Pancreatic rudiments from E12.5 mouse embryos undergo extensive development and differentiation when cultured in three-dimensional gels of extracellular matrix proteins for up to 12 days. Whereas collagen gels promote the formation of numerous exocrine acini and relatively small clusters of endocrine cells, in basement membrane (EHS) matrices the development of endocrine cells is dramatically favoured over that of acinar tissue. Buds embedded in a collagen gel contiguous to an EHS gel also fail to develop acini, suggesting the involvement of diffusible factor(s). Addition of cytokines to cultures of pancreatic buds in collagen gels modifies the relative proportions of the epithe-lial components of the gland. In the presence of EGF the proportion of the tissue occupied by ducts overrides that of acinar structures, whereas the endocrine portion of the tissue is not significantly modified. TGF-β1 partially mimicks the effect of EHS matrix in inhibiting the devel-opment of acinar tissue without decreasing the amount of ducts and mesenchyme; TGF-β1 also promotes the devel-opment of endocrine cells, in particular of insulin-contain-ing β cells and of cells expressing genes of the PP-fold family. These results show that cytokines can modulate the development of the pancreas and suggest a role for TGF-β1 in regulating the balance between the acinar and endocrine portions of the gland in vivo. More generally, they are compatible with the notion that, during organo-genesis, cytokines act as paracrine factors responsible for the development and maintenance of appropriate proportions of different tissue constituents.

In mammals the pancreas originates from the primitive gut through the fusion of two anlagen, which first appear as dorsal and ventral epithelial evaginations surrounded by caps of mesenchymal cells (Pictet et al., 1972). The original buds give rise to branching epithelial ducts from which clusters of endocrine cells and acini of exocrine cells progressively dif-ferentiate (Pictet and Rutter, 1972). The development of the mammalian pancreas represents an attractive model to study the molecular signals that direct the differentiation of epithe-lial cells along different lineages, since the same endodermal precursor cells are believed to differentiate along two divergent pathways to form the exocrine and endocrine portions of the definitive gland (Le Douarin, 1988). In addition, as yet unidentified mechanisms must operate to ensure that the different constituents of the tissue develop in appropriate proportions.

The identification of the molecular signals that trigger and control the differentiation of the various pancreatic cell types has been hampered by the lack of suitable model systems allowing in vitro manipulation and monitoring of these devel-opmental decisions. In particular, despite increasing knowledge of the mechanisms that regulate the development of other parenchymal organs such as salivary and mammary glands, lung and metanephric kidney (Bernfield et al., 1984; Schuger et al., 1990b; Nogawa and Takahashi, 1991; Weller et al., 1991; Warburton et al., 1992), very little is known about the regulatory factors that control the development of pancre-atic acini or islets of Langerhans. A mesenchymal factor that enhances the proliferation of exocrine and endocrine cells in pancreatic epithelial rudiments was described more than 20 years ago, but it has been only partially characterized (Ronzio and Rutter, 1973).

Culture of epithelial cells in reconstituted collagen or basement membrane matrices has been shown to promote his-totypic organization and functional differentiation in a variety of systems (see for example, Yang et al., 1980; Chambard et al., 1981; Montesano et al., 1983, 1991; Barcellos-Hoff et al., 1989). We have therefore embedded pancreatic rudiments in three-dimensional matrices and examined their fate during in vitro culture. In contrast to gels of type I collagen, which allowed the development of all pancreatic constituents, basement membrane (EHS) matrices selectively stimulated the development of the endocrine pancreas. Since this effect was due to a diffusible factor, we explored the effect of adding exogenous growth factors to pancreatic buds embedded in collagen gels. We have found that EGF affects the relative development of the different exocrine constituents of the pancreas, and that TGF-β1 enhances the formation of endocrine cells while completely inhibiting the development of acinar tissue. These results suggest an important modulatory role for cytokines and in particular TGF-β1 in pancreatic development and differentiation.

Culture of pancreatic rudiments in collagen gels

NMRI mice were mated overnight and the morning when the vaginal plug appeared was defined as day 0.5 of gestation (E 0.5). Pregnant mice with embryos of gestational age E11.5 or E12.5 were killed by cervical dislocation. Embryos were removed from the uterus and placed in Hank’s balanced salt solution (Gibco, Basel, Switzerland). Dorsal or ventral pancreatic buds were dissected, washed in sterile culture medium and embedded in three-dimensional collagen gels as follows: 8 volumes of rat tail collagen stock solution (approximately 1.5 mg/ml) were quickly mixed with 1 volume of 10× concentrated minimal essential medium (Gibco, Basel, Switzerland) and 1 volume of sodium bicarbonate (11.76 mg/ml) in a sterile flask kept on ice (Montesano et al., 1983). 200-300 µl of the solution were dispensed into 15 mm wells of four-well plates (NUNC) and allowed to gel at 37°C; pancreatic buds were placed on the surface of the collagen gel and covered with a further 200-300 µl of collagen solution. After the second layer of collagen had gelled, complete medium (400-600 µl per well), consisting of RPMI 1640 (Gibco, Basel, Switzerland) sup-plemented with 4 mM glutamine, 100 U/ml penicillin, 110 µg/ml streptomycin and 5.5 mM glucose, was added. When indicated, 10% heat-inactivated fetal calf serum (HI-FCS) was added.

Growth factors were added to both the collagen gels and the medium. Porcine platelet transforming growth factor-β1 (TGF-β1) was purchased from R & D Systems Europe (Oxon, UK); recombi-nant human epidermal growth factor (EGF) and porcine platelet-derived growth factor (PDGF) were from Boehringer Mannheim (Rotkreuz, Switzerland); nerve growth factor (NGF) was a generous gift from Dr L. Aloe (Istituto di Neurobiologia, CNR, Italy); rat insulin-like growth factor-II (IGF-II), human recombinant basic fibroblast growth factor (bFGF) and recombinant human hepatocyte growth factor (HGF) were kindly provided by Dr N. Yanaihara (Lab. of Bioorganic Chemistry, Shizuoka-shi, Japan), Dr P. Sarmientos (Farmitalia Carlo Erba, Milano, Italy) and Dr T. Nakamura (Bio-medical Research Center, Osaka, Japan), respectively.

Medium and growth factors were replaced every 48 hours and the buds were fixed, at the indicated times, in 2% glutaraldehyde in 0.1 M phosphate buffer or in Bouin’s solution, postfixed in osmium tetroxide, dehydrated and embedded in Epon 812, as described pre-viously (Bendayan et al., 1980).

Preparation of basement membrane-like matrix

The basement membrane-like matrix was obtained from the Engel-breth-Holm-Swarm (EHS) tumor, that was initially provided by Dr J.-M. Foidart (Liège, Belgium) and subsequently maintained in our lab- oratory by subcutaneous or intramuscular inoculation in C57 black mice. The procedures described by Kleinman et al. (1986) and Kramer et al. (1986) were followed with slight modifications. Briefly, 3-4 weeks after inoculation, the tumors were harvested and homogenized in 3.4 M NaCl, 50 mM Tris-HCl, pH 7.4, containing protease inhibitors (10 mM EDTA, 2 mM N-ethylmaleimide, and 1 mM phenylmethylsulfonylfluoride). After centrifugation at 500 g for 10 minutes, the supernatant was discarded, the pellet was resuspended in Tris-HCl buffer with protease inhibitors (see above), and the washing procedure was repeated three more times. The homogenate was then centrifuged at 30,000 g for 30 minutes, the supernatant discarded and the pellet resuspended in an equal volume (1 ml/g) of freshly prepared 2 M urea in Tris-HCl buffer with protease inhibitors. The suspension was stirred overnight at 4°C, and the insoluble material was removed by centrifugation at 30,000 g for 1 hour. The supernatant was dialyzed for 24-48 hours at 4°C under sterile conditions, first against 0.15 M NaCl in 50 mM Tris-HCl, and then overnight against a 1:10 dilution of MEM with Earle’s salts (Gibco). The dialyzed extract was centrifuged at 30,000 g at 4°C for 30 minutes to remove small amounts of insoluble material and the supernatant frozen at ‐20°C in small aliquots. For gel formation, 8:1:1 volumes of EHS extract, 10× MEM and sodium bicarbonate (11.76 mg/ml) were mixed on ice, dispensed into culture wells and allowed to gel for about 30 minutes at 37°C before adding complete culture medium. Commercially available EHS matrix (Matrigel™) was purchased from Collaborative Research Incorporated (Bedford, MA) and gelled by incubation at 37°C for 30 minutes. Pancreatic rudiments were suspended in EHS matrix or Matrigel before gelation occurred.

