Subcellular localization of epidermal growth factor receptor in human submandibular gland

Abstract

The subcellular distribution of the epidermal growth factor receptor (EGFr) was demonstrated in the normal human submandibular gland by means of immunogold cytochemistry. EGFr labelling appeared in both acinar and ductal cells, where strong immunoreactivity was associated with a tubulovesicular system near the basolateral surfaces. In addition, groups of reactive vesicles were highlighted among secretory granules of both serous and mucous cells and at the apex of ductal cells. Basolateral vesicles were interpreted as being a result of EGFr internalization after activation by an exogenous ligand, although the functional meaning of those located apically remains unclear.

Introduction

Salivary glands are influenced by exogenous circulating epidermal growth factor (EGF), which stimulates their development and causes changes in their activity (Purushotham et al. 1995; Ohlsson et al. 1997; Kashimata et al. 2000). Although we know this from experiments on rodent glands, many data suggest that human salivary glands also are controlled by EGF (Zhang et al. 2000; Yeh et al. 2005). Moreover, human glands produce EGF and release it into the saliva, where it is involved in mucosal repair in the mouth and in the gastro-oesophageal tract (Dubiel et al. 1992; Sarosiek & McCallum, 1995; Cossu et al. 2000). The biological effects of EGF are mediated by the activation of a tyrosine kinase receptor (EGFr) that recognizes both EGF and other factors. Activated EGFr undergoes dimerization, autophosphorylation and association with signal transductors (Boonstra et al. 1995); the cascade reactions end in the synthesis of molecules that modulate gene transcription. After binding, EGF, phosphorylated EGFr and adaptor protein are internalized in endosomes and recycled or degraded in lysosomes (Oksvold et al. 2000).

In the human parotid gland, we recently found EGF and EGFr in basolateral vesicles of both acinar and ductal elements (Lantini et al. 2001; Piludu et al. 2002). We interpreted this as a morphological sign of internalization of EGFr following binding to exogenous EGF. Moreover, we noticed that other EGFr-containing vesicles were grouped close to the secretory granules. This finding was intriguing because acinar granules contain the ligand EGF (Cossu et al. 2000; Lantini et al. 2001). It has been suggested that EGFr activation may occur not only on the surface but also inside the cell with endogenous ligands, giving weight to the belief that the EGF–EGFr system regulates cell activities in an autocrine fashion. We have extended the study of the intracellular localization of EGFr to other human salivary glands: and in this paper we show that cells of the submandibular gland also express EGFr both near the surface and in other cellular locations.

Materials and methods

Surgical samples of normal submandibular glands were obtained from eight consenting patients, five men and three women, aged 46–70 years, undergoing surgery at the Otorhinolaryngology Clinic, University of Cagliari. The procedures were approved by the Human Experimentation Commettee, University of Cagliari. The samples were cut into small pieces and fixed in a mixture of 3% formaldehyde and 0.1% glutaraldehyde in 0.1 m cacodylate buffer (pH 7.2) for 2 h. After washing with the same buffer, the pieces were dehydrated and embedded in Epon resin (Glycide Ether 100, Merck, Darmstadt, Germany).

Ultrathin sections collected on nickel grids were treated with 1% bovine serum albumin (BSA) and 5% normal goat serum (NGS) in phosphate-buffered saline (PBS) to block non-specific binding. The sections were incubated overnight at 4 °C with the primary antiserum, a mouse monoclonal antibody specific for the intracellular domain of human EGFr (SIGMA-Aldrich Srl, Milan, Italy) diluted 1 : 25 in the BSA–NGS–PBS solution. After rinsing with PBS, the grids were incubated for 60 min at room temperature with the secondary antiserum, goat anti-mouse IgG Auroprobe EM G10 (Amersham International plc, Little Chalfont, UK), diluted 1 : 50 in 1% BSA–PBS. The grids were washed with PBS and distilled water, stained with uranyl acetate and bismuth subnitrate, and observed and photographed with a JEOL 100S transmission electron microscope.

Controls were obtained with a non-immune mouse serum, with the primary serum absorbed with EGFr (SIGMA-Aldrich Srl), with an unrelated monoclonal antibody (mouse anti-human GH, SIGMA-Aldrich Srl), or omitting the primary antibody.

A few samples of the same glands were osmicated and routinely prepared as morphological controls.

Results

In all the glands examined, specific EGFr reactivity was observed in both acinar and ductal elements. Cell surfaces were occasionally decorated with gold particles, although they were expected to show the strongest staining. Instead, in both acinar and ductal cells, strong immunoreactivity was associated with a tubulovesicular system that was highlighted by the labelling. The most evident labelling was near the basolateral plasmalemma (Fig. 1A). In serous cells, groups of reactive vesicles were also located among secretory granules (Fig. 1B). By contrast, secretory granules, which are the most representative feature of acinar cells, were almost entirely unstained. The mucous cells also expressed EGFr, with gold particles deposited on isolated vesicles or on small clusters of these structures scattered among mucous droplets (Fig. 1C).

Fig. 1.

Epidermal growth factor receptor (EGFr) immunoreactivity in secretory cells of human submandibular gland. (A) Serous acinus. EGFr labelling appears within a row of small, pale vesicles located just beneath the plasma membrane. Inset: basal vesicles in an osmicated sample. (B) Serous acinus. Groups of labelled vesicles are highlighted among the secretory granules; this is better seen in osmicated samples (inset). (C) Mucous cell. Specific EGFr labelling is located in the cytoplasm near the mucous droplets. (D) Acinar cells. Note the presence of specific EGFr reactivity in the nuclei.

