Strain-Specific and Endocrine Control of Granular Convoluted Tubule Cells and Epidermal Growth Factor Expression in the Mouse Submandibular Gland.код для вставкиСкачать
THE ANATOMICAL RECORD 291:105–113 (2007) Strain-Speciﬁc and Endocrine Control of Granular Convoluted Tubule Cells and Epidermal Growth Factor Expression in the Mouse Submandibular Gland YUICHI MIYAJI, SHIGEO AIYAMA, AND SHINGO KURABUCHI* Department of Histology, The Nippon Dental University, School of Life Dentistry at Tokyo, Tokyo, Japan ABSTRACT Submandibular glands (SMGs) of 11-week-old mice from four strains, ICR, C57BL/6J, BALB/c, and C3H/HeN were examined by immunohistochemistry for epidermal growth factor (EGF). In addition to sex-related differences in granular convoluted tubules (GCTs), the GCT cells were signiﬁcantly larger in ICR mice than in other three strains. In males from each of the strains, almost all the GCT cells were strongly positive for EGF. The EGF-positive cells in the females, however, were markedly fewer in number, and were stained weaker in C57BL/6J, BALB/c, and C3H/HeN mice than in ICR mice. The GCT cells and their EGF expression in the F1 progeny from ICR and C3H/HeN strains were approximately intermediate between those of the parent strains of the same sex. T3 and/or dihydrotestosterone (DHT) enhanced the GCT phenotype in the C3H/HeN mice, and remarkably increased the EGF-positive cells in females. Electron microscopy revealed that gold-labeling of EGF was conﬁned to the secretory granules, and that the GCT cells in females, given T3 1 DHT, had a well-developed Golgi apparatus and net-like RER but few basal infoldings, whereas the equivalent cells in the untreated females had poor RER and prominent basal infoldings. These results suggest that the EGF concentration in SMGs is genetically high in ICR mice and low in other strain mice and that, considering the same response of GCT cells to T3 and/or DHT between the high and low EGF strains, the low EGF concentrations might be partly caused by a lesser sensitivity of the GCT cells to thyroid hormones. Anat Rec, 291:105– 113, 2007. Ó 2007 Wiley-Liss, Inc. Key words: epidermal growth factor (EGF); submandibular gland (SMG); granular convoluted tubule (GCT); thyroid hormone; mouse strain The granular convoluted tubule (GCT) cells in the submandibular glands (SMGs) of mice have been studied extensively in terms of morphological sexual dimorphism and their stored bioactive secretory products, which include growth factors, renin, and members of the kallikrein gene family (reviewed by Chrétien, 1977; Barka, 1980; Gresik, 1994; Kurabuchi and Hosoi, 2001). Several studies, many of which used ICR (CD-1) mice, have shown that GCT cells are under the multihormonal control of androgens, thyroid hormones, and adrenocortical Ó 2007 WILEY-LISS, INC. *Correspondence to: Shingo Kurabuchi, Department of Histology, The Nippon Dental University, School of Life Dentistry at Tokyo, Fujimi 1-9-20, Chiyoda-ku, Tokyo 102-8159, Japan. Fax: 81-3-3264-8399. E-mail: email@example.com Received 13 July 2007; Accepted 29 September 2007 DOI 10.1002/ar.20617 Published online in Wiley InterScience (www.interscience.wiley. com). 106 MIYAJI ET AL. hormones. Generally, these hormones act as strong promoters of the GCT phenotype and act as strong inducers of GCT-speciﬁc secretory products. Consequently, each GCT cell in males is large and contains abundant large secretory granules, whereas each GCT cell in females is small and contains fewer and smaller secretory granules than those observed in males, thus presenting a sexually dimorphic pattern. In accordance with the morphological differences of GCT cells according to sex, the concentration of many GCT-speciﬁc secretory products, including growth factors, renin and kallikrein gene family members (with the exception of a true tissue kallikrein, mK1) are also higher in the glands of males than in those of females. In addition, growth hormone has been recently suggested to up-regulate epidermal growth factor (EGF) expression and the granular cell phenotype in rodent salivary glands (Hiramatsu et al., 1994; Young et al., 2004). Variations in mouse strains are an important topic of study in this ﬁeld. SMG renin, but not kidney renin, is well known to be under genetic as well as hormonal control (Bing and Poulsen, 1971; Wilson et al., 1977; Bing et al., 1980), and inbred strains of mice have been divided into two groups according to whether they exhibited a low or high basal activity of SMG renin (Wilson et al., 1981). The SMG of C57BL/6J mice, which generally exhibit a low renin activity (Wilson et al., 1981), showed no speciﬁc staining for renin (Tanaka et al., 1980). The concentration of mK1 (recently renamed proteinase F) was higher in the SMGs of males than in those of females among the strains of mice with a low renin level, including BALB/CA, C3H/HeN, C57BL/10N, and DBA/2N (Murai, 1987). In contrast, the mK1 concentration in an outbred ICR strain, which was classiﬁed into the high renin group, was higher in the SMGs of females than in males (Hosoi et al., 1983). Furthermore, the response of mK1 expression to androgen was suggested to be opposite between ICR and C3H/HeN females, and such a difference may be caused by the lesser sensitivity of C3H/HeN females to thyroid hormone (Kurihara et al., 1999). To our knowledge, the renin regulator (Rnr) gene and the synthesis of GCT-speciﬁc secretory products are unrelated, other than the involvement of renin. However, if the low renin levels are due to a low sensitivity to hormones in the low renin strains, the