Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Fibrinogen (FBG) is a dimeric plasma glycoprotein that is composed of three pairs of non-identical disulfide-bonded polypeptide chains and is synthesized constitutively by hepatocytes (1, 2). FBG is also a component of the acute phase response to inflammation during which hepatic synthesis is upregulated 2-10-fold (3). Activation of the coagulation cascade after vascular injury results in the conversion of FBG to fibrin through the proteolytic action of thrombin, leading to formation of a hemostatic plug (4). Fibrin also plays an important role in the interaction of cells with the vessel wall in processes critical for wound repair and revascularization. Specific structural features of the provisional fibrin(ogen) matrix mediate cellular functions such as adhesion and spreading, proliferation, and migration of a variety of different cell types, including endothelial, fibroblasts, epithelial, and platelets (5).

Several extrahepatic sites of FBG synthesis have been identified, suggesting that this protein may function independently of hemostasis in cellular adhesive interactions or the maintenance of structural integrity of these tissues. While FBG A, B, and  chains are all expressed by hepatocytes (1),  chain expression has been demonstrated in several extrahepatic sources in vivo including brain, lung, and marrow (13, 14). Furthermore, in vitro studies have indicated that several nonhepatic epithelial cells synthesize and secrete FBG. Cultured uterine cervical carcinoma (15) and intestinal epithelial (Caco-2) cells express FBG constitutively (16). Upon stimulation with proinflammatory mediators, FBG expression is upregulated in Caco-2 epithelial cells (16). Recently, synthesis and secretion of fully assembled FBG from the lung alveolar epithelial cell line A549 was demonstrated (17). Although little constitutive FBG expression occurs in the lung epithelial cells, the FBG genes are transcriptionally upregulated 5-10-fold after induction with DEX and IL-6 (17), the proinflammatory mediator of an acute phase response (18). These results suggest that lung epithelium is a likely site of fully assembled FBG production during local inflammation (17).

Polarized epithelial cells, such as those of the lung, secrete proteins vectorially toward the apical or basolateral surface, directing them to domains specific for their function (19, 20). Moreover, microtubules function in cellular trafficking by mediating the transport and delivery of vesicles to the correct surface (21, 22). The present study is aimed at characterizing the polarity of FBG secretion from lung epithelial cells in vitro to aid in elucidating its function in vivo. As pneumocytes are exposed apically to the airways and basolaterally to the basement membrane, lung epithelial cell-derived FBG could be directed to either or both of two distinct extracellular environments. Evidence is presented that FBG secretion occurs predominantly in the basolateral direction and that such polarized secretion is dependent on an intact microtubular cytoskeleton.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cells and Culture Conditions

A549 human lung alveolar epithelial cells (CCL 185) were grown to confluent monolayers in Kaighn's Nutrient Mixture F12 medium (Irvine Scientific, Irvine, CA) containing 10% fetal bovine serum (Intergen, Purchase, NY), penicillin (100 U/ml) and streptomycin (0.1 mg/ml), and L-glutamine (2 mM) (Life Technologies, Gaithersburg, MD). HepG2 (HB 8065), a human hepatocarcinoma cell line, was cultured in Eagle's MEM supplemented with 10% FBS, penicillin (100 U/ml), and streptomycin (0.1 mg/ml), L-glutamine (2 mM), nonessential amino acids (0.1 mM), sodium pyruvate (1.0 mM), and Tricine (15 mM, pH 7.4). A549 and HepG2 cells were seeded onto 25 cm² flasks, 6-well plates, or 24 mm Transwell-Col^TM polycarbonate filters with 3.0 µm pores housed in 6-well cluster plates (Costar, Cambridge, MA). Culture media volumes in the Transwell-Col^TM filter chambers were 2.5 ml (basolateral) and 1.5 ml (apical). For induction of FBG synthesis, cells were stimulated as described with IL-6 (25 U/ml) and DEX (0.1 µM) (17). For studies involving microtubule depolymerization, cells were incubated with colchicine (Sigma, St. Louis, MO) at a final concentration of 10 µM (23).

