Sites of expression and induction of glandular kallikrein gene expression in mice

Molecular and Cellular Endocrinology, 0 1991 Elsevier Scientific Publishers

MOLCEL

135

81 (1991) 135-146 Ireland, Ltd. 0303.7207/91/$03.50

02610

Sites of expression and induction of glandular kallikrein gene expression in mice J.D. Penschow,

C.C. Drinkwater,

J. Haralambidis

and J.P. Coghlan

The Howard Florey Institute of Experimental Physiology and Medicine, lJnic,ersity of Melbourne, Parkrile, (Received

Key words: Glandular

kallikrein;

Hybridization,

14 June 1991; accepted

in situ; Salivary

gland;

21 June

Vie. 3052, Australia

1991)

Nasal gland

Summary

In order to provide a foundation for comparison across species of glandular kallikrein genes, we have studied the 12 functional mouse genes on the basis of expressing cell types, developmental patterns of expression and gene response to hormonal induction. We have shown expression of the renal kallikrein gene in the female anterior pituitary, the thick ascending limb of renal cortical distal tubules, nasal glands of neonatal mice and at varying levels throughout the duct tree of major salivary glands of immature and adult mice, except for intercalated ducts. This gene did not respond to hormonal induction in salivary glands. The other 11 of the 12 genes are expressed in androgen-responsive cells of granular convoluted tubules of the submandibular salivary gland from 22 days postnatal, when sexual dimorphism of expression first becomes apparent. Expression of these genes is induced prematurely in 22-day-old mice by treatment with testosterone or thyroxine. In the adult female mouse, estrogens also induce elev,ated levels of expression. One of the glandular kallikrein genes is expressed in Leydig cells of the testis as well as the submandibular gland. This study has extended the basis for cross-species comparison of glandular kallikrein genes.

Introduction

Glandular kallikreins constitute a family of structurally related serine protease enzymes encoded by separate genes in several mammalian species (Evans et al., 1988; Wines et al., 1989). The most highly conserved is a kininogenase, found at high levels in salivary glands, kidney and pancreas (Fritz et al., 1977; Swift et al., 1982; Fukushima et al., 1985; van Leeuwen et al., 1986).

Address for correspondence: Ms. Jennifer D. Penschow, Howard Florey Institute of Experimental Physiology and Medicine, University of Melbourne, Parkville, Vie. 3052, Australia. Fax (3) 348.1707.

This enzyme and its cleavage product, the vasodilator peptide bradykinin, are integral parts of the kallikrein-kinin system important in the regulation of local blood flow. There is little similarity between species in the number of glandular kallikrein genes identified or their nucleic acid sequences, although sites of expression are comparable in some instances. In the mouse, glandular kallikrein genes are encoded by 24 genes, of which 12 are expressed (Evans et al., 1987; Drinkwater et al., 1988a). In addition to the kininogenase, encoded by mGK-6 (van Leeuwen et al.. 19861, glandular kallikrein genes encode the y and CYsubunits of nerve growth factor (mGK-3 and mGK-4, respectively (Evans and

136

Richards, 1985)), epidermal growth factor binding protein types A, B, C (mGK-22, 13, and 9, respectively (Drinkwater et al., 1987)) and y-renin (mGK-16 (Drinkwater et al., 1988a)). There are five glandular kallikrein genes which encode unidentified proteins: mGK-1 (Mason et al., 1983), mGK-5 (Drinkwater and Richards, 19871, mGK-8 (Fahnestock et al., 1986), mGK-11 and mGK-21 (Evans et al., 1987; Drinkwater et al., 1988a). As in other species, cellular sites of expression of the kininogenase (renal kallikrein) gene in mice include distal convoluted tubules of the kidney (Coghlan et al., 1985) and striated ducts of salivary glands (van Leeuwen et al., 1987; CoghIan et al., 1989). Expression of the kininogenase gene was demonstrated in extracts of several other tissues from mice (van Leeuwen et al., 1987) and rats (Clements et al., 1990; Saed et al., 19901, including rat pituitary, where it is regulated by dopamine in the neurointermediate lobe and by estrogen in the anterior lobe (Clements et al., 1986; Powers, 1986). Other kallikrein genes are expressed in cells which are responsive to androgens, in the submandibular gland of mouse (van Leeuwen et al., 19871 and rat (Gerald et al., 1986) and in the prostate of several species (Chapdelaine et al., 1984; Clements et al., 1989). Expression of all 12 of the functional mouse glandular kallikrein genes is found in salivary glands (Drinkwater and Richards, 1988b) but none apart from that of kininogenase mGK-6 is reported in other tissues. A knowledge of the cellular sites of expression of glandular kallikrein genes would provide insight into the possible function of the enzymes encoded by these genes. Comparison across species of the response of glandular kallikrein genes to hormonal induction and of sites of expression may provide clues to the functional importance of a particular gene product. In this study we have addressed these questions. Materials

and methods

Animals Swiss outbred mice were used for all experiments. Animals were killed by cervical dislocation.