concentration of many secretory products, such as EGF studied herein, should differ between the low and high renin strains. On the basis of these ﬁndings, we examined the localization and distribution of EGF and the GCT phenotype in the SMGs of four mice strains, C57BL/ 6J, BALB/c, C3H/HeN, and ICR, using high-resolution immunocytochemical techniques for light and electron microscopy. Next, to conﬁrm whether EGF expression is under genetic control, the SMGs of F1 mice (the offsprings of ICR males 3 C3H/HeN females and of C3H/HeN males 3 ICR females) were also examined. Finally, the effect of thyroid hormone and androgen on the GCT phenotype and EGF expression in GCT cells was also examined in C3H/ HeN mice, and the results were compared with well-documented observations made in ICR mice. MATERIALS AND METHODS Animals and Experimental Procedures All animal experiments were conducted in compliance with the Nippon Dental University and the National Institutes of Health (NIH) Guidelines for the Care and Use of Laboratory Animals. Animals were housed under controlled environmental conditions (238C, 12-hr lightdark cycle) and were provided access to food and water ad libitum. Male and female mice of four strains, C57BL/6J, BALB/c, C3H/HeN, and ICR (CD-1), were purchased from Japan Clea (Tokyo, Japan). To examine whether the GCT phenotype and EGF expression are genetically regulated, two types of F1 progeny were prepared by mating male ICR mice 3 female C3H/HeN mice, and male C3H/HeN mice 3 female ICR mice. Furthermore, to examine hormonal control, male and female C3H/HeN mice were injected subcutaneously with 3,30 , 5 triiodo-Lthyronine (T3; ICN Biochemicals Inc., Aurora, OH) and/ or 5a-dihydrotestostrone (DHT; Wako Ltd., Osaka, Japan) as described below. At least ﬁve animals of each sex, strain or F1, and injected hormone combination were prepared. T3 was dissolved in a small amount of 0.1 N NaOH and then diluted with Ringer’s solution, while 5a-dihydrotestostrone (DHT: Wako Ltd.) was dissolved in sesame oil; these hormones were then injected subcutaneously at the following doses per kilogram of body weight: T3, 1 mg; DHT, 20 mg. The injections were performed every other day for 2 weeks (7 injections) before killing. At 11 weeks of age, the male and female mice of the four strains, the two types of F1 mice, and the hormoneinjected mice were subjected to an overnight fast before killing. Under Nembutal anesthesia (1.7 mg/kg body weight, intraperitoneally), the animals were killed in the morning by exsanguination, and then, a cold ﬁxative solution consisting of a 2% glutaraldehyde and 2% paraformaldehyde mixture in a 0.05 M cacodylate buffer, pH 7. 4, was perfused through the vasculature. The SMGs were removed, minced, and immersed in the same ﬁxative solution for 5 hr at 48C. The SMG fragments were then dehydrated with a graded ethanol series and embedded in resin: either Technovit 8100 (Heraeus Kulzer GmbH, Wehrheim, Germany) for light microscopy, or LR-white (London Resin Co. Ltd, Berkshire, England) for transmission electron microscopy. Light Microscopic Immunohistochemistry Sections of 2-mm-thick Technovit-8100-embedded tissue were mounted on silane-coated slides (Matsunami Co., Osaka, Japan). The sections were incubated in a solution of 0.3% H2O2 in methanol for 30 min to inhibit endogenous peroxidase activity, and washed with distilled water (DW) and phosphate-buffered saline (PBS: 0.01 M sodium phosphate buffer and 0.14 M NaCl, pH 7.5). They were then sequentially immunostained using the avidin–biotin–peroxidase complex method in a moist chamber at room temperature with the following reagents according to the kit manufacturer’s protocols (Vectastain Elite ABC Kit; Vector Lab., Burlingame, CA): normal goat serum (NGS, Vectastain Kit) for 1 hr; rabbit anti-mouse EGF antiserum (diluted 1:3,000; Upstate Biotechnology, Lake Placid, NY) for overnight; biotinylated goat anti-rabbit IgG (Vectastain Kit) for 1 hr; and ﬁnally peroxidase-conjugated streptavidin (Vectastain Kit) for 2 hr. Peroxidase activity was detected using 20 mg of 3,30 -diaminobenzidine tetrahydrochloride (DAB, Dojin Lab., Kumamoto, Japan) and 0.01% H2O2 TUBULE CELLS OF MOUSE SUBMANDIBULAR GLAND 107 Fig. 2. A,B: Bar graphs showing the size (cellular height) of granular convoluted tubule (GCT) cells (A) and the percentage of epidermal growth factor (EGF) -positive GCT cells occupying the GCT segments (B) in intact male (dark bars) and female (white bars) mice of four strains: C57BL/6J, BALB/c, C3H/HeN, and ICR. All values are the mean 6 SD of ﬁve animals. *P < 0.001, signiﬁcantly different from the ICR strain of the same sex. each sex, mouse strain, F1 progeny, and hormoneinjected mice (n 5 5 for each group). The signiﬁcance of the observed differences (P < 0.01 P < 0.001) was determined using Student’s t-test. Immunogold Staining for Electron Microscopy Fig. 1. a–h: Nomarski image of 2-mm-thick sections of submandibular glands (SMGs) of intact male and female mice of C57BL/6J (a,b), BALB/c (c,d), C3H/HeN (e,f), and ICR (g,h) mice; immunostaining with anti–epidermal growth factor (EGF). Almost all the granular convoluted tubule (GCT) cells are positively stained for EGF in the glands of males (a,c,e,g), whereas some of the GCT cells are EGF-positive, but others are negative, producing a cellular mosaic expression pattern of EGF in the glands of females (b,d,f,h). The arrows show EGF-negative GCT cells, and the arrowheads show transition granular cells. DAB color was augmented using 0.5% osmium tetroxide in b, d, and f because the original