EGFr reactivity appeared sporadically in intercalated ducts where it was intense in some cells, weak or even absent in others (Fig. 2A). In striated duct cells, labelled vesicles were grouped chiefly within the basolateral cytoplasmic folds that characterize these elements (Fig. 2B), and in lesser amounts apically (Fig. 2C) and near the nucleus. Diffuse reactive vesicles were present in most cells of the interlobular ducts (Fig. 2D).

Fig. 2.

Epidermal growth factor receptor (EGFr) immunoreactivity in ductal cells of human submandibular gland. (A) Intercalated duct. Only one cell shows a diffuse set of labelled vesicles. (B) Striated duct. The basolateral processes of the cells appear occupied chiefly by mitochondria and by numerous vesicles labelled for EGFr. Inset: osmicated sample. (C) Striated duct. The apical cytoplasm is filled with many vesicles and tubules, some of which are decorated with gold particles. (D) Interlobular duct. Apex of a superficial cell showing many labelled vesicles.

Finally, nuclei of acinar and ductal cells were often intensely labelled (Fig. 1D).

Control sections were always unstained.

Discussion

The pattern of labelling in acinar and ductal cells of human submandibular gland indicates that EGFr is chiefly intracellular and basal in location. The paucity of surface labelling does not contradict its designation as a transmembrane receptor: it probably reflects a short duration of EGFr within the membrane and an accumulation of internalized EGF–EGFr complex. The surface-bound receptor remains at low level, being slowly replaced by the recycled one. In addition, undifferentiated, neoplastic or stem-like cells seem to express high levels of surface-bound EGFr (Gusterson et al. 1984; Damjanov et al. 1986; Yamada et al. 1989; Aida et al. 1994), while a predominantly intracellular EGFr reactivity has been described in specialized cells such as the liver, pancreas and prostate cells (Damjanov et al. 1986), as in our specimens.

The most relevant result of our experiments concerned the labelling of EGFr within small vesicles grouped mainly near the basolateral cell surfaces, where previous experiments also indicated the presence of EGF. It might be deduced that what we highlighted were the vesicles that internalize the EGF–EGFr complex just after receptor activation. Indeed, we did not observe any labelled multivesicular body, where endocytosed materials are usually packed (Oksvold et al. 2000).

Our results suggest that both acinar and ductal cells of human submandibular gland respond to exogenous EGFr ligands with changes in their activity. The best characterized functional effects of EGFr activation are related to cell proliferation and differentiation (Boonstra et al. 1995). We can easily imagine intercalated duct cells under this type of effect as these elements were generally regarded as reservoir cells (Tandler, 1993). However, many data indicate that acinar and striated duct cells also undergo mitosis under appropriate conditions (Dardick et al. 1993; Denny et al. 1997; Man et al. 2001). Therefore, EGFr activation may represent a powerful means of regulating the renewal of salivary gland cells by stimulating proliferation of all parenchymal cell types. Besides this effect, EGFr activation may influence the normal activities of both acinar and ductal cells such as secretion or absorption; this possibility is indicated by data obtained with rodent glands and with human cell lines that concern salivary secretion (Slomiany & Slomiany, 2004; Yeh et al. 2005). In addition, an intriguing hypothesis is that the EGF–EGFr system helps to maintain cell shape, as suggested by recent studies documenting that exposure to EGF causes formation of cytoplasmic processes and membrane ruffling (Boonstra et al. 1995; Bailly et al. 2000). This may be related to the mechanism by which striated duct cells keep their peculiar membrane folds in order to concentrate basally the membrane enzymes involved in electrolyte transport (Cossu et al. 1984).

This interpretation appears appropriate for the labelled vesicles near the cell surface but does not explain the presence of those in the apical cytoplasm of both acinar and ductal cells. The simplest interpretation is that apical vesicles have a secretory fate, given that EGFr has also been found in saliva (Streckfus et al. 2000). However, certain observations must be taken into account. First, the vesicles at the apex of striated duct cells are partly secretory and partly absorptive (Hand, 1979; Tandler et al. 2001), so it is possible that EGFr is internalized even at this site. EGF also has been found here, suggesting that striated duct cells engage in re-uptake by transcytosis of the EGF released by acini (Lantini & Cossu, 1998; Tandler et al. 2001), but the presence of the receptor makes this hypothesis more doubtful. With regard to the acinar cells, the proximity between the labelled vesicles and secretory granules suggests that they take part in the secretory process. It appears unlikely that they fuse and combine their contents, because then the receptor would come into contact with the ligand EGF within the granules (Cossu et al. 2000) undergoing intragranular activation. Besides, EGFr-positive vesicles were also found around mucous droplets, which lack EGF (Piludu et al. 2003). It appears more probable that EGFr near acinar granules represents a regulatory tool for the last phases of secretion, for instance by changing the actin polymerization state (Boonstra et al. 1995). That EGFr activation influences the secretory process is documented by studies on amylase and mucin secretion, but most of these studies focused on the effects on the earliest transcriptional phases of mucin production (Purushotham et al. 1995; Nadel, 2001).

More recent data indicate that activated EGFr can influence gene transcription not only through the phosphorylative reaction cascade of the signal transduction, but also acting directly on DNA. The EGF–EGFr complex is able to migrate into the nucleus and bind to A- and T-rich sequences, although how it crosses the nuclear envelope is still obscure (Lin et al. 2001; Waugh & Hsuan, 2001). EGFr labelling within nuclei might be related to this function.