Differentiat ion
Differentiation (1989) 40:93-98
Ontogeny and Neoplasia
0 Springer-Verlag1989
Appearance of acinar-cell-specific mucin in prenatal mouse submandibular glands Patricia A. Denny, Watchara Pimprapaiporn, Barbara J. Bove, Yang Chai, Maureen S. Kim, and Paul C. Denny* Department of Basic Sciences, School of Dentistry, University of Southern California, Los Angeles, CA 90089, USA Abstract. The appearance of an acinar-cell-specific mucin was studied during fetal mouse submandibular gland development. The mucin was first detected in stage 23 and was quantitated through birth by radioimmunoassay (RIA). Quantitation results showed that the mucin accumulation was biphasic. Results from Western blotting and radioimmunoassay indicated that the mucin from the prenatal glands was similar both antigenically and in size to the mucin isolated from adult mice. Observations from light microscopy revealed a continuing progression of complexity throughout prenatal development, indicative of morphogenesis characteristic of differentiating exocrine tissues. When sections from various stages were compared morphometrically, it became clear that the overall ratio of epithelial cells to mesenchymal cells increased nearly 6-fold throughout the prenatal stages observed. The study suggests that acinar cell development in the mouse submandibular gland passes through a protodifferentiated stage. The proportions of epithelial and mesenchymal cells in the submandibular gland and the sensitivity of the RIA indicate that the mucin per cell actually increased to detectable levels at the onset of protodifferentiation, and this increase does not reflect a change in the relative proportions of epithelial and mesenchymal cells.
Introduction The developmental process for exocrine organ systems requires morphogenesis and cellular differentiation. Morphogenesis in the submandibular gland begins with a solid cord of epithelial cells, which subsequently branches to form the architecture of the secretory system. Studies of differentiation of acinar cells in the rat pancreas [20] have shown a biphasic pattern for the accumulation of cell-specific protein. The first phase, protodifferentiation, is characterized by a constant low level of acinar-cell-specific protein. The gland then passes through a transition to the cytodifferentiation phase, and culminates in cellular differentiation. This second phase results in a several-fold increase in the concentration of the cell-specific protein [I 71. Mouse submandibular mucin has been localized in the acinar cells of the adult mouse submandibular gland [8, 231. It has been recently reported that this mucin is present
* To whom offprint requests should be sent
in both proacinar cells and in secretory terminal-tubule cells of neonatal submandibular glands, prior to the differentiation of acinar cells [ll]. Proacinar cells are likely to be the immediate precursor of acinar cells in the neonatal gland, while the secretory terminal-tubule cell may represent an earlier stage in the sequence of differentiation of acinar cells [ 111. The biphasic accumulation of secretory proteins in the acinar cells of the pancreas [17, 211 and parotid [I], and the detection of mucin in the neonatal mouse submandibular gland [I 11, suggest that the acinar-cell-specific mucin may be present in the prenatal submandibular gland, perhaps even prior to cytodifferentiation. In this study we followed the pathway of acinar cell differentiation in the prenatal mouse submandibular gland by monitoring the appearance of and quantity of mucin, and correlated these results with histological observations. Methods Materials. Pregnant Swiss-Webster mice from Simonsen Labs (Gilroy, California) were sacrificed using sodium pentobarbital. Fetuses were quickly removed and kept on ice. Fetuses were staged by external features according to the methodology of Theiler I221 because there was as much as a three-stage variation within the same litter. The submandibular glands were carefully dissected, removed, washed in phosphate-buffered saline (PBS) and