in Experimental Rat Lung Fibrosis: Light, Fluorescence, and Electron Microscopy HEINZ FEHRENBACH,1* MICHAEL KASPER,2 MICHAEL HAASE,1 DIETER SCHUH,1 AND MARTIN MÜLLER1 1Institute of Pathology, University Clinics ‘‘Carl Gustav Carus,’’ Technical University of Dresden, Dresden, Germany 2Institute of Anatomy, University Clinics ‘‘Carl Gustav Carus,’’ Technical University of Dresden, Dresden, Germany ABSTRACT Vascular endothelial growth factor (VEGF) is a cytokine with main angiogenetic functions in embryonic development and tumor-formation. In the adult lung, reports of the localization of VEGF were controversial. A precise cell typing of VEGF-positive pulmonary cells is still lacking. Nothing is known about a potential role in pulmonary fibrosis. Immunohistochemistry (IH), double immunofluorescence microscopy (DIF), and immunoelectron microscopy (IEM) were used to study the differential distribution of VEGF in paraffin-embedded (IH, DIF) and in cryo-substituted, Lowicryl-embedded (IEM) specimens of normal rat and human lungs and fibrotic rat lungs. Fibrosis was induced by intratracheal bleomycin treatment. IH and DIF showed that VEGF was present in surfactant protein (SP) D-positive alveolar type II pneumocytes, bronchiolar Clara cells, smooth muscle (SM) cells, and ␣-SM actin-positive myofibroblasts of normal rat and human lungs. Fibrotic lesions in bleomycin-treated rat lungs were rich in VEGFpositive cells presenting with a heterogeneous phenotype (mainly SP-Dpositive type II pneumocytes, ␣-SM actin-positive myofibroblasts). There were no signs of angiogenesis. Post-embedding immunogold labeling using protein A-gold and IgG-gold technique revealed a specific localization of VEGF to mitochondria, Clara cell secretory granules, and capillary interendothelial cell junctions. The predominant localization of VEGF to bronchiolar and alveolar epithelial and ␣-SM actin-positive cells, and the marked increase of VEGF-positive type II pneumocytes and myofibroblasts in fibrotic lung lesions, indicate that in adult lungs VEGF is involved in processes other than angiogenesis. Anat Rec 254:61–73, 1999. r 1999 Wiley-Liss, Inc. Key words: lung; fibrosis; vascular endothelial growth factor; immunocytochemistry Vascular endothelial growth factor (VEGF), which was initially described as vascular permeability factor (VPF) (Senger et al., 1983), is a dimeric glycoprotein that specifically stimulated endothelial cell migration and proliferation, angiogenesis, and increased vascular permeability (Dvorak et al., 1995; Klagsbrun and D’Amore, 1996). Moreover, recent studies presented evidence that VEGF r 1999 WILEY-LISS, INC. Grant sponsor: Bundesminister für Bildung, Wissenschaft, Forschung und Technologie; Grant number: 01ZZ5904. *Correspondence to: H. Fehrenbach, PhD, Institute of Pathology, University Clinics ‘‘Carl Gustav Carus’’, TU Dresden, Fetscherstra␤e 74, D-01307 Dresden, Germany. Fax: 49 351 458 4328. E-mail: firstname.lastname@example.org Received 24 March 1998; Accepted 31 July 1998 62 FEHRENBACH ET AL. may have important, previously unsuspected roles in cell types other than the endothelium of blood vessels (Brown et al., 1997; Ergün et al., 1997; Horiuchi and Weller, 1997; Rooman et al., 1997; Yang and de Bono, 1997). VEGF protein and transcript have been shown to be present in various organs of human and rat origin (Berse et al., 1992; Monacci et al., 1993; Shifren et al., 1994; Tuder et al., 1995). VEGF was reported to be very abundant in the lung, with alveolar epithelial cells being the site of highest immunohistochemical reactivity and mRNA expression (Maniscalco et al., 1995; Monacci et al., 1993; Shifren et al., 1994). Evidence has accumulated that VEGF is involved in some pathogenetic and repair processes (Couffinhal et al., 1997; Maniscalco et al., 1995; Shinohara et al., 1996; Tolnay et al., 1998; Tuder et al., 1995; Voelkel and Tuder, 1995). However, to our knowledge nothing is known about a potential role of VEGF in pulmonary fibrosis. The bleomycin-treated rat lung is an established experimental model of pulmonary fibrosis (Fine and Goldstein, 1997). The structural and ultrastructural alterations associated with fibrogenesis, which include initial endothelial injury, proliferation of alveolar type II pneumocytes, increased synthesis and deposition of extracellular materials, and subsequent profound remodeling of the pulmonary parenchyma, have been described in detail (Adamson, 1984; Adamson et al., 1986; Adler et al., 1986; Kuhn et al., 1989). Immunohistochemical studies contributed to the understanding of the complex processes of remodeling and repair by a detailed specification of the various cell types involved as, e.g., the modulation of fibroblasts into myofibroblasts, or the modulation of structural and membrane proteins of the alveolar epithelium such as changes in the cytoskeleton, in adhesion molecules and lectin binding patterns (Kasper and Haroske, 1996; Kasper and Singh, 1995; Vyalov et al., 1993; Woodcock-Mitchell et al., 1984). In recent years, the focus of interest has been directed toward understanding the role of cytokines in pulmonary fibrosis (Wolff and Crystal, 1997). To gain a better understanding of the biology of VEGF in the lung, and of its role in pulmonary fibrosis, our study aimed at the detailed cellular and subcellular localization of VEGF to the components of both the gas-exchanging parenchyma and the gas-conducting tree of rat and human lungs. Double immunofluorescence staining was used in conjunction with a number of established cell markers (Kapanci et al., 1992; Kasper et al., 1995a; Kasper and Singh, 1995) to characterize the nature of VEGF-immunoreactive cells. And postembedding immunogold labeling was performed to study the subcellular localization of VEGF by means of electron microscopy. MATERIALS