¹²⁵I-Labeling

Purified human FBG (1 mg) (CalBiochem, La Jolla, CA) was radiolabeled with 1 mCi of carrier-free Na¹²⁵I (Dupont/NEN, Boston, MA) using 1.0 µg of Iodo-Gen (Pierce, Rockford, IL). Unbound ¹²⁵I was removed by gel filtration using a PD-10 column (Pharmacia, Piscataway, NJ). For studies of FBG diffusion between chambers, ¹²⁵I-FBG was added to culture medium either in the apical, basolateral, or both chambers of control and treated cells. Samples of the ¹²⁵I-FBG conditioned medium were removed following 2 h incubation at 37°C, the time utilized for metabolic pulse-chase labeling of the cells. Radioactivity was determined in a -counter.

Metabolic Labeling and Immunoprecipitation

Metabolic pulse-labeling of A549 cells was chosen since it results in secretion of FBG with high specific activity within a short (2 h) time period. After induction with IL-6 + DEX, the conditioned media was removed, and saved, then cell monolayers were washed and starved for 15 min in serum-free media. The cells were pulsed for 30 min with 1 mCi/ml of ³⁵S-methionine + cysteine Express Protein Labeling Mix (Dupont/NEN) diluted in complete media. A chase incubation with conditioned media followed for 1.5 h, at which time culture media was collected from apical and basolateral chambers. For analysis of cellular FBG, cells cultured in 6-well plates or on polycarbonate filters were lysed as previously described (23). FBG was immunoprecipitated from lysate or culture supernatant using rabbit anti-human FBG IgG (27) (Dako, Carpinteria, CA) coupled to Protein A-Sepharose (Sigma) (17, 23). This polyclonal anti-FBG IgG is free of cross-reactivity with normal human serum proteins, and recognizes all three polypeptide chains of FBG (23); however, as only intact FBG is secreted from A549 (17) and HepG2 cells (24), only intact FBG is immunopurified from culture media. After immunoprecipitation, samples were analyzed by SDS-PAGE and fluorography (27). Densitometry scanning of fluorographs was performed using the NIH Image 1.59 program. Because synthesis of the B chain is rate limiting in the synthesis of human FBG (24), the amount of labeled FBG in the different cellular compartments was determined by comparison of the intensities of the B chain only. P values were obtained using unpaired Student's t-test (two-tailed); P values < 0.05 indicate statistical significance. Total radiolabeled proteins were precipitated with cold 10% TCA, washed four times with cold 5% TCA, and the activity in cpm in the samples was determined by liquid scintillation.

Immunofluorescent Staining

Immunofluorescent labeling of cells cultured on glass coverslips was carried out as described (28, 29). Cells were fixed in 3.7% formaldehyde, permeabilized with 0.5% Triton X-100, and stained with 3 µg/ml anti- tubulin monoclonal antibody (mAb) (Amersham, Arlington Heights, IL), followed by rhodamine-conjugated rabbit anti-mouse IgG (1:10 dilution) (Cappel, Durham, NC). Actin staining was performed using rhodamine phalloidin (0.33 µM) (Molecular Probes, Eugene, OR). FBG staining was performed in a similar manner using 10 µg/ml mAb J88B against human FBG  chain (28). Coverslips were mounted onto glass slides with Fluoromount (Southern Biotechnology Associates, Birmingham, AL).

Immunoelectron Microscopy (IEM)