Immunocytochemistry and electron microscopy

Consecutive semithin sections (1 µm thick) of non-osmified pancre-atic buds were collected on glass slides. The Epon was removed (Maxwell, 1978) and the sections incubated with antibodies (see below) for 2 hours, washed 5 minutes, incubated with FITC-conju-gated anti-IgG antibodies for 1 hour at room temperature and coun-terstained with 0.003% (w/v) Evans blue. Sections were examined and photographed with a Zeiss Axiophot epifluorescence micro-scope.

Rabbit anti-porcine glucagon serum directed against the C-terminal region of glucagon was provided by Dr R. H. Unger (Dallas); guinea pig anti-porcine insulin serum by Dr P. Wright (Indianapolis); rabbit anti-bovine pancreatic polypeptide (anti-bPP) serum by Dr R. E. Chance (Indianapolis); rabbit anti-synthetic somatostatin sera by Dr R. Guillemin (San Diego, CA) and by Dr Y. Patel (Montreal, Canada). Rabbit anti-human pancreatic polypeptide (anti-hPP) and rabbit anti-synthetic porcine peptideYY (anti-PYY) were purchased respectively from Peninsula Laboratories (Belmont; CA) and Euro-Diagnostica (Malmö, Sweden); mouse monoclonal anti-BrdU IgGs were from Becton Dickinson & Co. (San José, CA). FITC-conjugated sheep IgG against guinea pig IgG, FITC-conjugated goat IgG against rabbit IgG, and FITC-conjugated rabbit IgG against mouse IgG were purchased from Biosys (Switzerland).

Specificity of anti-bPP, anti-hPP and anti-PYY antibodies was assessed with an ELISA assay (data not shown), using bovine PP, rat neuropeptide Y (NPY), porcine synthetic peptide YY (PYY) and porcine glucagon as antigens. The results indicate that anti-bPP rec-ognizes PP, NPY and PYY, but not glugagon; anti-bPP thus reveals all three members of the PP-fold family (Hazelwood, 1993). Anti-hPP and anti-PYY are specific for PP and PYY, respectively.

Electron microscopy was performed on ultrathin sections of Epon-embedded osmified samples, collected on 150-mesh copper grids and stained with uranyl acetate and Reynold’s lead citrate. Ultrathin sections of Epon-embedded non-osmified samples were mounted on nickel grids, and immunolabelled by the protein A-gold method (Roth et al., 1978). Sections were incubated for 2 hours at room tempera-ture with a 1:100 dilution of guinea pig anti-porcine amylase serum (Bendayan et al., 1980); the grids were then rinsed in distilled water and incubated for 1 hour at room temperature on a drop of a solution of protein A coupled to 10 nm gold particles and double stained with uranyl acetate (20 minutes) and Reynold’s lead citrate (10 minutes). All samples were examined using a Philips EM 300 transmission electron microscope.

Morphometry

The volume density (relative volume) of the different tissue compo-nents and of pancreatic hormone-containing cell types was determined using the point-counting method of morphometric analysis (Weibel, 1979) on positive prints. Groups, corresponding to different con-ditions tested, were compared by a Krushal-Wallis test; pairwise com-parisons were performed with the Bonferroni correction (Wallenstein et al., 1980).

To estimate the real volume of cultured buds, 20 rudiments in collagen gels, 26 rudiments in EHS matrix and 10 rudiments in collagen gels in the presence of TGF-β1 were measured after 8-10 days in culture and the following formula was used:
assuming that r2=r3. The absolute volumes of the different compo-nents (acinar cells, ducts, endocrine cells, etc.) were estimated using the following formula:

absolute volume = (total pancreatic volume × volume density)/100.

To determine the number of cells undergoing DNA synthesis in cultured buds, a final concentration of 10‐4 M bromo-deoxyuridine (BrdU; Sigma Chemical Co., St.Louis, MO) was added 12 hours before fixation in Bouin’s. The volume density of proliferating cells, demonstrated by anti-BrdU staining, was calculated by using the point-counting method of morphometric analysis (Weibel, 1979) on 3 randomly selected positive prints of at least 3 pancreatic rudiments from each group analyzed. For these two studies, groups were compared by an analysis of variance (ANOVA).

Quantification of apoptosis was performed on pancreatic rudiments cultured for 7 days in collagen gels and further incubated for 24 hours in the presence or absence of 1 ng/ml TGF-β1 (22 buds for each condition). The number of apoptotic bodies immediately adjacent to a measured length of basal acinar contour was counted, and the ratio between the number of apoptotic bodies and the length of acinar contour referred to as the frequency of apoptotic bodies. This parameter was determined on 10 randomly selected electron micro-graphs for each bud. Statistical evaluation was performed on the log-arithms of the frequencies by a nested ANOVA.

RNA extraction, cDNA synthesis and PCR

Total RNA was extracted from pancreata of E11.5 to newborn mice according to the method of Chomczynski (Chomczynski and Sacchi, 1987). cDNAs were synthesized from 1 µg of total RNA using oligo(dT) as primer for M-MLV reverse transcriptase (Promega). cDNAs were amplified by PCR with the following mouse TGF-β1 primers: 5′ primer, 5′-TCCCGTGGCTTCTAGTGCTG-3′, where 5′ = residue 396; 3′ primer (antisense), 5′-ATTTTAATCTCTGCAAGC-GCA-3′, where 5′ = residue 836 (numbering according to mouse TGF-β1 cDNA sequence, as described by Derynck et al., 1986); the expected size of the amplified cDNA is 440 bp. PCR was performed essentially as described (Saiki et al., 1988), in a 30 times repeated temperature cycle: 1 minute at 94°C, 1 minute at 56°C and 1 minute at 72°C. 10 µl of the PCR mixtures were loaded on 5% acrylamide gels that were stained with ethidium bromide after electrophoresis.

Development of pancreatic buds in collagen gels

When they first arise from the primitive gut, pancreatic buds consist of a central mass of epithelial cells containing narrow lumen-like intercellular spaces and surrounded by a mesenchymal cap (Fig. 1A). Immunofluorescence on mouse E12.5 pancreatic buds reveals a small proportion of epithelial cells containing glucagon and/or one or another member of the PP-fold family (i.e. PP, NPY or PYY; Herrera et al., 1991; Hazelwood, 1993; Teitelman et al., 1993; Upchurch et al., 1994).

Fig. 1.

