The Structure of Tight Junctions in Mouse Submandibular Gland


  • Kenichiro Kikuchi,

    1. Department of Histology, Nippon Dental University, Tokyo, Japan
    Search for more papers by this author
  • Jitesh Kawedia,

    1. Department of Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati College of Medicine, Cincinnati, Ohio
    Current affiliation:
    1. Pharmaceutical Science Department, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Anil G. Menon,

    1. Department of Molecular Genetics, Biochemistry and Microbiology, University of Cincinnati College of Medicine, Cincinnati, Ohio
    Search for more papers by this author
  • Arthur R. Hand

    Corresponding author
    1. Department of Craniofacial Sciences, University of Connecticut Health Center, Farmington, Connecticut
    2. Department of Cell Biology, University of Connecticut Health Center, Farmington, Connecticut
    • Department of Craniofacial Sciences, University of Connecticut Health Center, 263 Farmington Ave., Farmington, CT 06030-1610
    Search for more papers by this author
    • Fax: 860-679-4078


Salivary gland cells are joined by junctional complexes consisting of a tight junction (TJ), zonula adherens and one or more desmosomes. TJs regulate paracellular permeability, maintain separate apical and basolateral membrane domains, and serve as signaling centers. We examined TJs of mouse submandibular glands (SMG) in thin sections and freeze-fracture replicas. TJs between acinar cells and between intercalated duct cells had 2–6 parallel strands on the protoplasmic fracture face, with occasional branches, interconnections and free ends, and corresponding grooves on the extracellular face. Granular duct cell TJs had 2–30 strands, a depth of ≤0.5 μm, and occasional loops extending further basally. Where 3 or 4 cells met, the TJs extended basally ≤1 μm and consisted of 2 parallel boundary strands into which the apical strands inserted. Quantitative analyses showed significant differences in TJ complexity, measured by fractal geometry, and strand number of acinar compared to granular duct cells, and a greater number of strands in male compared to female granular ducts. Pilocarpine stimulation increased TJ strand number in female acinar cells, and increased complexity of male granular duct cell TJs. As the salivary gland water channel aquaporin 5 (AQP5) has been proposed to functionally interact with TJs to regulate salivary fluid composition, we also studied glands from AQP5 knock-out mice. In males lacking AQP5, granular duct TJs were more complex than those of wild-type mice, and exhibited more strands following pilocarpine stimulation. The results demonstrate specific gender, cell type and genetic differences in TJ structure and response to stimulation. Anat Rec, 2010. © 2009 Wiley-Liss, Inc.


Salivary glands secrete proteins, glycoproteins, water, and electrolytes to produce saliva. Protein and glycoprotein secretion occurs by the process of exocytosis, during which the contents of stored secretory granules are released, via membrane fusion and fission events, into the lumen of the acinus or secretory endpiece. Water secretion depends upon the establishment, by the acinar cells, of an osmotic gradient between the lumen and intercellular space, which draws water into the lumen. The initial, or primary, saliva produced by the acinar cells is isosmotic with plasma. As the saliva moves through the duct system toward the oral cavity, Na+ and Cl- are reabsorbed by the striated and excretory ducts, resulting in hypotonic final saliva.

Water, electrolytes, and other small molecules may enter the acinar lumen via transcellular or paracellular pathways. Aquaporin 5 (AQP5), located in the apical membrane, is the major transcellular water transporter in salivary gland acinar cells (Krane et al.,2001). In Aqp5-deficient (Aqp5−/−) mice the volume of saliva secreted by the salivary glands is significantly reduced (Ma et al.,1999; Krane et al.,2001). Paracellular fluid flow is regulated by tight junctions (TJs), or zonulae occludentes (Schneeberger and Lynch,2004; Van Itallie and Anderson,2006). In salivary glands, as in other epithelial tissues, the TJs separate the apical plasma membrane domain and luminal space from the lateral plasma membrane domain and intercellular space. In addition to their role in regulating paracellular fluid and electrolyte movement, TJs create a barrier to the free diffusion of plasma membrane components, maintaining the identity of the apical and lateral membrane domains. They also serve as important signaling sites for cell proliferation and differentiation, and may be targets for various infectious and pathological processes (Schneeberger and Lynch,2004; Matter et al.,2005).