Cells were grown on Transwell-Col^TM filters as described above. After induction, cells attached to the polycarbonate filters were fixed at 22°C for 1 h in 4% paraformaldehyde, 0.1% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.2. After washing, filters containing cells were cut from filter housings and dehydrated in ethanol. Filters containing cells were infiltrated into embedding resin with several changes of LR White (Electron Microscopy Sciences, Fort Washington, PA) onto 35-mm film caps covered with Aclar embedding film (Ted Pella, Inc., Redding, CA). Cell monolayers on filters were also fixed with a traditional EM fixation protocol of 2.5% glutaraldehyde in 0.1 M sodium phosphate buffer, pH 7.2 for 1 h. Post-fixation was done in 1% osmium tetroxide in 0.1 M sodium phosphate buffer. After rinsing, filters were cut from the filter housings and dehydrated in a graded ethanol series, followed by propylene oxide. Embedding of cells on filters was done in epoxy resin and polymerization at 60°C. Thin sections were cut from blocks and placed onto formvar-coated, slotted nickel grids. Immunolabeling was performed with normal rabbit serum IgG, which was purified over a column containing immobilized human plasma proteins (Dako), and with rabbit anti-FBG IgG purified as described (30). The IgGs were diluted to 0.1 mg/ml in 1% ovalbumin in PBS. Grids were placed section side down onto parafilm bearing drops of primary antibody. After washings, grids were incubated with 1:20 dilution of Protein A-gold (20 nm) (Goldmark Biologicals, Phillipsburg, NJ). Grids were washed, then stained with aqueous uranyl acetate and lead citrate (31). Photographs were taken at 60 kv on a Zeiss 10C transmission electron microscope.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

FBG Secretion from A549 Cells Is Polarized

To determine the time of peak FBG secretion, A549 cells were stimulated with IL-6 + DEX for times varying between 0 and 24 h, then metabolically pulse-labeled for 30 min followed by a 1.5-h chase. FBG immunopurified from culture medium was analyzed by SDS-PAGE and fluorography. The peak of FBG secretion occurred at 18 h of IL-6 + DEX treatment (Figure 1); therefore, this induction time was chosen for subsequent experiments. To analyze the polarity of FBG secretion, A549 and HepG2 cells were cultured on polycarbonate filters, which provide separation of apical and basolateral chambers. To ascertain whether mixing between chambers occurred, ¹²⁵I-FBG was added separately to apical or basolateral medium of control and 18 h IL-6 + DEX pretreated-cells for an additional 2 h, then the activity in cpm in each chamber was determined. Greater than 99% of ¹²⁵I-FBG remained in the chamber to which it was added for both control and IL-6 + DEX treated cells (not shown). Prior to analysis of FBG secretion, the distribution of total secreted protein between the two chambers was analyzed for both cell types. The total TCA-precipitable ³⁵S-protein secreted into apical and basolateral chambers was determined after metabolic pulse-labeling both IL-6 + DEX-treated and untreated cells. The treatment of cells with IL-6 + DEX did not affect the polarity of total protein secretion, as there was no significant difference in the distribution of TCA-precipitable protein before and after treatment for A549 (P = 0.26) or HepG2 (P = 0.65) cells (Table 1). While proteins secreted from HepG2 cells were nearly equally distributed between apical and basolateral chambers, a slightly greater, but statistically insignificant (P = 0.59), proportion of labeled proteins from A549 cells were secreted into the basolateral chamber (Table 1).

View larger version (38K): [in this window] [in a new window]   Figure 1.   Time course of FBG biosynthesis during induction. A549 cells untreated (con) or treated with IL-6 + DEX for 2, 6, 18, and 24 h. Conditioned media was collected following pulse-chase radiolabeling during the last 2 h of treatment. SDS-PAGE of immunoprecipitated samples confirms intact FBG A, B and  chain polypeptides. Note that B chain bands appear more intense than A and  chains, indicative of rapid labeling and secretion of B chain complexed with unlabeled free A and  chain polypeptides found in intracellular pools (24). Only intact FBG is secreted from A549 (17) and HepG2 cells (24). During a metabolic labeling experiment where the cells are pulsed for 30 min with 35S-cys + met, then chased with label-free complete medium, the B chain is the first polypeptide to be labeled. The FBG that is assembled will first contain unlabeled  and A chains assembled with newly synthesized, radioactive B chain. As the preexisting pool of unlabeled  and A chains are depleted, newly synthesized  and A polypeptides will be labeled and incorporated with the newly synthesized B chain, then secreted.