immunoreactions were faint. Scale bar 5 50 mm. in 0.05 M Tris-HCl buffer, pH 7.6. As the peroxidase reaction was very weak in the sections from female C57BL/6J, BALB/c, and C3H/HeN mice, these sections were incubated with 0.5% osmium tetroxide in PBS for 3 min to emphasize the DAB color, according to the protocol described by Johansson and Backman (1983). Finally, the sections were lightly counterstained with Mayer’s hematoxylin, dehydrated with ethanol, mounted in Canada Balsam (Wako, Tokyo, Japan) and examined under a BX-50 microscope equipped with Nomarski differential interference-contrast optics (Olympus, Tokyo, Japan). In the immunostained sections, the distance (cell height) between the apical and basal edge of more than 1,000 GCT cells per animal was measured. In addition, the total number of GCT cells and the number of EGFimmunopositive GCT cells was counted, and the percentage of EGF-immunopositive cells in GCT segments were calculated per gland in each animal. The mean 6 standard deviation (SD) of these values were calculated for Ultrathin sections of LR-white-embedded tissue were cut using an ultramicrotome (Ultrotome V: LKB, Sweden) and mounted on gold grids. The sections on the grids were etched in 3% H2O2 in DW for 30 sec. After rinsing with DW, the sections were incubated with NGS (Vectastain Kit) for 1 hr, and transferred into a droplet of rabbit anti-mouse EGF antiserum (diluted 1:2,000) and incubated overnight at room temperature. After rinsing with PBS, the sections were incubated with biotinylated goat anti-rabbit IgG (Vectastain Kit) for 1 hr. Finally, the sections were incubated with streptavidin conjugated with 20 nm gold particles (diluted 1:600; ICN Biochemicals) for 2 hr and then rinsed with PBS and DW. The immunolabeled sections were stained with uranyl acetate; after washing the sections with DW and drying, they were examined in a 2000 EX-II transmission electron microscope (JEM-2000EXII, JEOL, Tokyo, Japan) at an accelerating voltage of 80 kv. RESULTS Intact Male and Female Mice From the Four Strains Under light microscopy, the GCT cells in the SMGs of males appeared larger than those of females in each strain (Fig. 1). The difference in cell size was due to the number of secretory granules, which were more abundant in large cells than in small cells. In addition to such sexual differences, the GCT cells in the ICR strain were larger than those in the C57BL/6J, BALB/c, and C3H/HeN strains, especially in females (Fig. 1b,d,f,h). As summarized in Figure 2A, the cellular height (which corresponds to the cellular size in this study) of the GCT cells was approximately one and a half times larger in both sexes of ICR mice than in the other three strains, and no differences were observed between the latter 108 MIYAJI ET AL. Fig. 3. a–d: Nomarski image of 2-mm-thick sections of submandibular glands (SMGs) from two types of F1 progeny bred by crossing C3H/HeN males 3 ICR females (a,b), and ICR males 3 C3H/HeN females (c,d); immunostaining with anti-EGF. The GCT cells are larger, the immunoreactions are stronger, and the EGF-positive GCT cells are more abundant, especially in the female F1 (b,d), compared with the female C3H/HeN mice shown in Figure 1f. The arrows show the EGFnegative GCT cells, and the arrowheads show the transition granular cells. Scale bar 5 50 mm. three strains. Speciﬁc immunostaining for EGF was detected exclusively in the secretory granules of the GCT cells in male and female mice from each strain. Nomarsky differential-interference contrast microscopy revealed that almost all of the GCT cells in the males of each strain were immunostained for EGF, and EGF-negative GCT cells were rarely seen (Fig. 1a,c,e,g). Only a few of the negatively stained GCT cells were typical GCT cells, and many of them were slender in shape and contained only a few small secretory granules. We previously named this cell type ‘‘slender cell’’ (Kurabuchi et al., 1999), and such cells have been characterized as a transition type of granular duct cell (Hazen-Martin and Simson, 1987). EGF-positive GCT cells accounted for more than 98% of the population of GCT cells in males from each strain (Fig. 2B). In females, on the other hand, EGF-negative GCT cells were conspicuously seen in each strain, and a cellular mosaic expression pattern of EGF was prominent (Fig. 1b,d,f,h). Additionally, the EGF-staining reaction was very weak in C57BL/6J, BALB/c, and C3H/HeN mice (Fig. 1b,d,f). In ICR females, approximately half of the total population of GCT cells was negatively stained for EGF and the other half was EGF-positive (Fig. 2B). Compared with ICR females, this percentage was lower in the females of the other three strains: the EGF-positive cells accounted for approximately 20% of the total GCT cell population in C3H/HeN mice, and for less than 10% in C57BL/6J and BALB/c mice (Fig. 2B). Congenic Mice From ICR and C3H/HeN Strains The size of the GCT cells and the localization and distribution of EGF in the SMGs was almost the same between ICR male 3 C3H/HeN female offspring and C3H/HeN male 3 ICR female offspring, when compared Fig. 4. A,B: Bar graphs showing the size (cellular height) of granular convoluted tubule (GCT) cells (A) and the percentage of epidermal growth factor (EGF) -positive GCT cells occupying the GCT segments (B) in two types of F1 progeny bred by crossing C3H/HeN males 3 ICR females, and ICR males 3 C3H/HeN females. The dark bars represent F1 males, and the white bars represent F1 females. All values are the mean 6 SD of ﬁve animals. *P < 0.001, signiﬁcantly different from the ICR strain