Dorsal pancreatic buds from E12.5 mouse embryos. (A) Semithin section of a freshly dissected bud showing a central epithelial mass surrounded by a mesenchymal cap. The epithelial mass contains narrow lumen-like intercellular spaces (arrowheads). (B,C,D,E,F) Semithin sections of buds cultured in collagen gels for 3 (B), 4 (C,D), 9 (E) and 15 (F) days. During the first 3 days in culture epithelial cells actively proliferate (see Fig. 2A) and give rise to branching ducts (B). Exocrine cells, containing numerous zymogen granules and forming acinar structures, are first detected after 4 days in culture (C,D); the arrowhead in D (phase-contrast microscopy) points to a mitotic exocrine cell. Between 8 and 10 days, pancreatic buds cultured in a collagen gel show ducts, acini and clusters of endocrine cells (see Fig. 3A,B) in paraductal region (E). After about 15 days in culture acinar cells are undetectable (F). Bars, 30 µm (A,B) and 10 µm (D); magnification is the same in A, C, E and F.

Fig. 1.

Dorsal pancreatic buds from E12.5 mouse embryos. (A) Semithin section of a freshly dissected bud showing a central epithelial mass surrounded by a mesenchymal cap. The epithelial mass contains narrow lumen-like intercellular spaces (arrowheads). (B,C,D,E,F) Semithin sections of buds cultured in collagen gels for 3 (B), 4 (C,D), 9 (E) and 15 (F) days. During the first 3 days in culture epithelial cells actively proliferate (see Fig. 2A) and give rise to branching ducts (B). Exocrine cells, containing numerous zymogen granules and forming acinar structures, are first detected after 4 days in culture (C,D); the arrowhead in D (phase-contrast microscopy) points to a mitotic exocrine cell. Between 8 and 10 days, pancreatic buds cultured in a collagen gel show ducts, acini and clusters of endocrine cells (see Fig. 3A,B) in paraductal region (E). After about 15 days in culture acinar cells are undetectable (F). Bars, 30 µm (A,B) and 10 µm (D); magnification is the same in A, C, E and F.

When embedded in three-dimensional collagen gels, mouse E12.5 pancreatic buds underwent morphological changes rem-iniscent of normal development in vivo. After 3 days in culture, the epithelial portion of the tissue had developed into numerous branching ducts delimited by a single layer of epithelial cells (Fig. 1B), many of which were actively proliferating, as revealed by incorporation of BrdU (Fig. 2A). Clusters of epithelial cells apparently budding from the developing ducts were also observed; some of these cells could be revealed by anti-glucagon- and/or anti-bPP immunofluorescent staining. Rare epithelial cells containing a few zymogen granules could be detected by electron microscopy. Cells containing numerous zymogen granules and organized in acinar structures were first observed after 4 days in culture; the acini were embedded in a loose stroma containing mesenchymal cells (Fig. 1C,D). The presence of amylase, in zymogen granules, was confirmed by protein A-gold immunocytochemistry (data not shown). After 6 days, the relative volume occupied by proliferating epithe-lial cells was 4-fold lower than at 3 days (Fig. 2C; P<0.0001); this reduced rate of proliferation was maintained at least up to 10 days. After 8 days in culture in collagen gels, the amount of acini was increased and endocrine cells containing one or more islet hormone were present either as single cells in para-ductal regions or as clusters of variable size that sometimes acquired an islet-like organization (Figs 1E, 3). Surprisingly, the fraction of endocrine cells containing glucagon was quite low, even in cultured dorsal buds, in which at a corresponding stage in vivo, as well as in the adult, A cells are abundant (Orci et al., 1976; Orci, 1982; Herrera et al., 1991). During subse-quent culture in collagen gels, acinar cells progressively decreased in number and were undetectable after about 15 days in culture (Fig. 1F). In contrast, no obvious changes in the com-position and organization of the endocrine portion of the tissue were observed.

Fig. 2.

Incorporation of BrdU in E12.5 pancreatic buds cultured in collagen gels either in the absence (A,C) or in the presence of TGF-β1 (1 ng/ml; B,D). At 3 days (A,B) the proliferation rate in pancreata in the presence or absence of TGF-β1 is similar, whereas after 6 days (C,D) TGF-β1-treated buds contain less proliferating cells than the parallel control buds. Proliferating epithelial cells decrease between 3 and 6 days in both conditions. Bar, 30 µm.

Fig. 2.

Incorporation of BrdU in E12.5 pancreatic buds cultured in collagen gels either in the absence (A,C) or in the presence of TGF-β1 (1 ng/ml; B,D). At 3 days (A,B) the proliferation rate in pancreata in the presence or absence of TGF-β1 is similar, whereas after 6 days (C,D) TGF-β1-treated buds contain less proliferating cells than the parallel control buds. Proliferating epithelial cells decrease between 3 and 6 days in both conditions. Bar, 30 µm.

Similar results were obtained when E11.5 or E12.5 dorsal or ventral pancreatic buds were cultured in the same conditions. Taken together, these observations indicate that pancreatic primordia continue to develop when cultured in collagen gels, albeit somewhat more slowly than in vivo, yielding, after 8 days in culture, a histological pattern resembling that of an E16.5 pancreas. However, two differences are noteworthy in buds developing in culture: the relative paucity of glucagon-containing cells and the limited viability of the acinar tissue.

Development of pancreatic buds in basement membrane gels

There is considerable evidence that basement membranes play important roles in the develop-ment of epithelial tissues (Golosow and Grobstein, 1962; Wessells and Cohen, 1967; Bernfield et al., 1984; Schuger et al., 1990b; Takahashi and Nogawa, 1991), and numerous studies have shown that basement membrane-like matrices obtained from the EHS tumor promote tissue-specific cell organization and cytodif-ferentiation in various epithelial systems (Kleinman et al., 1986; Barcellos-Hoff et al., 1989; Schuger et al., 1990a). We therefore examined the development of pancreatic buds cultured in an EHS-derived matrix and compared it with that achieved in collagen gels.

After 8 days in culture in EHS-derived matrix, prominent clumps of endocrine cells were seen to bud from ductal epithelium. The majority of these cells could be stained using an anti-bPP antiserum, which reveals all three members of the PP-fold family (Fig. 3C; Teitelman et al., 1993; Upchurch et al., 1994; unpublished observations). Staining using specific anti-hPP and anti-PYY antibodies demonstrated the presence of both peptides, with PYY-containing cells being most abundant. Glucagon-, insulin- and somatostatin-(Fig. 3D) containing cells were frequently observed. A quantitative evaluation of the volume density of the different types of endocrine cells in pan-creatic buds cultured in collagen or in EHS-derived gels revealed that the development of the epithelial buds and of the endocrine portion of the tissue was significantly enhanced (Figs 4, 5; P<<0.0001) in the basement membrane-like extra-cellular matrix. A strikingly different pattern of development was observed in the case of the exocrine tissue: while the ductular portion of the tissue was similar in collagen and in EHS-derived gels, the acinar component was completely absent in buds cultured in the EHS matrix (Figs 3C,D, 4). In experiments in which pancreatic buds were grown in gels composed of a 1:1 mixture of EHS matrix and type I collagen, the pattern of development was similar to that achieved in EHS matrix alone.

Fig. 3.