Immunoprecipitation and gel analysis of metabolically labeled FBG following IL-6 + DEX induction showed that HepG2 cells secreted 51.0% of FBG basolaterally (Figure 2A; Table 1). No change in distribution of FBG was observed due to IL-6 + DEX treatment, as 53.1% of FBG was secreted basolaterally from control HepG2 cells (Figure 2A; Table 1). Since only disulfide-bonded FBG is secreted from A549 (17) and HepG2 cells (24), densitometric quantitation of total FBG is based only on intensity of the B chains. Comparison of the intensity of the B chain most accurately reflects the synthetic capacity of the cells and is the best way to monitor changes in partitioning of the FBG between apical and basolateral chambers. FBG secretion from IL-6 + DEX-treated A549 cells was more highly polarized, with 80.0% secreted into the basolateral chamber (Figure 2B; Table 1). Although much less FBG was secreted from control A549 cells (undetectable in the exposure of the fluorograph of Figure 2B), the apical/basolateral distribution was not different (P > 0.05; Table 1), indicating that IL-6 + DEX treatment enhanced FBG secretion, but did not alter its polarization.

View larger version (17K): [in this window] [in a new window]   Figure 2.   Polarity of FBG secretion from liver and lung epithelial cells. HepG2 (A) and A549 (B) cells grown on Transwell-ColTM polycarbonate filters were stimulated with IL-6 + DEX for 18 h then pulsed with 35S-cys + met. SDS-PAGE of immunoprecipitated apical media (up arrow) and basolateral media (down arrow) shows distribution of secreted FBG.

While the liver and the lung airways are composed primarily of epithelial cells, there are distinct differences in the morphological architecture and function of these polarized cell types. The cellular architecture of lung epithelium forms a highly polarized environment to separate the airway lumen from the abluminal basement membrane. The architecture of liver is represented by continually curving plates of hepatocytes bounded on the basolateral sides by sinusoids. The formation of tight junctions at the apical side of cells is a hallmark of highly polarized cells. Electron micrographs of HepG2 and A549 cells grown to confluent monolayers on Transwell-Col^TM filters indicated the presence of tight junctions between adjacent cells (Figures 3A and 3B), confirming a polarized phenotype. In addition, the morphological characteristics of the HepG2 and A549 cells grown on filters were similar to those in vivo (Figures 3C and 3D). The membranes of HepG2 cells grown on filters formed numerous microvilli (Figure 3C) that resemble the perisinusoidal space of Dissé, the sinusoid equivalent of a basement membrane through which plasma proteins are delivered to the bloodstream in vivo (32). The appearance of numerous loose and dense lamellar bodies in A549 cells grown on the polycarbonate filters, primarily at the apical face of the cells, is indicative of a polarized alveolar epithelial phenotype (Figure 3D). A549 cells were originally derived from a human lung carcinoma with properties of type II epithelium (33). However, as repair of injured alveolar epithelium comes from proliferation of type II cells to give rise to alveolar type I pneumocytes, cultured A549 cells treated with IL-6 + DEX are denoted as alveolar epithelium in this study.

View larger version (92K): [in this window] [in a new window]   View larger version (110K): [in this window] [in a new window]   Figure 3.   Transmission electron micrographs of tight junctions. The apical face of the cell layer is oriented toward the top of the micrograph. The polycarbonate filter on which the cells were grown was cropped out of each picture, showing the basal face of the cell layer at the bottom of each picture. Cells were fixed with a traditional electron microscopy stain to reveal cell structure and organelles. HepG2 (A) and A549 (B) cells form tight junctions (arrowheads) in cultures grown on filters, indicative of polarized epithelial cells. In a lower power view of HepG2 cells, a portion of the polycarbonate filter was left in the picture to show the orientation of growing cells (C). The A549 cells contain numerous lamellar bodies (D). The bar in panels A and B = 0.8 µm; the bar in panel C = 2 µm; the bar in panel D = 0.75 µm. Nu = nucleus; er = endoplasmic reticulum; pm = plasma membrane, LL = loose lamellar bodies; DL = dense lamellar bodies; DS = space of Dissé.

IEM of HepG2 and A549 cells was performed on cross-sectional slices of cells grown on filters to investigate directional FBG secretion. Immunocytochemical controls with normal rabbit serum IgG were consistently clear of any gold label for both cell types (Figures 4A and 4D). Cells were labeled with polyclonal anti-FBG IgG, followed by Protein A-gold. HepG2 cells displayed labeling in a pattern consistent with secretion basolaterally toward the filter, as well as through the microvilli into spaces between adjacent cells (Figures 4B and 4C). A549 cells revealed FBG-specific labeling at the basolateral surface (Figures 4E-4G), which was consistent with the predominantly basolateral secretion observed in metabolic labeling studies (Figure 2).