of the same sex shown in Figure 2A, B. **P < 0.01, signiﬁcantly different from the female C3H/HeN mice shown in Fig. 2A. according to sex (Figs. 3, 4). The GCT cells in males of both types of F1 progeny (Fig. 3a,c) were similar in appearance to those in ICR males (Fig. 1g) and C3H/ HeN males (Fig. 1e). Many GCT cells were immunostained for EGF in the males of both types of F1 progeny, and a few EGF-negative cells were sometimes seen, some of which were transition-type cells (Fig. 3a,c). As shown in Figure 4A, the GCT cells in both types of F1 males were larger than those of C3H/HeN males but were smaller than those of ICR males. More than 98% of the population of GCT cells was immunostained for EGF in both types of F1 progeny (Fig. 4B), similar to the results seen in mice of the same sex but from the parent strains (Fig. 2B). In females, the GCT cells in both types of F1 females (Fig. 3b,d) were a little larger than those in C3H/HeN females (Fig. 1f) but were smaller than those in ICR females (Fig. 1h). Nomarsky differential-interference contrast microscopy revealed that the cellular mosaic distribution of EGF was prominent and that EGF immunostaining was more intensive in the F1 females (Fig. 3b,d) compared with the observations in the C3H/HeN females (Fig. 1f). Approximately 30% of the total GCT cells were EGF-positive cells in both types of F1 females (Fig. 4B); this percentage was lower than that seen in ICR females but higher than that seen in CH3/HeN females (approximately 50% and 20%, respectively; Fig. 2B). Repeated Injections of T3 and DHT in C3H/HeN Mice After seven repeated injections of T3 alone, DHT alone, or a combination of these two hormones in C3H/ HeN mice on alternating days for 2 weeks, the body TUBULE CELLS OF MOUSE SUBMANDIBULAR GLAND 109 Fig. 6. A,B: Bar graphs showing the size (cellular height) of granular convoluted tubule (GCT) cells (A) and the percentage of epidermal growth factor (EGF) -positive GCT cells occupying the GCT segments (B) in T3-injected, 5a-dihydrotestostrone (DHT) -injected, and T3 1 DHT injected C3H/HeN mice. The dark bars represent males, and the white bars represent females. All values are the mean 6 SD of ﬁve animals. *P < 0.01, signiﬁcantly different from control C3H/HeN mice (Fig. 2A,B) of the same sex. Fig. 5. a–f: Nomarski image of 2-mm-thick sections of submandibular glands (SMGs) of T3-injected (a,b), 5a-dihydrotestostrone (DHT) injected (c,d), and T31DHT-injected (e,f) C3H/HeN mice; immunostaining with anti–epidermal growth factor (EGF) granular convoluted tubule (GCT) antiserum. GCT cells and their secretory granules were larger in each of the hormone-injected groups, compared with those in control, untreated C3H/HeN mice (Fig. 1e,f) in both sexes. EGF-positive GCT cells were more abundant and the immunoreactions were stronger in each of the hormone-injected group of female mice (b,d,f), compared with those in the control female mice (Fig. 1f). Arrows show EGF-negative GCT cells. Scale bar 5 50 mm. weight of the mice showed no signiﬁcant differences among T3 alone, DHT alone, T3 1 DHT, and control groups: among male mice, the body weights were 22.6 6 1.6 g, 23.0 6 1.0 g, 23.2 6 1.2 g, and 22.2 6 1.4 g for respective groups; among the female mice, the body weights were 18.6 6 0.8 g, 17.4 6 1.4 g, 18.8 6 1.2 g, and 17.5 6 1.8 g for the respective groups. On the other hand, the GCT cells in each of the hormone-injected groups were signiﬁcantly larger and contained numerous large secretory granules (Fig. 5), compared with control male and female mice of the same strain (Fig. 1e,f). In males, the GCT cells were approximately one and a half times larger in mice given T3 alone and mice given DHT alone (Fig. 6A), compared with untreated C3H/ HeN males (Fig. 2A). The concomitant injection of both hormones increased the size of the GCT cells to nearly twice that of the control GCT cells (Figs. 2A, 6A). Nomarsky differential-interference contrast microscopy revealed that almost all of the GCT cells in the males of the hormone injected groups were immunostained for EGF, and EGF-negative GCT cells were rarely seen, similarly to the results seen in control males (Figs. 1e, 5a,c,e). EGF-positive GCT cells accounted for more than 98% of the population of GCT cells in males from each hormone-injected group (Fig. 6B) and control untreated males (Fig. 2B). In females, the GCT cells of the hormone-injected groups (Fig. 5b,d,f) were larger than those of control C3H/HeN females (Fig. 1f). The GCT cells were 1.5–2 times larger in the DHT alone group, and T3 1 DHT groups than in the control untreated females (Fig. 6A). The average of the cell size in the T3 alone group (Fig. 6A) was a little higher than that of the control (Fig. 2A), but the signiﬁcant difference was not obtained. The EGF immunoreactivity was stronger in each of the hormone-injected groups, and EGF-positive GCT cells were markedly more abundant (Fig. 5b,d,f), compared with observations in control untreated females (Fig. 1f). More than 80% of the total GCT cells were EGF-positive in each of the hormone-injected groups (approximately 80%, 84%, and 87% for the T3 alone group, the DHT alone group, and the T3 1 DHT group, respectively); these percentages were signiﬁcantly higher than that observed for control untreated females (Figs. 2B, 6B). Electron Microscopic Immunocytochemistry Based on the light microscopy ﬁndings, an ultrastructural analysis using immunogold staining was used to examine comparatively the SMGs of untreated C3H/ HeN females, and those of C3H/HeN females, concomitantly injected with T3 1 DHT. The