In rodents, as in humans, the secretory units of the sublingual salivary gland consist of mucous cells forming acini or tubules which are capped by serous cells arranged as demilunes (Young and Van Lennep, 1978). In the adult gland, these two cell types are readily identified by their characteristic morphology. The mucous cells contain electron-lucent secretory granules that fill the apical cytoplasm and compress the nucleus against the basal cell membrane. In contrast, the nuclei of the serous cells are spherical, and their cytoplasm contains abundant rough endoplasmic reticulum and moderate numbers of electron-dense granules located near the intercellular canaliculi which convey the cells' secretions to the acinar lumen. Although it is recognized that certain aspects of this conventional description are artifacts of chemical fixation (Ichikawa and Ichikawa, 1987; Yamashina et al., 1999), it provides a convenient baseline against which data from experimental and developmental studies may be compared and evaluated.
Several secretory products of the rat sublingual gland, including neonatal submandibular gland secretory protein B (SMGB), parotid secretory protein (PSP), common salivary protein 1 (CSP-1) and sublingual mucin have been identified and characterized (Moschera and Pigman, 1975; Ball et al., 1988, 1993; Girard et al., 1993; Watson et al., 1997; Mirels et al., 1998b; reviewed in Denny et al., 1997). The biochemical composition of an additional sublingual secretory product, neonatal submandibular gland secretory protein D (SMGD) (Ball et al., 1991), has not yet been determined. Of these proteins, sublingual mucin is specific to the mucous acinar cells, and SMGD reactivity has been detected in both mucous acinar and serous demilune cells of adult glands. The remaining known sublingual secretory proteins are products of the serous demilunes. All known serous demilune-specific proteins—SMGB, PSP, and CSP-1 (as well as prolactin-inducible protein (PIP) (Mirels et al., 1998a) and peroxidase (Moriguchi et al., 1995; Redman et al., 1998), which were not examined in the current study)—are also products of the immature acinar cells of the developing parotid and submandibular glands (Strum, 1971; Yamashina and Barka, 1972; Redman and Field, 1993; Mirels et al., 1998a).
Development of both the submandibular and parotid glands includes a prolonged postnatal period of cytodifferentiation (Redman and Sreebny, 1971; Chang, 1974; Cutler and Chaudhry, 1974; Alvares and Sesso, 1975; Ball and Redman, 1984; Sivakumar et al., 1998). In the submandibular gland, SMGB, SMGD, PSP, and CSP-1 are products of the proacinar (type III) cells, which are present from day 19 in utero and complete their differentiation to adult acinar cells 3–4 weeks after birth (Ball et al., 1988, 1991, 1993; Moreira et al., 1990, 1991; Girard et al., 1993). The transition from submandibular proacinar (type III) to seromucous acinar cells involves an intermediate cell type (type IIIP), characterized by the presence of heterogeneous, morphologically distinct granules containing both neonatal (SMGB, PSP, and CSP-1) and adult secretory proteins (Ball et al., 1988; Moreira et al., 1990, 1991). At early stages, type IIIP cells may contain a mixture of these heterogeneous secretory granules and typical type III cell granules containing only neonatal proteins, whereas later in development, type IIIP cells with both type IIIP and adult MGs are found. Some proteins characteristic of the neonatal submandibular proacinar cells (PIP and peroxidase) are also produced by the mature seromucous acinar cells (Strum and Karnovsky, 1970; Mirels et al., 1998a), whereas the major proacinar cell products (SMGB, SMGD, CSP-1, and PSP) are present only in some intercalated duct cells in the adult gland.
Parotid gland cytodifferentiation also is incomplete until approximately 1 month of age. Although morphologically unique secretory granules are not a feature of immature parotid acinar cells, biochemical and immunocytochemical studies have demonstrated that, as in the submandibular gland, the complement of parotid acinar secretory proteins changes significantly during the third postnatal week. SMGB, PSP, CSP-1, PIP, and peroxidase are all products of the neonatal parotid acinar cells (Redman and Field, 1993; Sivakumar et al., 1998; Mirels et al., 1998a). By the fourth postnatal week, however, expression of both SMGB and CSP-1 is restricted to intercalated ducts, whereas adult acinar cells continue to produce PSP, PIP, and peroxidase, and initiate the synthesis of additional adult proteins (Lazowski et al., 1992; Charest et al., 1993; Sivakumar et al., 1998; Mirels et al., 1998a).