Development of pancreatic buds cultured in different conditions. Pairs of consecutive semithin sections from cultured E12.5 dorsal buds were stained with different anti-hormone antibodies. (A,B) Pancreatic bud grown for 9 days in a collagen gel and stained with anti-glucagon or anti-bPP (B) antibodies. The epithelial portion of the bud consists of ducts, acini (arrowheads) and clusters of endocrine cells adjacent to the ducts. (C,D) Pancreatic bud grown for 8 days in basement membrane-like extracellular matrix (EHS) and stained with anti-bPP (C) or anti-somatostatin (D) antibodies. The bud is characterized by an increased number of endocrine cells, which form clusters budding from the ducts, and by the absence of acini. (E,F) Pancreatic bud grown for 10 days in a collagen gel in the presence of TGF-β1 (1 ng/ml) and stained with anti-bPP (E) or anti-insulin (F) antibodies. TGF-β1 partly mimics the effects of the EHS matrix in increasing the number of endocrine cells (particularly of those containing the PP-fold family peptides and insulin) and in drastically reducing the development of acini. (G,H) Pancreatic bud grown for 9 days in a collagen gel in the absence of serum and stained with anti-glucagon (G) or anti-insulin (H) antibodies. The development of the endocrine component (that presents here an islet-like organization) is not affected by the absence of serum, whereas the amount of acinar cells (arrowhead), is increased (for quantification, see Fig. 4). Bar, 30 µm.

Fig. 3.

Development of pancreatic buds cultured in different conditions. Pairs of consecutive semithin sections from cultured E12.5 dorsal buds were stained with different anti-hormone antibodies. (A,B) Pancreatic bud grown for 9 days in a collagen gel and stained with anti-glucagon or anti-bPP (B) antibodies. The epithelial portion of the bud consists of ducts, acini (arrowheads) and clusters of endocrine cells adjacent to the ducts. (C,D) Pancreatic bud grown for 8 days in basement membrane-like extracellular matrix (EHS) and stained with anti-bPP (C) or anti-somatostatin (D) antibodies. The bud is characterized by an increased number of endocrine cells, which form clusters budding from the ducts, and by the absence of acini. (E,F) Pancreatic bud grown for 10 days in a collagen gel in the presence of TGF-β1 (1 ng/ml) and stained with anti-bPP (E) or anti-insulin (F) antibodies. TGF-β1 partly mimics the effects of the EHS matrix in increasing the number of endocrine cells (particularly of those containing the PP-fold family peptides and insulin) and in drastically reducing the development of acini. (G,H) Pancreatic bud grown for 9 days in a collagen gel in the absence of serum and stained with anti-glucagon (G) or anti-insulin (H) antibodies. The development of the endocrine component (that presents here an islet-like organization) is not affected by the absence of serum, whereas the amount of acinar cells (arrowhead), is increased (for quantification, see Fig. 4). Bar, 30 µm.

Fig. 4.

Volume densities (expressed as percentages) of the different tissue components in E12.5 buds cultured for 8-10 days in collagen gels (CTR), in EHS matrix (EHS), in collagen gels in the presence of either EGF (5 ng/ml) (EGF) or TGF-β1 (1 ng/ml) (TGF), or in collagen gels in the absence of serum (CTRSF). Results are presented as box-plots (Williamson, 1989) in which boxes symbolize the interquartile space of each distribution, the median (•) and the mean (×) being indicated within the box; triangles (▴) are the extreme values of each distribution. Numbers indicate the size of the sampling.

Fig. 4.

Volume densities (expressed as percentages) of the different tissue components in E12.5 buds cultured for 8-10 days in collagen gels (CTR), in EHS matrix (EHS), in collagen gels in the presence of either EGF (5 ng/ml) (EGF) or TGF-β1 (1 ng/ml) (TGF), or in collagen gels in the absence of serum (CTRSF). Results are presented as box-plots (Williamson, 1989) in which boxes symbolize the interquartile space of each distribution, the median (•) and the mean (×) being indicated within the box; triangles (▴) are the extreme values of each distribution. Numbers indicate the size of the sampling.

Fig. 5.

Volume densities of the different endocrine cell types, as revealed by anti-hormone staining, in E12.5 pancreatic buds cultured as in Fig. 4. The volume densities of total endocrine cells were calculated by adding the volume densities of each of the four endocrine cell types.

Fig. 5.

Volume densities of the different endocrine cell types, as revealed by anti-hormone staining, in E12.5 pancreatic buds cultured as in Fig. 4. The volume densities of total endocrine cells were calculated by adding the volume densities of each of the four endocrine cell types.

Surprisingly, although the results described above were consistently reproducible with several distinct batches of EHS tumor extract prepared in our laboratory, commercially available EHS matrix (Matrigel™) did not produce the same effect. Buds embedded in Matrigel gave rise to large cysts, which in semithin sections appeared as enlarged (ectatic) ducts delimited by flattened epithelial cells. Acinar and endocrine cells were rare and often poorly preserved (data not shown).

The striking effects of EHS-derived matrix on the develop-ment in cultured pancreatic buds might have been due to insoluble basement membrane components (such as laminin, type IV collagen or heparan sulfate proteoglycans) or to dif-fusible factors present in the EHS matrix, which is known to contain a number of cytokines (Vukicevic et al., 1992). To dis-tinguish between these possibilities, pancreatic buds were embedded in a collagen gel, which was subsequently covered with a layer of EHS matrix. This resulted in a pattern of pan-creatic development similar to that observed by culturing pan-creatic rudiments directly within the EHS matrix, suggesting that soluble mediators diffusing from the EHS matrix into the adjacent collagen gel were responsible for the increased devel-opment of endocrine tissue and the disappearence of acinar cells.

Effect of exogenous growth factors on the development of pancreatic buds in collagen gels

In view of the results described above, we investigated whether the addition of well characterized cytokines to pancreatic buds in collagen gels might affect their development, as determined after 10 days in culture. Addition of either bFGF (30 ng/ml), rhHGF (20 ng/ml), IGF-II (50 ng/ml), PDGF (50 ng/ml) or NGF (1, 10 or 100 ng/ml) did not detectably influence pancreatic differen-tiation (data not shown). rhEGF (5 ng/ml) stimulated the formation of duct-like struc-tures, while drastically decreas-ing the development of acini (Fig. 4); the formation of the endocrine part of the tissue was not affected (Fig. 5).

Pancreatic buds cultured for 8-10 days in the presence of TGF-β1 (1 ng/ml) were charac-terized by the absence of acinar cells and an abundance of epithelial buds neighbouring ductular structures; all types of endocrine cells were present in these epithelial buds, with a pre-ponderance of cells containing insulin and/or a member of the PP-fold family (Fig. 3E,F). The shape and arrangement of stromal cells was also modified by TGF-β1: as compared to stromal cells in control cultures, they appeared more elongated and oriented in parallel arrays. Similar effects were observed with TGF-β1 at concentrations of 0.1, 0.3 and 3 ng/ml; at the lower concentrations a few acinar cells were present. Specimens were also fixed at earlier time points (data not shown); this revealed that after the first 5 days in culture the buds cultured in the presence of TGF-β1 were not detectably different from those cultured in the absence of the growth factor, i.e. they contained well-developed acini formed by zymogen granule-containing cells. Incorporation of BrdU showed that the relative volume occupied by the prolif-erating epithelial cells after 3 days in culture was similar in the presence or absence of TGF-β1 (Fig. 2A,B); at 6 days, TGF-β1-treated cultures contained 2-fold less proliferating cells than parallel control cultures (Fig. 2D; P<0.001).

A quantitative analysis of the frequency of each cell type present in pancreatic buds after 10 days in culture confirmed that acinar cells were virtually absent in specimens cultured in the presence of TGF-β1 (Fig. 4), whereas the endocrine component was increased (Fig. 5). All endocrine cell types were significantly increased; the increase in abundance of A and D cells was less pronounced in TGF-β1-supplemented collagen gels than in EHS matrices (Fig. 5). The total volumes of the pancreatic rudiments cultured in the various experimen-tal conditions were also determined, and an estimation of volumes occupied by the different components was computed. In the presence of either EHS matrix or TGF-β1, the total volume of the endocrine compart-ment was increased by 4.5- and 3-fold (P<0.0001), respectively, as compared to control conditions.