AND METHODS Experimental Animals Bleomycin-induced pulmonary fibrosis was examined in female specific pathogen-free Wistar rats (200–300 g body weight) which received a single intratracheal dose of 7 units of bleomycin sulphate dissolved in 0.9% NaCl (0.25 ml/kg body weight). Animals were sacrificed at 5 days (n ⫽ 3), 24 days (n ⫽ 5), 4 weeks (n ⫽ 4), 5 weeks (n ⫽ 4), and 6 weeks (n ⫽ 5) following bleomycin application. The lungs of two untreated female Wistar rats were used as a control. Fixation and tissue processing was performed as described previously (Kasper et al., 1994). Briefly, after sacrifice the lungs were perfused via the right ventricle with phosphate- buffered saline (PBS, pH 7.4) prior to fixation of the lower lobe of the right lung by immersion into 4% formaldehyde in PBS. Subsequently, routine paraffin embedding was performed. Antibodies and Lectins Two affinity-purified polyconal antibodies, of rabbit and goat origin, and one monoclonal anti-VEGF antibody were used (Santa Cruz Biotechnology, Santa Cruz, CA). The monoclonal antibody was raised against a peptide corresponding to amino acids 1–191 of VEGF of human origin (with a deletion from amino acids 142–185) and is claimed to be non-crossreactive with VEGF-B, VEGF-C, or PlGF. The polyclonal antibodies were raised against a peptide corresponding to amino acids 1–20 mapping at the amino terminus of human VEGF. The rabbit antibody was applied at a 1:40 dilution of the original preparation (100 µg IgG/ml in PBS containing 0.1% sodium azide and 0.2% gelatin). The goat antibody was diluted to 1:100 in PBS containing 0.1% acetylated BSA to reduce unspecific background staining. Since the manufacturer offers the control peptide, preabsorption controls could be performed by using the anti-VEGF antibodies preincubated with excess control peptide. A polyclonal rabbit anti-human factor VIII antibody (DAKO Diagnostika, Hamburg, Germany) was used to assess vascularization of fibrotic lung lesions by immunohistochemistry (IH). For double immunofluorescence labeling, the rabbit anti-VEGF antibody was used in conjunction with the following secondary antibodies, lectins, and conjugates: 1) polyclonal mouse anti-rat alpha-smooth muscle (␣-SM) actin antibody (Dianova, Hamburg, Germany), used at a dilution of 1:50; 2) monoclonal mouse antibody E11 (courtesy of Dr. A. Wetterwald, Zürich, Switzerland), which has been shown by immunoelectron microscopy (IEM) to specifically label alveolar type I pneumocytes in rat lung (Wetterwald et al., 1996), used undiluted; 3) monoclonal mouse antibody against surfactant protein D (SP-D) of rat (Dr. S. Albrecht, Dresden, Germany), which has been characterized in detail previously (Kasper et al., 1995a), and was used undiluted; 4) Ulex europaeus agglutinin (UEA) (Vector Laboratories, Burlinghame, CA, USA), used at a dilution of 1:10; 5) Lycopersicon esculentum lectin (LEL) (Vector), used at a dilution of 1:200; 6) Maclura pomifera agglutinin (MPA) (Vector), used at a dilution of 1:100; 7) anti-rabbit antibody coupled to 4,6-dichlorotriazinyl-aminofluorescein (DTAF) (Dianova), used at a dilution of 1:40 for fluorescence labeling of VEGF antibody; 8) Texas red-coupled anti-mouse antibody (Dianova), used at a dilution of 1:80 for fluorescence labeling of E11 antibody and antibody against ␣-SM actin; 9) Texas red-coupled avidin (Vector), used at a dilution of 1:100 for fluorescence labeling of the lectins UEA, LEL, and MPA. For indirect IEM, 20 nm gold-labeled goat anti-rabbit IgG (Dianova) was used to detect primary rabbit antiVEGF antibody. Alternatively, the primary antibody was detected by means of protein A coupled to 15 nm gold (kindly provided by Dr. Posthuma, Utrecht). Fixation and Tissue Processing for IEM For IEM of rat lung, tissue blocks of 1 mm3 in size were fixed by immersion in a mixture of 4% paraformaldehyde and 0.05% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 60 min at room temperature. After infiltration with VEGF IN NORMAL AND FIBROTIC ADULT LUNG 2.3 M sucrose in PBS overnight, tissue blocks were frozen in liquid nitrogen. Subsequently, tissue blocks were transferred to the precooled methanol (⫺80°C) used as the medium for cryo-substitution. After 2–3 days at ⫺80°C and two changes of the substitution medium, low temperature embedding in Lowicryl HM20 and UV polymerization were performed using a CS Auto (Leica, Nussloch, Germany) as described previously (Kasper et al., 1995b). IH Staining For VEGF staining, paraffin sections were mounted on silane-coated slides, dried overnight, dewaxed, and irradiated with microwave in 0.01 M sodium citrate buffer (pH 6.0) at 750 W (twice for 5 min). The staining procedure was done at room temperature (RT), and comprised the following steps: 1) washing in PBS, 5 min; 2) 0.3% hydrogen peroxide, 30 min; 3) 50% fetal calf serum in PBS, 30 min; 4) affinity purified rabbit polyclonal anti-VEGF antibody (Santa Cruz Biotechnology), 1:40 dilution in PBS or affinity purified goat polyclonal anti-VEGF antibody (Santa Cruz Biotechnology), 1:100 dilution in PBS containing 0.1% acetylated BSA (Biotrend); 5) washing in PBS, 10 min; 6) peroxidase-coupled goat anti-rabbit IgG (HRP77; Dr. Grossmann, Dresden), 1:400 dilution in PBS containing 50% fetal calf serum, 60 min, or anti-goat IgG kit (Vector); 7) washing with PBS, 10 min; 8) diaminobenzidine (DAB), 8 min; 9) washing in PBS, 1 min; 10) counterstaining with hemalaun, 2–3 sec. For factor VIII staining (C. Lincoln, personal communication), dewaxed paraffin sections were incubated with polyclonal antibody (DAKO) diluted at 1:500 in PBS for 1 hr (RT), biotinylated anti-rabbit antibody at 1:200 (Vector) for 30 min, followed by ABC (Vector Elite) at a dilution of 1:100 for 30 min, and finally counterstained with hemalaun. Double Immunofluorescence