TGF-beta 1 stimulates cultured human fibroblasts to proliferate and produce tissue-like fibroplasia: a fibronectin matrix-dependent event.

During wound repair, fibroblasts accumulate in the injured area until any defect is filled with stratified layers of cells and matrix. Such fibroplasia also occurs in many fibrotic disorders. Transforming growth factor-beta (TGF-beta), a promotor of granulation tissue in vivo and extracellular matrix production in vitro, is expressed during the active fibroplasia of wound healing and fibroproliferative diseases. Under usual tissue culture conditions, normal fibroblasts grow to confluence and then cease proliferation. In this study, culture conditions with TGF-beta 1 have been delineated that promote human fibroblasts to grow in stratified layers mimicking in vivo fibroplasia. When medium supplemented with serum, ascorbate, proline, and TGF-beta was added thrice weekly to normal human dermal fibroblasts, the cells proliferated and stratified up to 16 cell layers thick within the culture dish, producing a tissue-like fibroplasia. TGF-beta stimulated both DNA synthesis as measured by 3H-thymidine uptake and cell proliferation as measured by a Hoechst dye DNA assay in these postconfluent cultures. The stratification was dependent on fibronectin assembly, as demonstrated by anti-fibronectin antibodies which inhibited both basal and TGF-beta-stimulated cell proliferation and stratification. Suppression of collagen matrix assembly in cell layers with beta-amino-proprionitrile (BAPN) did not inhibit basal or TGF-beta stimulated in vitro fibroplasia. BAPN did not interfere with fibronectin matrix assembly as judged by immunofluorescence microscopy. Thus, in concert with serum factors, TGF-beta stimulates postconfluent, fibronectin matrix-dependent, fibroblast growth creating a fibroplasia-like tissue in vitro.

Expansion of bronchial epithelial cell populations by in vitro culture of explants from dysplastic and histologically normal sites.

The genetic and phenotypic properties of cells which ultimately give rise to carcinoma of the lung are not well defined in part because of unavailability of preneoplastic cells from well-characterized dysplastic sites. In order to expand bronchial epithelial cell populations from patients at high risk for lung cancer, endobronchial biopsy specimens were explanted onto collagen- and fibronectin-coated dishes and cultured in serum-free, chemically defined media. One hundred forty-nine biopsy pairs were obtained from smokers and from healthy volunteers for culture and histologic evaluation. The histologic appearances of mucosa adjacent to the site of the cultured biopsies ranged from normal through varying degrees of noninvasive squamous dysplasia to invasive carcinoma. Confluent monolayers of pure epithelial cells were obtained from 68% of the cultured explants. Sites exhibiting high-grade dysplasia were 51% more likely to yield successful cultures than sites exhibiting normal histology (13 of 14 cultures successful versus 52 of 83 cultures successful, P < 0.02). Cultures had a maximum proliferative life span of 81 days and none of the cultures spontaneously became immortalized. Immunolabeling studies revealed that all cultured epithelial cells, regardless of the in situ histologic appearances of the mucosa at the biopsy site, strongly expressed keratin and epidermal growth factor receptor, weakly expressed transferrin receptor and human folate receptor, and were negative for neural cell adhesion molecule and human leukocyte antigen DR (HLADR). Ploidy and karyotypic analyses were performed in a limited number of explants from normal and dysplastic sites and all were found to be diploid without karyotypic abnormality. We conclude that pure bronchial epithelial cell populations can be routinely expanded from histologically normal and dysplastic sites by tissue culture of biopsy explants and that the expanded cell populations may represent a library of normal and preneoplastic cells which are suitable for immunophenotypic, ploidy, genetic, or functional analyses.

Neutral endopeptidase: variable expression in human lung, inactivation in lung cancer, and modulation of peptide-induced calcium flux.