Taken together, these observa-tions suggest that TGF-β1 induces the regression of the acinar com-partment of the developing pancreas and promotes differen-tiation of the endocrine tissue. In this respect, the effect of TGF-β1 mimics, at least in part, that of EHS matrix.

Development of pancreatic buds in serum-free medium

Besides extracellular matrices, another rich source of cytokines affecting growth and differen-tiation in tissue culture is serum. The experiments described thus far were all performed in medium containing 10% HI-FCS. Pancre-atic rudiments cultured for 10 days in collagen gels in serum-free medium showed a different pattern of development: while mesenchyme and ductular cells were less abundant (Fig. 4; P<0.0001), acinar cells organized in densely packed structures were particularly conspicuous (Figs 3G,H, 4; P<0.0001). The total volume density of endocrine cells was not significantly different in the presence or absence of FCS; however, A cells were more abundant (P<0.0001) in serum-free cultures (Fig. 5). The addition of TGF-β1 to serum-free cultures did not elicit effects comparable to those observed in the presence of serum (data not shown). These results suggest that components present in serum affect the differ-entiation of cultured pancreatic buds; they also show that the presence of serum is necessary for the effect of TGF-β1 on differen-tiation.

TGF-β1 enhances apoptosis of acinar cells

The striking disappearance of acinar cells during the incubation of pancreatic buds in the presence of TGF-β1 suggested that the cytokine might selectively prevent their survival. Thin sections of pancreatic buds cultured under different conditions were thus examined for signs of apoptosis. Analysis of cultures in collagen gels with serum, maintained for 10 days revealed the presence of apoptotic bodies containing fragmented nuclei with condensed peripheral chromatin and cytoplasmic organelles, including abundant endoplasmic reticulum and zymogen granules, which indicates that they were derived from acinar cells. These apoptotic bodies were either localized in the interstitium immediately outside the acinar basement membrane or phagocytosed by adjacent cells (Fig. 6), but they were not observed within the acinar epithelium itself. Mor-phological signs of apoptosis were not seen when buds were cultured for up to 15 days in the absence of serum. In serum-containing cultures, apoptotic bodies were more abundant in the presence of TGF-β1. Increased numbers of apoptotic bodies were also seen when TGF-β1 was added after 6 days in culture, i.e. at a time when acini are already conspicuous, and the cultures examined 24 hours later: a quantitative analysis revealed a 1.7-fold higher frequency of acinar apoptotic bodies in TGF-β1-treated cultures as compared to control cultures (P<0.0001; 95% confidence interval between 1,5 and 2.0). No obvious signs of apoptosis were noted, either in the absence or presence of TGF-β1, for the other cell types present in the cultures.

Fig. 6.

Apoptosis of acinar cells. Dorsal pancreatic buds at E12.5 were cultured in collagen gels for 6 days. 24 hours before fixation, TGF-β1 was added to a final concentration of 1 ng/ml. Fragments of cells containing condensed chromatin and zymogen granules can be seen in close proximity to acini (A). Apoptotic bodies appear to be engulfed by mesenchymal cells (B,C). Bars, 1 µm.

Fig. 6.

Apoptosis of acinar cells. Dorsal pancreatic buds at E12.5 were cultured in collagen gels for 6 days. 24 hours before fixation, TGF-β1 was added to a final concentration of 1 ng/ml. Fragments of cells containing condensed chromatin and zymogen granules can be seen in close proximity to acini (A). Apoptotic bodies appear to be engulfed by mesenchymal cells (B,C). Bars, 1 µm.

These results indicate that TGF-β1 selectively increases the level of apoptosis of acinar cells in serum-containing cultures, and that apoptosis may thus account for the TGF-β1-induced disappearance of acinar cells.

TGF-β1 mRNA is present in pancreatic buds throughout development

Previous studies have demonstrated the presence of TGF-β1 mRNA and protein in the adult human pancreas (Yamanaka et al., 1993). Since this cytokine affects development of cultured pancreatic buds, it was of interest to determine whether it is produced during embryogenesis. Total RNA extracted from mouse pancreatic buds from E11.5 to birth was used to generate oligo(dT)-primed cDNAs. PCR amplification of the cDNAs with specific oligonucleotide primers revealed the presence of TGF-β1 mRNA at all embryonic stages studied.

Among the most challenging issues in developmental biology is the understanding of how, within a given organ, a quantita-tively adequate contribution of its different constituents is achieved and maintained. In exocrine glands, for instance, the development of excretory ducts must be limited so as to leave space for the secretory units. The development of the pancreas is, in this respect, particularly intriguing since, in addition to the excretory ducts and secretory acini, a relatively fixed pro-portion (approximately 1%) of the tissue differentiates into the endocrine islets of Langerhans. Although growth factors or cytokines are reasonable candidates for the regulation of the relative proportions of an organ’s tissue constituents, the inves-tigation of these issues is difficult: in vitro studies allow the environment of the developing organ to be manipulated, but the conditions must be chosen so as to allow development to proceed in a manner that mimics in vivo ontogeny as closely as possible. Three-dimensional cultures of embryonic buds in extracellular matrix (ECM) gels provide a suitable experimental system to identify factors that control organogenesis (Takahashi and Nogawa, 1991). We have therefore explored the use of collagen and basement membrane-like gels to study the ontogeny of the murine pancreas in vitro. Gels formed from type I collagen were permissive for the development of both the exocrine and endocrine parts of the gland, albeit at a somewhat slower schedule than that observed in vivo; under these conditions a pattern compatible with that which occurs physiologically was observed. Cultures in collagen gels can thus be used to investigate the putative factors controlling pan-creatic development. Interestingly, the relative proportions of the different endocrine cell types formed during in vitro development were similar, irrespective of whether the cultures were initiated from dorsal or from ventral buds; this suggests that the differences in cell composition between juxta-duodenal and splenic islets of Langerhans in vivo (Orci et al., 1976; Orci, 1982) may be due to different environmental cues rather than to differences in intrinsic developmental potential of the buds themselves.

Earlier in vitro studies have pointed to the role of epithelial-mesenchymal interactions in pancreatic development and suggested the importance of diffusible mesenchyme-derived factors in determining the fate of the epithelial components (Golosow and Grobstein, 1962; Wessells and Cohen, 1967). Of particular interest was the demonstration that a ‘mesenchymal factor’ (MF) could induce isolated rat pancreatic epithelium to preferentially differentiate into acinar structures and B cells (Ronzio and Rutter, 1973). These observations set the stage for the present studies, in that they suggested that epigenetic phenomena might be involved in determining the relative pro-portions of tissues and cell types in developing organs. Under our own experimental conditions, we have found that buds (containing both epithelial and mesenchymal structures) cultured in a matrix composed of type I collagen develop all endoderm-derived pancreatic constituents, in approximately physiological proportions. In contrast, similar buds cultured in EHS matrix acquired a very different phenotype: islands of endocrine cells developed more extensively, while acini were almost completely absent. Mixing experiments suggested that the EHS matrix plays a dominant role in the acquisition of this abnormal phenotype, apparently through the release of dif-fusible substances. Surprisingly, whereas several batches of EHS matrix prepared in our laboratory yielded similar results, commercially available EHS matrix (Matrigel™) did not support the development of either acinar or endocrine cells and the only epithelial structures to form were enlarged, cyst-like ducts. Since minor components, such as growth factors contaminating the major ECM proteins, are most likely to vary between preparations of EHS matrix, this observation is also in accord with the notion that cytokines may influence the relative development of different pancreatic tissues. These results therefore led us to explore the possible effects of cytokines on embryonic buds in collagen gels.