GCT cells of the intact female mice contained electron-dense secretory granules, which were sparsely distributed in the apical cytoplasm (Figs. 7, 8). Golgi apparatuses with ﬂattened cisternae and several fragments of rough endoplasmic reticulum (RER) were poor, and numerous deep membrane infoldings associated with elongated mitochondria were usually present at the base (Fig. 8). Gold particles, indicating the presence of EGF, were restricted to the secretory granules in EGF-positive cells (Fig. 7). The secretory granules were generally larger in the EGFpositive cells, but they were sometimes the same in the EGF-positive and -negative cells (Fig. 7). On the other hand, the GCT cells in the glands of the DHT 1 T3 group were closely packed with abundant large electrondense secretory granules in the apical two-thirds of the 110 MIYAJI ET AL. Fig. 8. Electron micrograph of a cross section of a granular convoluted tubule (GCT) segment in the submandibular gland (SMG) of a control intact C3H/HeN female; immunostaining with anti–epidermal growth factor (EGF) antiserum. Two EGF-negative GCT cells are seen. Membrane infoldings are present in the base (arrowheads) of the cells and are associated with elongated mitochondria (M). BV, blood vessel; N, nucleus. Scale bar 5 2 mm. Fig. 7. Electron micrograph of a cross-section of a granular convoluted tubule (GCT) segment in the submandibular gland (SMG) of a control C3H/HeN female; immunostaining with anti–epidermal growth factor (EGF) antiserum. EGF-positive GCT cells (arrows) and EGF-negative GCT cells (an arrowhead) are visible. Almost all the secretory granules in the EGF-positive cells are labeled with numerous gold particles, indicating the presence of EGF. Few gold particles are seen on the secretory granules of the EGF-negative cells. L, lumen; N, nucleus. Scale bar 5 2 mm. cell and a round nucleus visible in the basal one-third (Fig. 9). There were, however, variations in the size of these secretory granules. The gold particles, indicating the presence of EGF, were restricted to these secretory granules and were labeled in many of the large secretory granules and in fewer small ones, producing an intergranular variation. Large Golgi apparatuses with swollen cisternae, and mesh-like cisternae on the RER were visible in the perinuclear region. A few short membrane infoldings were often observed in the basal membrane. DISCUSSION The results of this study suggest that GCT cells in the SMGs of ICR, C57BL/6J, BALB/c, and C3H/HeN mice exhibit a cellular mosaic expression pattern of EGF that is prominent in the females of each strain. Furthermore, some GCT cells were positively stained for EGF, but others were negatively stained, and both of these cells were ultrastructurally similar. The principal site of GCT-speciﬁc secretory products is the secretory granules of typical GCT cells (Bing et al., 1980; Watson et al., 1985), with the exception of a few dark cells, pillar cells, and transition type granular cells (reviewed by Mori et al., 1992; Gresik, 1994); the former two cell types are nongranulated, and the latter, which contain a few small secretory granules, may be in the process of differentiation into typical GCT cells (Hazen-Martin and Simson, 1987). The transition type cells are positively stained only for mK1 (Kurabuchi et al., 1999, 2001, 2002, 2004; Kurabuchi and Hosoi, 2001) and were conﬁrmed to be EGF-negative in the present study. Variations in EGF staining have been previously observed in SMG GCT cells from ICR and BKW strains (van Noorden et al., 1977; Gresik and Barka, 1977). Renin staining in the SMGs of SWR/J mice also showed considerable intercellular variation in the intensity of immunostaining, and some GCT cells apparently were negative (Tanaka et al., 1980, 1981). mK1 also showed an unusual sexually dimorphic mosaic pattern in GCT segments of SMGs in ICR mice (Kurabuchi et al., 1999). Together, this immunocytochemical evidence suggests that GCT cells, which are morphologically similar in appearance, are not homogeneous with respect to these secretory substances. We proposed in our previous report that variations in receptors for androgen, thyroid hormone, adrenocortical hormone, and growth hormone might divide the GCT cells into various cell types, resulting in the mosaic cellular distribution of mK1 (Kurabuchi et al., 2004). It is conceivable that the GCT cells might/might not have developed into full-ﬂedged granular cells and lost/not lost their ability for EGF synthesis depending upon the expression levels of these hormone receptors. The expression level of the androgen receptor has been shown to vary among GCT cells (Morrell et al., 1987; Sawada and Noumura, 1995). To the best of our knowl- TUBULE CELLS OF MOUSE SUBMANDIBULAR GLAND Fig. 9. Electron micrograph of a granular convoluted tubule (GCT) cells in the submandibular gland (SMG) of a female C3H/HeN mouse, after repeated injection with T315a-dihydrotestostrone (DHT); immunostaining with anti–epidermal growth factor (EGF) antiserum. A welldeveloped mesh-like rough endoplasmic reticulum (RER) and a large Golgi apparatus (G) with swelled cistern are visible close to the nucleus (N). A large apical portion of the cells is closely packed with numerous secretory granules, which are labeled with gold particles, indicating the presence of EGF. Short membrane infoldings are sometimes found in the base (arrowheads) of the cell. L, lumen; N, nucleus. Scale bar 5 2 mm. edge, however, no studies have investigated the expression of other hormone receptors in rodent salivary glands. Male mice are known