frozen at - 80" C until sufficient numbers of glands from each stage had been collected. Prior to stage 25, removal of the sublingual gland rudiments was attempted but could not be assured. No attempt was made to determine sex. Electrophoretically pure mouse submandibular sialomucin was prepared from adult females as previously described [lo]. Finally, dialyzed mucin following preparative polyacrylamide gel electrophoresis was purified by high-pressure liquid chromatography (HPLC) on a TSK-400 column (Bio-Rad, Richmond, California). Anti-mucin sera were raised in female New Zealand white rabbits by intradermal injections of purified mucin [8]. Affinity-purified antisera were prepared by utilizing an Afti-Gel 15 column (BioRad) . Assays. Pooled submandibular glands from each stage were homogenized in 0.01 M Tris (pH 7.6 at 25" C), 0.25 M NaCl
and prepared as previously described [I I]. Quantitation of mucin in the homogenate was performed by radioimmuno-
94
assay (RIA) [9]. The protein content of the homogenates was determined using bovine serum albumin as the standard [14], while DNA was quantitated fluorometrically using the fluorochrome Hoeschst 33258 [18]. Immunoblotting. Electrophoretically pure mouse submandibular mucin was purified from adult females and served as the standard mucin. Newborn submandibular glands were homogenized and centrifuged at 50000 g . The resulting s-50 was used for polyacrylamide gel electrophoresis (PAGE) without further purification. Glands from stages25 and -26 prenatal mice had to be purified further as the S-50 had a very low mucin to protein ratio. The S-50 was chromatographed using Sephacryl S-200 and subsequently dialyzed against 0.1 M sodium acetate (pH 4.85). Prepared samples were applied to 3.75% acrylamide sodium dodecyl sulfate (SDS)-urea tube gels for electrophoresis [lo]. The separated proteins were electrophoretically transferred at 200 mA for 4 h in 0.025 M Tris (pH 8.3 at 25" C), 0.2 M glycine from tube gels to Nylon 66 Plus membranes using a commercially available transfer apparatus (Hoefer Scientific Instruments, San Francisco, Calif.). The nylon membranes were blocked overnight at 45" C using 10% bovine serum albumin (BSA) in a Tris-buffered saline solution (TBS) containing 0.05 M Tris, pH 7.5 at 25" C; 0.2 M NaCI. Each membrane was reacted with affinity-purified antiserum diluted 1 :SO0 in TBS containing 1% BSA for 2 h at room temperature with gentle agitation. Controls were reacted with both normal rabbit serum and normal rabbit IgG. After two 10-min washes in TBS, we applied the second antibody, alkaline phosphatase-conjugated goat F(ab)2 fragment prepared against rabbit IgG (Cooper Biomedical, Malvern, Pennsylvania) for 60min at a dilution of 1 : 150. Following two 10-min washes in TBS and two 10-min washes in a buffer containing 0.1 M Tris (pH 9.3, 0.1 M NaCI, 0.005 M MgC12, the membranes were incubated in BCIP/NBT substrate (Bethesda Research Labs, Bethesda, Maryland) in the dark while monitoring color development. Tissue processing. Submandibular glands to be used for histology and immunofluorescent studies were removed, and rinsed quickly in 0.1 M sodium cacodylate buffer for 30 s and rapidly minced into approximately l-mm3 pieces. The tissues were then fixed by immersion in 4% formaldehyde (prepared immediately before use) and 0.5% glutaraldehyde, buffered to pH 7.2, in 0.1 M sodium cacodylate at 4 °C [12]. Tissue dehydration and embedding using LRGold resin are described in Denny et al. [ll]. A minimum of four glands for each stage were embedded within the
