Secretion of epidermal growth factor. The role of calcium in stimulcule during secretion.

Secretion of epidermal growth factor by mouse submandibular salivary glands was studied in vitro to determine the second messenger involved in stimulus-secretion coupling and also to determine whether epidermal growth factor is secreted in the molecular form in which it occurs within the glandular cells. The presence of Ca2+ in the extracellular fluid was found to be necessary for the normal secretion of epidermal growth factor that follows activation of alpha-adrenoceptors. Dibutyryl cyclic AMP (1 mM) did not evoke any secretion of the growth factor. Secreted epidermal growth factor retained its ability to bind to anti-epidermal growth factor antibodies and to epidermal growth factor-binding sites on liver cell membranes. However, during secretion the 74 000-dalton, 4-subunit complex, in which epidermal growth facto; occurs within the cells, dissociated. The adrenalin-stimulated submandibular salivary gland did not appear to release any material into the incubation medium which could modify the complex and produce the dissociation. It is suggested that the dissociation of the high molecule containing epidermal growth factor results from the loss of the arginyl residue from the C-terminus of epidermal growth factor at some time during the secretion process.     

An immunohistochemical study of secretagogue-induced secretion of epidermal growth factor in the submandibular salivary glands of mice

Routine histological and indirect immunofluorescence techniques were used to examine the histological details of changes in the distribution of epidermal growth factor (EGF) in the submandibular salivary glands of mice during secretion. Comparisons were made bewteen glands of normal mice and those of mice given one of a number of secretagogues at various times prior to sampling. Normal submandibular salivary glands in male mice had an extensive system of convoluted granular tubules (CGT), the cells of which contained EGF. When adrenaline of alpha-phenylephrine was administered, the CGT cells degranulated, and there was a concomitant loss of intracellular EGF-positive immunofluorescence. The excretory ducts were engorged with immunofluorescent material, indicating secretion of EGF into saliva, while the ductal cells themselves remained EGF-negative. The degranulation response could be blocked by phentolamine, but not by propranolol, and no changes in EGF distribution followed the administration of pilocarpine. It was concluded that EGF is secreted, at least partly into the saliva, following an alpha-adrenergic response, and that this occurs with degranulation of the cells of the CGT.     

Mouse submandibular salivary epithelial cell growth and differentiation in long-term culture: Influence of the extracellular matrix

Summary The adult mouse submandibular salivary gland provides a good model system to study gene regulation during normal and abnormal cell behavior because it synthesizes functionally distinct products ranging from growth factors and digestive enzymes to factors of relevance to homeostatic mechanisms. The present study describes the long-term growth and differentiation of submandibular salivary epithelial cells from adult male mice as a function of the culture substratum. Using a two-step partial dissociation procedure, it was possible to enrich for ductal cells of the granular convoluted tubules, the site of epidermal growth factor synthesis. Long-term cell growth over a period of 2 to 3 mo. with at least 3 serial passages was obtained only within three-dimensional collagen gels. Cells grew as ductal-type structures, many of which generated lumens with time in culture. Electron microscopic analysis in reference to the submandibular gland in vivo revealed enrichment for and maintenance of morphologic features of granular convoluted tubule cells. Reactivity with a keratin-specific monoclonal antibody established the epithelial nature of the cells that grew within collagen. Maintenance of cell differentiation, using immunoreactivity for epidermal growth factor as criterion, was determined by both cytochemical and biochemical approaches and was found to be dependent on the collagen matrix and hormones. Greater than 50% of the cells in primary collagen cultures contained epidermal growth factor only in the presence of testosterone and triiodothyronine. In contrast, cells initially seeded on plastic or cycled to plastic from collagen gels were virtually negative for epidermal growth factor. Biochemical analysis confirmed the presence of a protein with an apparent molecular weight of 6000 which comigrated with purified mouse epidermal growth factor. Epidermal growth factor was also present in detectable levels in Passage 1 cells. This culture system should permit assessment of whether modulation of submandibular gland ductal cell growth can be exerted via a mechanism that in itself includes epidermal growth factor and its receptor and signal transduction pathway. This work was supported by Public Health Service grant DE07766 from the National Institute of Dental Research, National Institutes of Health, Bethesda, MD.     

