Abstract
The action of androgens on the immunocytochemical distribution of mK1, a true tissue kallikrein, was examined in the submandibular gland (SMG) of developing and adult mice by indirect enzyme-labeled and immunogold-labeled antibody methods for light and electron microscopy, respectively. In both sexes at 3 weeks of age, essentially all of the immature granular convoluted tubule (GCT) cells were uniformly immunostained. At 4 weeks of age (the onset of puberty), morphological differences between the two sexes appeared in the GCTs, in which some cells became immunonegative. Thereafter, the immunonegative GCT cells became more abundant in the SMG of males than of females and considerable intercellular variation in staining intensity for mK1 was seen, especially in males. A few slender GCT cells with strong immunoreactivity appeared in GCT segments only in males. Castration of males resulted in an increase in the number of immunopositive GCT cells, whereas administration of dihydrotestosterone (DHT) decreased the number of immunopositive GCT cells in the SMGs of both sexes. Slender GCT cells immunoreactive for mK1 were seen in females treated with DHT for 6 days. However, there were no immunostained slender GCT cells in female SMGs after injection of DHT for 2 weeks. Immuno-electron microscopy disclosed this type of cell in male SMGs, which closely resembles immature GCT cells of prepubertal mice, with a few small secretory granules uniformly labeled with gold particles, a sparse Golgi apparatus and RER, and basal infoldings. In mature male SMGs and in SMGs of DHT-treated females and castrated males, typical GCT cells had a well-developed Golgi apparatus and a net-like RER but few to no basal infoldings, whereas in the female gland equivalent cells had moderately developed RER and some basal infoldings. These results suggest that mK1 is one of the enzymes characteristically present in immature GCT cells and that its synthesis is inhibited in part by androgens, resulting in decreased numbers of immunopositive cells.
Keywords
THE SEXUAL DIMORPHISM of the granular convoluted tubule (GCT) cells of the mouse submandibular gland (SMG) has been extensively documented by morphological studies at the light and electron microscopic levels and by the techniques of molecular biology, biochemistry, and immunocytochemistry (reviewed in Barka 1980; Gresik 1994). These studies show that the vast majority of the biologically active substances, such as epidermal growth factor (EGF), nerve growth factor (NGF), renin, tonin, and other proteinases, in addition to kallikrein gene family members (mK9, mK13, and mK22), are more abundant in the glands of males. Moreover, expression of these substances and the morphological appearance of GCT cells are upregulated by multihormonal control of androgens, thyroid hormones, and adrenocortical hormones (Pink-staff 1980; Gresik 1994). By contrast, the true tissue kallikrein mK1 (Hosoi et al. 1983b,1994) is more abundant in the female gland and its synthesis is inhibited by androgens (Hosoi et al. 1984).
Tissue kallikreins are a family of enzymes that produce the bioactive peptide kinin, which provokes pain, leukocyte migration, and vasodilatation. The SMG has long been known to produce large amounts of tissue kallikreins (Shacter and Barton 1979). Among various mouse tissue kallikrein enzymes, mK1 has been shown to be a true tissue kallikrein, exhibiting the strongest kinin-releasing activity. Protein sequence analysis indicates that it is the product of the mouse gene for true tissue kallikrein, mKlk-1 (Hosoi et al. 1994).
Because the kallikrein isozymes share high amino acid-sequence homology and the antisera against any particular kallikrein frequently crossreact with other members of this family (Orstavik 1980; Hosoi et al. 1983a,b), the exact distribution of the various isoforms has been problematic. In our previous study (Kurabuchi et al. 1999) we prepared an antiserum with restricted immunoreactivity against mouse mK1. A rabbit polyclonal anti-mK1 antiserum, which had been known to crossreact with other major kallikreins (Hosoi et al. 1983a,b), was preabsorbed with the purified kallikreins mK9, mK13, and mK22, enhancing its specificity against mK1. When the non-preabsorbed antiserum was used as the primary antibody, essentially all GCT cells in both sexes were immunostained. By contrast, using the preabsorbed anti-mK1 antibody, only a subset of GCT cells was immunostained in both sexes, demonstrating an unusual sexually dimorphic mosaic pattern of mK1 distribution in the SMG of the ICR mouse (Kurabuchi et al. 1999). In males, mK1-immunopositive GCT cells were composed of two morphologically different cell types. The most numerous type, which is the large typical GCT cell, is only occasionally and weakly stained for mK1. The other type, referred to as the slender GCT cell, has small apical secretory granules, sparse RER, and a poorly developed Golgi apparatus. This type of cell is much less abundant and stains strongly for mK1. In the female gland, on the other hand, a greater number of the typical GCT cells shows moderate to strong immunostaining. These results demonstrated that the lower content of mK1 in the male SMG is not due to lower synthesis of this enzyme uniformly in all GCT cells but to fewer GCT cells expressing mK1.