Litters of mice were weaned on postnatal day 17. Some animals (n = 3 each group) were injected subcutaneously behind the shoulders with a single dose of 2.5 mg/kg testosterone enanthate or 0.5 mg/kg estradiol valerate (Schering, Berlin, F.R.G.) in 0.04 ml peanut oil with 5% benzyl alcohol, or vehicle only. Littermates (n = 3) were injected i.p. daily for 5 days with 500 pg/kg thyroxine in normal saline. Adult female treatment controls (n = 2 per group) were given the equivalent doses of hormone in a volume of 0.1 ml, and additional groups (n = 4 each) received diethylstilbestrol dipropionate (DES) (Steraloids, Wilton, NH, U.S.A.) at 0.5 mg/kg in the oil vehicle, or vehicle only. Untreated adult male mice (n = 4) constituted another group. All treated and control animals were sacrificed 124 h after commencement of treatment. Littermates were sacrificed when 22 days old. Collection of tissues Tissues were freeze-embedded in OCT compound (Lab-Tek., Naperville, IL, U.S.A.) by immersion in hexane/dry ice for hybridization histochemistry as described previously (Penschow et al., 1986). Duplicate samples of tissue from untreated and experimental animals were fixed by immersion for 4 h at room temperature in 4% paraformaldehyde (Merck, Darmstadt, F.R.G.) in 0.1 M phosphate buffer, pH 7.2. Processing and sectioning of tissues for hybridization histochemistry Freeze-embedded tissues were stored at -20°C until sectioned. Fixed tissues were rinsed in buffer, rinsed in 70% ethanol, transferred through three changes of absolute ethanol for 2-16 h then processed to paraffin and sectioned at 5 Frn using conventional procedures. Sections were dewaxed, rinsed in three changes of ethanol and when dry, were post-fixed for 5 min at 4°C in 3% glutaraldehyde in 0.1 M phosphate buffer containing 20% ethylene glycol. The freeze-embedded specimens were sectioned at 6 wrn in a cryomicrotome and fixed as described for postfixation of wax sections. All sections were pre-hybridized, hybridized and autoradiographed using X-ray film and liquid emulsion as described previously (Penschow et al., 1986).

137

Preparation cleotides

and

labelling

of

oligodeoxyribonu-

Twelve gene-specific and one common kallikrein screening probe (van Leeuwen et al., 1986) and one control (rhodopsin) probe of 25-30 nucleotides were prepared and labelled as described previously (Penschow et al., 1989); at the 5’ end with [ y- “*P]ATP (5000 Ci/mmol, Amersham, Bucks., U.K.) or the 3’ end with [“H]dATP (85 Ci/‘mmol, Amersham). For maximum specificity of gene-specific probes, a region of maximum difference in the sequence in exon 4 of the respective genes was chosen for probes corresponding to mGK-3, mGK-4, mGK-5 and mGK-6 mRNAs (Drinkwater et al., 1988a) and a region in exon 3 (Evans et al., 1987) was chosen for the remainder (mGK-1, mGK-8, mGK-9, mGK-11, mGK-13, mGK-16, mGK-21, mGK-22). Sequences were selected for maximum specificity over 25-30 nucleotides. Hybridization histochemistry

Procedures using oligodeoxyribonucleotide probes have been described previously (Penschow et al., 1986, 1989). We have also shown that washing stringency at temperatures compatible with optimal preservation of tissue morphology have little effect for selective retention of hybrids in tissue sections (Penschow et al., 19871, and that carefully selected hybridization conditions are a preferable means of obtaining selective hybridization. Therefore we established hybridization conditions which would be specific for all 12 of the gene-specific probes. This was done by determining that each probe (except kidney-specific mGK6) did not hybridize with renal cortex and that mGK-6 did not hybridize with testis. The basis for this strategy is that the closest probe sequence homology was between the renal (mGK-6) and testis-specific (mGK-21) probes. Hybridization conditions that would discriminate these closely homologous mRNA sequences would also discriminate less homologous sequences. Kidney and testis were used in all experiments as specificity controls. Controls for non-specific binding of probes to tissue areas of interest were adjacent sections treated either with ribonuclease A (Penschow et al., 1989) or hybridized with a 30mer complementary to mouse mRNA for the retinal