same block. Blocks were cut at 1 pm using a Reichert-Jung Ultracut E ultramicrotome. Staining procedure and immunofluorescence. For visualization, sections were stained with 0.5% toluidine blue in 0.1 M phosphate buffer for 10 min at room temperature. Sections were then rinsed with distilled water, air-dried and mounted in Permount (Fisher Scientific, Pittsburgh, Pennsylvania). Sections for immunofluorescence were prepared essentially the same way as previously reported, (1 11 except that affinity-purified rabbit anti-serum to mouse submandibular gland mucin was additionally utilized at dilutions of 1 :20 and 1 :40 in 0.05 M Tris (pH 7.9, 0.15 M NaCI, 0.005 M ethylenediamine tetraacetic acid, 0.02% sodium azide and 0.05% NP-40. Morphametry. Sections from prenatal mice, which were staged 22 through 25, were stained with hematoxylin and eosin. Four randomly chosen submandibular gland crosssections from each stage were photographed and reproduced at x 562. The numbers of nuclei from epithelial and mesenchymal cells were counted and tabulated for each cross-section. The cross-sectional areas for epithelium and those areas occupied by mesenchyme were individually determined using a Jandel digitizing tablet, which was connected to an IBM PC-XT utilizing Sigma Scan software (Jandel Scientific, Sausalito, Calif.). A minimum of 3668 cells for each stage were counted. For determination of mesenchymal cell size, four randomly chosen submandibular gland cross-sections from each stage were photographed and reproduced at x 1780. The cell size was determined by randomly selecting 20 mesenchymal cells from each section, for a total of 80 cells per stage. Results
Quantitation The values shown in Table 1 confirm that mucin was present in the submandibular glands of prenatal mice from at least stage 23. We were unable to detect any mucin in pooled glands prior to this stage, at a level of assay sensitivity that could detect a minimum of 0.08 ng mucin per gland. Whereas there was a continual rise in the concentration of protein per gland, the mucin per gland showed three successive increases only after stage 24. A relatively stable level of mucin was maintained in stages 23 and 24. From stage 24 to stage 25, there was a 31-fold increase in mucin concentration per gland. From stage 25 to stage 26 there was a 21-fold increase, and again from stage 26 to birth,
Table 1. Quantitation of mucin. protein and DNA in pooled submandibular glands from prenatal stages through birth. Mucin values are means fstandard errors of the mean of assay error
Newborn Stagc26 Stage25 Stage24 Stage 23 Stage 22
Pooled gland no.
Mucin (ng)/gland
Protein (pg)/gland
DNA (pg)/gland
Mucinlprotein (w/w x 100)
Mucin/DNA (w/w x 100)
16 94 80 56 137 150
2340 +SO 301.5 k14.3 14.22& 0.90 0.46k 0.12 0.33+ 0.09 -
198 178.7 142.9 34.4 15.7 8.3
54.14 30.79 23.34 5.32 3.58 2.22
1.20% 0.168% 0.010%
4.3% 0.979% 0.061% 0.009% 0.009 Yo
0.001Yo
0.002% -
-
95 Yucin standard (ng) 3.13 6.25
12.5 25
50
12.5 6.25
3.13 1.57
A
B
C
D
E
0 50
-
U
c
240-
A
0)
eD
5
30-
g
n 20
50
25
Serial dilution of homogenatc
Fig. 1. Comparison by radioimmunoassay (RIA) of purified much (B) with submandibular gland homogenates from newborn mice (A) and stage 25 (C). The abscissa displays dilutions of homogenates for newborn and stage 25 tested in the RIA; the ordinate denotes the data obtained from the RIA. The newborn was tested at four appropriate serial dilutions and stage 25 was tested at three serial dilutions. Each dilution point represents the mean for at least duplicate assays. Only when the standard error of the means were greater than f 2 % of the percent bound, were they shown on the graph. The largest standard error obtained was *4.2% of the percent bound
another 8-fold increase. From the first detectable appearance of mucin at stage 23 until birth, there was a dramatic overall increase of much per gland of 7200-fold. The increase in the ratio of mucin per protein was 600-fold, with the largest gain occurring between stages 25 and 26. The concentration of mucin per cell increased 472-fold overall, with the largest increase seen again between stages 25 and 26. Serial dilutions of submandibular gland homogenates from stage-25, newborn and adult purified mucin were quantitated by RIA (Fig. 1). By comparing the resulting slopes, the slopes of the mucin from newborn (Fig. 1A) and stage 25 (Fig. lC), were very similar to the slope obtained from purified mucin of adult females (Fig. 1B). Thus it can be concluded that the mucins from the various ages were antigenically similar [I 13. Immunoblotting
