ErbB Signaling Regulates Lineage Determination of Developing ...

0013-7227/02/$15.00/0 Printed in U.S.A.

Endocrinology 143(11):4437– 4446 Copyright © 2002 by The Endocrine Society doi: 10.1210/en.2002-220382

ErbB Signaling Regulates Lineage Determination of Developing Pancreatic Islet Cells in Embryonic Organ Culture ¨ IVI J. MIETTINEN, JAAN PALGI, TARJA KOIVISTO, JARKKO USTINOV, MARI-ANNE HUOTARI, PA DANIEL HARARI, YOSEF YARDEN, AND TIMO OTONKOSKI Biomedicum Helsinki (M.-A.H., P.J.M., J.P., T.K., J.U., T.O.), Program for Developmental and Reproductive Biology and Haartman Institute, University of Helsinki, Helsinki 00014, Finland; Hospital for Children and Adolescents (M.-A.H., P.J.M., T.O.), Helsinki University Central Hospital, Helsinki 00029, Finland; and The Weizman Institute (D.H., Y.Y.), Rehovot 76100, Israel The neuregulin (NRG)/epidermal growth factor (EGF) family of growth factors consists of several ligands that specifically activate four erbB receptor-tyrosine kinases, namely erbB-1 (EGF-R), erbB-2 (neu), erbB-3, and erbB-4. We have previously shown that islet morphogenesis is impaired and ␤-cell differentiation delayed in mice lacking functional EGF-R [EGF-R (ⴚ/ⴚ)]. The present study aims to clarify which erbB ligands are important for islet development. Pancreatic expression of EGF, TGF-␣, heparin-binding EGF, betacellulin (BTC), and NRG-4 was detected as early as embryonic d 13 (E13). Effects of these ligands were studied in E12.5 pancreatic explant cultures grown for 5 d ex vivo. None of the growth factors affected the ratio of endocrine to exocrine cells. However, significant effects within the endocrine cell populations were induced by EGF, BTC, and NRG-4. ␤-Cell development was augmented by

T

HE MAMMALIAN PANCREAS is a mixed exocrine and endocrine gland that initially develops as dorsal and ventral buds from foregut endoderm. Understanding its development has progressed rapidly during recent years, mainly through the cloning and functional analysis of several tissue-specific transcription factors (1, 2). Endocrine differentiation of the pancreatic islets has classically been thought to be induced by interaction of mesenchymal cells with the adjacent epithelium (3, 4). More recently, it has been found that embryonic pancreatic epithelium forms islets even without contact to fetal mesenchyme, suggesting that the endocrine pancreas is developing by default. However, signals derived from the mesenchyme and extracellular matrix are important for the proper morphology and growth and development of the entire organ, including ducts and acinar components of the pancreas (5). Follistatin is one of the mesenchymal factors required for the development of exocrine tissue while exerting a repressive role on the differentiation of the endocrine cells (6). Ligands of the fibroblast growth factor-receptor 2b (FGF-R2b) are expressed in the developing pancreas and seem to stimulate the growth of the exocrine pancreas (7), whereas FGF-R1-mediated signaling is imporAbbreviations: BTC, Betacellulin; E13, embryonic d 13; EGF, epidermal growth factor; EGF-R, EGF receptor; FGF-R, fibroblast growth factor-receptor; GAPDH, glyceraldehyde-3-phosphate-dehydrogenase; HB, heparin-binding; IgB1, soluble Fc-conjugated erbB-1; NRG, neuregulin; PP, pancreatic polypeptide; RT, room temperature.

BTC, whereas the development of somatostatin-expressing ␦-cells was stimulated by NRG-4. Both ligands decreased the numbers of glucagon-containing ␣-cells. The effect of BTC was abolished in the EGF-R (ⴚ/ⴚ) mice. A soluble erbB-4 binding fusion protein totally inhibited the effects of NRG-4 but not of BTC. Neutralization of endogenous NRG-4 activity in the model system effectively inhibited ␦-cell development, indicating that this erbB4-ligand is an essential factor for delineation of the somatostatin-producing ␦-cells. Our results suggest that ligands of the EGF-R/erbB-1 and erbB-4 receptors regulate the lineage determination of islet cells during pancreatic development. BTC, acting through EGF-R/erbB-1, is important for the differentiation of ␤-cells. This could be applied in the targeted differentiation of stem cells into insulinproducing cells. (Endocrinology 143: 4437– 4446, 2002)