Expression of epidermal growth factor in salivary adenoid cystic carcinoma

This study used an immunohistochemical technique to assess the expression of epidermal growth factor (EGF) in 40 specimens of salivary adenoid cystic carcinoma (ACC), 7 specimens of labial glands adjacent to mucocele, and 5 specimens of normal submandibular glands. In normal submandibular glands, immunohistochemically detectable EGF was demonstrated in all ductal segments, including intercalated, striated, and excretory duct cells. No EGF positive staining was found in acinar compartments. including serous and mucous acinar cells. In degenerated labial glands adjacent to mucocele, no EGF staining was detected in the remaining acinar and ductal cells. In salivary ACCs, positive EGF immunostaining was observed in one of the 5 (20%) ACCs with a solid pattern and in 13 of the 35 (37.1%) ACCs with a tubular-cribriform pattern. The overall EGF expression rate in 40 salivary ACCs was 35%. Positive EGF staining was predominantly found in tubular structures in the tubular ACCs and in duct-like structures in large cribriform patterns or in the stroma of the cribriform ACCs. There was no significant correlation between EGF expression in salivary ACCs and any of the clinicopathological parameters including patient age and sex, cancer location, TNM status, clinical stage, histologic type, perivascular or perineural invasion, focal necrosis of tumor, and cellular atypia. We conclude that the duct segments of the normal submandibular gland are the sites of EGF synthesis and secretion. In degenerated labial glands adjacent to mucocele, EGF synthesis is completely inhibited. Furthermore, EGF is mainly biosynthesized in cells forming tubular or duct-like structures in tubular or cribriform salivary ACCs, and EGF may play a biologic role, particularly as a mitogen in salivary ACC growth.     

Immunocytochemical localization of nerve growth factor in mouse salivary glands

Summary The submandibular glands of female mice and the sublingual and parotid glands of adult male and female mice have been examined by light microscopical immunocytochemistry for nerve growth factor (NGF). In female submandibular glands, staining for NGF was observed in granular convoluted tubule and striated duct cells. Sublingual glands of the mouse contained relatively few granular cells staining for NGF compared with submandibular glands. However, such granular cells appeared to be more numerous in male sublingual glands than in female glands. The remainder of the intralobular duct cells in both male and female sublingual glands exhibited apical subluminal staining for NGF as well as light basal plasmalemmal staining. Parotid glands in both male and female mice exhibited a similar pattern of staining for NGF in striated duct cells. However, the glands did not contain granular cells nor did they exhibit any pattern of staining which reflected a sexual dimorphism. Immunodot staining of salivary gland extracts confirmed the presence of immunoreactivity for NGF in all three of the major salivary glands.     

Altered secretion and processing of epidermal growth factor in adrenergic-induced growth of the rat submandibular gland

The granular convoluted tubule (GCT) cells of the submandibular glands represent a major production site for epidermal growth factor (EGF). This study investigates EGF production in the submandibular glands in relation to beta-adrenergic stimulation. Rats were treated with isoproterenol (beta-agonist), which caused up to a 400% increase in submandibular tissue weight after 3 weeks. The weight increase coincided with marked morphologic changes, with degranulation and an apparent decrement in the number of the GCT cells. Immunostaining against EGF revealed a reduction in the number of EGF-immunoreactive cells. Concomitantly, the glandular contents of 6-kDa EGF decreased from 12.86+/-3.42 nmol/gland (mean+/-S.E.M.) in controls to 0.26+/-0.03 nmol/gland. EGF mRNA levels, expressed relative to total RNA levels, only tended to be reduced after 3 weeks as judged from RT-PCR and in situ hybridization (ISH). The isoproterenol-treated rats had increased output of EGF in the saliva, but the salivary secretion of protein was also increased. In both glandular tissue and saliva, gel filtration revealed partially processed high molecular weight forms of EGF in the isoproterenol-treated rats. These data indicate that isoproterenol treatment leads to a hyperstimulatory state of the GCT cells, which then causes depletion of the cellular stores of mature EGF, and most likely due to a shortened posttranslational transit, incomplete peptide processing.     

Immunocytochemical localization of epidermal growth factor in mouse submandibular gland.

The cellular and subcellular localization of epidermal growth factor in the submandibular glands of male and female adult mice was established by immunoperoxidase techniques. In light microscopic preparations epidermal growth factor was found exclusively in the granular convoluted tubules of the gland. The intensity of staining for epidermal growth factor varied from cell to cell, and some cells apparently were negative. The pattern of staining was similar in the glands of male and female mice; however, the granular convoluted tubules are androgen-responsive, and thus more extensive and composed of larger cells in males. In thin sections epidermal growth factor was most heavily concentrated in the secretion granules of the granular convoluted tubule cells. Within a given cell there was variation in intensity of staining of individual secretion granules, with some granules appearing minimally reactive or negative. The only other cell component with deposits of reaction product was the ribosomes.     