The migration on SDS-polyacrylamide gels of purified mucin obtained from adult submandibular glands (Fig. 2 A) was similar to the migration of mucin identified by immunoblotting. Results from Western blotting indicated that the mucin purified from adult submandibular glands (Fig. 2B) was similar antigenically and in charge and size to preparations from newborn (Fig. 2C) and prenatal glands from stages 26 (Fig. 2D) and 25 (Fig. 2E). Control blots showed no reaction, indicating that there was no nonspecific binding. Histology
Observations from light microscopy (Fig. 3 A-E) revealed a progressive complexity occurring in the developing mouse
Fig. 2. Identification of mucin by staining on a sodium dodecyl sulfate (SDS)-urea polyacrylamide gel and by immunoblotting. Samples of adult submandibular glands were electrophoretically purified prior to running on analytical gels. Approximately 10 pg purified adult mucin, or 1/30 of an adult gland, is represented in the polyacrylamide gel stained with Stains-all shown in A and in the immunoblot in B. Mucin from approximately one newborn gland is shown in C.The immunoblot shown in D is 36 submandibuldr glands from stage 26 and E represents 169 submandibular glands from stage 25. Samples for immunoblot (B-E) were first electrophoresed on 3.75% SDS-urea polyacrylamide tube gels, electrophoreticdlly transferred onto nylon filters, and the immunoblotted proteins were reacted with antibodies against mucin. The gel for staining A was washed for 24 h in 7.5% acetic acid to rid the gel of SDS, and then stained with Stains-all
submandibular gland. In stage 22 (Fig. 3A), there was a densely packed mesenchymal capsule surrounding undifferentiated epithelial cells. The epithelial cells were arranged in cords with end buds that were beginning to branch. In a section from stage 23 (Fig. 3 B), the mesenchyme was still densely packed, branching continued and the end buds were smaller. The mesenchyme was becoming less densely packed in stage 24 (Fig. 3C). Lumena formation within the ductal cords was first seen, with the lumen extending near the end buds. Stage 25 (Fig. 3D) showed extensive branching and revealed the first appearance of granules in secretory terminal-tubule cells. By stage 26 (Fig. 3E), secretory terminal-tubule cells and proacinar cells could be distinguished by toluidine blue staining. This stage also revealed the first appearance of secretory granules within the proacinar cells and the first appearance of acini. The level of sensitivity was such that immunolocalization of mucin was not detectable in the prenatal stages observed. Morphometry
The morphometric results shown in Table2 suggest that there are differences with age in both epithelial cells and mesenchymal cells per unit area. However, when tested by a single classification ANOVA between stages, there was a significant difference ( P < O . O l ) only between stages 24 and 25 for the density of epithelial cells within the epithelium. For mesenchymal cells, the only significant change (P
96
Fig. 3A-E. Sections of prenatal mouse submandibular glands stained with toluidine blue: A Stage 22. B Stage 23. C Stage 24. D Stage 25. E Stage 26. c, represents proacinar cells; e , epithelium; m, mesenchyme; f, secretory terminaltubule cell. Light microgruphs x 150. Inserts x 1300
rapid rate than was the mesenchymal cell number. The overall ratio of epithelial cells to mesenchymal cells increased nearly 6-fold throughout the ages observed. While the difference in the ratios of stages 22 and 23 was not significant, the increases in the ratios that occurred between 23 and 24 and also between 24 and 25 were significant (P<0.05). Statistical analysis by single classification ANOVA of the data presented in Table 3 indicated that there was no signif-
icant change in mesenchymal cell size despite the decrease in number of mesenchymal cells per unit area.
Discussion Results from the RIA have shown that acinar-cell-specific mucin is present in the submandibular glands from prenatal mice. The mucin could be detected as early as stage 23.