to have larger GCT cells, containing abundant secretory granules, in their SMGs, and stereologic analyses have shown that the GCTs in the glands are much larger in males than in females (reviewed by Gresik, 1994). This ﬁnding agrees with the much higher concentration of secretory products in males, compared with in females (Barka, 1980). These sex-related differences were also conﬁrmed in each strain, C57BL/6J, BALB/c, C3H/HeN, and ICR, used in this study. In addition, the GCT cells were much larger in both sexes of ICR mice than in both sexes of C57BL/ 6J, BALB/c, and C3H/HeN mice. Furthermore, most of the GCT cells were EGF-positive in the males of each strain, whereas EGF-negative GCT cells were prominent in the females. Considering the size of the GCT cells, the percentage of EGF-positive cells, and the EGF-immunoreactivity, the EGF concentration was remarkably lower in both sexes of C57BL/6J, BALB/c, and C3H/HeN mice than in both sexes of ICR strain. This ﬁnding suggests 111 that the basal concentration of EGF in SMGs is strainspeciﬁc as well as sex-speciﬁc, and that C57BL/6J, BALB/c, and C3H/HeN mice belong to a low EGF group, whereas ICR mice belong to a high EGF group. Incidentally, inbred strains of mice have been classiﬁed into two distinct groups, one with a much higher SMG renin activity than the other (Wilson et al., 1978, 1981). C57BL/ 6J, BALB/c, and C3H/HeN mice belong to the low renin group, whereas the ICR, an outbred strain, belongs to the high renin group (Wilson et al., 1981). SMG renin is controlled by the renin regulator (Rnr) gene, and the allele carried by strains in the high SMG renin group is symbolized as Rnrs, while the allele carried by strains in the low SMG renin group is termed Rnrb (Wilson et al., 1977). To our knowledge, however, no relationship exists between the renin regulator gene and the production of GCT-speciﬁc secretory products, other than the involvement of renin. The results of mating between parents derived from ICR, a high EGF strain, and C3H/HeN, a low EGF strain, suggest that no histological or immunohistological differences were present between the two types of F1 progeny, one type obtained from ICR males 3 C3H/HeN females and the other type from C3H/HeN males 3 ICR females, in either sex. However, the size of the GCT cells and the ratio of EGF-positive cells in the F1 progeny suggest that the EGF concentration in the glands of the F1 progeny might be intermediate between the parent strains, when compared within the same sex. Similar results were obtained in a congenic strain derived from high and low SMG renin strains. Concretely, when the allele Rnrs from a donor AKR/J, a high SMG renin strain, was included on the genetic background of C57L/ J, a low SMG renin strain, SMG renin activity in the F1 progeny was intermediate between that of the parent strains, suggesting that SMG renin is controlled by genes located in the same region of chromosome 1 (Wilson et al., 1981). Furthermore, using the same mating pairs as those used in the present study, either type of F1 progeny inherited all of the kallikrein isozymes from either of the parent strains (Kurihara et al., 1999). These two previous mating experiments suggest that EGF is directly/indirectly under genetic regulation similar to that seen for SMG renin and that the F1 progeny seem to inherit genetic characteristics from both parent strains. Repeated injection with T3 and/or DHT in male and female C3H/HeN mice, a low EGF strain, enlarged the GCT cells and increased the number of secretory granules in both sexes, compared with the untreated controls of the same sex; the number of EGF-positive cells also increased remarkably, especially in females. Furthermore, the present immunoelectron microscopy examination revealed that the RER and Golgi apparatuses in the GCT cells of the hormone-injected mice were well developed and that basal infoldings were less abundant, compared with the GCT cells in the untreated controls. These ﬁndings show that GCT cells in both sexes of untreated C3H/HeN mice were not as well developed as those observed in ICR mice and that repeated hormone injections resulted in a marked enhancement of the GCT phenotype, promoting EGF synthesis in many GCT cells especially in females. An even stronger effect was obtained in the group that received both of the hormones: the GCT phenotype and EGF expression in the 112 MIYAJI ET AL. males nearly reached the levels seen in ICR males, while the levels in females were much higher than the levels seen in ICR females. It has been established in ICR mice that the additive and/or synergistic actions of androgen and thyroid hormone enhance the GCT phenotype and produce more abundant GCT-speciﬁc secretory products, including EGF (Hosoi et al., 1992; Gresik, 1994; Kurabuchi and Hosoi, 2001). These facts suggest that the response of SMG GCT cells in both sexes of C3H/HeN mice to thyroid hormone and androgen was similar to that seen in ICR mice, ruling out the possibility that the low and high concentration of EGF are caused by the opposite actions of these hormones in ICR and C3H/HeN mice. It is suspected that a low EGF concentration in SMGs may be caused by a lower level of hormones in blood. However, according to previous studies, serum T3 levels do not vary signiﬁcantly between low and high EGF strains in mature adults: approximately 0.45 ng/ml and 0.52 ng/ml in C3H/HeN males and females, respectively, and 0.53 ng/ml and 0.45 ng/ml in ICR males and female, respectively (Kurihara et al., 1999), 0.66 ng/ml in C57BL/6J males and 0.61 ng/ml in C3H/HeN