Growth factor regulation of the amylase promoter in a differentiating salivary acinar cell line

Salivary glands contain two major epithelial cell types: acinar cells which produce the primary salivary secretion, including amylase, and ductal cells which reabsorb electrolytes but also secrete kallikrein. Here we investigated salivary acinar cell differentiation in vitro using the activity of the salivary amylase and tissue kallikrein promoters as markers of acinar cell and ductal cell differentiation, respectively. Each of the promoter sequences was cloned into a replication-deficient adenoviral vector containing the luciferase reporter gene. Previous studies showed that a human submandibular gland cell line (HSG) differentiated into acinar cells when cultured on a reconstituted basement membrane matrix (Matrigel). The luciferase activity of the amylase promoter vector (AdAMY-luc) was low in HSG cells cultured on plastic, where they grow as an epithelial monolayer. The promoter activity increased approximately tenfold when HSG cells were cultured on Matrigel and developed an acinar phenotype. Under the same conditions, the luciferase activity of the kallikrein promoter (AdKALL-luc) was not induced. Because HSG cells demonstrate acinar cell morphology, but not amylase gene expression, when cultured on laminin-1, certain soluble components of Matrigel were tested for their ability to induce the amylase promoter during in vitro differentiation of acinar cells. We find that epidermal growth factor (EGF) and transforming growth factor-alpha (TGF-alpha), which are present in the basement membrane, and hepatocyte growth factor (HGF) increase activity of the amylase promoter. Other basement membrane-derived growth factors such as TGF-beta, basic fibroblast growth factor (bFGF), and platelet-derived growth factor (PGDF), as well as tumor necrosis factor (TNF-alpha), keratinocyte growth factor (KGH), nerve growth factor (NGF) and interferon gamma (IFN-gamma) were inactive. This system will be further exploited to study the mechanisms by which extracellular matrix molecules and growth factors regulate salivary acinar cell differentiation.     

Hydroxyurea inhibits degradation of internalized epidermal growth factor in HeLa cells

Because epidermal growth factor stimulates DNA synthesis in cultured cells, five inhibitors of DNA synthesis were tested in HeLa cells to see whether the inhibition of DNA synthesis has any effect on the metabolism of the growth factor. Among these, only hydroxyurea inhibited the degradation of ¹²⁵I-labeled epidermal growth factor strongly. The reversal of hydroxyurea-induced inhibition of DNA synthesis by deoxyribonucleosides did not result in a recovery from the inhibition of the degradation. From these findings, it might be concluded that the inhibitory effect of hydroxyurea on the degradation is distinct from that on DNA synthesis.     

Epidermal growth factor: Characteristics of specific binding in membranes from liver,placenta, and other target tissues

Epidermal growth factor, a 6,400-dalton polypeptide from the mouse submaxillary gland, binds specifically to cells and membranes derived from a variety of human, rat, mouse, and bovine tissues. Liver, placenta, skin, cornea, and cultured chondrocytes, Hela cells, and Chang liver cells bind large amounts of epidermal growth factor, whereas fat cells, resting and lectin-stimulated human peripheral lymphocytes, mouse thymocytes, cultured rat hepatoma cells, and mammary cells from virgin and pregnant mice bind little or no epidermal growth factor. The binding site for epidermal growth factor is distinct from receptors for other anabolic peptides such as insulin, nerve growth factor, and growth hormone. The binding of epidermal growth factor is rapid and reversible. The rate constant of association is approximately 10⁶ mole^?1 sec^?1, the rate constant of dissociation is about 6 × 10^?4 sec^?1, and the apparent equilibrium dissociation constant is about 10^?9m. Trypsin at low concentrations (50–200 μg/ml) destroys the receptor site for epidermal growth factor. The binding of epidermal growth factor by membranes is not accompanied by appreciable degradation of the peptide present in the medium or of that bound to the membranes. Use can be made of the high affinity and specificity of membranes for epidermal growth factor to measure by a competitive binding assay as little as 200 pg of EGF per ml (3 × 10^?11m).     