In thin sections observed in the transmission electron microscope (TEM), TJs are seen as very close appositions or apparent punctate fusions of the cell membrane at the apical or luminal extent of the lateral cell membranes. In freeze-fracture replicas, TJs appear as branching and interconnecting strands on the protoplasmic face (P-face), and as corresponding grooves on the extracellular face (E-face; Goodenough and Revel,1970). The number of TJ strands varies in different cells and tissues; a general correlation exists between the number of TJ strands and the “tightness” of the junction, that is, the extent to which water and small molecules may pass through the junction (Claude and Goodenough,1973). However, several exceptions have been reported (e.g., Stevenson et al.,1988), and it is known that the molecular composition of the TJs also is important in determining junctional permeability (Van Itallie et al.,2003; Schneeberger and Lynch,2004; Van Itallie and Anderson,2006). The degree of branching and interconnection of the TJ strands, or “complexity,” also may vary in different tissues and in different physiological conditions (Kniesel et al.,1994,1996).

Autonomic agonists and parasympathetic nerve stimulation have been shown to alter the permeability of TJs in salivary glands (Garrett et al.,1982; Mazariegos et al.,1984; Kawabe and Takai,1990; Segawa,1994; Takai et al.,1995). Our recent studies, in concordance with Hill's hypothesis, also suggest that the functions of the paracellular and transcellular pathways are linked (Hill and Shachar-Hill,2002; Kawedia et al.,2007). In Aqp5−/− mice, paracellular fluid flow is decreased by 50%. However, the mechanisms that coordinate fluid secretion via AQP5 water channels in the apical membrane and via the TJs are unknown.

This study was designed to examine the ultrastructure of TJs in the mouse submandibular gland (SMG), using thin sections and freeze-fracture replicas. We also compared TJ structure in the SMG of Aqp5−/− mice with that of wild type mice, and assessed the structural changes in TJs after stimulation of fluid secretion by the parasympathomimetic agonist, pilocarpine. We show that there are significant differences in TJ structure in different segments of the gland and between male and female mice, and that there are specific differences between wild type and Aqp5−/− mice and their response to pilocarpine stimulation.

Abbreviations used: D = desmosome; E = E-face; Lu = lumen; P = P-face; TJ = tight junction; ZA = zonula adherens.



Age and sex-matched 5-month-old wild type (+/+) and Aqp5−/− (knock-out) littermates from Aqp5 recombinant inbred 129SvJ/Black Swiss (line 187) were used in all experiments. Mice null for Aqp5 were generated as previously described (Krane et al.,2001). Animal husbandry followed the NIH Guidelines for the Care and Use of Laboratory Animals. Mice were bred and housed in a pathogen-free environment in the animal care facilities of the University of Cincinnati College of Medicine. The mice were provided food and water ad libitum, and room lights were on a 12-hr light and dark cycle. The six male and six female mice used in this study were shipped to the Center for Laboratory Animal Care at the University of Connecticut Health Center, and used within 5 days of arrival. All experiments involving the use of live mice were approved by the Institutional Animal Care and Use Committee review board of the University of Cincinnati College of Medicine, and the Animal Care Committee of the University of Connecticut Health Center.

Tissue Fixation

The SMG were fixed by vascular perfusion of anesthetized animals (Ketamine/Xylazine, 100 mg/10 mg per kg body weight) with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4. Salivary secretion was stimulated in some animals by pilocarpine-HCl (Sigma-Aldrich, St. Louis, MO; 10 mg/kg, i.p.), 30 min before perfusion. The fixed glands were excised and placed in fresh fixative solution for 2–4 hr, then rinsed and stored in 0.1 M cacodylate buffer at 4°C.

Thin Section Preparation

For routine thin section analysis, the SMG were cut into small pieces, ∼1 mm on a side, postfixed in 1% osmium tetroxide and 0.8% potassium ferricyanide in 0.1 M cacodylate buffer, rinsed in distilled water, and treated with 1% aqueous uranyl acetate. The samples were dehydrated in a graded ethanol series, substituted with propylene oxide, infiltrated and embedded in PolyBed epoxy resin (Polysciences, Warrington, PA). Thin sections were cut using a Reichert Ultracut E ultramicrotome, collected on 200 mesh copper/rhodium grids, stained with uranyl acetate and lead citrate, and examined in a JEOL 100CX or Philips CM10 TEM.