In accord with our hypothesis, we have found that two cytokines, EGF and TGF-β1, cause profound changes in the fate of cultured buds. EGF increased the proportion of tissue occupied by ducts, at the expense of acinar structures; the endocrine portion of the tissue did not appear to be affected. TGF-β1 induced a complete involution of acini with morpho-logical signs of apoptosis, and also induced a marked increase in development of the endocrine tissue. It may be relevant that EGF and TGF-β1 are among the cytokines that have been iden-tified in certain preparations of EHS matrix (Vukicevic et al., 1992): the different patterns of development that we have observed in cultures embedded in EHS gels or Matrigel™ could conceivably be due to differences in the concentrations of these cytokines in the different batches of ECM used. In any event, our experiments indicate that EGF and TGF-β1 are mod-ulators of pancreatic organogenesis in vitro and that cytokines can play a part in controlling the relative development of the different tissues that compose the mature gland.

Previous studies on the effects of EGF (or TGF-α, which is structurally related to EGF and binds to the same receptor) and TGF-β on pancreatic and other tissues have yielded results relevant to our observations. Transgenic mice overexpressing TGF-α in their pancreas exhibit extensive acinoductular meta-plasia, resulting in an increased proportion of duct-like struc-tures, similar to what we have observed in EGF-treated buds (Jhappan et al., 1990; Sandgren et al., 1990). In cell culture, TGF-β1 has been shown to inhibit the growth of adult mouse pancreatic acinar cells and also to induce apoptosis of various epithelial cell types, including rabbit uterine epithelial cells and rat hepatocytes (Rotello et al., 1991; Logsdon et al., 1992; Oberhammer et al., 1992). The effect of TGF-β1 on embryonic acinar cells in buds developing in vitro is thus compatible with these earlier findings. It is however puzzling that acinar cells are only eliminated after an initial few days in culture during which time they form at an apparently normal rate; this could be due to a relatively late appearance of TGF-β receptors during development of the buds. An alternate possibility, i.e. that the TGF-β effect on acinar structures takes several days to become manifest because it induces progressive changes in the composition of the ECM, is less likely: addition of TGF-β1 to cultures containing well-developed acini leads to their very rapid involution. Another intriguing aspect of acinar apoptosis in our organotypic cultures is the localization of the apoptotic bodies: essentially all were present immediately adjacent to but clearly separated from the epithelium itself. Apoptotic bodies could be unambiguously identified as originating from acinar cells by the presence of well-preserved typical organelles such as zymogen granules. This suggests either that cells undergo-ing apoptosis are very rapidly ‘rejected’ from the epithelium, or that the primary effect of TGF-β1 is to induce epithelial cells to leave the epithelium. In the latter situation, ‘misplaced’ epithelial cells would subsequently eliminate themselves through apoptosis; this would resemble anoikis, a form of apoptosis induced by denied anchorage (Frisch and Francis, 1994; Ruoslahti and Reed, 1994).

Besides inducing involution of acinar cells, the EHS matrix and TGF-β1 had a striking effect on the development of the endocrine portion of the pancreatic buds. Both the relative pro-portion and the total volume of tissue occupied by endocrine cells were increased as compared to control cultures in collagen gels; hence this increase does not simply reflect the loss of part of the exocrine compartment, but must be due to increased proliferation and/or differentiation of endocrine cells or their precursors. To our knowledege this positive influence of TGF-β1 on the endocrine pancreas has not been previously reported; on most epithelial cell types TGF-β1 exerts an inhibitory effect on proliferation (Moses, 1992). No dramatic changes in the relative proportions of the different endocrine cell types were noticed in response to TGF-β1, although the increase in B and PP gene family expressing-cells tended to be more pronounced than the increase in A cells. Whether the effect of TGF-β1 on the developing islets of Langerhans is mediated by a direct effect of the cytokine on endocrine cells or their precursors, whether it is due to the induction of another growth-stimulating cytokine or whether it is secondary to changes in the ECM through effects on stromal cells, cannot be decided at this time.

Our observations demonstrate that exogenous cytokines can modify pancreatic development. Could the same cytokines be endogenous physiological modulators of this process? TGF-α and the EGF receptor have been identified in fetal pancreas (Miettinen and Heikinheimo, 1992); TGF-βs are present in adult pancreatic tissue (Yamanaka et al., 1993), and by RT-PCR we have demonstrated the presence of TGF-β1 mRNA in pancreatic buds at all stages of development. These cytokines could thus play a paracrine role in organogenesis of the pancreas. The lack of reported effects of TGF-β1 gene inacti-vation on pancreatic development does not invalidate this hypothesis, since other members of the TGF-β family may also be physiologically relevant factors or may substitute for the missing cytokine (Shull et al., 1992). It is also interesting to note that endogenous cytokines, in particular TGF-βs, may play a part in controlling the in vitro development of cultured buds. The evidence in favor of this is indirect: in long-term cultures maintained in presence of serum, acinar structures undergo slow involution, similar to that which occurs, albeit with much more rapid kinetics, in the presence of exogenous TGF-β1. As in other systems (Rotello et al., 1991), the effect of exogenous TGF-β1 on pancreatic buds requires the presence of serum, and cultures performed in the absence of both TGF-β1 and serum show increased development of acini, which persist even in long term cultures. The use of antisense oligonucleotides or antibodies capable of preventing TGF-β production or activity may help to decide whether endogenous TGF-β, either synthesized in situ or present in serum, influ-ences the development of cultured buds.

Given that EGF and TGF-β1 affect the development of epithelial derivatives in the developing pancreas, is one of these the long sought MF? It has been proposed that MF plays a role in pancreatic ontogeny by determining the proportion of endocrine and acinar cells; it increases the development of B cells and decreases the yield of A cells, but it also enhances the differentiation of acinar cells (Ronzio and Rutter, 1973). Thus, in view of our results, neither EGF nor TGF-β1 alone can account for the effects of MF. However, the data available on MF suggest that it probably encompasses more than just one active principle (Filosa et al., 1975); it is possible that some of its effects are due to the presence of EGF and/or TGF-β, in combination with other modulatory factors.

In conclusion, we have shown that two cytokines, EGF and TGF-β1, influence the in vitro differentiation of pancreatic buds, in particular with respect to the relative development of the endodermal derivatives. The notion that cytokines can act as paracrine mediators responsible for achieving the appropri-ate balance of different tissue constituents during organogene-sis is compatible with the observed effects of transgene-encoded TGF-β1 in the mammary gland: overexpression of the cytokine leads to alterations in the relative development of lob-uloalveolar versus ductal tissue (Pierce et al., 1993). It will be of interest to determine whether molecular ablation of members of TGF-β family in knock out mice, or targetted over-expression of these cytokines, alters the structure of the pancreas, and whether certain developmental pathologies, such as nesidioblastosis (Heitz et al., 1977; Creutzfeldt, 1985), or species-related characteristics, such as the relative abundance of endocrine tissue in the acomys mouse (Gonet et al., 1965), are due to differences in the cytokine-mediated regulation of pancreatic ontogeny.