Staining Paraffin sections of normal rat lungs (n ⫽ 2) and of experimental rat lungs fixed at 24 days (n ⫽ 2) and 5 weeks (n ⫽ 2) after bleomycin treatment, respectively, were investigated using the polyclonal rabbit anti-VEGF antibody. Dewaxed and microwave-treated sections were stained at RT as follows: 1) washing in PBS containing 50 mM glycine, 15 min; 2) PBS containing 0.2% gelatine (PBG) and 0.5% bovine serum albumin (BSA), twice, 10 min each; 3) primary antibody, 1:40 dilution in PBG, 45 min; 4) rinsing with PBG, four times, 2 min each; 5) washing with PBS, twice, 2 min each; 6) DTAF-coupled goat anti-rabbit IgG (Dianova), 1:200 dilution in PBS; 7) rinsing with PBS, six times, 2 min each; 8) secondary antibody (see above) in PBG or lectin (see above) in PBS, 45 min unless stated otherwise; 9) rinsing in PBS, six times, 2 min each; 10) Texas red-coupled goat anti-mouse IgG (Dianova) or Texas red-coupled avidin (Vector); 11) washing with PBS, six times, 2 min each; 12) mounting in glycerol-PBS (9:1) containing 2.5% 1,4-Diazobicyclo2.2.2.Octane (DABCO, Janssen, Beerse, Belgium) to reduce fading of fluorescent dyes. Double-stained sections were examined with an Olympus BH-2 microscope equipped with a reflected light fluorescence device (Olympus, Tokyo, Japan). Dye specific fluorescence was selected with standard fluorescence filter sets (Olympus, Tokyo, Japan) to demonstrate fluorescein and Texas red staining, respectively. 63 Immunogold Labeling For ultrastructural localization of VEGF, ultrathin sections were collected on 200-mesh nickel grids coated with 3% collodion. Labeling was performed on a sheet of Parafilm in a humid chamber at RT unless indicated otherwise, and comprised: 1) washing in PBS containing 50 mM glycine, 15 min; 2) PBG containing 0.5% BSA, twice, 10 min each; 3) primary antibody, i.e., polyclonal rabbit anti-VEGF Ab, 1:20 dilution in PBG, overnight at 8°C; 4) rinsing with PBG, four times, 2 min each; 5) washing with PBS, twice, 2 min each; 6) 20 nm gold-coupled goat antirabbit IgG (Dianova), 1:50 dilution in PBS; alternatively, 15 nm gold-coupled protein A (Aurion) was used at a dilution of 1:50 in PBS; 7) rinsing with PBS, six times, 2 min each; 8) counterstaining with 2% uranyl acetate and Reynolds lead citrate. Labeled ultrathin sections were examined with a Zeiss EM 900 operated at 80 kV. Specificity Controls Three specificity controls were performed with each secondary detection system. The primary antibody against VEGF was substituted by 1) primary antibody absorbed to control peptide (Santa Cruz Biotechnology); 2) an irrelevant polyclonal rabbit IgG at an equivalent concentration; or 3) primary antibody was omitted. The lectin binding was specifically blocked by the corresponding sugar (for details, see Kasper and Singh, 1995). Western Blotting Frozen lung tissue was pulverized under liquid nitrogen. After addition of lysis buffer (20 mM HEPES, pH7.9, 400 mM NaCl, 1mM EGTA, 1 mM EDTA, 1% Triton X-100, 0.5% sodium desoxycholate, 0.5% NP-40, 1 mM DTT, 0.5 mM PMSF) the suspension was homogenized for 1 min using an Ultraturrax (IKA Labortechnik, Staufen, Germany). After an incubation time of 10 min, cellular debris was pelleted at 12,000g for 5 min. The supernatant was stored at -80°C. Proteins were fractionated with denaturing polyacrylamide electrophoresis (Laemmli, 1970). Proteins were blotted onto nylon membranes (Pall, Dreieich, Germany). Polyclonal rabbit anti-VEGF antibody was used at a dilution of 1:220. A horseradish peroxidaselabeled donkey anti-rabbit secondary antibody (Amersham, Braunschweig, Germany) was used at a dilution of 1:1000. Super Signal Ultra chemiluminescent substrate (Pierce, via Merck, Dresden, Germany) was applied according to the manufacturer’s instructions. Membranes were exposed to Hyperfilm-ECL (Amersham) for 5–10 min. RESULTS Immunohistochemistry Irrespective of the antibody used, IH staining of paraffinembedded normal rat and human lung, revealed the presence of VEGF in cells not only of the gas-exchanging parenchyma, but also of the gas-conducting bronchioles (Fig. 1A–D). Bronchiolar staining comprised the nonciliated epithelial Clara cells and smooth muscle cells of the peribronchiolar arterioles. The bronchiolar tunica muscularis exhibited moderate staining in human lung and only a trace amount of staining in rat lung. Endothelial cells were devoid of immunohistochemical reaction product. The most prominent alveolar staining was seen in type II pneumocytes (Fig. 1C,D). Alveolar macrophages exhibited Figure 1. 65 VEGF IN NORMAL AND FIBROTIC ADULT LUNG TABLE 1. Immunohistochemical labeling pattern of rat pulmonary cells during the development of bleomycin-induced fibogenesis Time after treatment Cell type Gas-conducting region Bronchial epithelium ciliated nonciliated Vascular endothelium Smooth muscle cells Gas-exchanging region Alveolar epithelium type I cells type II cells Alveolar macrophages Interstitial cells Capillary endothelium Control (n ⫽ 2) 5 days (n ⫽ 3) 24 days (n ⫽ 5) 4 weeks (n ⫽ 4) 5 weeks (n ⫽ 4) 6 weeks (n ⫽ 5) ⫺ ⫹⫹ ⫺ ⫹ ⫺ ⫹⫹ ⫺ ⫹ ⫺ ⫹⫹ (⫺) ⫹ ⫺ ⫹⫹ (⫺) ⫹ ⫺ ⫹⫹ (⫺) ⫹ ⫺ ⫹⫹ (⫺) ⫹ (⫺) ⫹⫹ (⫹) (⫹) (⫺) (⫺) ⫹⫹/> ⫹/> (⫹) (⫺) (⫺) ⫹⫹/> ⫹/> ⫹/⫹⫹> (⫺) (⫺) ⫹⫹/> ⫹/> ⫹/⫹⫹> (⫺) (⫺) ⫹⫹/> ⫹/> ⫹/⫹⫹> (⫺) (⫺) ⫹⫹/> ⫹/> ⫹/⫹⫹> (⫺) ⫺ ⫽ Not stained. (⫺) ⫽ Weak staining of single cells cannot be excluded. (⫹) ⫽ Weak to moderate staining restricted to some cells. ⫹ ⫽ Weak to moderate, generalized staining. ⫹⫹ ⫽ Strong, generalized staining. > ⫽ Cell proliferation. / ⫽ In normal/fibrotic regions, respectively. only weak if any staining for VEGF in rat lungs (Fig. 1C), but were regularly stained in the human lung specimens (Fig. 1D). In the experimental rat lungs treated with bleomycin, interstitial fibrotic regions were observed in the alveolar parenchyma 24 days or more after