Small pieces of glutaraldehyde-fixed SMG were cryoprotected with 30% glycerol, placed in gold specimen carriers, and rapidly frozen by immersion in liquid nitrogen-cooled propane. The samples were stored at liquid nitrogen temperature until they could be fractured. Fracturing and replication with platinum and carbon was done in a Balzers 400D freeze-fracture machine (Reading, UK). The replicas were cleaned in bleach and acid, rinsed in distilled water, collected on Formvar coated grids, and examined in the TEM. The TEM negatives were scanned at 1,200 ppi using an Agfa Duoscan or Epson Perfection V750 Pro scanner, and levels and contrast were adjusted in Adobe Photoshop.

Quantitative Analyses

The mean number of junctional strands (Stevenson et al.,1988) and the complexity (Kniesel et al.,1994) of the junctions were analyzed to assess differences in TJ structure between acinar cells and granular duct cells, between males and females, between wild type and Aqp5−/− mice, and between unstimulated and pilocarpine stimulated mice. The number of junctional strands intersecting a line approximately perpendicular to the luminal surface was counted at roughly 0.5-μm intervals along the TJ. The complexity of the TJs was determined by fractal geometry. The magnification of each image was adjusted to 100,000× in Photoshop, and a series of grids forming box sizes of 0.2, 0.1, 0.05, 0.025, and 0.0125 μm2 were superimposed over the image. The number of boxes of each size containing elements of the TJ were counted, and the fractal dimension was calculated as the slope of the linear regression of log(N)/log(1/S) for each box size, where N = number of boxes containing TJ elements and S = area of the box. The significance of the results was assessed by one-way ANOVA and two-tailed t-tests using the statistical functions in Microsoft Excel.


In the SMG, as in other salivary glands, TJs join adjacent cells and form a junctional complex with belt-like adherens junctions, or zonulae adherentes, and spot-like desmosomes, or maculae adherentes, at the apical extent of the lateral cell membrane (Fig. 1). Junctional complexes also form a boundary between the intercellular space and intercellular canaliculi, extensions of the lumen along the lateral cell membrane. In thin sections through acini and intercalated ducts, the tight junctions typically are 0.1–0.15 μm in depth (Fig. 1A), whereas in granular and striated ducts, the TJs extend much further basally, often more than 0.5 μm (Fig. 1B). In the acini and all portions of the duct system, a zonulae adherens is located just basal to the TJ, and one or more desmosomes may be located basal to the zonulae adherens. In the TJs, the interactions of claudins, occludin and other transmembrane molecules obliterate the intercellular space at the points of contact, and microfilaments may be present in the adjacent cytoplasm (Schneeberger and Lynch,2004). In the zonulae adherentes, the transmembrane proteins E-cadherin and nectin mediate cell-cell adhesion (Niessen and Gottardi,2008); the intercellular space is 10–15 nm wide, and microfilaments associated with the cytoplasmic face of the junction extend into the apical terminal web. The intercellular space in the desmosome is ∼20 nm wide, and usually exhibits a dense central line, thought to represent the interacting domains of the cadherin family members, desmogleins and desmocollins (Dusek et al.,2007). Plakoglobins, plakophilins and desmoplakins are present in the dense cytoplasmic plaque of the desmosome, and intermediate filaments are anchored to the plaques.

Figure 1.

Thin section TEM of wild-type male mouse SMG TJs. A: Intercellular canaliculus between two acinar cells. B: Junctional complex between two granular duct cells. The depth of the TJ is ∼0.75 μm. TJ, tight junction; ZA, zonula adherens; D, desmosome; Lu, lumen. Scale bars, main figures = 0.5 μm; insets = 0.1 μm.