photoreceptor protein rhodopsin, which is not expressed in any of the tissues studied. Hybridization conditions selected for specificity were 48°C in buffer containing 40% formamide, deionized (Merck) 0.6 M sodium chloride, 50 mM sodium phosphate, 5.0 mM EDTA, 0.02% Ficoll, 0.02% bovine serum albumin, 0.02% poly(vinyl)pyrrolidone and 0.1% crude oligonucleotides (Sigma, St. Louis, MO, U.S.A.). Sections were hybridized for 18 h, washed in 2 x SSC (1 X SSC is 0.15 M sodium chloride and 0.15 M sodium citrate) at room temperature for 15 min then in 1 X SSC at 50°C for 45 min, rinsed in absolute ethanol and allowed to dry. Autoradiography

Sections labelled with 32P were exposed overnight to XAR-5 film (Kodak, Rochester, NY, U.S.A.) and hybridization signals evaluated. Some sections were then exposed to high resolution X-ray film (MRF-34, DuPont, Wilmington, DE, U.S.A.) and others were coated with K5 emulsion (Ilford, Essex, U.K.) diluted 1:2 with distilled water then exposed for l-20 days. “H-labelled sections were coated with the same emulsion and exposed for 2-16 weeks. Further details of this procedure have been described elsewhere (Penschow et al., 1989). Results

The cellular sites of expression of the 12 glandular kallikrein genes are summarized in Table 1. mGK-6 expression in kidney and pituitary

mGK-6 mRNA was located in renal convoluted distal tubules as previously reported (Coghlan et al., 1985). The 3H-labelled mGK-6 oligodeoxyribonucleotide probe used in the present study provided accurate resolution, which permitted comparison of relative mRNA levels in regions of tubules identified by morphological criteria (Kriz and Kaissling, 1985). The level of mGK-6 expression was considerably higher in the convoluted distal tubule than in the distal thick ascending limb (Fig. 1) or connecting tubule. There was no expression observed at sites at the vascular poles comparable to those reported in the rat by Xiong et al. (1989).

Testicular teydig cells

Striated ducts Granular convoluted tubule Excretory ducts

S~~~l,~~glands

+

+

f

mGK-6

Ducts of nasal ~IU~~~

c

mCK-5

+

+

mGK-4

Anterior pituitary

c

mGK-3

KALLIKREIN

f f

t

mCK- 1

OF GLANDULAR

Distal convoluted tubules Thick ascending limb

RemI cartex

Site

CELLULAR LOCATION OF EXPRESSION

TABLE L

+

mGK-8

+

mG#-9

+

mGK-11

GENES IN MOUSE TISSUE

+

mGK-13

+

mGK-16

i

-I-

mGK-21

+

mGK-22

Fig. 1. Photomicrographs in bright-field (a) and dark-field (b) of the same field of a liquid emulsion autoradiograph of a frozen section of mouse renal cortex after hybridization with the “H-labelled mGK-6 oligodeoxyribonucleotide probe. (a and b) A strong hybridization signal in distal convoluted tubule (D) compares to a weak signal in thick ascending limb (arrowed). Collecting duct (0, glomerulus (G) and the numerous proximal tubules are unlabelled. Stain: hematoxylin and eosin. Bar = 170 +m.

A low level of mGK-6 expression was detected in the anterior pituitary of a female mouse (Fig. 2). There was no detectable expression in the neurointermediate lobe, in which renal kallikrein is expressed in the rat at a lower level than in the anterior lobe (Chao et al., 1987).

could be obtained. The ducts containing mGK-6 mRNA were located on the lateral walls and septum of the nasal cavity (Fig. 3a). The ducts of the lateral gland, shown in Fig. 3a and b, resemble striated ducts of the major salivary glands.

mGK-6 expression in nasal glands

Expression of mGK-6 salicary glands

(the non-inducible gene) in

mGK-6 expression was detected in ducts of nasal glands (Fig. 3). In the adult mouse, the situation of the nasal bones precluded accurate dissection of the complete glandular area so a neonatal mouse head was embedded in paraffin and serially sectioned in the horizontal plane. The sections were hybridized with the mGK-6 probe so that an overview of the glandular area

Expression of mGK-6 was at a high level in striated ducts of the submandibular, sublingual and parotid glands of untreated 22-day-old male and female mice (Fig. 4a and b). These results correspond with previous studies in the adult mouse (van Leeuwen et al., 1987). We also observed mGK-6 expression in inter-lobular and excretory ducts (Fig. 4a and b). Expression was

b Fig. 2. Autoradiographs on high-resolution X-ray film of serial frozen sections of pituitary gland from a female mouse after hybridization with the “*P-labelled oligodeoxyribonucleotide 30mer probes complementary to mRNAs for mGK-6 (a) or mouse rhodopsin as control (b). (a and b) There is a stronger hybridization signal in the anterior lobe (A) than in the neurointermediate and posterior lobes. Bar = 0.3 mm.