tant for the proper development and function of the islet ␤-cells (8). EGF family ligands bind to and activate four erbB tyrosine kinase receptors (EGF-R/erbB-1, neu/erbB-2, erbB-3, and erbB-4) (9, 10). It has previously been demonstrated that EGF-R is expressed throughout the developing fetal human pancreas (11), and we have previously shown that pancreatic islet development is impaired in the EGF-R-deficient mice [EGF-R (⫺/⫺)] (12). In particular, the development of ␤-cells is delayed and the migration of developing islet cells is disturbed (12). Expression of all other erbB receptors (erbB-2, -3, and -4) is detectable in the developing mouse pancreatic ducts (13). It has been difficult to assess the specific roles of erbB-2, -3, and -4 because their inactivation by gene targeting results in early embryonic lethality (14 –16). ErbB-3 (⫺/⫺) mice survive through early pancreatic organogenesis and display signs of attenuated pancreatic development but no evidence of arrested islet differentiation (14). Several lines of evidence point to a role for neuregulin (NRG)/EGF growth factors in islet differentiation and growth. EGF stimulates the formation of duct-like structures in E12.5 mouse pancreatic rudiments cultured in collagen gels (17). Transient up-regulation of EGF, TGF-␣, and EGF-R is detected in the acini undergoing differentiation into ductlike structures during the islet regeneration of ␥-interferon transgenic mice (18). Targeted overexpression of EGF under the insulin promoter increases islet size and the frequency of

4437

4438

Endocrinology, November 2002, 143(11):4437– 4446

insulin-positive cells in the pancreatic ducts (19). TGF-␣-like immunoreactivity is detected in the ducts, acini, and islets, where it colocalizes with insulin (11). Heparin-binding EGF (HB-EGF) expression pattern has been shown to be similar to that of the islet-specific transcription factor pancreatic duodenal homeobox-1, with early expression in primitive ducts and later in the endocrine pancreatic cells (20). Betacellulin (BTC), which was initially discovered from a mouse tumoral ␤-cell line (21), is expressed in both ␣-cells and duct cells in the adult human pancreas and in primitive ducts of the human fetal pancreas (22). In cell culture studies, BTC induced the conversion of exocrine AR42J cells into insulin producing endocrine cells (23). BTC has also been shown to induce proliferation of insulin-producing INS-1 cells (24). A novel ligand of erbB-4, NRG-4, was recently identified and found to be highly expressed in the pancreas (25). Based on this evidence, it is apparent that EGF-R-mediated signaling is important for islet cell development. The major aim of the present study was to identify the important EGF-R ligands in this respect. Furthermore, we wanted to assess the role of erbB-4 and its ligands. Our experimental results suggest that both EGF-R/erbB-1 and erbB-4 ligands have an important role in regulating endocrine pancreatic differentiation. Materials and Methods Materials Recombinant human EGF was obtained from Upstate Biotechnology, Inc. (Lake Placid, NY). Recombinant human BTC, recombinant human TGF-␣, recombinant human HB-EGF and antihuman BTC antibody were obtained from R&D Systems Europe, Ltd. (Oxford, UK). Nonspecific goat IgG was kindly provided by docent Ilkka Seppa¨ la¨ (Haartman Institute, University of Helsinki, Helsinki, Finland). NRG-4 was synthesized as described (25). NRG-4 inhibitory antibody was generated by injecting New Zealand White rabbits (bred and housed at the Weizmann Institute animal facilities) five times with a purified, refolded synthetic peptide comprising the EGF domain of NRG-4 (25), with antigenicity bolstered by standard Freund’s adjuvant protocol. IgB4, a fusion protein between the extracellular ligand binding portion of erbB-4 and the Fc portion of human IgG, was synthesized as described (26).

Generation and genotyping of the EGF-R (⫺/⫺) mice EGF-R was disrupted as described previously (27). Mice used for generation of EGF-R (⫺/⫺) were outbred Black Swiss strain [originally bought from the Jackson Laboratory (Bar Harbor, ME), bred and housed at Haartman Institute animal facilities], which have a milder phenotype than inbred strains used (28, 29). Embryos used in this study were derived from intercrosses between EGF-R (⫹/⫺) mice. E12.5 embryos were genotyped by PCR and Southern blot analysis as described previously (12, 27). The study was approved in the Haartman Institute Ethical Committee for Animal Studies.