In contrast to the submandibular and parotid glands, sublingual cytodifferentiation is thought to be completed during prenatal development. Sublingual gland development has been investigated previously (Laj et al., 1971; Redman and Ball, 1978; Moriguchi et al., 1995; Taga and Sesso, 1998). The rat sublingual rudiment is initiated at age 14 days in utero, in the same mesenchymal capsule as the submandibular gland (Redman and Sreebny, 1970). As described by Redman and Ball (1978), at ages 18–20 days in utero, the cells in the terminal buds are arranged around distinct lumina and contain secretory granules exhibiting three different morphological appearances. These secretory granule types are: 1) dense (apparently serous), 2) empty-looking (apparently mucous), or 3) “mixed” (mucous-appearing granules with dense cores). At birth, cells containing mixed granules are infrequent, the mucous and serous cells have a typical adult morphology, and the terminal buds have reorganized to form mucous acini with serous demilunes.
The characterization of sublingual gland cell-specific marker proteins has made it possible to revisit sublingual gland development, with emphasis on two potential parallels with submandibular and parotid development. The goals of this study were: 1) to investigate whether the serous demilune products SMGB, PSP, CSP-1, and SMGD, which are among the earliest-expressed and most abundant proteins of the developing submandibular and parotid glands, are also developmentally regulated in the sublingual gland; and 2) to determine whether the cells containing mixed granules observed in the prenatal sublingual acini represent an intermediate or transitional cell type in sublingual cytodifferentiation, similar to the immature acinar cells seen in submandibular and parotid development.
MATERIALS AND METHODS
All procedures involving animals were approved by the Animal Care Committee of the University of Connecticut Health Center, and carried out according to institutional and NIH guidelines. Timed-pregnant female Sprague-Dawley rats were obtained from Harlan (Indianapolis, IN) on the 13th day of gestation, housed singly in plastic microisolator cages, and provided with laboratory chow and water ad libitum.
Developing glands were taken from fetuses at ages 16, 18, 19, and 20 days in utero, and from postnatal animals at 0, 1, 5, 9, 14, 18, 25, 40, and 60 days of age. To obtain fetal glands, the mothers were narcotized in a CO2 chamber and then killed by cardiac transsection, and the fetuses were removed and placed in ice-cold phosphate-buffered saline (PBS). The glands were quickly dissected and placed in cold fixative solution. Newborn, 1-, and 5-day-old pups were chilled on ice prior to dissection of their glands. All older animals were anesthetized by an intraperitoneal injection of Ketamine/Xylazine (90 mg/10 mg/kg body weight). The sublingual glands were then either dissected, cut into small (∼1 mm3) pieces, and immersed in fixative solution, or were fixed by vascular perfusion of fixative solution through a cannula placed in the ascending aorta.
For light microscopy, the sublingual glands were fixed overnight at 4°C in 4% paraformaldehyde in 0.1 M sodium cacodylate buffer, pH 7.4. After rinsing in cacodylate buffer, glands to be used for immunofluorescence labeling were embedded in Histoprep (Fisher Scientific, Fairlawn, NJ), rapidly frozen in isopentane cooled on dry ice, and sectioned at 8–10 μm in a cryostat. The sections were collected on SuperFrost Plus slides (Fisher Scientific) and stored at 4°C for up to 2–3 days prior to use. Glands to be used for immunogold silver staining were dehydrated in cold methanol, embedded in LR Gold resin (London Resin Co., Basingstoke, UK) and polymerized in UV light (365 nm) at –20°C. One-micrometer sections of LR Gold-embedded tissue were cut with glass knives and placed on SuperFrost Plus slides. For electron microscopy, glands were fixed in 1% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, rinsed in buffer, dehydrated in ethanol, substituted with propylene oxide, and embedded in Polybed epoxy resin (Polysciences, Warrington, PA). Tissue samples for each age were also dehydrated in cold methanol and embedded in Lowicryl K4M (Chemische Werke Lowi, Waldkraiburg, Germany) at –20°C. For routine morphological study, some tissues were postfixed in 1% OsO4/0.8% potassium ferricyanide in 0.1 M cacodylate buffer, stained in block with 0.5% aqueous uranyl acetate, and embedded in Polybed resin. Thin sections were cut with a diamond knife and collected on bare or formvar-coated nickel grids for immunogold labeling, or on copper grids for morphological study.