treatment. Notably, fibrotic regions exhibited strong immunoreactivity for VEGF (Fig. 1E,F,G; Table 1). In the early phase, characterized by hyperplasia of the alveolar epithelium, type II pneumocytes were easily identifiable (Fig. 1E). With progressing fibrosis, however, interstitial fibrotic lesions were densely populated by VEGF-positive cells, which could not unequivocally be ascribed to a distinct cell type (Fig. 1F,G). There were no signs of increased vascularization of the fibrotic lesions as indicated by factor VIII staining (not shown). Immunofluorescence Microscopy The staining patterns observed by double immunofluorescence labeling experiments are summarized in Table 2. Fig. 1. Abbreviations for figures: A: alveolar space; AM: alveolar macrophage; BE: bronchiolar epithelium; C: capillary lumen; E: endothelium; G: secretory granule; L: lamellar body; M: mitochondria; Mu: tunica muscularis; Nu: nucleus; P1: type I pneumocyte; P2: type II pneumocyte; PL: pleura; SM: smooth muscle cell; V: vessel. IH staining for VEGF of normal (A, C) and bleomycin-treated rat lung (E, F) 24 days and (G, H) 6 weeks after treatment, and of normal (B, D) human lung. In the gas-conducting system (A, B, G), VEGF-staining is seen in Clara cells (arrowheads) and smooth muscle cells of the tunica muscularis (Mu). In normal pulmonary parenchyma (C, D), VEGF-staining is prominent in alveolar type II pneumocytes (P2); note difference in staining intensity between rat and human alveolar macrophages (AM). In fibrotic rat lung 24 days after bleomycin-treatment (E), metaplasia of alveolar epithelium with numerous VEGF-positive type II pneumocytes is seen; note focal staining of thin leaflet of transformed alveolar epithelium (arrowhead) and of capillary endothelium (arrow). Areas of interstitial fibrotic lesions (F, G) present with numerous, heterogeneous VEGF-staining cells. Staining is completely abolished by preincubation of the antibody with control peptide (H). Scale bar represents 50 µm (A–F), and 100 µm (G, H). Double staining for VEGF and Ulex europaeus agglutinin (UEA) confirmed the absence of VEGF from vascular endothelial cells of normal rat lung specimens (Figs. 2A, 4A,B). Double staining with anti-VEGF antibody and Lycopersicon esculentum lectin (LEL) confirmed the presence of VEGF-immunoreactivity in surfactant protein D (SP-D)-positive Clara cells and its absence from ciliated cells of the bronchiolar epithelium (Figs. 2D, 3A,E). Looking at the parenchyma, double immunofluorescence staining for SP-D and VEGF revealed that far more alveolar cells stained positive for VEGF than for SP-D, which is specific for type II pneumocytes (Fig. 2C). Double staining for VEGF and the type I pneumocyte-specific monoclonal antibody E11 indicated that VEGF was absent from the thin type I cell leaflets of the air–blood-barrier (Fig. 2B). Double staining for VEGF and Maclura pomifera agglutinin (MPA), which also labels the entire population of alveolar macrophages (Kasper et al., 1994), confirmed that in the rat VEGF was weakly present in this cell type (not shown). Colocalization of ␣-SM actin and VEGF was seen around small vessels (Fig. 3B,F), and at alveolar entrance rings (Fig. 3C,G), which are known to contain ␣-SM actin positive pericytes and ring muscles, respectively (Kapanci et al., 1992). In all cell types, staining for VEGF was largely cytoplasmatic and frequently of granular appearance. In the experimental lungs, double immunofluorescence staining revealed a remarkable heterogeneity of VEGFimmunoreactive cells in fibrotic regions. The population comprised cells that stained for SP-D (Fig. 2E,F), and for ␣-SM actin (Fig. 3D,H), respectively. Flattened epithelial cells lining the alveolar wall stained more intensely for SP-D (Fig. 2E) than those cells within fibrotic lesions (Fig. 2F). While cells staining for the lectins MPA and UEA also contributed to the VEGF-positive cell population in fibrotic regions, cells double-labeled for the type I pneumocytespecific antibody E11 did not (data not shown). Double labeling for VEGF and UEA revealed that, in contrast to control lungs, individual vascular endothelial cells of fibrotic rat lungs were VEGF-immunoreactive (Fig. 4C,D). 66 FEHRENBACH ET AL. TABLE 2. Results of immunofluorescence double labeling of normal rat lung Cell marker Cell type Gas-conducting region Bronchial epithelium ciliated nonciliated Vascular endothelium Smooth muscle cells Gas-exchanging region Alveolar macrophages Alveolar epithelium type I cells type II cells Interstitial cells Myofibroblasts Pericytes Capillary endothelium Antibodies Lectins VEGF E11 SP-D ␣-SMA LEL MPA UEA ⫺ ⫹⫹ ⫺ ⫹ n.e. n.e. n.e. n.e. ⫺ ⫹⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹⫹ ⫹⫹ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ ⫺ ⫹ ⫹ ⫹⫹ ⫺ (⫹) ⫺ (⫹) ⫺ ⫺ ⫹⫹ ⫹ ⫺ ⫹⫹ ⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫺ ⫺ ⫹⫹ ⫹ ⫹ (⫹) (⫹) ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫺ n.e. n.e. ⫺ n.e. n.e. ⫺ n.e. n.e. ⫹ ⫺ ⫽ Not labeled. (⫹) ⫽ Labeling not generalized. ⫹ ⫽ Labeling generalized, but weak. ⫹⫹ ⫽ Strong generalized labeling. n.e. ⫽ Not specifically examined. Immunoelectron Microscopy Indirect immunogold labeling of freeze-substituted parenchymal tissues embedded in Lowicryl HM20 with polyclonal rabbit anti-VEGF antibody revealed a predominant staining of mitochondria of type II pneumocytes, both in rat (Fig. 5A–C) and human lung (Fig. 5D). Identical labeling patterns were seen using gold-coupled IgG (Fig. 5A,D) or protein A-gold (Fig. 5C) to detect the primary antibody. Few gold granules were observed over the nucleus, and even less over endoplasmic reticulum, cytosol, and the surrounding extracellular matrix. Staining was absent from the surfactant storing lamellar bodies. However, gold particles were not restricted to alveolar type