View larger version (194K): [in this window] [in a new window]   Figure 4.   Immunoelectron micrographs of HepG2 and A549 cells grown on filters and induced with IL-6 + DEX. HepG2 cells tend to pile atop neighboring cells (A), while A549 cells form a distinct monolayer (D). The Transwell-ColTM filter at the bottom edge is denoted in panels B and F (arrow), but was cropped out of the pictures of panels A, C-E, and G; thus, the cells are oriented with the apical surface toward the top and the basolateral surface at the bottom of each photomicrograph. The immunodetection was performed with affinity purified normal rabbit serum IgG in place of the primary antibody step, followed by Protein A-gold labeling; less than one gold particle per field (negative shot at ×4,000) was noted in either cell type (A and D). Incubation with anti-FBG antibodies and Protein A-gold demonstrated polarized secretion of FBG from HepG2 cells (B and C) and A549 cells (E-G) in the basolateral direction. Panel C represents a higher power magnification of gold particles in a secretory vesicle of HepG2 cells derived from the same field represented by the boxed portion of panel B. Panel G represents a 10-fold higher magnification of the circled region of the same field of A549 cells in panel F. The bars in panels A, B, D, and G = 2.5 µm; the bar in panel E = 1.3 µm; the bar in panel C = 0.75 µm; the bar in panel G = 0.25 µm. Nu = nucleus.

Microtubule Depolymerization Alters FBG Secretion

The role of microtubules in polarized secretion of FBG from A549 cells was explored using colchicine, an inhibitor of microtubule assembly. Treatment of cells for 2 h with 10 µM colchicine resulted in depolymerization of microtubules (Figure 5B compared with 5D), but did not affect the actin cytoskeleton (Figures 5A and 5C). To verify that colchicine treatment of cells did not produce leakage through the monolayers, the diffusion of ¹²⁵I-FBG between chambers of IL-6 + DEX-treated cells that were further treated with or without colchicine for 2 h was measured. In the time course of the experiment, colchicine treatment had no significant effect on the rate of diffusion of ¹²⁵I-FBG between chambers (< 3% diffusion; P = 0.08) (not shown). These results demonstrated the integrity of the polarized phenotype of A549 cell monolayers grown on the polycarbonate membrane filters and treated with colchicine. After IL-6 + DEX induction for 16 h, cells were incubated in the presence or absence of colchicine in addition to IL-6 + DEX for 2 h, metabolically pulse- labeled, then FBG immunopurified from apical and basolateral culture media and from the cell-associated fraction attached to the filter was analyzed by gel electrophoresis. The addition of colchicine significantly reduced the amount of FBG secreted basolaterally from A549 cells. Whereas colchicine treatment resulted in a slight statistically, but most likely not biologically, significant increase of apical FBG secretion, it resulted in a dramatic increase in cell- associated FBG (Table 2; Figure 6). These data suggest that colchicine treatment introduces a shift in the distribution of protein from secretion into the medium to retention within the cells.

View larger version (123K): [in this window] [in a new window]   Figure 5.   Colchicine treatment of A549 cells and immunofluorescent detection of microtubular and actin structures. Untreated (A and B) and colchicine treated (C and D) A549 cells were immunofluorescently stained with rhodamine phalloidin (A and C) and anti- tubulin antibodies (B and D). Colchicine treatment specifically alters tubulin structure, with actin and overall cellular morphology remaining as in untreated cells. Bar = 20 µm.

View larger version (39K): [in this window] [in a new window]   Figure 6.   Microtubule dependence of polarized FBG secretion from A549 cells. Triplicate sets of cells grown on Transwell-ColTM filters were stimulated with IL-6 + DEX for 18 h, treated in the absence (panel A) or presence (panel B) of colchicine during the last 2 h, then metabolically pulse-labeled. FBG in the conditioned media from apical (up arrow) and basolateral (down arrow) chambers and cell-associated (C) FBG from the filters was immunoprecipitated and analyzed by SDS-PAGE and fluorography. Molecular weight markers (Mr) in kD are from top to bottom: 220, 97, 66, 46 and 30.