In freeze-fracture replicas, identification of the different cell types was achieved by application of specific morphologic criteria. Acinar cells had numerous large apical granules that tended to cross-fracture, and relatively small luminal profiles. Occasionally, it was difficult to differentiate intercalated duct cells from acinar cells. Observation of a long, straight luminal profile with a slightly greater diameter than typical acinar lumina was taken as indicative of an intercalated duct. Granular duct cells contained granules that tended to fracture in the plane of their lipid bilayer membrane, numerous basally located mitochondria and infolded cell membranes, and a large lumen. Mainly due to the influence of androgens (Gresik,1994), the granular ducts are much more highly developed in male mice than in female mice. Thus, fractures through granular duct cells were observed more commonly in male glands, whereas fractures of acinar cells were more common in female glands.

In freeze-fracture replicas, acinar cell TJs typically were comprised of 2–6 P-face strands or corresponding E-face grooves (Fig. 2). The strands and grooves generally paralleled the luminal surface and each other, but showed occasional branches and interconnections, as well as free ends that extended basally, away from the lumen. Desmosomes were recognized as clusters of particles, most prominent on the P-face, located on the lateral cell membranes basal to the TJs (Figs. 2A, 3). Tight junctions between intercalated duct cells (Fig. 3) were very similar in structure to those between acinar cells. Preliminary analyses of TJ strand number and complexity showed no differences between acinar and intercalated duct cells. Therefore, the data from these two cell types were combined for the quantitative analyses shown in Tables 1 and 2.

Figure 2.

Freeze-fracture TEM of wild-type female acinar cell TJs. A: Unstimulated; intercellular canaliculus. Three to four strands with occasional perpendicular extensions (arrowheads) are seen on the P-face (P). D, desmosome. B: Pilocarpine stimulated; acinar lumen. Three to five E-face (E) grooves, with a few perpendicular extensions (arrowheads), and P-face strands separate the particle-rich luminal (Lu) surface from the lateral membrane. Scale bars = 0.25 μm.

Figure 3.

Freeze-fracture TEM of wild-type female intercalated duct. Tight junction structure is very similar to that of the acinar cells. Lu, lumen; D, desmosome. Scale bar = 0.25 μm.

Table 1. Number of SMG tight junction strands
 Mean number of TJ strands ± SE (n)
  1. +/+, wild type; −/−, Aqp5−/−.

  2. Values with the same symbol are significantly different: §,*,‡P < 0.001; P < 0.01; P < 0.05.

Acinar/intercalated duct cells
 +/+ male3.86 ± 0.22 (22)*
 +/+ female4.07 ± 0.15 (143)†,‡
 +/+ female + pilocarpine4.80 ± 0.19 (30)
Granular duct cells
 +/+ male10.43 ± 0.36 (82)§,*
 +/+ female7.44 ± 0.36 (34)§,‡
 +/+ male + pilocarpine10.06 ± 0.43 (63)
 +/+ female + pilocarpine8.25 ± 0.61 (16)
 −/− male10.31 ± 0.60 (48)
 −/− male + pilocarpine12.19 ± 0.82 (36)
Table 2. Fractal analysis of SMG tight junctions
 Mean fractal dimension ± SE (n)
  1. +/+, wild type; −/−, Aqp5−/−.

  2. Values with the same symbol are significantly different: †,‡P < 0.01; *P < 0.05.

Acinar/intercalated duct cells
 +/+ female−0.582 ± 0.017 (29)*
 +/+ male−0.595 ± 0.030 (8)
 +/+ (male + female)−0.585 ± 0.015 (37)
 +/+ (male + female) + pilocarpine−0.607 ± 0.036 (14)
 −/− (male + female) + pilocarpine−0.679 ± 0.023 (6)
Granular duct cells
 +/+ female−0.650 ± 0.023 (16)*
 +/+ female + pilocarpine−0.634 ± 0.030 (9)
 +/+ male−0.642 ± 0.018 (35)‡,§
 +/+ (male + female)−0.644 ± 0.014 (51)
 −/− male−0.713 ± 0.014 (22)
 +/+ male + pilocarpine−0.694 ± 0.016 (21)§
 −/− male + pilocarpine−0.718 ± 0.029 (16)

There was no difference in the mean number of TJ strands between wild-type male (3.86 ± 0.22) and female (4.07 ± 0.15) acinar/intercalated duct cells (Table 1). Pilocarpine stimulation resulted in a small but significant increase in the number of acinar TJ strands in female mice (Table 1). However, there were no obvious differences between wild type, Aqp5−/− or unstimulated and pilocarpine stimulated male and female mice in the arrangement (Fig. 2) or complexity of the acinar TJs (Table 2).

