Caveolin-1 is an integral membrane protein, with both the N- and C-terminal domains being cytoplasmic, and is the principal component of caveolar membranes. Caveolae are specialised vesicular microdomains of the plasma membrane that are formed by localised accumulation of cholesterol, glycosphingolipids, and caveolin. Caveolae can be found individually or in clusters and are found in most cell types, but they are most abundant in adipocytes, endothelial cells, fibroblasts, and muscle cells. Functionally, caveolae have been implicated in endothelial transcytosis, potocytosis, and signal transduction. Caveolin-1 can play a role in intracellular vesicular and cholesterol trafficking.

Recently, caveolin-1 has been found unexpectedly in the cytoplasm, mitochondria and elements of the secretory pathways of exocrine secretory cells. These data suggest that caveolin-1 may also exist in a soluble form as a cytosolic protein or associated with secretory products. However, the function of secreted caveolin-1 is still unknown, but a role in extracellular lipid transport has been hypothesised.

To better understand this role of caveolin-1, we have immunohistochemically localised caveolin-1 in secretory cells of the gastro-intestinal tract of an amphibian, the red-legged frog Rana aurora aurora. We have previously described the gastro-oesophageal mucosa of this anuran. In the red-legged frog, as in other Ranidae, such as Rana esculenta, pepsinogen in the gastric juice is for the larger part synthesised by oesophageal glandular cells, whereas hydrochloric acid is produced by oxyntic cells in fundic glands. Thus, there is a gradient in the production of proteolytic enzymes and hydrochloric acid along the oral–aboral axis of the gastro-oesophageal tract, as has been reported for other non-mammalian vertebrates. Since peptic glands are particularly voluminous and secrete abundant amounts of pepsinogen, the oesophageal mucosa of this frog seems to represent an appropriate model to investigate the presence and functions of caveolin-1 in the secretory pathway.

Specimens of R. a. aurora from western North America were obtained from Minervini (Bari, Italy).

Four adult specimens of R. a. aurora were sacrificed after ether anaesthesia and their digestive tracts were quickly removed and processed for embedding in Technovit 8100 (Heraeus-Kulzer, Wehrheim, Germany). Small samples were fixed with 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS), pH 7.4, for 4 h at 4°C. After several rinses in PBS, the samples were incubated overnight at 4°C in PBS containing 6.8% sucrose. They were then dehydrated at 4°C in a graded series of acetone. Specimens were incubated in the Technovit 8100 monomer solution for 6 h at 4°C under gentle stirring. Polymerisation was carried out on ice for 3 h. Semithin sections (3 ?m thick) were cut with glass knives using an Ultratome (LKB, Uppsala, Sweden). Before histochemical and immunohistochemical staining, sections were mounted on microscope slides coated with polylysin and were incubated for 5 min at 37°C in 0.01% trypsin and 0.1% CaCl₂, in PBS, pH 7.8, for antigen retrieval.

Zymogen granules were identified with a modified Bowie’s staining method according to, or with a combined PAS-Bowie staining method to demonstrate both mucus and zymogen granules simultaneously. All sections were counterstained with haemallum.

Some sections were stained using the Masson–Fontana silver method to detect argentaffin endocrine cells.

Caveolin-1 and pepsinogen were localised using the peroxidase–antiperoxidase (PAP) method. Endogenous peroxidase was blocked by incubating sections in PBS containing 1% H₂O₂ for 10 min at room temperature. After several rinses in PBS, sections were incubated for 2 h at 37°C in a humid Petri dish with the primary antibodies (anti-caveolin-1: ABR, Golden CO, USA; anti-pepsinogen-1: DPC Biermann, Bad Nauheim, Germany) diluted 1:100 in PBS containing 1% normal goat serum (blocking buffer). Sections were washed in blocking buffer and incubated for 1 h at 37°C with goat anti-rabbit IgG (Sigma, St. Louis MO, USA) diluted 1:100 in blocking buffer. After several washes in blocking buffer, sections were incubated with horseradish PAP complex (Sigma) at a dilution of 1:100 for 1 h at 37°C. Finally, immunolabelling was visualised by incubation with 3-3?-diaminobenzidine–H₂O₂ medium for 10 min at room temperature.

Controls were performed by using antibodies pre-adsorbed with the respective antigens or by omitting the primary antibodies.

Images were captured using an E600 photomicroscope equipped with a digital camera DMX1200 (Nikon, Kawasaki, Japan).

Small samples of gastro-oesophageal mucosa obtained from four adult specimens of R. a. aurora were fixed in a mixture of 3% paraformaldehyde and 1% glutaraldehyde in 0.1 M PBS, pH 7.4, for 4 h at 4°C, and then postfixed in 1% OsO₄ in PBS for 30 min at 4°C. Fixed specimens were washed in several changes of PBS and dehydrated in a graded series of ethanol. Finally, samples were embedded in Epon (Taab, Redding, UK). Sections were mounted on formvar-coated nickel grids and stained routinely with uranyl acetate and lead citrate. Finally, grids were observed with an EM 109 electron microscope (Zeiss, Oberkochen, Germany).

For immunoelectron microscopy, ultrathin non-osmicated sections, mounted on formvar-coated gold grids, were treated with 0.05 M glycine in PBS buffer for 15 min at room temperature. Grids were incubated for 30 min at room temperature with 1% BSA in PBS containing 0.2% gelatin (PBG) and then placed on a drop of anti-caveolin-1, diluted 1:100 in PBG overnight at 4°C. The grids were rinsed with PBG, and then incubated in a dilution of 1:10 of 10-nm gold-conjugated anti-rabbit IgG (Sigma) in PBG for 1 h at room temperature. After several rinses in PBG and distilled water, the grids were lightly stained with uranyl acetate and lead citrate.

Immunolabelling controls were performed by using antibodies pre-adsorbed with antigens or by omitting the primary antibodies.

The oesophagus of the frog R. a. aurora was found to be lined with a pseudostratified ciliated epithelium containing abundant amounts of mucous goblet cells. Two different types of goblet cells were found in the epithelium. Type-I goblet cells were characterised by supranuclear cytoplasm containing large PAS-positive secretory vesicles that were often confluent. Type-II goblet cells were thinner with ovoid basal nuclei and numerous small secretory granules in the supranuclear cytoplasm. The granules, which were PAS positive, often showed weak immunostaining of pepsinogen.

Figure 1. Gastro-oesophageal mucosa of the frog R. a. aurora stained with different histochemical methods. (A–C) Oesophageal mucosa. (A) The ciliated epithelium contains two types of goblet cells (g1, and g2). Oesophageal glands consist of serous (peptic) cells (s) and mucous cells (m). PAS-Bowie-haemallum staining. (B) Numerous pepsinogen-positive granules (arrows) are present in serous cells (s) in oesophageal glands. Anti-pepsinogen staining and haemallum. (C) Secretory granules (arrows) are stained for caveolin-1. (s) Serous cells. Anti-caveolin-1 and haemallum staining. (D–F) Gastric mucosa. (D) The mucosa is lined with a single layer of mucus-secreting cells (sc). Gastric glands mainly consist of mucous neck cells (nc) and oxyntic cells (ox). PAS-Bowie-haemallum staining. (E) Oxyntic cells (ox) are not positive after anti-pepsinogen staining. Anti-pepsinogen and haemallum staining. (F) Immunostaining is not present in oxyntic cells (ox) after staining with anti-caveolin-1. Anti-caveolin-1 and haemallum staining. Bars: (A,D–F) 30 ?m; (B,C) 25 ?m.