140

considerably lower in cells of the main submandibular duct, located in the hilus, than in striated ducts (Fig. 4~). There was no expression detectable in intercalated ducts (Fig. 4d). Thus, it appears that mGK-6 is expressed throughout the duct tree of the major salivary glands, from -the junction of intercalated and striated ducts up to and including the main ducts.

Expression of the 11 inducible genes (mGK-1, 3, 4, 5, 8, 9, 11, 13, 16, 21, 22) in salivary glands Expression of all of these genes was tested on the tissues described. The results shown in Figs. 4, 5 and 6 for inducible genes are representative of the overall pattern of expression for these 11 genes. Some results for mGK-6 are included in Figs. 4, 5 and 6 for comparison, as this is the only gene which showed a different pattern of expression in salivary glands. Expression of these 11 kallikrein genes was first detectable in a few cells within striated ducts in the submandibular glands of 20- to 21-day-old male mice (data not shown) and 23- to 24-day-old female mice. This was the first sign of the testosterone-induced sexual dimorphism which is a characteristic of these glands in mature mice (Chretien et al., 19791, whereby the granular convoluted tubules (GCT) differentiate from striated ducts to a greater extent in male than in female glands, resulting in higher levels in male glands of mRNAs encoding the numerous GCT cell proteins (van Leeuwen et al., 1986). The difference between males and females in the level of mRNAs encoded by these genes increased with the maturity of the mice (Fig. 5a and b) as the extent of the expressing GCT cell population increased. By 6-7 weeks of age, levels of mRNAs encoded by these 11 genes were maximal. Previous studies have shown levels of mGK-3, 4 and 5 to be at least lo-fold greater in the adult male than in the female. Thus, all of these 11 genes, now referred to as the inducible genes, exhibited the sexual dimorphism of expression previously shown for

Fig. 3. Liquid emulsion autoradiographs of paraffin sections of the nasal region from a l-day-old mouse, cut in the horizontal plane, after hybridization with the ‘*P-labelled mGKh-specific oligodeoxyribonucleotide probe. Photographs hy bright-field (a, h) and dark-field (c) microscopy. (0) An overview of the glandular region at the lateral edges and septum (S) of the nasal cavity showing the location of some of the ducts (arrowed) which contain mGK-6 mRNA. Stain: hematoxylin and eosin. Bar = 0.8 mm. (b) A higher magnification of the lateral nasal gland shown in the upper portion of (a). In (b) the same labelled ducts are arrowed as in (a). Bar = 0.2 mm. (c) A control showing the same field as in (b) in a serial section treated prior to hybridization with ribonuclease A. Unlabeled ducts in (c) (arrowed) are the same as those arrowed in (b).

141

those with known products (Drinkwater et al., 1987, 1988a; van Leeuwen et al., 1987). Of the 11 inducible genes, the cellular site of expression of only one, mGK-21, has been located outside the salivary glands, in the testis (see below). Ejjects of testosterone, thyroxine, estradiol and diethylstilbestrol on expression of kallikrein genes in prepubertal and adult female mice

Expression of the 11 inducible kallikrein genes, but not mGK-6, was induced by treatment of adult female or 17-day-old male and female mice (Fig. 6c and d) for 5 days with testosterone or thyroxine. Male and female 17-day-old mice showed the same response (mice were 22 days old when sacrificed). Expression was induced also by estradiol or DES in adult female but not in

prepubertal mice (Fig. 6b, d, and f ). Computerized densitometry (Sierra Scientific, Sunnyvale, CA, U.S.A.) was used to compare the degree of induction obtained at the cellular level after the various hormone treatments. The optical density of high resolution X-ray film autoradiographs of sections hybridized with the various probes, such as those shown in Fig. 6, was measured and the background density over negative tissue subtracted. There are inherent inaccuracies in this type of measurement due to variations in section thickness, degree of film exposure, etc. Thus, extensive analysis of the raw data, with its inherent flaws, would produce statistics of doubtful accuracy. The range of values we obtained showed that the level of induction compared to control mice was greatest for testosterone; 2- to 9-fold in