Tight junctions between granular duct cells (Fig. 4) were much more elaborate than those between acinar cells or intercalated duct cells, exhibiting a greater number of strands, extending further basally, and often showing complex configurations. At the borders where three or four cells came together and were joined by TJs, the P-face exhibited a ridge with two closely adjacent, parallel, “boundary” strands extending basally up to about 1 μm. Strands separating the lateral and luminal membrane domains inserted into the boundary strands, whereas other strands originated from and inserted back into the boundary strands (Fig. 4B,C). Similar, although less elaborate, configurations were occasionally seen where three or four intercalated duct cells shared a common TJ (not shown). The mean number of strands in granular duct cell TJs was significantly greater than in acinar/intercalated duct TJs for both males (10.43 ± 0.36 vs. 3.86 ± 0.22) and females (7.44 ± 0.36 vs. 4.07 ± 0.15) (Table 1). The mean number of strands between male granular duct cells was significantly greater than between female granular duct cells (Table 1).

Figure 4.

Freeze-fracture TEM of wild-type granular duct TJs. A: Unstimulated male. The TJs consist of numerous, anastomosing P-face strands extending basally up to 0.5 μm in depth. B: Unstimulated female. Where three or four cells join (arrowheads), the TJs exhibit an elaborate structure extending basally more than 1 μm. C: Pilocarpine stimulated male. Higher magnification of a TJ between three or four cells; the lumen is at the right. Two parallel strands extend along the boundary between the cells. Strands separating the lateral and luminal membrane domains curve basally near the boundary and insert almost perpendicularly into the parallel boundary strands. Other strands originate from the boundary strands, loop out in the lateral membrane, and insert back into the boundary strands. D, E: Pilocarpine stimulated male. In some cases, the TJ strands form very elaborate, looping configurations, extending basally up to 4 μm or more. Lu, lumen. Scale bars, A and C = 0.25 μm; B, D and E = 0.5 μm.

The complexity of TJ structure was assessed using fractal geometry (Table 2). Comparison of male vs. female acinar/intercalated duct TJs, or male vs. female granular duct TJs, revealed no significant differences in complexity. The complexity of female granular duct TJs was significantly greater than that of female acinar/intercalated duct TJs (Table 2). However, the complexity of male granular duct TJs was not significantly different from that of male acinar/intercalated duct TJs. This may be due to the small number of observations for male acinar cells. When the data for males and females were combined, a significant difference was seen between granular duct and acinar/intercalated duct TJs.

Following pilocarpine stimulation, granular duct cells exhibited numerous looping TJ strands that extended basally up to 4 μm or more (Fig. 4D,E). Quantitative analyses revealed a significant increase in complexity of male granular duct TJs as a result of pilocarpine stimulation of wild type mice (Table 2). However, there was no difference in the number of granular duct TJ strands between unstimulated and pilocarpine stimulated animals.

The complexity of granular duct TJs was significantly greater in Aqp5−/− mice than in wild type mice (Table 2). Although the number of strands between granular duct cells after pilocarpine stimulation was greater in Aqp5−/− mice than in wild type mice (Table 1), there was no difference in complexity of the TJs following pilocarpine stimulation (Table 2).

Occasional gap junctions with P-face particles and E-face pits were observed on the lateral surfaces of the acinar and intercalated duct cells (Fig. 5). Gap junctions were not observed between granular duct cells.

Figure 5.

Freeze-fracture TEM of wild-type female intercalated duct cell (A) and acinar cell (B). Gap junctions (arrowheads), with P-face (P) particles and E-face (E) pits are present on the lateral surfaces of the cells. Scale bars = 0.25 μm.