The GCT compartment develops postnatally from striated duct cells (Gresik 1980,1994). Immature GCT cells strongly resemble the slender granular cells found in mature males (Chabot et al. 1987). Developing GCT cells are also sensitive to multihormonal control. The effects of androgens in regulating expression of mK1 in individual GCT cells of developing and mature mice have not been defined. In this study we examined the immunocytochemical distribution of mK1 by light and electron microscopy in SMGs of postnatally developing mice of both sexes and the effect of the administration of dihydrotestosterone (DHT) on the distribution of mK1 in the GCT compartment of castrated male mice and of intact males and females.
Materials and Methods
Animals
ICR strain mice were housed under controlled conditions of temperature (22C) and lighting (light 12 hr: dark 12 hr), and allowed food and water ad libitum. The GCT compartment was studied in the following groups, each of which consisted of four mice. Postnatal developmental changes were studied in mice of both sexes at 1, 2, 3, 4, 6, 8, 11, 14, 18, and 24 weeks of age. Effects of androgen withdrawal were examined in castrated adult male mice (14 weeks old), orchiectomized at 6 weeks of age. To examine the direct effect of androgens, 14-week-old male and female mice and castrated males were injected SC at a dose of 20 μg/g body weight with 5α-dihydrotestosterone (DHT) dissolved in sesame oil, seven times (during 2 weeks) before sacrifice. In addition, one group of females was sacrificed after only three injections, 6 days after the start of the experiment. Before examination, all animals were fasted overnight. Under anesthesia with sodium pentobarbital (30 mg/kg body weight), SMGs were quickly removed and processed for light and electron microscopic immunocytochemical (ICC) examination.
Antiserum
The antiserum against mK1, previously generated in rabbits, was found to be weakly crossreactive with three other major kallikreins, mK9, mK13, and mK22 (Hosoi et al. 1983b). Therefore, we preabsorbed this antiserum with the above three kallikreins to eliminate this crossreactivity. Western blotting analysis using this preabsorbed antibody preparation demonstrated a complete lack of crossreactivity with these enzymes (Kurabuchi et al. 1999). Briefly, the rabbit polyclonal anti-mK1 antiserum, diluted 1:300, was absorbed with a mixture of purified mK9, mK13, and mK22 (Hosoi et al. 1983a), each at a concentration of 100 μg/ml. The mixture was incubated at 4C for at least 1 day before its use in ICC.
Immunocytochemistry
Small pieces of the SMG were fixed by immersion in a mixture of 2% glutaraldehyde and 2% paraformaldehyde in 0.05 M cacodylate buffer, pH 7.4, for 2 hr at 4C, then dehydrated through a graded ethanol series, cleared in propylene oxide, and embedded in an Epon/Araldite mixture (Kurabuchi et al. 1999).
For light microscopy, 2-μm-thick sections were mounted on silane-coated slides (Matsunami; Tokyo, Japan), treated with methanol saturated with NaOH to remove the resin, and incubated in a solution of 0.3% H2O2 in methanol for 30 min to inactivate endogenous peroxidase. The sections were washed in distilled water and then in PBS (10 mM sodium phosphate buffer, 140 mM NaCl, pH 7.5) and then incubated for 2 hr at room temperature with the preabsorbed anti-mK1 antiserum diluted 1:40,000. After three rinses in PBS, the sections were immunostained by the avidin-biotin-peroxidase complex method according to the manufacturer's protocol (Vectastain Elite ABC kit; Vector Labs, Burlingame, CA). They were then dehydrated with ethanol and mounted in Entellan (Merck; Gibbstown, NJ), and examined with a BX-50 microscope equipped with Nomarski differential interference contrast optics (Olympus; Tokyo, Japan). In the immunostained sections, the percentage of mK1-immunopositive cells in the GCT segments was determined by counting a total of at least 1000 cells for each mouse in each group. The means ± standard deviation (SD) for each group (consisting of four animals) were calculated.