For immunofluorescence, sections were rinsed in PBS, treated with 1% bovine serum albumin (BSA)/5% normal goat serum (NGS) in PBS to block nonspecific binding, and incubated with primary antibody diluted in 1% BSA/5% NGS in PBS for 30 min at room temperature. After rinsing with PBS, the sections were incubated with FITC-labeled goat-anti-rabbit IgG (Cappel, Durham, NC) for 30 min and rinsed with PBS, and coverslips were mounted with Vectashield (Vector Laboratories, Burlingame, CA). The sections were examined and photographed in a Leitz Orthoplan microscope using epifluorescence illumination.
For immunogold silver staining, nonspecific binding was blocked as above and the sections were incubated with primary antibody for 90 min at room temperature. The sections were then rinsed with PBS and incubated for 60 min with gold-labeled goat-anti-rabbit IgG (Amersham, Arlington Heights, IL) diluted in 1% BSA in PBS. After rinsing thoroughly with PBS and distilled water, the bound gold was visualized by silver enhancement (British BioCell International, Cardiff, UK). The sections were lightly stained with 1% methylene blue/1% azure II, and coverslips were mounted with DPX. The sections were examined and photographed using brightfield illumination.
Electron microscopic immunogold labeling was done as previously described (Hand, 1995). Nonspecific binding was blocked with 1% BSA/1% instant milk in PBS, then the grids were incubated with primary antibody diluted in 1% BSA/5% NGS in PBS for 1 hr at room temperature or overnight at 4°C, and rinsed with PBS. The bound antibodies were visualized by incubation with gold-labeled goat anti-rabbit IgG (10 nm diameter; Amersham). After thorough rinsing with PBS and distilled water, the sections were stained with uranyl acetate and lead citrate and examined in a JEOL 100CX transmission electron microscope.
Immunocytochemical controls included omission of the primary antibody from the labeling sequence, and substitution of preimmune or nonimmune serum for the primary antibody.
Antibodies to rat CSP-1 (Girard et al., 1993) and SMGB (Mirels et al., 1998b) were prepared and characterized as described earlier. Antibodies to rat submandibular gland protein SMGD (Ball et al., 1991) and rat parotid secretory protein (PSP) (Ball et al., 1993) were obtained from Dr. William D. Ball, Howard University. Antibody to rat sublingual gland mucin (Man et al., 1995) was provided by Dr. David J. Culp, University of Rochester.
The ultrastructure of the developing rat sublingual gland was similar to that described for mouse and rat in previous studies (Laj et al., 1971; Redman and Ball, 1978; Moriguchi et al., 1995). At age 18 days in utero, most terminal buds lacked a lumen and consisted mainly of relatively undifferentiated cells. These cells typically had abundant rough endoplasmic reticulum and a prominent Golgi apparatus, but lacked secretory granules. A few terminal buds did have a lumen, however, and the apical ends of cells adjacent to the lumen occasionally contained small granules of variable appearance (Figs. 1A and 2A). Some granules had an electron-dense content, others were electron-lucent, and some had a mixed content, with a dense core and a pale peripheral region (Fig. 2A).
By age 19 days in utero, more cells of the terminal buds contained granules, and larger numbers of granules were present in the cells. Cells containing electron-lucent MGs and cells containing electron-dense serous granules (SGs) were present around a common lumen (Fig. 1B). The MGs usually had a fine fibrillar content; their membranes were indistinct, and adjacent granules often were fused (Fig. 2B). The size and density of SGs varied from granule to granule, although the content of individual granules was generally of uniform density (Fig. 2C). Increased numbers of mixed granules with a dense core and electron-lucent periphery were present.
From age 20 days in utero through birth, the proportion of serous cells located at the periphery of the terminal buds increased, resulting in the typical configuration of mucous acini and serous demilunes (Fig. 1C). The size of the mucous cells and the number of granules increased during the late fetal and early postnatal period. Cells containing mixed granules were present in decreasing numbers through 5 days of age. The mixed granules had a variable density, and some contained a dense core located either peripherally or centrally within the granule (Fig. 3A and B). These cells also frequently contained MGs of typical appearance (Fig. 3A and B). From 9 days onward, cells containing mixed granules were rarely seen, and the morphology of the MGs was similar to that seen in adult glands (Fig. 3C).