We thank Mrs D. Ben-Asr and Mrs G. Moussard for their skillful technical assistance, Mr J.-P. Gerber, B. Favri and Mr G. Negro for photographic work and Mr G. Andrey for computer work. We are grateful to Dr M. Castellucci for preparing extracts of EHS tumor, to Ms B. Mermillod for help and advice for all statistical analyses and to Dr M. Pepper for valuable suggestions. This work was supported by a grant from the Fonds National Suisse de la Recherche Scien-tifique (31-34088.92).

Barcellos-Hoff
,
M. H.
,
Aggeler
,
J.
,
Ram
,
T. G.
and
Bissell
,
M. J.
(
1989
).
Functional differentiation and alveolar morphogenesis of primary mammary cultures on reconstituted basement membrane
.
Development
105
,
223
235
.
Bendayan
,
M.
,
Roth
,
J.
,
Perrelet
,
A.
and
Orci
,
L.
(
1980
).
Quantitative immunocytochemical localization of pancreatic secretory proteins in subcellular compartments of the rat acinar cell
.
J. Histochem. Cytochem
.
28
,
149
160
.
Bernfield
,
M.
,
Banerjee
,
S. D.
,
Koda
,
J. E.
and
Rapraeger
,
A. C.
(
1984
).
Remodeling of the basement membrane as a mechanism of morphogenetic tissue interaction
. In
The Role of Extracellular Matrix in Development
(ed.
R. L.
Trestald
), pp.
545
572
.
New York
:
Alan.R.Liss, Inc
.
Chambard
,
M.
,
Gabrion
,
J.
and
Mauchamp
,
J.
(
1981
).
Influence of collagen gel on the orientation of epithelial cell polarity: follicle formation from isolated thyroid cells and from preformed monolayers
.
J. Cell Biol
.
91
,
157
166
.
Chomczynski
,
P.
and
Sacchi
,
N.
(
1987
).
Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction
.
Analyt. Biochem
.
162
,
156
159
.
Creutzfeldt
,
W
(
1985
).
Endocrine tumors of the pancreas
. In
The Diabetic Pancreas
(ed.
B.W.
Volk
and
E.R.
Arquilla
), pp.
543
-
586
. New York and London: Plenum Medical Book Company.
Derynck
,
R.
,
Jarrett
,
J. A.
,
Chen
,
E. Y.
and
Goeddel
,
D. V.
(
1986
).
The murine transforming growth factor-β precursor
.
J. Biol. Chem
.
261
,
4377
4379
.
Filosa
,
S.
,
Pictet
,
R.
and
Rutter
W. J.
(
1975
).
Positive control of cyclic AMP on mesenchymal factor controlled DNA synthesis in embryonic pancreas
.
Nature
257
,
702
704
.
Frisch
,
S. M.
and
Francis
,
H.
(
1994
).
Disruption of epithelial cell-matrix interactions induces apoptosis
.
J. Cell Biol
.
124
,
619
626
.
Golosow
,
N.
and
Grobstein
,
C.
(
1962
).
Epitheliomesenchymal interaction in pancreatic morphogenesis
.
Dev. Biol
.
4
,
242
255
.
Gonet
,
A. E.
,
Stauffacher
,
W.
,
Pictet
,
R.
and
Renold
,
A. E.
(
1965
).
Obesity and diabetes mellitus with striking congenital hyperplasia of the islets of Langerhans in spiny mice (Acomys cahirinus)
.
Diabetologia
1
,
162
171
.
Hazelwood
,
R. L.
(
1993
).
The pancreatic polypeptide (PP-fold) family: gastrointestinal, vascular, and feeding behavioral implications
.
Proc. Soc. Exp. Biol. Med
.
202
,
44
63
.
Heitz
,
P. U.
,
Kloppel
,
G.
,
Hacki
,
W. H.
,
Polak
,
J. M.
and
Pearse
,
A. G.
(
1977
).
Nesidioblastosis: the pathologic basis of persistent hyperinsulinemic hypoglycemia in infants. Morphologic and quantitative analysis of seven cases based on specific immunostaining and electron microscopy
.
Diabetes
26
,
632
642
.
Herrera
,
P.-L.
,
Huarte
,
J.
,
Sanvito
,
F.
,
Meda
,
P.
,
Orci
,
L.
and
Vassalli
,
J.-D.
(
1991
).
Embryogenesis of the murine endocrine pancreas; early expression of pancreatic polypeptide gene
.
Development
113
,
1257
1265
.
Jhappan
,
C.
,
Stahle
,
C.
,
Harkins
,
R. N.
,
Fausto
,
N.
,
Smith
,
G. H.
and
Merlino
,
G. T.
(
1990
).
TGFα overexpression in transgenic mice induces liver neoplasia and abnormal development of the mammary gland and pancreas
.
Cell
61
,
1137
1146
.
Kleinman
,
H. K.
,
McGarvey
,
M. L.
,
Hassell
,
J. R.
,
Star
,
V. L.
,
Cannon
,
F. B.
,
Laurie
,
G. W.
and
Martin
,
G. R.
(
1986
).
Basement membrane complexes with biological activity
.
Biochemistry
25
,
312
318
.
Kramer
,
R. H.
,
Bensch
,
K. G.
and
Wong
,
J.
(
1986
).
Invasion of reconstituted basement membrane matrix by metastatic human tumor cells
.
Cancer Res
.
46
,
1980
1989
.
Le Douarin
,
N. M.
(
1988
).
On the origin of pancreatic endocrine cells
.
Cell
53
,
169
171
.
Logsdon
,
C. D.
,
Keyes
,
L.
and
Beauchamp
,
R. D.
(
1992
).
Transforming growth factor-β (TGF-β1) inhibits pancreatic acinar cell growth
.
Am. J. Physiol
.
262
,
G364
G368
.
Maxwell
,
M. H.
(
1978
).
Two rapid and simple methods used for the removal of resins from 1.0 micron thick epoxy sections
.
J. Microsc
.
112
,
253
255
.
Miettinen
,
P. J.
and
Heikinheimo
,
K.
(
1992
).
Transforming growth factor- alpha (TGF-α) and insulin gene expression in human fetal pancreas
.
Development
114
,
833
840
.
Montesano
,
R.
,
Mouron
,
P.
,
Amherdt
,
M.
and
Orci
,
L.
(
1983
).
Collagen matrix promotes reorganization of pancreatic endocrine cell monolayers into islet-like organoids
.
J. Cell Biol
.
97
,
935
939
.
Montesano
,
R.
,
Schaller
,
G.
and
Orci
,
L.
(
1991
).
Induction of epithelial tubular morphogenesis in vitro by fibroblast-derived soluble factors
.
Cell
66
,
697
711
.
Moses
,
H. L.
(
1992
).
TGF-β regulation of epithelial cell proliferation
.
Mol. Reprod. Dev
.
32
,
179
184
.
Nogawa
,
H.
and
Takahashi
,
Y.
(
1991
).
Substitution for mesenchyme by basement-membrane-like substratum and epidermal growth factor in inducing branching morphogenesis of mouse salivary epithelium
.
Development
112
,
855
861
.
Oberhammer
,
F. A.
,
Pavelka
,
M.
,
Sharma
,
S.
,
Tiefenbacher
,
R.
,
Purchio
,
A. F.
,
Bursch
,
W.
and
Schulte Hermann
,
R.
(
1992
).
Induction of apoptosis in cultured hepatocytes and in regressing liver by transforming growth factor β1
.
Proc. Natl. Acad. Sci. USA