Fig. 4. Liquid emulsion autoradiographs of frozen sections from salivary glands from a 22-day-old mouse after hybridization with the “H-labelled mGK-6 30mer oligodeoxyribonucleotide (a, c, d) or 32P-labelled m. rhodopsin 30mer (b). Photographs by dark-field microscopy. (a and b) Areas occupied by submandibular (SM), sublingual (SL), parotid (P) glands and lymph node (L) are identified in (b) and are clearly visible in serial section (a). The heavily labelled striated (S), interlobular (I) and sublingual main excretory duct (E) are arrowed in (a). There is no labelling in control section (b). Bar = 1 mm. Cc) The strong hybridization signal over striated ducts (RHS) can be compared to the weak labelling over the main excretory duct (arrowed) of the submandibular gland, situated in the hilus adjacent to the sublingual gland (top left portion). Bar = 0.2 mm. Cd) The junction (arrowed) between an unlabelled intercalated duct and a labelled striated duct of the submandibular gland. Bar = 50 lrn.

142

27

23

26

K 27

34

F

b

after estradiol there was no induction in 22-dayold and 2- to 5-fold in adults and after DES there was a 2-fold induction in adults. This comparison is consistent with the visual impressions in Fig. 6 of the relative levels of induction. Expression was located in GCT cells, which constituted a greater portion of submandibular ducts in treated animals than in controls. Conversely, for mGK-6 there was no apparent difference in cellular mRNA levels between the treated and untreated groups.

Fig. 5. Autoradiographs on high-resolution X-ray film of sections of salivary glands and kidney from mice of various ages (indicated in days) after hybridization with “zP-labelled 3Omer oligodeoxyribonucleotides specific for mGK-9 (a) or mGK-11 (/I) mRNAs. (a) Tissue at top left corresponds to a piece of kidney (K) which has folded at the edges, thereby causing probe entrapment and an apparent hybridization signal. There are also three sets of salivary glands from animals at the ages designated. Male glands (M) are on the left and female (F) on the right, as indicated. Exposure. IO days. Bar = 5 mm. (h) Glands from male mice at the ages indicated. Exposure, 5 days. Bar = 4 mm. Each set of tissues represents one frozen section through a composite block comprising the tissues shown.

22-day-old and 4- to 1Zfold in adults. The degree of induction after thyroxine treatment was 2- to 7-fold in 22-day-old and 2- to 7-fold in adults;

Fig. 6. Autoradiographs on high-resolution X-ray film of sections of salivary glands from mice treated with testosterone CT), thyroxine (TJ), estradiol (E), vehicle cr.3 or DES CD). Untreated adult male salivary glands (A) and kidney (K) are hybridization controls. Tissues were hybridized with 32Plabelled oligodeoxyribonucleotides specific for mGK-6 tu, c), mGK-9 (b), mGK-22 (d) or mGK-21 tf) mRNAs. (a, h) Autoradiographs from serial sections of a set of tissues from 22-day-old mice. Bar = 4 mm. (c, d) Autoradiographs from serial sections of a set of tissues from adult female mice and a male control. Bar = 6 mm. (e) Sections from another set of treated adult female and control mice. (f) The autoradiograph obtained after hybridization of the sections shown in (e). Bar = 8 mm. Each autoradiograph represents one frozen section of 5-6 tissues embedded together, as shown in te). Each autoradiograph was exposed optimally so that signals over each tissue could be compared, thus comparisons between signal intensities of autoradiographs rr, b, c, d, and f are not relevant. The identification of tissues shown in ((0 applies also to (/I), in tc) applies to Cd) and in (e) applies to

d

143

mGK-21 expression in the testis Using the common kallikrein screening probe, we observed hybridization to Leydig cells of the testis in mature Swiss male mice. Stringent hybridization with the 12 gene-specific kallikrein probes identified this mRNA as mGK-21 (Fig. 7). None of the other gene-specific probes hybridized to these cells under the conditions employed. Discussion Expression of mGK-6 in adenohypophysis and renal distal convoluted and connecting tubules corresponds with sites of kininogenase expression in rat (Figueroa, 1984; Chao et al., 1987; Wines et al., 1989). However, there are differences between our results and reported sites of expression

in rat kidney. We were unable to detect mGK-6 mRNA at vascular poles as Xiong et al. (1989) showed in the rat with a ““S-labelled rat renal kallikrein cDNA probe. The signals they obtained at the vascular poles were stated to be stronger than those over cortical distal tubules. As we also showed strong signals over distal convoluted tubules in the mouse kidney (Fig. l), the absence of labelling over vascular poles cannot be explained simply by possible differences in sensitivity of their ‘“S-labelled cDNA versus our 3H- or 32P-labelled oligodeoxyribonucleotide probes. The accurate resolution of the ‘H-labelled oligomer showed mGK-6 expression in the cortical thick ascending limb at a much lower level than in distal convoluted tubules (Fig. 1). This site of kininogenase gene expression was not reported in other studies or species. Expression in