males (Maia et al., 1995), and 0.72 ng/ml in ICR males (Burgi et al., 1986). These facts suggest that serum T3 levels are not responsible for the strain-speciﬁc low or high concentration of EGF in SMG. On the other hand, serum testosterone levels are known to be strain-speciﬁc, and age-speciﬁc (Eleftheriou and Lucas, 1974; Nelson et al., 1975; Jean-Faucher et al., 1982). The serum levels in mature adults were measured approximately 1.3 ng/ ml in C57BL/10J and 4.6 ng/ml in DBA/2J (Bartke, 1974) and were also reported to be approximately 2.2 ng/ml and 0.22 ng/ml in male and female C3H/HeN, respectively (Angele et al., 1998), and 3.0 ng/ml in ICR males (Shapiro et al., 1989). These data proved that serum level of androgen is remarkably lower in C57BL and C3H/HeN males, compared with males of the other strains. A low serum level of androgen may be one source of the low EGF expression in SMGs and for a less developed GCT phenotype. However, the present study showed that sexual differences were evident in GCT cells even in low EGF strains, despite the lower serum androgen levels, and a signiﬁcant strain-speciﬁc difference in EGF expression was observed especially in females. Therefore, androgen is unlikely to be a main cause of this difference. Walker et al. (1982) demonstrated that excessive exogenous T3 resulted in a marked increase in the SMG EGF concentration, and concluded that T3 receptor expression might be signiﬁcantly modulated by T3. Androgen receptors have also been reported to be upregulated by thyroid hormones in the SMGs of developing and adult mice (Minetti et al., 1986, 1987) and a marked increase in the nuclear binding of androgen in the SMGs of mice of both sexes has been shown using repeated testosterone injection (Katsukawa et al., 1983). Based on the ﬁndings of all of these previous studies, the low concentration of EGF in the SMGs of the low EGF group may be caused by fewer receptors for several hormones, including thyroidal and androgenic hormones, with fewer receptors leading to a lower hormonal sensitivity in the low EGF group. We believe that external androgen and/or thyroid hormone may also increase additively/synergistically the expression levels of each other’s receptors in each GCT cell, resulting in the enhancement of the GCT phenotype in many GCT cells and the production of more EGF in the low EGF group. As described above, SMG renin is known to be under genetic and hormonal control (Bing and Poulsen, 1971; Wilson et al., 1977). On average, the renin activity in 8-week-old females is known to be 200 times higher in high renin strains than in low renin strains, and the renin activity in females from low renin strains is not induced by androgen to the intact level observed in females from high renin strains (Wilson et al., 1981). In the contrast, EGF expression in SMGs in the three low renin strains used in this study was relatively low but was restored to a high level by the injection of thyroid hormone and/or androgen. Therefore, EGF regulation differs from that of SMG renin in inbred mice strains. If the low EGF level in such low renin strains was caused by a low sensitivity to hormones, the expression of many GCT-speciﬁc secretory products, including nerve growth factor (NGF), the gene families of kallikrein, mK9 (EGFbinding protein), 13 (prorenin converting enzyme), 22 (bNGF endopeptidase), as well as EGF will be low in the low renin strains. Furthermore, whether the hypothetical low sensitivity to hormones proposed in this study may be applied only to the low renin group of inbred mice or to all inbred mice, including the high renin group, remains to be elucidated. LITERATURE CITED Angele MK, Ayala A, Ciofﬁ WG, Blank KI, Chaudry IH. 1998. Testosterone: the culprit for producing splenocyte immune depression after trauma hemorrhage. Am J Physiol 274:C1530–C1536. Barka T. 1980. Biologically active polypeptides in submandibular glands. J Histochem Cytochem 28:836–859. Bartke A. 1974. Increased sensitivity of seminal vesicles to testosterone in a mouse strain with low plasma testosterone levels. J Endocr 60:145–148. Bing J, Poulsen K. 1971. The renin system in mice. Acta Pathol Microbiol Scand (A) 79:134–138. Bing J, Poulsen K, Hackenthal E, Rix E, Taugner R. 1980. Renin in the submaxillary gland: a review. J Histochem Cytochem 28:874– 880. Burgi U, Feller C, Gerber AU. 1986. Effects of an acute bacterial infection on serum thyroid hormones and nuclear triiodthyronine receptors in mice. Endocrinology 119:515–521. Chrétien M. 1977. Action of testosterone on the differentiation and secretory activity of a target organ: submaxillary gland of the mouse. Int Rev Cytol 50:333–396. Eleftheriou BE, Lucas LA. 1974. Age-related changes in testes, seminal vesicles and plasma testosterone levels in male mice. Gerontologia 20:231–238. Gresik EW. 1994. The granular convoluted tubule (GCT) cell of rodent submandibular glands. Microsc Res Tech 27:1–24. Gresik EW, Barka T. 1977. Immunocytochemical localization of epidermal growth factor in mouse submandibular gland. J Histochem Cytochem 25:1027–1035. Hazen-Martin DJ, Simson JAV. 1987. Immunocytochemical localization of nerve growth factor in mouse salivary glands. Histochem J 19:210–216. Hiramatsu M, Kashimata M, Takayama F, Minami M. 1994. Developmental changes in and hormonal modulation of epidermal growth factor concentration in the rat submandibular gland. J Endocrinol 140:357–363. Hosoi K, Tanaka I, Ishii Y, Ueha T. 1983. A new esteroproteinase (proteinase F) from the submandibular glands of female mice. Biochim Biophys Acta 756:163–170. Hosoi