89
,
5408
5412
.
Orci
,
L.
,
Baetens
,
D.
,
Ravazzola
,
M.
,
Stefan
,
Y.
and
Malaisse Lagae
,
F.
(
1976
).
Pancreatic polypeptide and glucagon: non-random distribution in pancreatic islets
.
Life Sci
.
19
,
1811
1815
.
Orci
,
L.
(
1982
).
Macro- and micro-domains in the endocrine pancreas
.
Diabetes
31
,
538
565
.
Pictet
,
R. L.
,
Clark
,
W. R.
,
Williams
,
R. H.
and
Rutter
,
W. J.
(
1972
).
An ultrastructural analysis of the developing embryonic pancreas
.
Dev. Biol
.
29
,
436
467
.
Pictet
,
R. L.
and
Rutter
,
W. J.
(
1972
).
Development of the embryonic endocrine pancreas
. In
Handbook of Physiology. Section
8
, vol.
1
The Endocrine Pancreas’ (ed.
D.
Steiner
,
N.
Freinkel
), pp.
25
-
66
. American Physiological Society, Washington Baltimore Maryland: Williams & Wilkins.
Pierce
,
D. F. J.
,
Johnson
,
M. D.
,
Matsui
,
Y.
,
Robinson
,
S. D.
,
Gold
,
L. I.
,
Purchio
,
A. F.
,
Daniel
,
C. W.
,
Hogan
,
B. L.
and
Moses
,
H. L.
(
1993
).
Inhibition of mammary duct development but not alveolar outgrowth during pregnancy in transgenic mice expressing active TGF-β1
.
Genes Dev
.
7
,
2308
2317
.
Ronzio
,
R. A.
and
Rutter
,
W. J.
(
1973
).
Effects of a partially purified factor from chick embryos on macromolecular synthesis of embryonic pancreatic epithelia
.
Dev. Biol
.
30
,
307
320
.
Rotello
,
R. J.
,
Lieberman
,
R. C.
,
Purchio
,
A. F.
and
Gerschenson
,
L. E.
(
1991
).
Coordinated regulation of apoptosis and cell proliferation by transforming growth factor β1 in cultured uterine epithelial cells
.
Proc. Natl. Acad. Sci. USA
88
,
3412
3415
.
Roth
,
J.
,
Bendayan
,
M.
and
Orci
,
L.
(
1978
).
Ultrastructural localization of intracellular antigens by the use of protein A-gold complex
.
J. Histochem. Cytochem
.
26
,
1074
1081
.
Ruoslahti
,
E.
and
Reed
,
J. C.
(
1994
).
Anchorage dependence, integrins, and apoptosis
.
Cell
77
,
477
478
.
Saiki
,
R. K.
,
Gelfand
,
D. H.
,
Stoffel
,
S.
,
Scharf
,
S. J.
,
Higuchi
,
R.
,
Horn
,
G. T.
,
Mullis
,
K. B.
and
Erlich
,
H. A.
(
1988
).
Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase
.
Science
239
,
487
491
.
Sandgren
,
E. P.
,
Luetteke
,
N. C.
,
Palmiter
,
R. D.
,
Brinster
,
R. L.
and
Lee
,
D. C.
(
1990
).
Overexpression of TGFα in transgenic mice: induction of epithelial hyperplasia, pancreatic metaplasia, and carcinoma of the breast
.
Cell
61
,
1121
1135
.
Schuger
,
L.
,
O’Shea
,
K. S.
,
Nelson
,
B. B.
and
Varani
,
J.
(
1990a
).
Organotypic arrangement of mouse embryonic lung cells on a basement membrane extract: involvement of laminin
.
Development
110
,
1091
1099
.
Schuger
,
L.
,
O’Shea
,
S.
,
Rheinheimer
,
J.
and
Varani
,
J.
(
1990b
).
Laminin in lung development: effects of anti-laminin antibody in murine lung morphogenesis
.
Dev. Biol
.
137
,
26
32
.
Shull
,
M. M.
,
Ormsby
,
I.
,
Kier
,
A. B.
,
Pawlowski
,
S.
,
Diebold
,
R. J.
,
Yin
,
M.
,
Allen
,
R.
,
Sidman
,
C.
,
Proetzel
,
G.
,
Calvin
,
D.
,
Annunziata
,
N.
, and
Doetschman
T.
(
1992
).
Targeted disruption of the mouse transforming growth factor-β1 gene results in multifocal inflammatory disease
.
Nature
359
,
693
699
.
Takahashi
,
Y.
and
Nogawa
,
H.
(
1991
).
Branching morphogenesis of mouse salivary epithelium in basement membrane-like substratum separated from mesenchyme by the membrane filter
.
Development
111
,
327
335
.
Teitelman
,
G.
,
Alpert
,
S.
,
Polak
,
J. M.
,
Martinez
,
A.
and
Hanahan
,
D.
(
1993
).
Precursor cells of mouse endocrine pancreas coexpress insulin, glucagon and the neuronal proteins tyrosine hydroxylase and neuropeptide Y, but not pancreatic polypeptide
.
Development
118
,
1031
1039
.
Upchurch
,
B.H.
,
Aponte
,
G.W.
and
Leiter
,
A.B.
(
1994
).
Expression of peptide YY in all four islet cell types in the developing mouse pancreas suggests a common peptide YY-producing progenitor
.
Development
120
,
245
252
.
Vukicevic
,
S.
,
Kleinman
,
H. K.
,
Luyten
,
F. P.
,
Roberts
,
A. B.
,
Roche
,
N. S.
and
Reddi
,
A. H.
(
1992
).
Identification of multiple active growth factors in basement membrane Matrigel suggests caution in interpretation of cellular activity related to extracellular matrix components
.
Exp. Cell Res
.
202
,
1
8
.
Wallenstein
,
S.
,
Zucker
,
C. L.
and
Fleiss
,
J. L.
(
1980
).
Some statistical methods useful in circulation research
.
Circ. Res
.
47
,
1
9
.
Warburton
,
D.
,
Seth
,
R.
,
Shum
,
L.
,
Horcher
,
P. G.
,
Hall
,
F. L.
,
Werb
,
Z.
and
Slavkin
,
H. C.
(
1992
).
Epigenetic role of epidermal growth factor expression and signalling in embryonic mouse lung morphogenesis
.
Dev. Biol
.
149
,
123
133
.
Weibel
,
E. R.
(
1979
).
Stereological Methods
, p.
415
.
London
:
Academic Press
.
Weller
,
A.
,
Sorokin
,
L.
,
Illgen
,
E.M.
and
Ekblom
,
P.
(
1991
).
Development and growth of mouse embryonic kidney in organ culture and modulation of development by soluble growth factor
.
Dev. Biol
.
144
,
248
261
.
Wessells
,
N. K.
and
Cohen
,
J. H.
(
1967
).
Early pancreas organogenesis: morphogenesis, tissue interactions, and mass effects
.
Dev. Biol
.
15
,
237
270
.
Williamson
,
D. F.
,
Parker
,
R. A.
and
Kendrick
,
J. S.
(
1989
).
The box plot: a simple visual method to interpret data
.
Ann. Intern. Med
.
110
,
916
921
.
Yamanaka
,
Y.
,
Friess
,
H.
,
Buchler
,
M.
,
Beger
,
H. G.
,
Gold
,
L. I.
and
Korc
,
M.
(
1993
).
Synthesis and expression of transforming growth factor β-1, β-2, and β-3 in the endocrine and exocrine pancreas
.
Diabetes
42
,
746
756
.
Yang
,
J.
,
Richards
,
J.
,
Guzman
,
R.
,
Imagawa
,
W.
and
Nandi
,
S.
(
1980
).
Sustained growth in primary culture of normal mammary epithelial cells embedded in collagen gels
.
Proc. Natl. Acad. Sci. USA
77
,
2088
2092
.