Fig. 7. Frozen sections of adult mouse testis after hybridization with “*P-labelled oligodeoxyribonucleotide probes, (a, b) 25mer, corresponding to mGK-21 mRNA; (c, d) 30mer, corresponding to mouse rhodopsin mRNA. Bar = 0.4 mm. (a and b) show the same field photographed by bright-field (a) or dark-field (b) microscopy. The same applies to (c and d). Due to the small size of the silver grains at this magnification, they are not visible in bright-field photomicrographs. (a, b) Groups of Leydig cells, shown in (b) to contain mGK-21 mRNA, can be located in (a) by corresponding arrows. These cells are situated between the seminiferous tubules (TX (c, d) Control, showing that Leydig cells (arrowed) in fd) have not hybridized to the mouse rhodopsin probe. These can be located in (c) by the corresponding arrow.

144

cells of connecting tubules (not shown), but not in collecting ducts (Fig. 1) corresponds with sites of kallikrein immunostaining in the rat and human (Figueroa et al., 1984, 1988). Another site of mGK-6 expression was in ducts of lateral nasal glands (Fig. 31, which appear to be the murine equivalent of Steno’s gland, first described in the rat by Steno in 1662. The watery secretion from these glands is discharged at the entrance to the nasal airway, thereby adding moisture to inspired air. Morphological studies by Moe and Bojsen-Moller (1971) have shown that nasal glands in the rat are structurally very similar to the major serous salivary glands, with secretory acini and a duct system consisting of intercalated, striated and excretory ducts. The striated and excretory ducts appear to be the sites of mGK-6 expression in the lateral glands in the mouse. Ducts expressing mGK-6 in the mouse nasal septum would appear, from descriptions of this area in rat by Bojsen-Moller (1964) to be excretory ducts which originate in underlying glands in the septum and open into the vestibule. Most sites of mGK-6 expression in mice parallel sites of kininogenase gene expression reported in other species. Although kininogenase mRNA was not identified in nasal glands of other species, an immunoreactive glandular kallikrein with kininogenase activity was measured in nasal secretions from humans challenged with allergens (Baumgarten et al., 1986). The site of secretion of this kallikrein was not identified and the lack of structural similarity between the human and rodent nose precludes direct comparison with the mouse. However, considering the structure of the human nose, the seromucous glands and ducts underlying the respiratory mucosa are the most likely source of the nasal kininogenase. Expression of mGK-6 at a high level in a 12 h neonatal mouse (Fig. 3) raises the question of a possible role for this kininogenase in fetal life. Analysis of patterns of expression of the 12 functional glandular kallikrein genes through puberty in salivary glands of mice showed that expression of only one, mGK-6, was independent of the androgen-mediated differentiation of striated ducts to granular convoluted tubules which occurs at puberty. This corresponds with the rat gene family, in which only the renal kininogenase

gene has been shown to be androgen-independent (Clements et al., 1988). We had located previously expression of this gene in striated ducts of the submandibular and sublingual glands (van Leeuwen et al., 1987). Further analysis shows mGK-6 to be expressed at varying levels throughout the duct trees of submandibular, sublingual and parotid glands, with the exception of intercalated ducts (Fig. 4). mRNA levels in the submandibular main duct were considerably lower than in striated ducts (Fig. 4~). The maintenance of the GCT component of the submandibular gland in the male mouse by testosterone and thyroxine has been shown by the extensive regression which occurs after castration or thyroidectomy (Chretien, 1977). Although GCT development in the female is not as extensive as in the male, it also begins at puberty, which suggests regulation by reproductive hormones. Morphological studies have shown that ovariectomy reduces the level of granulation in GCTs and estrogen replacement increases granulation beyond that of controls (Harvey, 1952). This is consistent with the morphological appearance of submandibular tubules and the elevation in expression of the 11 kallikrein genes induced by estradiol treatment (Fig. 6). The failure of estradiol treatment to induce expression in the submandibular gland of prepubertal mice may be due to a lack of estrogen receptors on striated duct cells of mice 17-22 days old. The induction of expression of responsive genes in the adult female submandibular gland by DES, comparable to a low-level induction by estradiol (Fig. 6f), suggests either that estradiol is not acting through androgen receptors, or only partially, as DES is known not to elicit androgenic effects. The androgen-dependent pattern of expression of the 11 hormone-inducible genes in the submandibular gland is similar to that shown for several rat kallikrein genes (Clements et al., 1988), one of which (Pl) has been detected also in the testis of the mature rat (Clements et al., 1990). This rat gene has not been characterized and the cellular site of expression has not been reported. The gene mGK-21, which shows a similar pattern of androgen-dependent expression in the submandibular gland as well as testicular expression in Leydig cells (Fig. 71, may be the mouse equiva-