The TJs between acinar (and intercalated duct) cells of the mouse SMG typically consist of 2–6 P-face strands and corresponding E-face grooves, with occasional branches, interconnecting strands and free ends. Their overall structure is very similar to that previously reported for acinar and intercalated duct TJs in the rat SMG (Inoue et al.,1987; Hashimoto et al.,2003) and in other salivary glands [e.g., rat parotid (DeCamilli et al.,1976; Mazariegos et al.,1984; Simson and Bank,1984) and mouse parotid (Kawedia et al.,2007)]. The acinar TJs also are similar to those described for rat sublingual gland mucous and serous endpiece cells (Shimono et al.,1980), however the intercalated ducts of the sublingual gland were reported to have a greater number of strands than those of the SMG. Although the SMG acinar cells of male and female mice exhibit significant sexual dimorphism (Jayasinghe et al.,1990; Menghi et al.,1998; Señorale-Pose et al.,1998), the quantitative analyses indicate that there are no differences between males and females in regard to the number of acinar TJ strands or complexity.

A freeze-fracture analysis of rodent granular duct TJs has not been reported previously. The present study shows that the TJs between granular duct cells of the mouse SMG are more highly developed than those between acinar cells. The number of junctional strands in mouse granular duct TJs varies considerably, from as few as 2 to nearly 30, and the pattern of branches and interconnections may be quite complex. The depth of the junction along the lateral cell surface may reach 0.5 μm or more, and looping strands often extend even further basally. Quantitative analyses show that the number of strands present in granular duct TJs is significantly greater than in acinar TJs of both males and females. The complexity of TJs, as measured by fractal geometry, also is significantly greater for granular ducts than acini in females, and for males and females combined. The complexity of granular duct and acinar TJs in males was not significantly different, however, possibly due to the relatively few observations for male acinar cells. Consistent with the well-known differences in granular duct development in male and female mice (Gresik,1994), the number of TJ strands was significantly greater in males.

In regions where 3 or 4 granular duct cells meet and contribute to the TJ, junctional strands form a complex structure consisting of 2 closely parallel P-face strands, into which the strands separating the luminal and lateral membrane domains insert. Additional strands originate from these parallel “boundary” strands, loop out into the lateral membrane, and reinsert on the boundary strands. These elaborate boundary structures extend up to 1 μm basally from the luminal surface, and presumably provide increased sealing capacity and resistance to mechanical stress at vulnerable sites where several cells meet. We observed similar structures in intercalated ducts of the SMG, and they probably occur as well in striated ducts [e.g., see Fig. 7 in DeCamilli et al.,1976 and Fig. 12 in Simson and Bank,1984).

In many tissues, the number of TJ strands or TJ complexity is correlated with junctional “tightness” (Claude and Goodenough,1973; Humbert et al.,1976). In salivary glands, acini and intercalated ducts are considered to have “leaky” junctions, whereas striated and excretory ducts have “tight” junctions (Shimono et al.,1980; Simson and Bank,1984). The greater number of TJ strands and complexity of granular duct TJs suggest that these junctions are tighter, or less permeable, than acinar cell TJs. Granular ducts differentiate from the proximal portion of striated ducts, and the structure of their TJs is consistent with that shown in the few previously published images of striated ducts (DeCamilli et al.,1976; Shimono et al.,1980; Mazariegos et al.,1984; Simson and Bank,1984). In general, the structure of acinar and duct TJs seen in this and previous studies is consistent with interpretations of data from physiological studies of salivary glands (Schneyer et al.,1972; Augustus et al.,1978).

On the other hand, recent studies have emphasized the role of the molecular components of TJs in establishing and regulating junctional permeability. Claudins are transmembrane proteins that bind either homotypically or heterotypically with their counterparts in the adjacent cell to create the tight junctional seal (Schneeberger and Lynch,2004; Van Itallie and Anderson,2006). Claudins appear to be the main component of the intramembranous strands observed in freeze-fracture electron micrographs. The claudin protein family comprises 24 members, which exhibit different permeability properties, mainly through regulation of paracellular selectivity to small ions, especially cations (Angelow et al.,2008). Claudins-1, -3, -4 and -7 have been detected in rodent salivary glands by immunological and molecular methods (Peppi and Ghabriel,2004; Kawedia et al.,2007); claudin-3 is present mainly in acini and intercalated ducts, whereas claudin-4 is located in striated and excretory ducts. Expression of claudin-4 in MDCK II cells increases transepithelial electrical resistance and decreases Na+ permeability (Van Itallie et al.,2003); these properties are consistent with the localization of claudin-4 and the physiological characteristics of striated and excretory ducts.