For electron microscopy, ultrathin sections on gold grids were etched with 3% H2O2 for 5 min, rinsed well with distilled water, and then subjected to an immunogold procedure. They were first incubated with 20% normal goat serum for 2 hr, then transferred into a drop of 1:40,000 preabsorbed anti-mK1 antiserum and kept overnight at 4C. After three rinses in PBS, sections were incubated with biotinylated goat anti-rabbit IgG (Vectastain Elite ABC kit) for 1 hr. They were finally incubated for 2 hr with 1:60 streptavidin conjugated with 10-nm gold particles (Zymed; San Francisco, CA). After staining with uranyl acetate and lead citrate, sections were viewed in a transmission electron microscope (JEM-2000EXII; JEOL, Tokyo, Japan) at 80 kV accelerating voltage.
Results
Postnatal Development
At 1 week of age in both sexes, very faint immunoreactivity for mK1 was restricted to the subluminal rim of intralobular ducts of the SMG (data not shown). At 2 weeks of age the subluminal immunoreactive zone increased in extent and intensity in both sexes (Figures 1a and 1b). Some segments of the intralobular ducts contained cells, presumably developing GCT cells, with small secretory granules that were strongly and uniformly stained for mK1. These granules were densely accumulated in the subluminal cytoplasm and more dispersed in the perinuclear regions. At 3 weeks of age the secretory granules increased in number and size in developing GCT cells, all of which were immunostained (Figure 2).
At 4 weeks of age (onset of puberty) sexual dimorphism was clearly evident: GCT segments in males were more extensive and their secretory granules were larger and more abundant (Figures 1c and 1d). Moreover, from this age onward the GCT cells showed intercellular variation in immunostaining for mK1, especially in males. The staining was moderate in some cells and strong in others, and immunopositive and immunonegative secretory granules were sometimes present in the same GCT cell. In addition, a few (approximately 30%) GCT cells were completely unstained; such cells appeared first in males (Figure 2). In females, almost all GCT cells were immunoreactive at this age (Figures 1d and 2).
By 6 weeks of age the GCT cells had a clear sexually dimorphic mosaic ICC distribution of mK1-positive secretory granules (Figures 1e and 1f). In males, about 70% of GCT cells with large secretory granules showed no immunostaining for mK1, whereas the remaining GCT cells had secretory granules with strong to moderate immunoreactivity (Figures 1e and 2). Occasional slender GCT cells were observed next to immunonegative or weakly stained typical GCT cells (see Figure 3 below). In females, the GCT segments were smaller and had many typical GCT cells with moderate to strong immunostaining, but a few completely lacked immunoreactivity (Figure 2).
At older ages the sexually dimorphic mosaic pattern of immunoreactive cells was even more pronounced. In mature males, remarkably few cells in the GCT segments were stained for mK1 (Figure 2), and many of these immunostained cells were slender GCT cells (Figure 1g). In females, about two thirds of the GCT cells were immunostained for mK1, and this proportion remained relatively constant between 11 and 24 weeks of age (Figures 1h and 2). Immunopositive slender GCT cells were commonly found in male GCTs; at least one cell of this type was routinely found in every lobule of male SMGs. However, no such cells were found in the female GCTs, even after sections of at least 40 SMGs of untreated adult females of 11–24 weeks of age were examined.
Fine structural immunolocalization of mK1 demonstrated that gold particles were essentially confined to secretory granules in the GCT cells. At 2 weeks of age, small irregularly shaped and uniformly dense secretory granules occupied the subluminal region of all ductal cells, and they were strongly labeled with gold particles (Figure 3). The secretory granules became rounded and increased in number at 3 weeks of age (data not shown). In these immature GCT cells, only a few segments of RER with narrow cisternae and small Golgi fields were seen (Figure 3). The basal plasmalemma was usually elaborately infolded and associated with large mitochondria.