Neutral endopeptidase (NEP; CALLA, CD10, EC 3.4.24.11) is a cell surface endopeptidase that hydrolyses bioactive peptides, including the bombesin-like peptides, as well as other neuropeptides. Bombesin-like peptides and other neuropeptides are autocrine growth factors for both small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). Low expression of NEP has been reported in SCLC and NSCLC cell lines. NEP inhibition has been shown to increase proliferation in one cell line. To date, NEP expression has not been quantitatively evaluated in normal adult lung, SCLC or NSCLC tumors, paired uninvolved lung from the same patient, or in other pulmonary neoplasms such as mesotheliomas and carcinoids. We examined the expression of NEP in these tissues and human cell lines using immunohistochemistry, flow cytometry, enzyme activity, ELISA, Western blot, and reverse transcription (RT)-PCR. Uninvolved lung tissue from different individuals displayed considerable variation in NEP activity and protein. By immunohistochemistry, NEP expression was detectable in alveolar and airway epithelium, fibroblasts of normal lung, and in mesotheliomas, whereas it was undetectable in most SCLC, adenocarcinoma, squamous cell carcinoma, and carcinoid tumors of the lung. NEP activity and protein levels were lower in all SCLC and adenocarcinoma tumors when compared to adjacent uninvolved lung, often at levels consistent with expression derived from contaminating stroma. NEP expression and activity were reduced or undetectable in most SCLC and lung adenocarcinoma cell lines. NEP mRNA by RT-PCR was not expressed or was in low abundance in the majority of lung cancer cell lines. The majority of lung tumors did not express NEP by RT-PCR as compared with normal adjacent lung. In addition, recombinant NEP abolished, whereas an NEP inhibitor potentiated, the calcium flux generated by neuropeptides in some lung cancer cell lines, demonstrating potential physiological significance for low NEP expression. NEP, therefore, is a signal transduction and possibly a growth modulator for both SCLC and NSCLC, emphasizing the role of neuropeptides in the pathogenesis of the major histological forms of lung cancer.

Platelet isoforms of platelet-derived growth factor stimulate fibroblasts to contract collagen matrices.

Fibroplasia and angiogenesis are essential components of tissue repair when substantial tissue has been lost at a site of injury. Platelets and monocyte/macrophages accumulate at these sites and release a variety of growth factors that are thought to initiate and sustain the repair. Often the involved tissue contracts, a process that can markedly reduce the amount of fibroplasia and angiogenesis necessary for the reestablishment of organ integrity. Such tissue contraction occurs over hours or days, a much slower time course than the rapid, reversible contraction of muscle tissue. Fibroblasts, which are rich in f-actin bundles, appear to be responsible for wound contraction. However, the signals that stimulate contraction are not known. Using cultured fibroblasts, which are also rich in f-actin bundles, we demonstrate the platelet and monocyte isoforms of platelet-derived growth factor (PDGF; AB and BB) but not PDGF-AA, can stimulate fibroblasts to contract collagen matrix in a time course similar to that of wound contraction. In addition, PDGF appears to be the predominant fibroblast/collagen gel contraction activity released from platelets. Vasoactive agonists known to stimulate smooth and striated muscle contraction do not stimulate fibroblast-driven collagen gel contraction.

Optimization of immunohistochemical techniques to detect extracellular matrix proteins in fixed skin specimens.