97
Table 2. Qudntitation of epithelial and mesenchymal cells per unit area from cross-sections of prenatal mouse submandibular glands at various stages. All values are meansf standard errors Stage Epithelial cells/ unit area*
Mesenchyrnal cells/
Epithelial/
unit area**
mesenchyrnal***
22 23 24 25
3.53 f0.20 3.92f0.12 1.36k0.19 1.1 2 k0.09
0.5Ok 0.11 0.58 k0.14 1.63k0.24 2.95 k0.27
5.65k0.37 5.16k0.30 4.58k0.25 3.47k0.13
* Represents cell no. per unit area of epithelium ** Represents cell no. per unit area occupied by mesenchyme *** Total no. of epithelial cells occupying epithelium divided by total no. of mesenchymal cells occupying mesenchyme
Table 3. Comparison of mesenchymal cell cross-sectional areas from various stages of prenatal mouse submandibular glands. All values are means k standard errors Stage
Mesenchymal cells (PZ)*
22 23 24 25
12.08k 1.46 12.55k0.60 12.29k0.55 11.55f0.26
* Values have been reduced to actual size Although stages 22 through 24 may have been contaminated by sublingual tissue, it has been previously shown that the antisera shows no cross-reactivity with proteins from the sublingual gland in the RIA [9]. The pattern of accumulation of mucin with age is similar to the biphasic pattern for other secretory proteins observed in the rat pancreas [17], mouse pancreas [21] and in the rat parotid gland [I]. Though immunofluorescence is not sensitive enough to localize the mucin to a specific cell type, results from serial dilutions in the RIA and Western blotting provide additional evidence that the mucin, which is present in the prenatal submandibular glands, is similar antigenically and in size and charge to the mucin isolated from adult submandibular glands. The similarity in charge and size of mucin from stage 25 compared with the adult suggests that the mucins, share a common level of glycosylation. Thus it is likely that the glycosylation apparatus is functional by the onset of cytodifferentiation, prior to the appearance of proacinar cells. Differentiation of the submandibular gland is more protracted than that of the pancreas, with the submandibular gland reaching maturity at 3-4 months of age [16]. Postnatal development occurs more rapidly in the mouse submandibular gland than in the rat [IS]. This may be true for prenatal development as well, since the mouse pancreas emerges 1-2 days earlier than in the rat [19]. Though the correlation of exact staging of submandibular gland development between mice and rats has not been undertaken, it is important to note that the sequence of morphological events that we have observed in the mouse at the level of the light microscope does parallel those observed in the rat [6, 7, 251. The anlage of the submandibular gland first appears on day 14 of gestation in the rat [6] and stage
21 in the mouse. Branching is initiated in the bulb portion of the rudiment on day 15 in the rat and stage 22 in the mouse, and continues on day 16 in the rat and stage 23 in the mouse. At the ultrastructural level, Cutler and Chaudhry [7] reported that the appearance of the rough endoplasmic reticulum was consistent with the pattern displayed during protein synthesis, though no secretory granules were apparent on either day 16 or 17. Although we present no ultrastructural data, it is of interest that this interpretation may be consistent with the biochemical detection of mucin first appearing at stage 23 and remaining relatively stable through stage 24. While observations in the 3 %day prenatal rat reveal the coincident appearance of lumena and of secretory granules, which are characteristic of secretory tenninal-tubule cells [7, 241, only lumena formation was seen in the stage-24 mouse. Secretory granules were not observed in terminal-tubule cells until stage 25 in the mouse. During the 20th day of gestation in the rat (stage 26 for the mouse), a second granule type appeared, which was indicative of proacinar cells [7]. This appearance occurred 2 days prior to birth in the rat and the day prior to birth in the mouse. The appearance of granules is consistent with high-level synthesis, which is characteristic of the cytodifferentiated phase. In the pancreas, there was no cytological correlation associated with