K, Maruyama S, Ueha T, Sato S, Gresik EW. 1992. Additive and/or synergistic effects of 5-a-dihydrotestosterone, dexametha- TUBULE CELLS OF MOUSE SUBMANDIBULAR GLAND sone, and triiod-L-thyronine on induction of proteinases and epidermal growth factor in the submandibular glands of hypophysectomized mice. Endocrinology 130:1044–1055. Jean-Faucher C, Berger M, de Turckheim M, Veyssiere G, Jean C. 1982. Plasma and testicular testosterone and dihydrotestosterone in mice: effect of age and HCG stimulation. IRCS Med Sci 11:26–27. Johansson O, Backman J. 1983. Enhancement of immunoperoxidase staining using osmium tetroxide. J Neurosci Methods 7:1217–1223. Katsukawa H, Tanabe Y, Funakoshi M. 1983. Effects of androgens on the activity of stroid-metabolizing enzymes in the murine submaxillary gland. J Dent Res 62:725–727. Kurabuchi S, Hosoi K. 2001. Sexual dimorphism and hormonal regulation of granular convoluted tubule (GCT) cells of mouse submandibular gland. Immunocytochemical analysis. (In Japanese). J Masticat Health 10:61–70. Kurabuchi S, Tada J, Gresik EW, Hosoi K. 1999. An unusual sexually dimorphic mosaic distribution of a subset of kallikreins in the granular convoluted tubule of the mouse submandibular gland detected by an antibody with restricted immunoreactivity. Histochem J 31:19–28. Kurabuchi S, Hosoi K, Gresik EW. 2001. Androgen regulation of the cellular distribution of the true tissue kallikrein mK1 in the submandibular gland of the mouse. J Histochem Cytochem 49:801– 802. Kurabuchi S, Hosoi K, Gresik EW. 2002. Developmental and androgenic regulation of the immunocytochemical distribution of mK1, a true tissue kallikrein, in the granular convoluted tubule of the mouse submandibular gland. J Histochem Cytochem 50:135–145. Kurabuchi S, Gresik EW, Hosoi K. 2004. Additive and/or synergistic action (downregulation) of androgens and thyroid hormones on the cellular distribution and localization of a true tissue kallikrein, mK1, in the mouse submandibular gland. J Histochem Cytochem 52:1437–1446. Kurihara K, Maruyama S, Nakanishi N, Sakagami H, Ueha T. 1999. Thyroid hormone (3,5,30 -triiodo-L-thyronine) masking/inversion of stimulatory effect of androgen on expression of mK1, a true tissue kallikrein in the mouse submandibular gland. Endocrinology 140:3003–3011. Maia AL, Kieffer JD, Harney JW, Larsen PR. 1995. Effect of 3,5,30 triiodothyronone (T3) administration on dio 1 gene expression and T3 metabolism in normal and type 1 deiodinase-deﬁcient mice. Endocrinology 136:4842–4849. Minetti CA, Valle LB, Fava-de-Moraes F, Romaldini J, OliveiraFilho R. 1986. Ontogenesis of androgen receptors in the mouse submandibular gland: correlation with the development proﬁles of circulating thyroid and testicular hormones. Acta Endocrinol 112:290–295. Minetti CA, Valle LB, Romaldini J, Fava-de-Moraes F, OliveiraFilho R. 1987. Thyroidal modulation of androgenic expression in mice submandibular gland. Horm Metab Res 19:146–151. 113 Mori M, Takai Y, Kunikawa M. 1992. Review: biologically active peptide in the submandibular gland—role of the granular convoluted tubule. Acta Histochem Cytochem 25:325–341. Morrell JI, Gresik EW, Barka T. 1987. Autoradiographic localization of dihydrotestosterone binding in the major salivary glands and other androgen-responsive organs of the mouse. J Histochem Cytochem 35:1053–1058. Murai T. 1987. Studies on proteinase F in major salivary glands of inbred mice—sex difference of proteinase F. (In Japanese). Jpn J Oral Biol 29:416–427. Nelson JF, Latham KR, Finch CE. 1975. Plasma testosterone levels in C57BL/6J male mice: effects of age and disease. Acta Endocrinol 80:744–752. Sawada K, Noumura T. 1995. Developmental pattern of androgen receptor immunoreactivity in the mouse submandibular gland. Zool Sci 12:243–248. Shapiro BH, Niedermeyer TM, Babalola GO. 1989. Serum androgen levels in senescent Crl: CD-1R(ICR) BR mice: effects of castration and testosterone treatment. J Gerontol 44:B15–B19. Tanaka T, Gresik EW, Michelakis AM, Barka T. 1980. Immunocytochemical localization of renin in kidneys and submandibular glands of SWR/J and C57BL/6J mice. J Histochem Cytochem 28:1113–1118. Tanaka T, Gresik EW, Barka T. 1981. Epidermal growth factor and renin in mouse submandibular glands. J Histochem Cytochem 29:1229–1231. van Noorden S, Heitz P, Kasper M, Pearse AGE. 1977. Mouse epidermal growth factor: light and microscopical localization by immunocytochemical staining. Histochemistry 52:329–340. Walker P, Coulbome P, Dussault JH. 1982. Time- and dose-dependent effect of triiodthyronine on submaxillary epidermal growth factor concentration in adult female mice. Endocrinology 111:1133–1139. Watson AY, Anderson JK, Siminoski K, Mole JE, Murphy RA. 1985. Cellular and subcellular colocalization of nerve growth factor and epidermal growth factor in mouse submandibular glands. Anat Res 213:365–376. Wilson CM, Erdos EG, Dunn JF, Wilson JD. 1977. Genetic control of renin activity in the submaxillary gland of the mouse. Proc Natl Acad Sci USA 74:1185–1189. Wilson CM, Erdos EG, Wilson JD, Taylor. 1978. Location on chromosome 1 of Rnr, a gene that regulates renin in the submaxillary gland of the mouse. Proc Natl Acad Sci USA 75:5623–5626. Wilson CM, Cherry M, Taylor BA, Wilson JD. 1981. Genetic and endocrine control of renin activity in the submaxillary gland of the mouse. Biochem Genet 19:509–523. Young WG, Ramirez-Young GO, Daley TJ, Smid JR, Coshigano KT, Kopchick JJ, Waters MJ. 2004. Growth hormone and epidermal growth factor in salivary glands of giant and dwarf transgenic mice. J Histochem Cytochem 52:1191–1197.