Complete antigen visualization in the context of well-preserved tissue architecture is the goal of all immunohistochemical techniques. Frozen tissue section techniques achieve optimal antigen visualization but preserve tissue architecture poorly. On the other hand, formalin-fixed tissue section techniques preserve tissue architecture very well but result in antigen masking. Enzymatic digestion or salt extraction of formalin-fixed sections has been used to reestablish antigen expression. Recently acid-alcohol-fixed tissue has been used as a successful compromise between tissue architecture preservation and the visualization of cytoskeletal antigens. In an attempt to find an improved immunohistochemical process for non-cytoskeletal antigens, we compared avidin-biotin immunofluorescence staining in frozen, formalin-fixed, and acid-alcohol-fixed tissues. The fixed tissues were either untreated or treated with enzyme digestion or salt extraction. For this study, we examined healing cutaneous wounds in Yorkshire pigs with antibodies to fibronectin, laminin, von Willebrand factor VIII, and keratin. Although tissue architecture was poor, frozen sections provided the best antigen visualization and were therefore used as the standard for complete antigen expression. Formalin-fixed tissues had excellent tissue architecture, but most antigens were completely masked. Pre-treatment technique only partially overcame the antigen masking caused by formalin. In contrast, acid-alcohol fixation preserved tissue architecture almost as well as formalin and sometimes allowed complete antigen visualization; however, laminin and fibronectin were partially masked. Total recovery of the expression of these antigens could be obtained by pre-treating the acid-alcohol-fixed tissue with either hyaluronidase or 1 M NaCl. Therefore, acid-alcohol-fixed tissue appears best for extracellular matrix (ECM) protein immunostaining as well as for cytoskeletal staining. However, certain ECM antigens require hyaluronidase or 1 M NaCl treatment for optimal visualization.

Fibronectin mediates adherence of rat alveolar type II epithelial cells via the fibroblastic cell-attachment domain.

The lung alveolar surface is composed of types I and II epithelial cells. Extremely attenuated type I cells cover 90% of the surface and are prone to necrosis during acute lung injury. After denudation of type I cells, the alveolar epithelium is restored by proliferation of type II cells. During reepithelialization in vivo the type II cells have been observed to reorganize on an extracellular matrix that contains fibronectin. We thus sought to determine whether type II cells would adhere to purified fibronectin. Adherence assays of primary rat type II cells were performed in protein-coated bacteriologic microtiter wells for 24 h at 37 degrees C. Concentrations of fibronectin from 1 to 300 micrograms/ml mediated type II cell adherence, 10 micrograms/ml gave maximal adherence, and 4 micrograms/ml gave 50% maximal adherence. Adherence progressively increased from 1 to 72 h. Adherence on fibronectin was at least 50% greater than adherence on laminin, types I and III collagen, or IV collagen. Little or no adherence was observed on bacteriologic plastic or albumin. Spreading on these various substrata paralleled adherence. Adherence to fibronectin, laminin, and fibrinogen was specifically blocked by their respective polyclonal antibodies. Monoclonal antibodies (MoAb) to the fibronectin cell-attachment domain blocked adherence to fibronectin, whereas MoAb to other domains did not. From the data reported here and the previously mentioned in vivo study we propose that fibronectin is an important functional component of the extracellular matrix that supports type II cells during alveolar reepithelialization.

Either exogenous or endogenous fibronectin can promote adherence of human endothelial cells.

Recently, we have presented evidence that proliferating blood vessels produce and deposit fibronectin in situ during the angiogenesis of wound repair. This report extends these observations by demonstrating that human endothelial cells from both large and small vessels depend on fibronectin for their adherence in vitro. Endothelial cells were grown from human umbilical veins (HUVEC) by the method of Gimbrone and from the microvasculature of human omentum by the method of Kern, Knedler and Eckel. Second-passage cells were plated into microtitre wells that had been coated with 100 micrograms ml-1 of fibronectin, types I and III collagen, type IV collagen or laminin. After a 3-h incubation, adherent cells were solubilized with Zap-Isoton and quantified on a Coulter Counter. Under normal culture conditions HUVEC showed no preference for fibronectin substrates while microvascular cells always demonstrated a striking preference for fibronectin substrates. However, when HUVEC were exposed to 2.5 or 25 micrograms ml-1 of cycloheximide for 4 h before and during the adherence assays, the adherence to fibronectin was 50-200% greater than to types I and III collagen. Immunofluorescence studies showed that while HUVEC expressed a large quantity of surface fibronectin, microvascular cells expressed very little. Metabolic labelling studies confirmed that HUVEC cultures had substantial quantities of fibronectin in their cell layer while microvascular cells did not. In antibody blocking experiments, preincubation of fibronectin-coated surfaces with anti-fibronectin antibodies totally blocked microvascular cell adhesion but only abrogated HUVEC adherence by 50%, presumably since these latter cells were able to deposit additional fibronectin onto the surface during the 3 h assay period. In the presence of cycloheximide anti-fibronectin antibodies totally blocked HUVEC adherence. These results demonstrate that both endothelial cell types rely, at least in part, on fibronectin for adherence in vitro. HUVEC can synthesize, secrete and deposit enough fibronectin for their adherence in vitro, while microvascular cells rely on an exogenous source of fibronectin under these culture conditions. Thus, the increased blood vessel fibronectin observed during angiogenesis in vivo may mediate adherence of the proliferating and migrating endothelial cells.