Global analysis of gene expression profiles in the submandibular salivary gland of klotho knockout mice

Salivary dysfunction commonly occurs in many older adults and is considered a physiological phenomenon. However, the genetic changes in salivary glands during aging have not been characterized. The present study analyzed the gene expression profile in salivary glands from accelerated aging klotho deficient mice (klotho?/?, 4 weeks old). Microarray analysis showed that 195 genes were differentially expressed (z‐score > 2 in two independent arrays) in klotho null mice compared to wild‐type mice. Importantly, alpha2‐Na⁺/K⁺‐ATPase (Atp1a2), Ca²⁺‐ATPase (Atp2a1), epidermal growth factor (EGF), and nerve growth factor (NGF), which have been suggested to be regulators of submandibular salivary gland function, were significantly decreased. When a network was constructed from the differentially expressed genes, proliferator‐activated receptor‐γ (PPAR γ), which regulates energy homeostasis and insulin sensitivity, was located at the core of the network. In addition, the expression of genes proposed to regulate various PPAR γ‐related cellular pathways, such as Klk1b26, Egfbp2, Cox8b, Gpx3, Fabp3, EGF, and NGFβ, was altered in the submandibular salivary glands of klotho?/? mice. Our results may provide clues for the identification of novel genes involved in salivary gland dysfunction. Further characterization of these differentially expressed genes will be useful in elucidating the genetic basis of aging‐related changes in the submandibular salivary gland.

     

Absence of the alpha and gamma subunits of 7S nerve growth factor in denervated rodent iris: immunocytochemical studies

Immunocytochemical studies were performed to determine if denervated rodent iris produces nerve growth factor (NGF) in a form chemically similar to that of the 7S NGF complex in mouse submandibular glands. Antisera to the alpha, beta, and gamma subunits of 7S NGF were raised in rabbits and characterized on immunoblots of SDS-containing polyacrylamide gels. Antisera were applied to stretch preparations of rat and mouse irides that were cultured for periods of 2 to 6 days or sympathetically denervated by superior cervical ganglionectomy and left in situ 4 days. Antibody binding was visualized by indirect immunofluorescence. In control studies done on plastic sections of mouse submandibular glands, antisera co-localized the three subunits of 7S NGF within secretory granules of granular tubule cells. In denervated rat iris, beta NGF immunoreactivity was evident in a cellular plexus that resembled in distribution and morphology nerve fibers in the normal iris, in agreement with a previous study (R.A. Rush (1984). Nature (London) 312, 364-367). Identical staining patterns were observed in mouse iris. In neither rat or mouse, however, did the nerve-like processes stain with antibodies suggests that the NGF-like protein in denervated rodent iris is not synthesized as part of the 7S NGF complex. Iris also did not react with antibodies to epidermal growth factor, a protein co-localized with NGF in mouse submandibular glands and in guinea pig prostate.     

Vasoactive intestinal polypeptide and acetylcholine stimulate exocrine secretion of epidermal growth factor from the rat submandibular gland

The effect of vasoactive intestinal polypeptide (VIP) and acetylcholine on secretion of epidermal growth factor (EGF) from the rat salivary glands was investigated. VIP in doses of 3 X 10(-10) to 3 X 10(-8) mol/kg per h stimulated secretion of saliva and total output of EGF dose-dependently. Acetylcholine also stimulated salivation and output of EGF. VIP in a dose of 3 X 10(-11) to 3 X 10(-10) mol/kg per h enhanced the stimulatory effect of acetylcholine, but this effect disappeared when the dose of VIP was increased. Adrenalectomy decreased acetylcholine stimulated total output of EGF by approximately 50%, but only by 20% when acetylcholine plus VIP was administered. EGF was localized to the convoluted granular tubules in the submandibular gland, whereas EGF could not be detected in the remaining salivary glands. The results suggest that VIP and acetylcholine cooperate in the control of exocrine secretion from the rat salivary glands. The effect of acetylcholine, however, seems to be partly dependent on circulating catecholamines.     

Immunohistochemical demonstration of epidermal growth factor and nerve growth factor in experimental carcinogenesis in the mouse submandibular gland

Immunohistochemical demonstration of epidermal growth factor (EGF) and nerve growth factor (NGF) was made during chemical carcinogenesis in the mouse submandibular gland. The granular convoluted tubule cells in the normal male submandibular gland contained larger amounts of EGF and NGF than in the female. The initial phase and early stages in chemical carcinogenesis showed degranulation of the granular convoluted tubule cells with a marked decrease in EGF and NGF. Premalignant lesions such as duct-like structures and multicystic lesions showed variable staining for EGF and were usually negative for NGF. Material secreted into the luminal spaces revealed increased staining for EGF and NGF. Scattered tumor cells of the poorly differentiated squamous-cell carcinoma type and desquamated tumor cells contained abundant EGF, but not NGF. No positive reaction for EGF or NGF was found in the induced squamous-cell carcinoma cells.     