145

lent of rat Pl. It is interesting to speculate on the substrate for the protein encoded by mGK-21. It may be pro-opiomelanocortin (POMC), which is also synthesized by Leydig cells both in mice (Gizang-Ginsberg and Wolgemuth, 1985) and rats (Pintar et al., 1984). There is evidence that in rat pituitary neurointermediate lobe, one or more glandular kallikreins may act as processing enzymes for POMC, as kallikrein activity is coordinately regulated with POMC synthesis (Powers, 1986) and also with the activity of a known POMC processing enzyme (Millington et al., 1986). At least one of the rat glandular kallikrein genes expressed in the neurointermediate lobe is unidentified as yet (Clements et al., 1989). With our screening probe we were unable to detect glandular kallikrein mRNAs in the neurointermediate lobe of mouse pituitary sections; however, we obtained a weak signal in the anterior lobe, similar to that shown in Fig. 2 for mGK-6. Perhaps in mouse pituitary, as in rat (Chao et al., 19871, there is a lower level of kallikrein mRNA in the neurointermediate lobe than in the anterior lobe, which was undetectable in our sections. The developmental patterns of expression of the five characterized glandular kallikrein genes in rat are similar to those in mouse, in that the kininogenase is apparently the only gene expressed prior ,to the elevation of testosterone levels associated with puberty. Premature induction of expression in the submandibular gland by testosterone or thyroxine depends upon the cellular site of expression. The 11 inducible genes are expressed in GCT cells, which develop at puberty, whereas the striated duct cells expressing mGK-6 develop at an earlier age. The same sexual dimorphism of submandibular gland structure is also a feature of the rat. The cellular site of expression of the one androgen-independent and four androgen-dependent genes in the rat is presumably the basis for this characteristic and for their developmental expression profiles. Further studies of cellular sites of expression, hormonal responsiveness and developmental profiles of glandular kallikrein genes should clarify the identities of genes which have little sequence homology across species and for which we have no basis for comparison at present.

Acknowledgements We thank Bruce Thorley and Scott Pownall for labelling of oligodeoxynucleotides, Paula Darling for excellent technical assistance and Dr. Daine Alcorn from the Department of Anatomy, University of Melbourne for expert advice. This work was supported by grants-in-aid from the National Health and Medical Research Council of Australia, the Myer Family Trusts, the Ian Potter Foundation and the Howard Florey Biomedical Foundation. References Baumgarten, C.R., Nichols, R.C., Naclerio, M. and Proud, D. (1986) J. lmmunol. 137, 1323-1328. Bojsen-Moller, F. (1964) Anat. Rec. 150, 11-24. Chao, J., Chao, L., Swain, CC., Tsai, J. and Margolius, H.S. (1987) Endocrinology 120, 475-482. Chapdelaine, P., Paradis, G., Tremblay, R. and Dub& J.Y. (1988) Fed. Eur. Biochem. Sot. 236, 205-208. Chretien, M. (1977) lnt. Rev. Cytol. 50, 333-396. Clements, J.A., Fuller, P.J., McNally, M., Nikolaidis, I. and Funder, J.W. (1986) Endocrinology 268-273. Clements, J.A., Matheson, B.A., Wines, D.A., Brady, J.M., MacDonald, R.J. and Funder, J.W. (1988) J. Biol. Chem. 263, 16132-16137. Clements, J.A., Matheson, B.A., MacDonald, R.J. and Funder, J.W. (1989) J. Neuroendocrinol. 1, 199-203. Clements, J.A., Matheson, B.A. and Funder, J.W. (1990) J. Biol. Chem. 265, 1077-1081. Coghlan, J.P., Aldred, P., Haralambidis, J., Niall, H.D., Penschow, J.D. and Tregear, G.W. (1985) Anal. Biochem. 149, l-28. Coghlan, J.P., Penschow, J.D., van Leeuwen, B.H. and Richards, R.I. (1989) in Recent Advances in Endocrinology and Metabolism, Vol. 3 (Edwards, C.R.W. and Lincoln, D.W., eds.), pp. l-27, Churchill Livingstone, Edinburgh. Drinkwater, C.C. and Richards, R.I. (1987) Nucleic Acids Res. 15, 10052-10053. Drinkwater, C.C., Evans, B.A. and Richards, R.I. (1987) Biochemistry 26, 6750-6756. Drinkwater, C.C., Evans, B.A. and Richards, R.I. (1988a) J. Biol. Chem. 263, 8565-8568. Drinkwater, C.C., Evans, B.A. and Richards, R.I. (1988b) Trends Biol. Sci. 13, 169-172. Evans, B.A. and Richards, R.I. (1985) EMBO J. 4, 133-138. Evans, B.A., Drinkwater, C.C. and Richards, R.I. (1987) J. Biol. Chem. 262, 8027-8034. Fahnestock, M., Brundage, S. and Shooter, E.M. (1986) Nucleic Acids Res. 14, 4828-4835. Figueroa, C.D., Caorsi, I. Subiabre, J. and Vio, C.P. (1984) J. Histochem. Cytochem. 32, 117-121.