From early pubertal (6 weeks of age) to fully adult (24 weeks) stages, the fine structural features of the cells of the GCT compartment remained relatively constant (Figure 4). Most GCT cells had the characteristic sexually dimorphic morphology previously described by several authors (see reviews by Mori et al. 1992; Gresik 1994; Kurabuchi et al. 1999). In both sexes there was a wide range of immunogold labeling for mK1 over variably sized secretory granules. In some cells, many secretory granules were labeled but there was significant intergranular variation in the intensity of staining, whereas in adjacent cells no granules were stained. However, as previously noted (Kurabuchi et al. 1999), no differences in fine structure were discernible between the positively and negatively immunoreactive GCT cells.
Nomarski differential interference microscopy of Epon/Araldite-embedded 2-μm sections of SMGs from developing and adult mice. Sections were stained with preabsorbed anti-mK1 antiserum. Males (a,c,e,g); females (b,d,f,h). A subluminal rim of immunoreactivity is seen in intralobular ducts of both sexes at 2 weeks of age (a,b). After the onset of puberty (c,d, 4 weeks; e,f, 6 weeks) the GCT cells are larger in males than in females and a few GCT cells lack immunoreactivity in both sexes (c,f, arrows). In adult males (24 weeks of age), only slender SG cells show strong immunoreactivity, but most typical GCT cells are immunonegative in males (g), whereas in the females about half of the GCT cells are immunopositive (h). Bar = 50μm.
Percentage of mK1-immunopositive GCT cells during postnatal development of male and female mice. Asterisks indicate statistically significant differences of p<0.001 determined by Student's t-test between the means of the two sexes at the same age.
In addition, from 6 weeks onward the GCT compartment of males became hypertrophic and contained cells that were slender in shape with morphological features similar to immature GCT cells (Kurabuchi et al. 1999). This type of cell had a more electron-lucent cytoplasm with sparse RER, a small Golgi apparatus, moderate basal infoldings, and small apical secretory granules that were uniformly strongly immunoreactive for mK1 (Figure 4). These slender GCT cells were not seen in mature females.
Effect of Castration and Administration of DHT on the SMG
GCTs of castrated adult males resembled those of normal mature females (Figures 5a, 5c, and 5e). However, the GCT cells in the castrated males (Figure 5c) were smaller, and the secretory granules were somewhat less abundant, compared to those of the normal females (Figure 5e). In castrated males, about two thirds of the GCT cells showed strong to moderate immunostaining, while the others were immunonegative (Figure 6). Both positive and negative cells showed the same morphological features. Slender GCT cells were not detected in the GCT segments of castrated males.
Electron microscopy demonstrated that the secretory granules varied in size and density (Figure 7). In some cells, uniformly small, dense apical secretory granules were common, but in other cells lucent granules (usually large) and variably sized dense granules were present along with several vacuoles (Figure 7). Perinuclear RER and Golgi apparatus were sparse, but basal infoldings associated with large mitochondria were prominent. In immunopositive GCT cells there was significant intergranular variation in the number of gold particles: both large and small dense secretory granules usually were labeled well, but large lucent granules were weakly labeled or unlabeled. The fine structural feature of the positively and negatively stained cells were essentially the same.
DHT had a strong trophic effect on GCT cells of intact males (Figure 5b), castrated males (Figure 5d), and intact females (Figure 5f). Although DHT reduced immunostaining for mK1, the pattern of immunostaining was not identical in the three groups. In intact males treated with DHT the percentage of positively stained cells was even lower than in untreated normal males (Figure 6). After administration of DHT for 2 weeks to castrated males (Figure 5d) or intact females (Figure 5f), the percentage of GCT cells positive for mK1 declined (Figure 6), and considerable intergranular variation in staining intensity persisted in those cells that were positive. However, more positively stained GCT cells were present than in intact or DHT-treated normal males (Figures 5a-5f). Strongly stained slender GCT cells reappeared in castrated males after treatment with supraphysiological doses of DHT. These slender GCT cells were also seen in females treated with DHT for 6 days (three injections) (Figure 5f) but were never seen in females treated for 2 weeks (seven injections) (Figure 5g) or in intact females. Finally, in all three groups many large typical GCT cells were completely unstained for mK1.