To determine whether the colchicine-induced alteration in FBG secretion from A549 cells resulted in intracellular retention, cells grown in 6-well tissue culture dishes were exposed to IL-6 + DEX, then treated for 2 h with colchicine prior to metabolic pulse-labeling and analysis of secreted or cell-associated FBG. The presence of colchicine resulted in a reduction in the percentage of newly synthesized FBG secreted from cells grown on a solid surface, with an average of 63.7 ± 1.9% from control cells to 39.8 ± 1.8% from colchicine-treated cells (P = 0.002, n = 3) (Figure 7a). Similarly, colchicine treatment resulted in a reduction in the total amount of FBG secreted from A549 cells cultured on polycarbonate filters with an average of 67.9 ± 1.1% secreted from control cells and 47.6 ± 1.8% secreted from treated cells (P = 0.002). To verify that FBG synthesis was not affected by colchicine treatment, densitometry scanning was performed on total FBG synthesized during metabolic pulse-labeling. The total amount of FBG synthesized (secreted + cell-associated) by control and colchicine-treated A549 cells was nearly equivalent (not shown), indicating that the shift in FBG distribution was not attributed to inhibition of FBG synthesis. Statistical analysis of the total amounts of FBG secreted from cells grown on filters compared with cells grown in culture dishes indicated that there was not a statistically significant difference in the culture method (control cells grown on dish versus filter, P = 0.198; colchicine-treated cells grown on dish versus filter, P = 0.065). Therefore, to determine whether the cell-associated FBG was retained intracellularly, indirect immunofluorescent staining with anti-FBG mAb J88B was performed. A549 cells induced with IL-6 + DEX were incubated in the absence or presence of colchicine for 2 or 16 h. FBG staining was seen in sparse, cytoplasmic vesicular structures in the IL-6 + DEX control cells (Figure 7b, panel A); however, FBG staining in a diffuse cytoplasmic pattern was greatly enhanced in both the 2 and 16 h colchicine-treated cells (Figure 7b, panels B and C), indicative of intracellular accumulation of FBG. The data suggest that disruption of microtubules results in inhibition or retardation of the FBG secretory pathway.

View larger version (39K): [in this window] [in a new window]   View larger version (108K): [in this window] [in a new window]   Figure 7.   Cellular and secreted FBG after colchicine treatment. Panel a: Monolayer A549 cells were induced with IL-6 + DEX then treated with (+) or without () colchicine for 2 h. After metabolic pulse-labeling, media was collected and cells were lysed and scraped from wells. Immunoprecipitated samples of FBG (secreted = media; intracellular = cells) were analyzed by SDS-PAGE and fluorography. Panel b: Following IL-6 + DEX stimulation, noncolchicine treated (A) and cells treated with colchicine for 2 h (B) or 16 h (C) were immunofluorescently stained with anti-FBG mAb J88B. Cells treated with colchicine for 2 h showed negative staining with an irrelevant antibody (D). Positive fluorescence was obtained for cells stained with mAb J88B IgG, indicating increased amounts of FBG retained intracellularly in colchicine-treated (B and C) over noncolchicine-treated cells (A). Bar = 20 µm.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Prior studies demonstrated that the lung epithelial cell line, A549, synthesizes and secretes fully-assembled FBG after gene expression is dramatically upregulated by exposure to DEX and the proinflammatory mediator, IL-6 (17). This in vitro observation likely reflects the in vivo situation since ubiquitous FBG  chain gene expression (34) is upregulated in lung epithelium during infection with Pneumocystis carinii (35). In the present study, the polarity of secretion of FBG from A549 cells was explored using cells cultured on porous polycarbonate supports (Transwell-Col^TM filter chambers). Confluent monolayer cultures formed tight junctions and effectively blocked exchange of ¹²⁵I-labeled FBG between the apical and basolateral chambers whether or not the cells were treated with IL-6 + DEX ± colchicine. Metabolic labeling studies indicated that FBG synthesized by A549 cells was secreted in a highly polarized fashion with  80% in the basolateral direction. This result was confirmed by IEM, which revealed the presence of FBG in vesicles in proximity to the basolateral cell membrane.