146

Figueroa, C.D., MacIver, A.G., Mackenzie, T.C. and Bhoola, K.D. (1988) Histochemistry 89, 437-442. Fritz, H., Fiedler, F., Dietl, T., Warwas, M., Truscheit, E., Kolb, G., Mair, G. and Tschesche, H. (1977) in Kininogenases, Vol. 4 (Haberland, G.L., Rohen, J.W. and Suzuki, T., eds.), pp. 15-25, Schattauer, Stuttgart. Fukushima, D., Kitamura, N. and Nakanishi, S. (1985) Biochemistry 24, 8037-8043. Gerald, W.L., Chao, J. and Chao, L. (1986) Biochim. Biophys. Acta 867, 16-23. Gizang-Ginsberg, E. and Wolgemuth, D.J. (1985) Dev. Biol. 11 I, 293-305. Harvey, H. (1952) Physiol. Zool. 25, 205-222. Mason, A.J., Evans, B.A., Cox, D.R., Shine, J. and Richards, RI. (1983) Nature 303, 300-307. Kriz, W. and Kaissling, B. (1985) in The Kidney: Physiology and Pathophysiology (Seldin, D.W. and Giebisch, G., eds.), pp. 265-306, Raven Press, New York. Millington, W.R., O’Donohue, R.L., Chappell, M.C. and Roberts, J.L. (1986) Endocrinology 118, 2024-2029. Moe, H. and Bojsen-Moller, F. (1971) J. Ultrastruct. Res. 36, 127-148. Penschow, J.D., Haralambidis, J., Aldred, P., Tregear, G.W. and Coghlan, J.P. (1986) Methods Enzymol. 124, 534-548.

Penschow, J.D., Haralambidis, J., Darling, P.E., Darby, I.A., Wintour, E.M., Tregear, G.W. and Coghlan, J.P. (1987) Experientia 43, 741-750. Penschow, J.D., Haralambidis, J., Pownall, S. and Coghlan, J.P. (1989) Methods Neurosci. 1, 222-238. Pintar, J.E., Schachter, B.S., Herman, A.B., Durgerian, S. and Krieger, D.T. (1984) Science 225, 632-634. Powers, C.A. (1986) Neuroendocrinology 43, 368-376. Saed, G.M., Carretero, O.A., MacDonald, R.J. and Scicli, A.G. (1990) Circ. Res. 67, 510-516. Steno, N. (1662) in Observationes Anatomicae, Lugduni Batavorum. Swift, G.H., Dagorn, J.-C., Ashley, P.L., Cummings, S.W. and MacDonald, R.J. (1982) Biochemistry 79, 7623-7627. van Leeuwen, B.H., Evans, B.A., Tregear, G.W. and Richards, RI. (1986) J. Biol. Chem. 261, 5529-5535. van Leeuwen, B.H., Penschow, J.D., Coghlan, J.P. and Richards, RI. (1987) EMBO J. 6, 1705-1713. Wines, D.R., Brady, J.M., Pritchett, D.B., Roberts, J.L. and MacDonald, R.J. (1989) J. Biol. Chem. 264, 7653-7662. Xiong, W., Chao, L. and Chao, J. (1989) Kidney Int. 35, 1324-1329.