II pneumocytes, but were observed to be present over mitochondria of vascular smooth muscle cells (not shown), bronchiolar Clara cells (Fig. 6A), and of virtually every cell type of the alveolar septum. The Clara cells were characterized by additional specific labeling of their secretory granules (Fig. 6A), and the capillary endothelium was unique in that VEGF-labeling additionally localized to intercellular junctions (Fig. 6B). In control sections, which were concurrently performed with every labeling sequence, only slight unspecific labeling was seen (Fig. 5B). Western Blotting Based on Western blots of total lung protein extracts, which were performed using the polyclonal rabbit antibody against VEGF, we could show that a single protein band was present at about 22 to 26 kD (Fig. 7). This corresponds to Western blot data reported from extracts of human testis and seminiferous tubules, which were characterized by a protein band at 24 kD (Ergün et al., 1997). A second protein band at 49 kD indicative of a dimeric form of VEGF in human testis, was not observed in the extracts of rat lung tissue. DISCUSSION The present study investigated the differential distribution of VEGF by means of cell targeting using double immunofluorescence in normal and fibrotic rat and normal human lung. The intracellular localization of VEGF was examined by IEM. A remarkable heterogeneity of the VEGF-immunoreactive cell population was observed by double immunofluorescence microscopy both in normal and fibrotic lungs. VEGF in Normal Adult Lungs A number of studies have shown that VEGF is abundantly present in alveolar type II pneumocytes of normal adult lungs (Christou et al., 1998; Maniscalco et al., 1995, 1997; Monacci et al., 1993; Shifren et al., 1994; Tuder et al., 1995). In our study, VEGF-immunoreactivity was observed not to be restricted to this alveolar cell type, but to be present in bronchiolar Clara cells, smooth muscle cells, and alveolar myofibroblasts as well, while ciliated bronchiolar epithelial cells and vascular endothelial cells did not stain. These observations extend previous reports of the presence of VEGF in pulmonary cell types other than alveolar type II pneumocytes. The results are in line with the studies of Monacci et al. (1993), who, from in situ hybridization experiments, reported that VEGF mRNA was present in virtually every pulmonary alveolar cell of adult rat lung. Shifren et al. (1994) in their study of fetal and adult human tissues observed VEGF-immunoreactivity in pulmonary epithelial cells, myocytes (including smooth muscle cells of vessels), but not in vascular endothelial cells. Tuder et al. (1995) reported VEGF mRNA to be present in type II pneumocytes and alveolar macrophages in control rats, while hypoxia induced additional expression in bronchiolar epithelial cells and in vascular smooth muscle cells. In contrast, Maniscalco et al. (1995), studying rabbit tissues, reported that VEGF mRNA and peptide were enriched in type II pneumocytes, but absent from alveolar macrophages, fibroblasts, and endothelial cells. Further, these authors reported only little or no VEGFexpression in large vessel endothelial cells, airway cells, smooth muscle cells, and type I pneumocytes. We have shown that VEGF is present in ␣-SM actin-positive pericytes and vascular smooth muscle cells of normal adult Fig. 2. Double immunofluorescence staining of normal (A–D) and fibrotic (E, F) rat lung. Green (fluorescein staining) indicates VEGFimmunoreactivity; red (Texas Red staining) indicates UEA (A), E11 (B), and SP-D (C, D) labeling; orange to yellow indicates colocalization of both markers. (A) While only smooth muscle cells stain for VEGF (green), with some cells also staining for UEA (yellow), the endothelial cells are only positive for UEA (arrowheads). (B) VEGF is seen in type II pneumocyte (arrowheads), but is absent from E11-labeled type I pneumocytes (red). (C) Type II pneumocytes staining for SP-D are always VEGF-positive as indicated by yellow to orange fluorescence, while numerous other cells of the alveolar septum only stained for VEGF (green). (D) Clara cells stained positively for both VEGF and SP-D. (E, F) After 24 days of bleomycintreatment, flattened alveolar epithelial cells stained strongly for SP-D and VEGF (E), while only weak staining for SP-D was seen in cells within fibrotic lesions (F). Scale bar ⫽ 100 µm (A, C), and 20 µm in (B, D–F). 68 FEHRENBACH ET AL. Figure 3. VEGF IN NORMAL AND FIBROTIC ADULT LUNG 69 Fig. 4. Double immunofluorescence staining of arterial vessel wall of normal (A, B) and bleomycin-treated (C, D) rat lung to demonstrate immunoreactivity for VEGF (A, C) and UEA (B, D), respectively. While in control lung tissue UEA-positive endothelium (B) is devoid of VEGF- immunoreactivity (A), individual UEA-positive vascular endothelial cells (D) were seen in fibrotic lungs to be immunoreactive for VEGF (arrowheads in C). Arterial smooth muscle cells (SM) were immunoreactive for VEGF in both normal and fibrotic lungs. Scale bar ⫽ 20 µm. lungs, which is in line with in situ hybridization studies performed to localize VEGF mRNA (Shifren et al., 1994; Tuder et al., 1995). Although the precise role of VEGF in these cell types still remains to be determined, one may follow Senger et al. (1993), who proposed that vascular smooth muscle-VEGF might be involved in the maintenance of normal vascular function. In general, VEGF-staining has been reported to be cytoplasmatic (Couffinhal et al., 1997; Ergün et al., 1997; Shifren et al., 1994). However, Park et al. (1993) presented evidence that there may be differences in the localization of the different isoforms. By means of a preembedding immunogold technique and electron microscopy, immuno- reactivity for VEGF189/206 has been localized to extracellular matrix-like