Sprague–Dawley rats were used. The animals were deeply anesthetized with chloral hydrate (600 mg/kg body weight, i.p.) and perfused transcardially with 0.02 M phosphate-buffered saline (PBS; pH 7.2) followed by 4 % paraformaldehyde in 0.1 M phosphate buffer (PB; pH 7.4). For prenatal experiments, the day on which a vaginal plug was recognized was considered as “embryonic day (E) 0”. Pregnant mothers at E18, and E20 were sacrificed by an overdose injection of chloral hydrate (800 mg/kg), and the fetuses were extracted by caesarian surgery. The whole heads were fixed in 4 % paraformaldehyde in 0.1 M PB (pH 7.4) for 3 days. All postnatal pups were decalcified with 7.5 % ethylene diamine tetraacetic acid (EDTA) for 1–4 weeks at 4 °C. After decalcification, the head was cut into exact halves along the medial plane. For frozen sections two heads from each group were transferred to PBS containing 20 % sucrose. For paraffin sections, two heads from each group were post-fixed in 4 % paraformaldehyde in 0.1 M PB (pH 7.4) overnight or for longer. Specimens were then dehydrated by ethanol, cleared in xylene and embedded in paraffin. In accordance with previous reports [7, 8], at E18, gland buds and epithelial cords with a terminal bulb at the distal end were elongated from the epithelial basement membrane (ectoderm) into the stromal connective tissue (mesenchyme). At E20, branching and lumenization had taken place, and immature acini and ducts were formed. We applied lectin histochemistry to frozen sections to avoid false-negative errors. Parasagittal frozen sections were prepared at a thickness of 14 μm, and mounted on MS-coated glass slides (Matsunami, Osaka, Japan), rinsed with PBS, dried and processed for lectin histochemistry. For the identification of specific carbohydrate residues, tissue sections were incubated for 30 min with 0.3 % H₂O₂ in methanol to block endogenous peroxidase activity. The sections were then incubated for 12–14 h with one of seven different biotinylated lectins, Glycine max (SBA), Dolichos biflorus (DBA), Vicia villosa (VVA), Ulex europaeus (UEA-1), Triticum vulgare (WGA), Succinyl WGA (sucWGA) or Arachis hypogaea (PNA). Sections were then washed three times in PBS, followed by incubation with ABC (Vector Laboratories) for 60 min and then washed again three times with PBS. The horseradish peroxidase was visualized by incubating slides with 0.05 M Tris–HCl buffer (pH 7.5) containing 0.08 % diaminobenzidine and 0.003 % H₂O₂. Sections were then lightly counterstained with methyl blue, dehydrated and cover slips were mounted using Permount (Fisher, NJ, USA). Sections were examined by light microscopy.

At the prenatal stage, particularly at E18, lectin histochemistry showed considerable variety in the extent or presence of staining among different animals and even among different buds in the same gland. In general, at E18, N-acetylgalactosamine residues visualized by DBA, SBA and VVA showed heterogeneous staining that was negative to weakly positive in the terminal buds. At E21 this staining became progressively more moderate, in the cytoplasm and lumen of ducts, but VVA staining was confined to apical cytoplasm. PNA showed a heterogeneously positive reaction to cell membranes in the epithelial cord but little or no reaction to cells in the terminal buds. At E20, the PNA reaction was located near the lumen. At E18, UEA staining was similar to that of PNA; however, at E20 UEA staining was moderate in the apical cytoplasm and cell membrane as well as on the lumen surface of ducts. WGA reacted strongly at E18, while at E20 staining was moderate on cell membranes and in apical cytoplasm. sucWGA did not show any reaction at E18, but at E20 it stained the cell membrane but not the apical cytoplasm, where secretory granules reside.

In prenatal developmental differentiation, our data showed the importance of terminal sialic acid rather than N-acetylglucosamine, as indicated by intense binding of WGA to the cell membrane and stromal cells and the lack of sucWGA binding [34]. sucWGA labeling appeared largely during postnatal development. Other lectins showed heterogeneous patterns of staining with high affinity of PNA and UEA-1 observed in the epithelial cord at the bud stage. This pattern of reaction is similar to that of developing human labial and lingual minor salivary glands [35]. The varieties in the reactivity of lectins during epithelial budding and bud migration are indicative of a differentiation-dependent alteration in cell surface carbohydrates [36].

Sprague–Dawley rats aged PN 0–7 were used. We used the same methods for histological analysis and lectin histochemistry as for postnatal animals. The classification of the developmental periods used in this study is based upon the physiological and nutritional stages of animal development [37]. Histological analysis of suckling stage rats showed that the acini of newborns had the general overall appearance of the adult; they contained basophilic nuclei located basally within basal eosinophilic cytoplasm, and pale apical cytoplasm underlined by clear lumen. At day 0, the secretory units of palatine glands consisted of immature acini and ducts sparsely distributed in the connective tissue, and no extending epithelial cord was observed. The nuclei were round in shape and flattened progressively with age. Gradually with maturation, the glands arranged into lobules, and the number and size of acini increased. Lectin histochemistry at the suckling stage showed that all lectins examined, except PNA, bound to the luminal border and showed generally similar reaction patterns. The staining tended to be diffuse or reticular in the apical cytoplasm and apical membrane, as well as in the basolateral membrane of acini in newborns. Moreover, SBA and to a lesser extent PNA showed more positive reactions at supranuclear membranes. The staining by all lectins progressively increased to become moderately distributed. UEA reacted in a similar pattern but was more intense and diffuse, particularly in the acini located on the anterior area of the soft palate where the mother’s nipple sits during suckling.

Palatine glands of newborns exhibited the same general overall appearance as those of the adult; therefore, unlike the parotid gland [37], and to a lesser extent the submandibular gland [38], no dramatic histological or morphological changes occurred during postnatal development. Lectin labeling in the newborn was located in the apical portion of the cytoplasm, with slight variation, and with maturation progressively increased and spread out basally corresponding with the progressive enlargement of the apical eosinophilic cytoplasm where secretory granules exist.

Sprague–Dawley rats aged PN 10–14 were used. In this stage, both suckling and feeding on solid food occur after eruption of teeth and before weaning. The histological sections showed enlargement of the glandular lobules, while the interlobular connective tissue was slightly reduced. Some acini had formed the tubuloacinar and a few scattered serous cells were observed. Histochemically, the distribution of lectin staining remained similar to that in the first week, although the extent of reactions was slightly increased and a granular pattern of staining was noticed in most acini. In addition to mucous cells, UEA showed high affinity to serous cells. sucWGA, however, rarely reacted to serous cells.