Fibronectin, as well as other extracellular matrix proteins, mediate human keratinocyte adherence.

During the reepithelialization of cutaneous wounds, the migratory epidermis transits over a provisional matrix of fibronectin and fibrin in the absence of laminin and type IV collagen as well as ultrastructurally identifiable basement membrane. Since significant quantities of fibronectin occur at these sites of reepithelialization, we surmised that fibronectin is a suitable substrate for keratinocyte adherence and therefore undertook the in vitro investigations reported here. Purified human plasma fibronectin precoated on bacteriologic microtiter wells was demonstrated to mediate human keratinocyte adherence when concentrations greater than 10 micrograms/ml fibronectin were used. Maximal keratinocyte adherence was obtained in wells precoated with 100 micrograms/ml fibronectin and when cells were incubated with substrate for 60 min or longer at 37 degrees C. Both primary and second-passaged human keratinocytes adhered as well or better to fibronectin than to types I and III collagen, laminin, or type IV collagen under both high- and low-Ca++ culture conditions. However maximal adherence to all substrates occurred when second-passaged keratinocytes were assayed in low-Ca++ medium. Under these latter culture conditions, keratinocyte phenotype resembles the phenotype of cells in the migrating epidermis. To determine specificity of these adherence reactions, antifibronectin antibodies were shown to block keratinocyte adherence to fibronectin but not to laminin substrates. Conversely, antilaminin antibodies blocked adherence to laminin but not fibronectin substrates. Thus, human keratinocytes demonstrate specific adherence to fibronectin in a time- and dose-dependent fashion and this adherence relies on de novo protein synthesis. We take these observations to support our hypothesis that the provisional fibronectin matrix observed beneath the migrating epithelium during tissue repair plays a functional role in the reepithelialization process.

Human keratinocytes that have not terminally differentiated synthesize laminin and fibronectin but deposit only fibronectin in the pericellular matrix.

Fibronectin and laminin production by human keratinocytes cultured in serum-free, low-calcium medium without a fibroblast feeder layer were examined by several techniques. By indirect immunofluorescence, fibronectin but not laminin appeared as short radial fibrils between the cells and the substratum, and in the pericellular matrix. Synthesis of fibronectin and laminin by 7-day keratinocyte cultures was determined by 18 hr 35S-methionine metabolic labeling followed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and autoradiography. Fibronectin accounted for 2.9% of total synthesized protein, 26.5% of fluid phase protein secretion, and 4.3% of deposited ECM protein. In contrast, only 0.1% of the total synthesized protein was laminin, little (6.3%) of this product was secreted, and none of this product was deposited in the ECM. Our results indicate that human keratinocytes under culture conditions that prevent terminal differentiation in vitro can synthesize, secrete, and deposit fibronectin in the extracellular matrix. Although these cells synthesize laminin, they secrete very little and deposit no detectable laminin in the matrix under these culture conditions. From these data we believe that fibronectin may play an important role in the interaction of epidermal cells with connective tissue matrix during wound healing or morphogenesis in in vivo situations in which the epidermis is not terminally differentiated.