material in transfected human embryonic kidney CEN4 cells expressing the respective isoform (Park et al., 1993). Using postembedding immunogold and protein A-gold techniques, we could demonstrate by electron microscopy that VEGF-immunoreactivity was enriched in mitochondria. Specificity of the labeling was confirmed by preabsorption of the primary antibody to the peptide used for immunization, which always abolished mitochondrial staining. Labeling was most prominent in type II pneumocytes and bronchiolar Clara cells, but was observed in virtually every cell type. This again corresponds to in situ hybridization studies showing the widespread presence of VEGF mRNA in rat lung (Monacci et al., 1993). While the strong immunoreactivity of the secretory granules of Clara cells may indicate that VEGF is released along this secretory pathway, the significance of the high level of VEGF-immunoreactivity in mitochondria remains unclear. By analogy, another growth factor, TGF-␤1, has been localized to mitochondria of rat and mouse cardiac myocytes and rat hepatocytes (Heine et al., 1991). As with VEGF, the functional role of TGF-␤1 in mitochondria is still unknown. One may speculate that mitochondria represent a potential intracellular site of phosphorylation. In agreement with previous studies (Maniscalco et al., 1995; Shifren et al., 1994; Tuder et al., 1995), vascular Fig. 3. Double immunofluorescence staining of normal (A–C, E–G) and bleomycin-treated (D, H) rat lung. (A–D) VEGF immunofluorescence labeling with corresponding staining for (E) LEL, and (F–H) ␣-SM actin. (A, E) Intense fluorescence of VEGF-positive granules is confined to apical portion of Clara cells (arrowheads). (B, F) VEGF-positive staining is seen to colocalize with ␣-SM actin-positive pericytes at capillary-venular junction (arrowheads); asterisks indicate two VEGF-positive type II pneumocytes. (C, G) VEGF-positive, ␣-SM actin-positive myofibroblasts at an alveolar entrance ring. (D, H) Fibrotic lesions contain VEGF-positive, ␣-SM actin-positive cells (black arrowheads), VEGF-positive, ␣-SM actinnegative cells (arrow) as well as VEGF-negative, ␣-SM actin-positive cells (white arrowhead). Scale bar ⫽ 10 µm (A–C, E–G), and 50 µm (D, H). 70 FEHRENBACH ET AL. Fig. 5. Indirect immunogold labeling of cryo-substituted, Lowicrylembedded rat (A, B, C) and human (D) lung. Mitochondria (M) of alveolar type II pneumocytes are specifically stained by the affinity-purified, polyclonal rabbit anti-VEGF antibody, while lamellar bodies (L) are devoid of labeling irrespective of using gold-conjugated anti-rabbit IgG (A, D) or proteinA-gold (C) as secondary detection system. Staining of mitochondria and interendothelial cell junctions (arrowhead) is not seen with the primary VEGF-antibody being preabsorbed to control peptide (B). Scale bars ⫽ 1 µm. endothelial cells of normal rat lungs did not stain for VEGF by IH. However, IEM revealed discrete immunogold labeling for VEGF of interendothelial cell junctions of alveolar capillaries in normal rat lungs. This specific localization of VEGF to a structure actively involved in the regulation of capillary permeability may indicate a poten- VEGF IN NORMAL AND FIBROTIC ADULT LUNG 71 Fig. 6. Indirect IgG-gold labeling of cryo-substituted, Lowicryl embedded normal rat lung. Beside mitochondrial staining, specific labeling was seen (A) in secretory granules (G) of bronchiolar Clara cells, and (B) to be associated with interendothelial junctions (arrowheads) of capillary endothelium (E). Scale bars ⫽ 1 µm. a dose- and time-dependent increase in permeability (Wang et al., 1996). In the frog mesentery, exposure to 1 nM VEGF rapidly and transiently increased microvessel hydraulic conductivity within 30 sec and returned to control within 2 min (Bates and Curry, 1996). VEGF in Pulmonary Fibrosis Fig. 7. Western blot of protein extract of total lung tissue. A single protein band at about 24 kD was detected by means of the rabbit polyclonal antibody against VEGF used in most of the immunostaining experiments. tial role of VEGF in the regulation of normal capillary function. While the involvement of VEGF in the induction of hyperpermeability of tumor-associated neovessels is widely accepted (Dvorak et al., 1995; Feng et al., 1996; Qu et al., 1995; Senger et al., 1993), evidence supporting a role for VEGF in the physiological regulation of vascular permeability has been presented only recently. In cultured bovine brain microvessel endothelial cells, VEGF induced In fibrotic regions of bleomycin-treated rat lungs, VEGFimmunoreactive cells were abundantly present. By double immunofluorescence microscopy, alveolar type II pneumocytes as identified by their staining for surfactant protein D (SP-D), and myofibroblasts as identified by ␣-SM actinstaining, were prominent contributors to the VEGFimmunoreactive population of cells in rat lung fibrotic regions. Notably, SP-D staining of cells within fibrotic regions was less intense compared with cells facing the alveolar air space. This may correspond to the finding that type II pneumocytes containing abundant VEGF mRNA have little if any SP-C mRNA, and vice versa (Maniscalco et al., 1995). Early morphometric studies have shown that at 28 days after bleomycin-instillation, the largest increase in any cell population was in contractile interstitial cells, which have subsequently been shown to be predominantly ␣-SM actin-positive myofibroblasts (Adler et al., 1986, 1989; Vyalov et al., 1993). Proliferation of alveolar myofibroblasts was observed as early as 24 hr to 3 days after intratracheal bleomycin administration, and preceded the formation of collagen fibers (Vyalov et al., 1993). Knowing that hypoxia is a potent pathological stimulus for the upregulation of VEGF (Namiki et al., 1995; Stavri et al., 1995; Tuder et al., 