The persistence of slender GCT cells was confirmed by electron microscopy (Figure 8). The cytoplasm of these cells was lighter than in typical GCT cells; the RER was sparse, and basal infoldings were well developed. The apical secretory granules were strongly labeled with gold particles for mK1. The electron microscopic appearance of the hypertrophied GCT cells of DHT-treated females and of DHT-treated castrated males was essentially the same (data not shown).
Electron micrograph of a GCT segment of a 2-week-old female. Positive staining for mK1 is seen in small apical secretory granules of immature GCT cells. L, lumen; M, mitochondria. Bar = 2 μm. (
Electron micrograph of a slender and a large GCT cell from a 6-week-old male, showing positive staining for mK1 on small secretory granules. Cisternae of RER are flattened and sparse, and basal infoldings are associated with elongated mitochondria (M). L, lumen; N, a nucleus of the slender cell. Bar = 2 μm. (
Discussion
The ICC data in this study indicate that all immature GCT cells in the mouse SMG at prepubertal ages synthesize mK1. It has been documented previously by isoelectric focusing that mK1 appears earliest, by Day 15 of age, and other members of the kallikrein family—mK9, mK13, and mK22—follow thereafter (Hosoi et al. 1990). Indeed, among the various biologically active substances present in the mouse SMG, mK1 appears to be the first of these secretory proteins produced in the developing GCT cells; immunocytochemically detectable EGF (Gresik and Barka 1978; Chabot et al. 1987) and NGF (Kusakabe et al. 1986) appear at about 3 weeks of age only in some GCT cells, and the expression of renin (Gresik et al. 1978b) is detected at far later ages. In fact, GCT cells immunoreactive for these substances (except mK1) subsequently increase in number. The present study, however, shows that after the onset of puberty GCT cells immunoreactive for mK1 decreased in number in both sexes. Moreover, this study shows that the rapid and remarkable decrease of GCT cells expressing mK1 occurs in early puberty in males. This stage corresponds to the period in which sexual dimorphism of the GCT cells first appears (reviewed by Pinkstaff 1980; Gresik 1994). We also confirm our previous report that only a few cells express mK1 abundantly in fully matured animals (Kurabuchi et al. 1999). In females the cells expressing mK1 decreased later, and approximately two thirds remained immunoreactive even in fully matured animals. These results suggest that the difference in mK1 expression between the two sexes is probably caused by the inhibitory effect of androgens on mK1 expression.
Nomarski differential interference microscopy of Epon/Araldite-embedded 2-μm sections of SMGs of adult mice (14 weeks of age), stained with preabsorbed anti-mK1 antiserum. Intact male (
In castrated males, GCTs regressed and resembled those of normal females, as noted previously (Kaiho et al. 1975). Correspondingly, the GCTs of castrated males showed a mosaic distribution of cells immunoreactive for mK1, very similar to that seen in normal females. Multihormonal support by thyroid hormones, adrenocortical hormones, and androgens is necessary for full development and maintenance of the GCT of the SMG (Sato et al. 1981; Gresik 1994). The synthesis of many biologically active substances (except mK1) in the atrophied GCT cells of hypophysectomized mice is induced by DHT, T3, and dexamethasone; however, the synthesis of mK1 is inhibited by these hormones (Hosoi et al. 1992; Maruyama et al. 1993). Therefore, the mosaic pattern of immunopositive and -negative GCT cells in the SMG of normal females and castrated males is likely the result of the action of thyroid and adrenal hormones in the absence of the repressive action of androgen.
Percentage of mK1-immunopositive GCT cells in SMGs of male, castrated male, and female mice treated with DHT for 2 weeks. Each value indicates the mean ± SD for four animals. ∗ p<0.05; ∗∗ p<0.01, significantly different from control normal males. ¶¶ p0.01, significantly different from castrated males. §§ p<0.01, significantly different from control normal females. Differences were calculated by Student's t-test.
The present study found that some structural features typical of female GCT cells did not completely disappear after androgen treatment, i.e., a few basal infoldings remained even after hormone treatment. The present ICC data indicate that the immunopositive GCT cells decrease in number in both female mice and castrated male mice exposed to DHT, and show that androgens inhibit the synthesis of mK1 in many GCT cells. This is in good agreement with the previous report that DHT treatment decrease mK1 content in homogenates of the SMG (Hosoi et al. 1984).