Light Microscopic Immunocytochemistry
At age 18 days in utero, some reactivity for CSP1 was detected by immunofluorescence in cells of the developing sublingual gland, but little or no reactivity was seen with antibodies to the other proteins. By age 19 days in utero, both immunofluorescence and immunogold silver staining showed reactivity for CSP-1, SMGD, and mucin in the apices of cells in the terminal buds. By age 20 days in utero, differentiation of the mucous acinar and serous demilune cells was well under way. The acinar cells were labeled for mucin (Figs. 4A and 5D), and the demilune cells showed strong labeling for CSP1 (Figs. 4D and 5A) and SMGB (Figs. 4G and 5C). This pattern of labeling was maintained in newborn animals (Fig. 4B, E, and H) and at all older ages examined (Fig. 4C, F, and I). Reactivity for SMGD was observed in both demilune cells and mucous acinar cells (Figs. 4J–L, and 5B). Mucous cell reactivity for SMGD was stronger in cryostat sections labeled using immunofluorescence (Fig. 4K and L) than in plastic sections labeled by immunogold silver staining (Fig. 5B). Reactivity with anti-PSP also was present at age 19 days in utero and in early postnatal glands (Fig. 5E), but was undetectable in 9-day-old (Fig. 5F) and older postnatal glands.
Electron Microscopic Immunogold Labeling
A low level of reactivity for SMGB was observed in a few electron-dense granules in cells of the developing sublingual gland at age 18 days in utero (Fig. 6A). These presumably were SGs, since the dense granules present in cells with mucous-like granules and mixed granules were unreactive (Fig. 6B). These mucous-like granules sometimes labeled with anti-mucin, and the dense cores of the mixed granules occasionally labeled with anti-SMGD.
By age 19 days in utero, consistent labeling of serous and mucous secretory granules was seen. Small, irregularly-shaped as well as spherical granules, with dense content in the apical cytoplasm of some cells, were labeled with anti-CSP1 and anti-SMGB antibodies (Fig. 7A and B). Mucous-like granules and mixed granules sometimes showed reactivity with anti-mucin (Fig. 7C), with gold particles present over both the electron-lucent and electron-dense regions. Mucin granules with the appearance of typical mature granules also were labeled (Fig. 7D). Occasionally, the electron-dense cores of mixed granules labeled with anti-SMGD (Fig. 7E).
At age 20 days in utero, the labeling pattern was similar to that seen on the previous day. In addition, at this age, SGs also were labeled by the anti-SMGD antibody. The distribution of gold particles indicating the localization of SMGD was similar to that seen previously in postnatal rat sublingual gland (Ball et al., 1991), i.e., associated with the content at the periphery of the granules (Fig. 8A). Reactivity for PSP also was observed in SGs of fetal (Fig. 8B) and newborn animals, but little or no labeling was detected in the demilune cells of older animals.
In newborn and postnatal animals through 60 days of age, the labeling patterns observed with anti-mucin (Fig. 9A), anti-CSP1 (Fig. 9C), and anti-SMGB antibodies were similar to those seen in the glands of fetal rats. In the postnatal rats, as in the glands of younger animals, SMGD had a peripheral distribution in the granules of serous demilune cells. Mucous cells also continued to label with anti-SMGD. In Lowicryl sections, labeling was seen in the rough endoplasmic reticulum and the Golgi saccules, and up to 9 days of age, in the dense cores of the MGs (Fig. 9B).
In control preparations, no labeling was observed when the primary antibody was omitted from the labeling sequence (Fig. 9D).
Consistent with earlier studies of the prenatal development of the sublingual gland (Laj et al., 1971; Redman and Ball, 1978; Moriguchi et al., 1995), our morphological and immunocytochemical data reveal that cytodifferentiation of the mucous and serous cells, and the accumulation of secretory proteins occur very rapidly during the late fetal period. The terminal buds of the developing gland acquire lumina and exhibit the first indications of secretory activity at age 18 days in utero. Small amounts of mucin, CSP-1, and SMGB are present in the cells at this age, but within 1 day the levels of these proteins increase and reactivity for SMGD is also present. Through age 19 days in utero, the differentiating mucous and serous cells are typically located around a common lumen. By age 20 days in utero, serous cells have begun to move peripherally to form demilunes, and the lumen is lined mainly by mucous cells. This process continues over the next day or two, and by the time of birth the basic morphological and immunocytochemical characteristics of the adult gland are already established.
Initially, three types of secretory granules are evident in the developing gland: empty-looking, electron-lucent mucous-like granules; electron-dense, serous-like granules; and mixed granules with a peripheral lucent region and a dense core of variable size. The mucous-like granules were labeled with the anti-mucin antibody, whereas the electron-dense granules labeled first with anti-CSP-1 and anti-SMGB, and by age 19 days in utero, also with anti-SMGD. These results clearly indicate, as suggested by their morphological appearance, that the cells containing the mucous-like granules are the precursors of the adult mucous acinar cells, and the cells containing the serous-like granules are precursors of the adult serous demilune cells.