1995), the increase in VEGF-positive cells in fibrotic lung regions may be a response to an impaired oxygen supply of the thickened fibrotic septa. VEGF was shown to induce several rapid cellular responses in cultured human endothelial cells as, e.g., a 72 FEHRENBACH ET AL. twofold increase in the release of Von Willebrand factor (vWF) (Brock et al., 1991). In radiation-induced pulmonary fibrosis, the number of vWF-positive cells per parenchymal unit decreased during the first 2 months after irradiation and showed a marked increase thereafter (Kasper et al., 1996). The fact that we could not observe a similar increase in vascularization in our bleomycin model may be explained by the shorter period of 6 weeks of observation in this study. Recently, VEGF and one of its receptors (flt-1) have been demonstrated to be elevated in activated macrophages in pulmonary sarcoidosis (Tolnay et al., 1998). Further, VEGF has been shown to act as a chemoattractant of mast cells (Gruber et al., 1995), which are well known to immigrate into fibrotic lung lesions (Goto et al., 1984). Thus, the elevated level of VEGFpositive type II pneumocytes and myofibroblasts may induce immigration and aggregation of activated macrophages and mast cells in fibrotic lung lesions. We conclude that the predominant localization of VEGFimmunoreactivity to bronchiolar and alveolar epithelial cells, and to ␣-SM actin-positive cells in adult rat and human lungs, and the marked increase of VEGF-positive cells in the absence of an induction of vascularization in fibrotic lesions of bleomycin-treated rat lungs, indicate that VEGF exerts its action in processes other than angiogenesis in the adult lung. ACKNOWLEDGMENTS The expert technical assistance of I. Peterson, B. Rudolph, S. Petzold, S. Langer, and H. Seidel is acknowledged with thanks. Professor Dr. K. Kayser (Thoraxclinic, University of Heidelberg, Germany) kindly provided the human material used in this study. We thank Professor Dr. K.-W. Wenzel and Dr. R. Koslowski (Institute of Physiological Chemistry, TU Dresden, Germany) who provided the bleomycin rat lung model. REFERENCES Adamson IYR. 1984. Drug-induced pulmonary fibrosis. Environ Health Perspect 55:25–36. Adamson IYR, Orr FW, Young L. 1986. Effects of injury and repair of the pulmonary endothelium on lung metastasis after bleomycin. J Pathol 150:279–287. Adler KB, Callahan LM, Evans JN. 1986. Cellular alterations in the alveolar wall in bleomycin-induced pulmonary fibrosis in rats. An ultrastructural morphometric study. Am Rev Respir Dis 133:1043– 1048. Adler KB, Low RB, Leslie KO, Mitchell J, Evans JN. 1989. Contractile cells in normal and fibrotic lung. Lab Invest 60:473–485. Bates DO, Curry FE. 1996. Vascular endothelial growth factor increases hydraulic conductivity of isolated perfused microvessels. Am J Physiol 271:H2520–H2528. Berse B, Brown LF, Van de Water L, Dvorak HF, Senger DR. 1992. Vascular permeability factor (vascular endothelial growth factor) gene is expressed differentially in normal tissues, macrophages, and tumors. Mol Biol Cell 3:211–220. Brock TA, Dvorak HF, Senger DR. 1991. Tumor-secreted vascular permeability factor increases cytosolic Ca2⫹ and von Willebrand factor release in human endothelial cells. Am J Pathol 138:213–221. Brown LF, Detmar M, Tognazzi K, Abu-Jawdeh G, Iruela-Arispe ML. 1997. Uterine smooth muscle cells express functional receptors (flt-1 and KDR) for vascular permeability factor/vascular endothelial growth factor. Lab Invest 76:245–255. Christou H, Yoshida A, Arthur V, Morita T, Kourembanas S. 1998. Increased vascular endothelial growth factor production in the lungs of rats with hypoxia-induced pulmonary hypertension. Am J Respir Cell Mol Biol 18:768–776. Couffinhal T, Kearney M, Witzenbichler B, Chen D, Murohara T, Losordo DW, Symes J, Isner JM. 1997. Vascular endothelial growth factor/vascular permeability factor (VEGF/VPF) in normal and artheriosclerotic human arteries. Am J Pathol 150:1673–1685. Dvorak HF, Brown LF, Detmar M, Dvorak AM. 1995. Vascular permeability factor/vascular endothelial growth factor, microvascular hyperpermeability, and angiogenesis. Am J Pathol 146:1029– 1039. Ergün S, Kilic N, Fiedler W, Mukhopadhyay AK. 1997. Vascular endothelial growth factor and its receptors in normal human testicular tissue. Mol Cell Endocrinol 131:9–20. Feng D, Nagy JA, Hipp J, Dvorak HF, Dvorak AM. 1996. Vesiculovacuolar organelles and the regulation of venule permeability to macromolecules by vascular permeability factor, histamine, and serotonin. J Exp Med 183:1981–1986. Fine A, Goldstein RH. 1997. Animal models of pulmonary fibrosis. In: The Lung. Scientific Foundations. Crystal RG, West JB, Barnes PJ, Weibel ER, editors. Philadelphia, New York: Lippincott/Raven Publishers. p 2525–2536. Goto T, Befus D, Low R, Bienenstock J. 1984. Mast cell heterogeneity and hyperplasia in bleomycin-induced pulmonary fibrosis of rats. Am Rev Respir Dis 130:797–802. Gruber BL, Marchese MJ, Kew R. 1995. Angiogenic factors stimulate mast-cell migration. Blood 86:2488–2493. Heine UI, Burmester JK, Flanders KC, Danielpour D, Munoz EF, Roberts AB, Sporn MB. 1991. Localization of transforming growth factor-beta 1 in mitochondria of murine heart and liver. Cell Regul 2:467–477. Horiuchi T, Weller PF. 1997. Expression of vascular endothelial growth factor by human eosinophils: Upregulation by granulocyte macrophage colony-stimulating factor and interleukin-5. Am J Respir Cell Mol Biol 17:70–77. Kapanci Y, Ribaux C, Chaponnier C, Gabbiani G. 1992. Cytoskeletal features of alveolar myofibroblasts and pericytes in normal human and rat lung. J Histochem Cytochem 40:1955–1963. Kasper M, Haroske G. 1996. Alterations in the alveolar epithelium after injury leading to pulmonary fibrosis. Histol Histopathol 11:463– 483. Kasper M, Singh G. 