Intergranular variation in immunostaining for mK1 observed by light and electron microscopy was most obvious in the GCT cells of young males (about 6 weeks of age) and of females and castrated males treated with excess DHT. The present electron microscopic analysis confirms our previous report (Kurabuchi et al. 1999) that immunogold labeling for mK1 varies from cell to cell and among secretory granules within a single cell, even though these cells appear the same in their fine structure. It is conceivable that the GCT cells showing frequent intergranular variation may be transient cells shifting from mK1-immunopositive to mK1-immunonegative cells. In these hyper-trophic GCT cells the amount of mK1 was decreased by the action of androgen. Slight and infrequent intercellular and intergranular variations in staining intensity have been reported for EGF, NGF, and renin in GCT cells (Gresik and Barka 1977; Gresik et al. 1978a; Tanaka et al. 1981; Kusakabe et al. 1986), but not to the degree seen here for mK1. Moreover, these variations in immunostaining are not sexually dimorphic and are clearly different from the sexual variation in mK1 distribution.
Previously we suggested that the slender GCT cells may represent a cell type that is distinct from the typical large, mature GCT cells (Kurabuchi et al. 1999). However, the results of the present study indicate that this cell may be an immature GCT cell. The fine structure of slender GCT cells is almost identical to that of immature GCTs in prepubertal mice of both sexes. Moreover, the strong uniform staining for mK1 of their small secretory granules is the same as that seen in prepubertal GCT cells. The immediate precursor of the immature GCT cell is the striated duct cell (reviewed in Gresik 1994). Very few striated duct cells bind androgen; moreover, although most GCT cells bind androgen, a few do not (Morrell et al. 1987; Sawada and Noumura 1995). Immature GCT cells are also nonresponsive to androgen but they are responsive to thyroid hormone. In fact, expression of the androgen receptor in the mouse SMG is induced by thyroid hormone (Minetti et al. 1986,1987). Slender GCT cells were not observed in castrated males, normal females, or females treated with androgen for 2 weeks. However, they were seen in females treated for 6 days (Figure 5f). These observations are consistent with the idea that the slender GCT cell, intensely immunoreactive for mK1, is an immature GCT cell that is insensitive to androgen. During its maturation to a typical large GCT cell it becomes responsive to androgen (possibly by the inductive influences of thyroid and other hormones) and gradually represses expression of mK1, resulting in GCT cells with secretory granules of variable mK1 content and, eventually, in cells completely lacking mK1. Such transitional cells were noted in segments of the GCT nearest to the striated and intercalated ducts of the SMG (Kurabuchi et al. 1999). Their presence in normal and androgen-treated males and in females treated for short periods with androgen correlates with an expanding GCT cell population, whereas their apparent absence from the glands of normal females and castrated males represents a more quiescent state. Their absence from the female treated for long periods with androgen suggests that expansion of the GCT has become stabilized under these conditions.
Electron micrograph of GCT cells of the SMG from a castrated male. The two cells whose nuclei (N) are visible are immunopositive, whereas the two upper cells are immunonegative, but show no fine structural differences from the former cells. L, lumen; V, vacuole; M, mitochondria. Bar = 2 μm. (
Electron micrograph of an immunopositive slender cell with apical small secretory granules, situated among immunonegative typical GCT cells of a DHT-treated castrated male. Only a few dilated cisternae of RER are seen in the basal cytoplasm of the slender cell, and several basal infoldings are seen. G, Golgi complex; M, mitochondria; N, nucleus of slender cell. Bar = 2 μm. (
In summary, in this study we have followed the postnatal development of the sexually dimorphic mosaic distribution of mK1 in mature GCT cells from the uniformly immunopositive immature GCT cells and have characterized the role of androgen status on this peculiar distribution pattern. Lastly, our data suggest that the slender GCT cells seen in adults may represent newly forming GCT cells that have an immature phenotype as they develop into typical large GCT cells.
Footnotes
Acknowledgements
This work was supported in part by a grant-in-aid for scientific research (11671820) from the Ministry of Education, Science and Culture, Japan to S. Kurabuchi, and NIH Grant DE10858 to E. W. Gresik.