Production and characterization of antibody blocking epidermal growth factor: Receptor interactions

Membranes were prepared from the human epithelioid carcinoma cell line A-431 which has approx. 2 · 10⁶ epidermal growth factor receptors per cell. This membrane preparation which retained a high epidermal growth factor binding specific activity was used as an antigen to produce antisera in rabbits. Double-immunodiffusion experiments demonstrated that the immune serum contained precipitating antibodies to several components of detergent solubilized A-431 membranes.The immonoglobulin G fraction of this immune sera inhibited ¹²⁵I-labeled epidermal growth factor binding to receptors in: (1) intact human and mouse cells; (2) membrane preparations from A-431 cells and human placenta, and (3) solubilized A-431 membranes. Inhibition of ¹²⁵I-labeled epidermal growth factor binding was observed with divalent and monovalent fragments of immunoglobulin G prepared from the immunoglobulin G fraction. Also, the immunoglobulin G fraction blocked growth factor binding to membranes at low temperature (5°C).Anti-A-431 antibody blocked the induction of DNA synthesis in quiescent fibroblasts by epidermal growth factor in a manner similar to that of anti-epidermal growth factor antibody. Addition of either anti-A-431 or anti-epidermal growth factor antibodies to fibroblasts at times up to 5 h after the addition of epidermal growth factor completely reversed the hormone's mitogenic potential. At later times (after 12 h) addition of either antibody was without effect on the stimulation of DNA synthesis by epidermal growth factor. Anti-A-431 antibody did not block the induction of DNA synthesis in fibroblasts by fibroblast growth factor or serum.     

Immunocytochemical localization of epidermal growth factor during the postnatal development of the submandibular gland of the mouse.

The time of appearance and the pattern of localization of epidermal growth factor (EGF) in submandibular glands of mice was studied during postnatal development immunocytochemically. EGF was first detectable in the granular convoluted tubule (GCT) cells in the glands of males at 20 days of age and of females at 30 days of age. Development of GCT cells containing EGF was rapid in males, approaching adult conditions by 45 days of age. In females EGF- containing GCTs developed more slowly and irregularly, and did not reach adult status by 45 days of age. It is concluded that EGF is restricted during postnatal development to the GCT cells, and that these cells and the distribution of EGF are represented dimorphically from their first appearance in the submandibular glands of both sexes.     

Neuroblastoma cells express c-sis and produce a transforming growth factor antigenically related to the platelet-derived growth factor.

Mouse neuroblastoma Neuro-2A cells produce transforming growth factors during exponential growth in a defined hormone-free medium, which, on Bio-Gel columns in 1 M HAc, elute at a molecular size of 15 to 20 kilodaltons (kDa). These neuroblastoma-derived transforming growth factors have strong mitogenic activity, but they do not compete with epidermal growth factor for receptor binding (E. J. J. van Zoelen, D. R. Twardzik, T. M. J. van Oostwaard, P. T. van der Saag, S. W. de Laat, and G. J. Todaro, Proc. Natl. Acad. Sci. U.S.A. 81:4085-4089, 1984). In this study approximately 80% of the mitogenic activity was immunoprecipitated by antibodies raised against platelet-derived growth factor (PDGF). Immunoblotting indicated a true molecular size of 32 kDa for this PDGF-like growth factor. Analysis of poly(A)+ RNA from Neuro-2A cells demonstrated the expression of the c-sis oncogene in this cell line, whereas in vitro translation of the RNA yielded a 20-kDa protein recognized by anti-PDGF antibodies. Separation by reverse-phase high-pressure liquid chromatography demonstrated the presence of two distinct mitogenic activities in neuroblastoma-derived transforming growth factor preparations, one of which is antigenically related to PDGF. Both activities had the ability to induce anchorage-independent growth in normal rat kidney cells, both in the presence and in the absence of epidermal growth factor. It is concluded that Neuro-2A cells express c-sis with concomitant production and secretion of a PDGF-like growth factor, which plays a role in the induction of phenotypic transformation on normal rat kidney cells.     