the protodifferentiated state [19]. The uniform appearance of exocrine cells in the protodifferentiated state and the absence of zymogen granules during this period support the view that low levels of cell-specific protein are not the result of a small number of differentiated cells [19, 251. In this study, the increase in DNA per gland from stage 22 until birth, together with the change in the ratio of epithelial cells compared with mesenchymal cells, indicates that not only is there an increase in total cell number but that the epithelial cells increase at a more rapid rate. It is important to note that the ratio of epithelial to mesenchymal cells was the same in stage 22 and 23, yet we were unable to detect mucin in glands from stage 22. Thus, the onset of the protodifferentiated phase appears to be signaled by an increase in the mucin per cell and not by an increase in the number of one cell type relative to the other. The drop in mucin per protein between stage 23 and stage 24 may indicate asynchronous synthesis of mucin and other proteins by the protodifferentiated acinar cells. Mesenchyme plays an important role in the induction of epithelial morphogenesis in the mouse submandibular gland [2, 131. Cutler and Chaudhry [6] observed direct epithelial-mesenchymal cell-to-cell contact or epithelial-nerve contacts during day 15 and day 16 when primary branching had already occurred in the prenatal rat submandibular gland, but prior to the onset of cytodifferentiation. These contacts were apparent only at the interface of developing end buds, areas which give rise to acinar cells, myoepithelial cells, intercalated ducts and convoluted granular ducts. Thus, the investigators hypothesized that these epithelialmesenchymal interactions may influence epithelial cytodifferentiation. In the mouse submandibular gland, epithelialmesenchymal cell-to-cell contacts were seen to occur during this same phase of morphogenesis [3]. Further work by Cutler [4] indicated that two types of epithelial-mesenchyma1 interactions were necessary to attain a fully differentiated gland. The first requires a prolonged period of interaction to initiate and maintain branching [13]. The second interaction described above is more transient in nature, last-
98
ing for only 2 days, but is critical for cytodifferentiation to occur [4]. It was also shown that following the critical cell-to-cell interaction, cytodifferentiation but not morphogenesis could proceed in the absence of mesenchyme. Spooner earlier observed that once pancreatic acinar cells were protodifferentiated, then cytodifferentiation can proceed in the absence of mesenchyme [21]. More recently, Culter has shown that ongoing synthesis of extracellular matrix is not a requirement for epithelial cell cytodifferentiation [5]. This study helps to unify these observations by suggesting that the protodifferentiated phase in the mouse submandibular gland occurs during stages 23 and 24. During stage 22 of the mouse submandibular gland development, which is equivalent to the first day of the critical cell-to-cell contact period observed for the rat submandibular gland [6], mucin was not detectable. One day later, during the latter half of the critical period, mucin was detected. Thus, there is a strong possibility that the critical epithelial-mesenchyma1 cell-to-cell interaction described by Cutler and Chaudhry [6]is the external event triggering the regulatory changes that initiate protodifferentiation. The reduction in density of mesenchymal cells, which coincides with the end of the critical period at stage 24, is also consistent with the observation that close contact with mesenchymal cells is no longer required for epithelial cytodifferentiation [4]. Finally, the increase in mucin, which begins at stage 25, is accompanied by the appearance of secretory granules. These provide evidence for the onset of cytodifferentiation. In this study, mesenchymal cell size showed no correlation with events signaling cytodifferentiation; however, this result does not preclude the possible influence of mesenchyma1 cell proteins on protodifferentiation. Acknowledgemenis. This study was supported by National Institutes of Health grant DE-04960.