Hepatocytes are the primary source of the plasma pool of FBG, thus vectorial secretion from this cell type would be expected to direct the majority of FBG through the perisinusoidal space of Dissé into the bloodstream (32). Therefore, the polarity of secretion of FBG from the lung epithelial cell line, A549, was compared with that secreted from the liver carcinoma cell line, HepG2. HepG2 cells cultured on Transwell-Col^TM filter chambers formed tight junctions; however, they did not form a typical monolayer, as multiple cell layers were present. FBG secretion from HepG2 cells, as analyzed following metabolic labeling, was polarized in the basolateral direction, albeit less dramatically than from A549 cells, and this polarized distribution was unaltered by the presence of IL-6 + DEX (P > 0.25). IEM revealed FBG within vesicular structures in proximity to the basolateral membrane, consistent with vectorial secretion into the space of Dissé through which FBG would enter the bloodstream (32).

Microtubules and microtubule motor proteins may be responsible for delivering transport vesicles to the correct membrane destination (36, 37). Disassembly of microtubules by colchicine treatment altered the pattern of FBG secretion, resulting in an increase in metabolically labeled, cell-associated FBG and a dramatic increase in intracellular FBG as shown by immunofluorescence staining in A549 cells cultured on filters, glass coverslips, and plastic tissue culture dishes. Thus, regardless of the culture conditions of the cells, the dependence on intact microtubules for FBG secretion was similar. This phenomenon was previously described in hepatocytes grown on a solid surface, wherein colchicine treatment resulted in sequestration of FBG within Golgi vesicles (38). The intracellular accumulation of FBG in colchicine-treated A549 cells grown on filters and culture dishes was likely due to retardation of the basolateral component of the FBG secretory pathway, since residual apical secretion was unaffected. Similar microtubule-dependence has been described for basolateral secretion of the adhesive glycoprotein, laminin, from kidney epithelial cells (36), and for apical insertion of plasma membrane proteins (39, 40).

Synthesis of FBG by lung epithelium and polarized secretion toward the epithelial basement membrane raises questions as to the role of FBG in normal and diseased lung. In light of our results demonstrating vectorial secretion of FBG toward the extracellular matrix, it is likely that, in vivo, FBG associates with basement membrane proteins and cells underlying the alveolar epithelium or denuded epithelium and thus may function as a substrate for cell adhesion or migration during wound healing. Consistent with this concept, FBG has been implicated as a substratum for epithelial cell migration during wound repair (9, 41). Furthermore, FBG and fibrin degradation products stimulate proliferation of fibroblasts, which are a direct source of matrix components and inflammatory cytokines (7, 42). FBG and fibrin are chemotactic, acting respectively as a mitogen in hematopoiesis (43) and as an enhancer of IL-1 production by mononuclear cells (44). Gene expression of FBG A and B chains does not appear to occur in epithelial cells of normal lung in vivo, but inducible expression of FBG A, B, and  chain genes is clearly evident during Pneumocystis carinii infection (35). The dramatic upregulation in response to inflammatory stimuli suggests that FBG may contribute to pathophysiologic events associated with fibrotic lung disease. Combined with the increased permeability of alveolar capillaries characteristic of an inflammatory response, increased procoagulant activity in lung occurring during P. carinii pneumonia (45, 46) could potentiate the deposition of fibrin in the alveolar sacs and interstitium. Plasma proteins such as FBG can decrease the activity of surfactant in acute respiratory distress syndrome, as levels of surfactant drop concomitantly with the progression of the disease (47). Because polymerized fibrin plays a significant role as a provisional matrix in wound healing, this study supports the possibility that FBG secreted basolaterally into the basement membrane may function in early wound repair during lung injury. However, excessive production of fibrin(ogen) by lung epithelium during inflammation may hinder adequate resolution of extensive lung injury. Thus, elucidation of the function of lung derived FBG will provide further understanding of its role in homeostasis as well as hemostasis.