Several studies have shown increased permeability and altered structure of salivary gland TJs following secretory stimulation, either by treatment with sympathomimetic and parasympathomimetic drugs (Mazariegos et al.,1984; Segawa,1994; Hashimoto et al.,2003) or by autonomic nerve stimulation (Garrett et al.,1982; Kawabe and Takai,1990; Takai et al.,1995). Pilocarpine stimulation of the SMG resulted in an increased number of acinar TJ strands in female mice; the number of observations of acini in stimulated males was insufficient to determine if changes in strand number had occurred. In the parotid glands of the same animals, pilocarpine stimulation resulted in an increase in acinar TJ strand number in males, but not in females (Kawedia et al.,2007) suggesting that there may be both gland-specific and gender-specific differences in the structure and composition of TJs. Isoproterenol stimulation also caused an increase in the number of TJ strands between rat SMG acinar cells (Inoue et al.,1987). Although the number of TJ granular duct strands was not increased by pilocarpine stimulation, the complexity of these TJs was significantly increased, including extension of the junctional strands further basally and the formation of numerous loops. Tight junction loops and free ends also were described during the early developmental of the rat SMG, when significant changes occur in TJ structure (Shimono et al.,1981). It has been reported that the function of claudins, especially their permeability properties, is altered due to phosphorylation by protein kinases A and C and other kinases (Angelow et al.,2008), which may occur as a result of autonomic stimulation of salivary glands. These observations suggest that tight junction structure, function and, possibly, composition in the granular duct cells may be regulated via post-translational modification.

Our previous study of the mouse parotid gland suggested that AQP5 and TJ components may communicate to control fluid secretion pathways (Kawedia et al.,2007). In vivo estimates of paracellular transport indicated decreased paracellular fluid flow in Aqp5−/− mice. Additionally, the expression of claudin-7 and occludin was decreased by about 50% in both males and females, and a similar decrease in claudin-3 expression occurred in females. Whether other salivary gland claudins also are affected in Aqp5−/− mice is unknown, as are the signaling pathways responsible for these effects. The increased complexity of granular duct TJ strands in Aqp5−/− mice, and increased number of strands after pilocarpine stimulation, suggest that AQP5 also may modulate TJ structure in these cells.

Gap junctions are present between acinar cells and between intercalated duct cells in the mouse SMG, but not between granular duct cells. Their appearance in freeze-fracture replicas is similar to that described for gap junctions in numerous other animal tissues, including salivary glands (Larsen,1983; Shimono et al.,1992). Gap junctions were previously identified in thin sections of mouse SMG acini by Yohro (1971), and in thin sections, freeze-fracture replicas, and immunocytochemical studies of connexin expression in the rat SMG (Inoue et al.,1987; Shimono et al.,1992; Muramatsu et al.,1996). As in the present study, gap junctions were not observed in the granular, striated and excretory ducts of the rat SMG (Muramatsu et al.,1996). The ionic and metabolic communication afforded by gap junctions suggests that individual acini, and perhaps clusters of acini connected by branches of the same intercalated duct, function as a common unit in response to secretory stimulation.

In summary, the present results emphasize the significant differences in TJ structure between acinar cells and granular duct cells in the mouse SMG, and demonstrate that the sexual dimorphism in granular duct structure and function includes the TJs. The results also show that TJ structure may be altered by secretory stimulation with the parasympathomimetic agonist pilocarpine, and that the water channel protein AQP5 influences TJ structure and response to pilocarpine stimulation. Finally, a clear description is provided of the elaborate “boundary” structure of TJs where 3 or 4 cells come together. Taken together these results suggest that the TJs play a significant role in the structural organization and physiological functions of the salivary glands.


The authors thank Dr. John Aghajanian for his advice, Ms. Maya Yankova for expert technical support and Ms. Kareen Elder for assistance in data analysis. They are grateful to Mr. Dale Callaham and the Central Microscopy Facility, Department of Biology, University of Massachusetts, Amherst, for the use of their laboratory facilities and freeze-fracture equipment.