References 1. Ball WE (1974) Development of the rat salivary glands. IV. Amylase and ribonuclease activity during embryonic and neonatal development of the parotid and submaxillary glands. Dev Biol41: 267-277 2. Borghese E (1950) The development in vitro of the submandibular and sublingual glands of Mus musculus. J Anat 84: 287-302 3. Coughlin M (1975) Early development of parasympathetic nerves in the mouse submandibular gland. Dev Biol43 :123-1 39 4. Cutler LS (1980) The dependent and independent relationships between cytodifferentiation and morphogenesis in developing salivary gland secretory cells. Anat Rec 196:341-347 5. Cutler LS (1988) The role of proteoglycans in morphogenesis and cytodifferentiation of the submandibular gland. J Dent Res 67 :622 6. Cutler LS, Chaudhry AP (1973) Intercellular contact at the epithelial-mesenchymal interface during the prenatal-development of the rat submandibular gland. Dev Biol 33:22Y-240
7. Cutler LS, Chaudhry AP (1974) Cytodifferentiation of the acinar cells of the rat submandibular gland. J Morphol 41:3141 8. Denny PA, Denny PC (1982) Localization of a mouse submandibular sialomucin by indirect immunofluorescence. Histochem J 14:403-408 9. Denny PC, Denny PA (1984) Diurnal study of a sialomucin in female mouse submandibular glands as measured by radioimmunoassay. Arch Oral Biol29: 1033-1040 10. Denny PA, Denny PC, Jenkins K (1980) Purification and biochemical characterization of a mouse submandibular sialomucin. Carbohydrate Res 87 :265-274 1 1 . Denny PA, Pimprapaiporn W, Kim MS. Denny PC (1988) Quantitation and localization of acinar cell-specific much in submandibular glands of mice during postnatal development. Cell Tissue Res 251 :381-386 12. Ellinger A, Pavelka M (1985) Post-embedding localization of glycoconjugates by means of lectins on thin sections of tissues embedded in LR white. Histochem J 17:1321-1336 13. Grobstein C (1953) Epithelio-mesenchymal specificity of the morphogenesis of mouse submandibular rudiments in vitro. J Exp Zoo1 124:383-404 14. Hartree EF (1972) Determination of protein: a modification of the Lowry method that gives a linear photometric response. Anal Biochem 48:422427 15. Jacoby F (1959) Observations on the post-natal developmental of the mouse submaxillary gland. J Anat 93: 579 16. Jacoby F, Leeson CR (1959) The post-natal development of the rat submaxillary gland. J Anat 93:201-216 7. Kemp JD, Walther BT, Rutter WJ (1972) Protein synthesis during thc secondary developmental transition of the embryonic rat pancreas. J Biol Chem 247 13941-3952 8. LaBarca C, Paigen K (1980) A simplc. rapid, and sensitive DNA assay procedure. Anal Biochem 102:344-352 9. Pictet RL, Clark WR, Williams RH, Rutter WJ (1972) An ultrastructural analysis of the developing embryonic pancreas. Dev Biol29:4 3 W 6 7 20. Rutter WJ, Ball WD, Bradshaw WS, Clark WR, Sanders TG (1967) Levels of regulation in cytodifferentiation. In: Hagen E, Wechsler W, Zilliken F (eds) Experimental Biology in Medicine. Karger, Basel Vol 1, pp 110-124 21. Spooner BS. Cohen HI, Faubion J (1977) Development of the embryonic mammalian pancreas : The relationship between morphogenesis and cytodifferentiation. Dev Biol 61 :119-130 22. Theiler K (1972) The House Mouse. Springer Heidelberg Berlin New York 23. Vreugdenhil AP, Nieuw Amerongen AV, De Lange GL, Roukema PA (1982) Localization of amylase and mucins in the major salivary glands of the mouse. Histochem J 14:767-780 24. Yamashina S, Barka T (1972) Localization of peroxidase activity in the developing submandibular gland of normal and isoproterenol-treated rats. J Histochem Cytochem 20: 855-872 25. Yamashin S, Barka T (1973) Development of endogenous peroxidase in fetal rat submandibular gland. J Histochem Cytochem 21 :42-50
Accepted in revised form January 9, 1989