The labeling pattern exhibited by the mixed granules is more difficult to interpret. In some cases, these granules labeled with the anti-mucin antibody, suggesting that they were present in developing mucous cells. In some cases the mixed granules were reactive with anti-SMGD and, rarely, with anti-SMGB. As demonstrated here and reported earlier (Ball et al., 1991), SMGD reactivity is present in both serous and mucous cells of the sublingual gland. However, the early SG marker CSP-1, which appears prior to SMGD in cells with serous-like granules, was never seen in the mixed granules. This also is consistent with the suggestion that the cells with mixed granules are developing mucous cells. Similarly, the lack of CSP-1 labeling and infrequent SMGB labeling of mixed granules would argue against the possibility that the cells with mixed granules are serous cells differentiating into mucous cells. Redman and Ball (1978) previously suggested that the cells with mixed granules are mucous cell precursors and that the mixed granules contain incompletely glycosylated apomucin, synthesized prior to the development of the enzymatic mechanisms required for formation of carbohydrate side chains. The lack of reactivity with the anti-mucin antibody would be consistent with this possibility, as this antibody is known to react mainly with carbohydrate epitopes (Man et al., 1995).
The pattern of secretory protein expression in the fetal sublingual gland also suggests that mucous and serous cells differentiate from separate lineages. From the earliest times observed in the present study, mucin reactivity was found only in cells containing mucous and/or mixed granules. Additionally, the distribution and timing of serous protein expression suggests that the serous cells arise independently. This conclusion is supported by the occurrence of a mutation in the NFS/sld mouse strain in which differentiation of mucous cells of the sublingual gland is markedly delayed, but the serous cells appear to develop normally (Hayashi et al., 1988) (D.J. Culp and A.R. Hand, unpublished observations).
The secretory protein expression pattern of the sublingual gland differs substantially from those of the submandibular and parotid glands. In the latter two glands, CSP-1 and SMGB are expressed at their highest levels in the immature cells of the perinatal glands, and are restricted to a subset of intercalated duct cells in the adult (Ball et al., 1988; Girard et al., 1993; Sivakumar et al., 1998). SMGD and PSP are similarly regulated in the submandibular gland (Ball et al., 1991, 1993; Man et al., 1995). The dramatic changes in CSP-1, SMGB, PSP, and SMGD transcript and protein levels that accompany submandibular and parotid acinar cell maturation are readily observed in developmental Northern (Girard et al., 1993; Mirels et al., 1998b) and Western (Ball et al., 1991, 1993) blots. In contrast, Northern blots of sublingual gland RNA from postnatal days 5–30 showed consistent high levels of SMGB and CSP-1, and low levels of PSP transcripts (Mirels, unpublished observations).
Reactivity for PSP was detected at low levels in the serous cells of the fetal sublingual gland, but little or no PSP was found in the adult gland. The expression of the two other known secretory products of the sublingual serous cells, peroxidase and PIP, follows a similar pattern. Peroxidase is detectable in serous cells at age 18–20 days in utero (Moriguchi et al., 1995; Redman et al., 1998), and PIP is present in newborn animals (Mirels et al., 1998a), although fetal glands were not examined in the latter study. Like PSP, these two proteins are present at very low levels or are absent in adult glands, suggesting that some modest, developmentally regulated restriction of sublingual gland gene expression or protein synthesis occurs postnatally.
In conclusion, the results of this study indicate that the expression of mucin and the serous cell proteins CSP-1 and SMGB in the rat sublingual gland is initiated at age 18 days in utero, and that synthesis of the serous and mucous cell protein SMGD begins at age 19 days in utero. PSP was detected only in fetal glands and may, with peroxidase and PIP, form a group of proteins that are expressed in neonatal submandibular gland, but only transiently in sublingual serous demilunes. The results also suggest that mucous and serous cells follow separate pathways of differentiation, and that the cells containing mixed granules are developing mucous cells. In contrast to the submandibular and parotid glands, the phenotypes of the adult sublingual gland cells are established during fetal or early postnatal development. Among the questions that remain are those concerning the nature of the factors that control differentiation and are involved in regulating secretory protein expression.
The expert technical assistance of Ms. Mary W. Goss is gratefully acknowledged.