Immunocytochemical localization of transforming growth factor α, epidermal growth factor and epidermal growth factor receptor in the cat endometrium and placenta

This study was undertaken to determine the immunocytochemical localization of transforming growth factor α, epidermal growth factor and epidermal growth factor receptor in the endometrium of ovariectomized cats treated with oestradiol-17β and/or progesterone and in the endometrium and placenta of pregnant cats. Specific immunostaining was observed for all three antibodies. Moderate immunostaining for transforming growth factor α was observed in the epithelium of ovariectomized and oestrogen-treated cats. Dark epithelial staining was observed throughout pregnancy. The epithelial cells in progesterone-treated and peri-implantation animals contained dense deposits of reaction product, which were not reduced in intensity when immunoabsorbed antiserum was used. For epidermal growth factor, light--moderate epithelial staining was observed in ovariectomized and steroid-treated animals, and this increased in pregnant cats. Stromal staining for both the transforming and the epidermal growth factors was limited in steroid-treated animals and increased as pregnancy continued. Dark staining for epidermal growth factor receptor was observed in the epithelium and stroma in all the animals studied. The tips of surface epithelial convolutions in the non-implantation sites were always more darkly stained than in other regions of the surface epithelium. Staining in the placental trophoblast was limited to the syncytiotrophoblast for the two growth factors and the cytotrophoblast for the receptor during most of pregnancy and was absent late in pregnancy. The placental maternal giant cells contained specific immunoreactivity for all the immunogens from the middle of pregnancy to term. This study demonstrates that the two growth factors and the epidermal growth factor receptor are present in the endometrium and placenta of cats and suggests that these growth factors may play an autocrine/paracrine role during reproduction     

Nitric oxide modulates salivary amylase and fluid, but not epidermal growth factor secretion in conscious rats

The involvement of the L-arginine/NO pathway in the control of salivary fluid, amylase and epidermal growth factor (EGF) secretion was investigated in conscious rats. For the collection of saliva, an oesophageal cannula was implanted. To obtain steady secretion, submaximal carbachol background infusion was given. Different treatments included NO synthase inhibitor N(G)-nitro-L-arginine (NOLA; with or without phentolamine, propranolol), L-arginine, D-arginine and NO donor 3-morpholinosydnonimine (SIN-1) administration. Volume, amylase activity and EGF output in the secreted fluid were determined in 30 min mixed saliva samples. Carbachol infusion alone produced a modest, sustained salivary fluid and amylase secretion. NOLA (30 mg/kg) further increased both fluid (p<0.001) and amylase outputs (p<0.001). These latter effects were prevented by L-arginine but not by D-arginine or by phentolamine. Propranolol administration decreased both fluid and amylase secretion below the carbachol plateau, and NOLA did not modify this suppressed secretory rate. SIN-1 did not alter either volume or amylase secretion. Interestingly, NOLA given without carbachol did not modify salivation. Neither carbachol nor NOLA changed salivary EGF output. The present results suggest that the L-arginine/NO pathway has a modulatory role in the cholinergic control of salivary amylase secretion, but not in EGF output. The mechanisms of inhibitory action of NO on salivary fluid and amylase secretion remain to be identified.     

Basic fibroblast growth factor in rat salivary glands

We studied the occurrence and localization of basic fibroblast growth factor (bFGF) in rat salivary glands using a specific monoclonal antibody. It was shown that the extract of rat salivary glands has a pronounced stimulatory activity on the growth of bovine capillary endothelial cells, which is blocked by the addition of an antibody against bFGF. The concentration of bFGF in the submandibular/sublingual gland, as determined by radioimmunoassay, was 80% that in the brain. Immunocytochemistry revealed bFGF-immunoreactivity localized primarily in the epithelial cells lining the striated ducts and excretory ducts of the parotid, sublingual and submandibular glands. In addition, intense bFGF-immunoreactivity was observed in the granular convoluted tubule of the submandibular gland, localized predominantly in the agranular pillar cells, which lay in small numbers among the majority of weakly immunostained cells containing many apical secretory granules. At the electron-microscopic level, the immunoreactive material was distributed diffusely in the cytoplasmic matrix and nuclei of all immunoreactive cells, whereas it was absent from all cytoplasmic organelles including the secretory granules. These results indicate that bFGF is localized in different cellular and subcellular compartments from those of other growth factors in the duct system of rat salivary glands.