1995. Epithelial lung cell marker: Current tools for cell typing. Histol Histopathol 10:155–169. Kasper M, Sakai K, Koslowski R, Wenzel KW, Haroske G, Schuh D, Müller M. 1994. Localization of surfactant protein A (SP-A) in alveolar macrophage subpopulations of normal and fibrotic rat lung. Histochemistry 102:345–352. Kasper M, Albrecht S, Grossmann H, Grosser M, Schuh D, Müller M. 1995a. Monoclonal antibodies to surfactant protein D: Evaluation of immunoreactivity in normal rat lung and in a radiation-induced fibrosis model. Exp Lung Res 21:577–588. Kasper M, Huber O, Grossmann H, Rudolph B, Tränkner C, Müller M. Immunocytochemical distribution of E-cadherin in normal and injured lung tissue of the rat. 1995b. Histochem Cell Biol 104:383– 390. Kasper M, Schöbl R, Haroske G, Fischer R, Neubert F, Dimmer V, Müller M. 1996. Distribution of von Willebrand factor in capillary endothelial cells of rat lungs with pulmonary fibrosis. Exp Toxicol Pathol 48:283–288. Klagsbrun M, D’Amore PA. 1996. Vascular endothelial growth factor and its receptors. Cytokine Growth Factor Rev 7:259–270. Kuhn C, Boldt J, King TE Jr, Crouch E, Vartio T, McDonald JA. 1989. An immunohistochemical study of architectural remodeling and connective tissue synthesis in pulmonary fibrosis. Am Rev Respir Dis 140:1693–1703. Laemmli UK. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. Maniscalco WM, Watkins RH, Finkelstein JN, Campbell MH. 1995. Vascular endothelial growth factor mRNA increases in alveolar epithelial cells during recovery from oxygen injury. Am J Respir Cell Mol Biol 13:377–386. Maniscalco WM, Watkins RH, D’Angio CT, Ryan RM. 1997. Hyperoxic injury decreases alveolar epithelial cell expression of vascular endothelial growth factor (VEGF) in neonatal rabbit lung. Am J Respir Cell Mol Biol 16:557–567. VEGF IN NORMAL AND FIBROTIC ADULT LUNG Monacci WT, Merrill MJ, Oldfield EH. 1993. Expression of vascular permeability factor/vascular endothelial growth factor in normal rat tissues. Am J Physiol 264:C995–C1002. Namiki A, Brogi E, Kearney M, Kim EA, Wu T, Couffinhal T, Varticovski L, Isner JM. 1995. Hypoxia induces vascular endothelial growth factor in cultured human endothelial cells. J Biol Chem 270:31189–31195. Park JE, Keller GA, Ferrara N. 1993. The vascular endothelial growth factor (VEGF) isoforms: Differential deposition into the subepithelial extracellular matrix and bioactivity of extracellular matrixbound VEGF. Mol Biol Cell 4:1317–1326. Qu H, Nagy JA, Senger DR, Dvorak HF, Dvorak AM. 1995. Ultrastructural localization of vascular permeability factor/vascular endothelial growth factor (VPF/VEGF) to the abluminal plasma membrane and vesiculovacuolar organelles of tumor microvascular endothelium. J Histochem Cytochem 43:381–389. Rooman I, Schuit F, Bouwens L. 1997. Effect of vascular endothelial growth factor on growth and differentiation of pancreatic ductal epithelium. Lab Invest 76:225–232. Senger DR, Galli SJ, Dvorak AM, Perruzzi CA, Harvey VS, Dvorak HF. 1983. Tumor cells secrete a vascular permeability factor that promotes accumulation of ascites fluid. Science 219:983–985. Senger DR, Van De Water L, Brown LF, Nagy JA, Yeo KT, Yeo TK, Berse B, Jackman RW, Dvorak AM, Dvorak HF. 1993. Vascular permeability factor (VPF, VEGF) in tumor biology. Cancer Metastasis Rev. 12:303–324. Shifren JL, Doldi N, Ferrara N, Mesiano S, Jaffe RB. 1994. In the human fetus, vascular endothelial growth factor is expressed in epithelial cells and myocytes, but not vascular endothelium: Implications for mode of action. J Clin Endocrinol Metab 79:316–322. Shinohara K, Shinohara T, Mochizuki N, Mochizuki Y, Sawa H, Kohya T, Fujita M, Fujioka Y, Kitabatake A, Nagashima K. 1996. Expression of vascular endothelial growth factor in human myocardial infarction. Heart Vessels 11:113–122. Stavri GT, Zachary IC, Baskerville PA, Martin JF, Erusalimsky JD. 73 1995. Basic fibroblast growth factor upregulates the expression of vascular endothelial growth factor in vascular smooth muscle cells. Synergistic interaction with hypoxia. Circulation 92:11–14. Tolnay E, Kuhnen C, Voss B, Wiethege T, Müller KM. 1998. Expression and localization of vascular endothelial growth factor and its receptor flt in pulmonary sarcoidosis. Virchows Arch 432:61–65. Tuder RM, Flook BE, Voelkel NF. 1995. Increased gene expression for VEGF and the VEGF receptors KDR/Flk and Flt in lungs exposed to acute or to chronic hypoxia. Modulation of gene expression by nitric oxide. J Clin Invest 95:1798–1807. Voelkel NF, Tuder RM. 1995. Cellular and molecular mechanisms in the pathogenesis of severe pulmonary hypertension. Eur Respir J. 8:2129–2138. Vyalov SL, Gabbiani G, Kapanci Y. 1993. Rat alveolar myofibroblasts acquire alpha-smooth muscle actin expression during bleomycininduced pulmonary fibrosis. Am J Pathol 143:1754–1765. Wang W, Merrill MJ, Borchardt RT. 1996. Vascular endothelial growth factor affects permeability of brain microvessel endothelial cells in vitro. Am J Physiol 271:C1973–C1980. Wetterwald A, Hoffstetter W, Cecchini MG, Lanske B, Wagner C, Fleisch H, Atkinson M. 1996. Characterization and cloning of the E11 antigen, a marker expressed by rat osteoblasts and osteocytes. Bone 18:125–132. Wolff G, Crystal RG. 1997. Biology of pulmonary fibrosis. In: The Lung. Scientific Foundations. Crystal RG, West JB, Barnes PJ, Weibel ER, editors. Philadelphia, New York: Lippincott/Raven Publishers. p 2509–2524. Woodcock-Mitchell J, Adler KB, Low RB. 1984. Immunohistochemical identification of cell types in normal and in bleomycin-induced fibrotic rat lung. Cellular origins of interstitial cells. Am Rev Respir Dis 130:910–916. Yang W, de Bono DP. 1997. A new role for vascular endothelial growth factor and fibroblast growth factors: Increasing endothelial resistance to oxidative stress. Febs Lett 403:139–142.