Pancreatic explant cultures Pancreatic explant cultures of both normal and EGF-R (⫺/⫺) d 12.5 (E12.5) embryonic mice were used to study the effect of EGF family growth factors during pancreatic organ culture. The appearance of the vaginal plug was noted as E0.5. The duodenal loop along with the pancreatic anlage and stomach were microdissected. The tissues were then cultured by a technique originally designed for embryonic kidney (30). Tissue explants, consisting of the dorsal and ventral pancreatic buds (mesenchyme plus epithelium), stomach, and duodenal loop, were placed on Nucleopore filters (1.0 ␮m pore size, Costar, Corning, NY) on metal grids and cultured at the air-liquid interphase in serum-free improved MEM (Life Technologies, Inc., Gaithersburg, MD) supplemented

Huotari et al. • ErbB Signaling in Islet Development

with transferrin (30 ␮g/ml), penicillin (100 IU/ml), and streptomycin (100 ␮g/ml) and with either EGF (20 ng/ml), TGF-␣ (20 ng/ml), HB-EGF (20 ng/ml), BTC (20 ng/ml), NRG-4 (1 ng/ml), antihuman BTC antibody (1 ␮g/ml), anti-NRG-4 antibody (1:100), IgB4 (10 ␮g/ml), nonspecific goat IgG (10 ␮g/ml), or cultured in plain medium (control group). The media were changed every second day. After 5 d in culture, explants were fixed for 4 h at room temperature (RT) in Bouin⬘s fixative. After rinsing with 50% alcohol, tissues were stored in 70% alcohol before dehydration and paraffin embedding.

NRG-4 antibody and BTC neutralizing antibody binding To test specificity of the NRG-4 antibodies, NRG-4 and EGF (as negative control) were both radiolabeled with 125I (2 ⫻ 106 cpm/␮l) using Iodogen (Pierce Chemical Co., Rockford, IL), and cleaned on a G25 Sepharose column. Ten microliters of NRG-4 antisera from two independently injected rabbits were preadsorbed to protein A-Sepharose beads in HNTG buffer at 4 C for 30 min. Beads were washed three times in HNTG buffer, blocked in HNTG ⫹ 0.1% BSA for 30 min at 4 C, before spinning the beads and incubating for a further 2 h with 5 ␮l of radiolabeled ligand at 4 C. The same protocol was repeated for control antibodies and soluble Fc-conjugated erbB-1 (IgB1). Beads were then washed four times with HNTG, before being boiled in reducing sample buffer, proteins resolved by SDS-PAGE (15% acrylamide) and bound ligand detected by autoradiography. Ability of antihuman BTC antibody to block BTC effects was tested on A431 cell line (ATCC, Manassas, VA; no. CRL-1555) (31). For this purpose, cells were plated in six-well plates and allowed to attach in DMEM (Life Technologies, Inc.) supplemented with 10% heat-inactivated fetal calf serum (Life Technologies, Inc.), 100 IU/ml penicillin, 100 ␮g/ml streptomycin, and 2 mm l-glutamine for 24 h. After overnight culture in serum-free DMEM, the cells were incubated in serum-free DMEM supplemented with either 1 or 10 ng/ml BTC, 1 ␱r 10 ␮g/ml antihuman BTC or no additions (control group) for 5 min at 37 C. One or 10 ng/ml of BTC was then added to cells treated with antihuman BTC and the cells were then incubated for another 5 min at 37 C. Cells were then lysed in ice-cold lysis buffer (25 mm HEPES, 5 mm EDTA, 150 mm NaCl, 1% Triton X-100, leupeptin 20 ␮g/ml, aprotinin 20 ␮g/ml, 1 mm phenylmethylsulfonylfluoride, and 0.1 mm Naorthovanadate) and left on ice for 10 min. The whole cell extract was then cleared by centrifugation, boiled in reducing gel sample buffer, and resolved by 10% SDS-PAGE before transfer onto nitrocellulose. Filters were blocked in TBST buffer [10 mm Tris-HCl (pH 8), 150 mm NaCl, 0.1% Tween 20] containing 5% milk overnight at 4 C, blotted with anti-EGF-R [pY1086] phosphospecific antibody [1:1000 dilution; Biosource International (Camarillo, CA), catalog no. 44-790] in blocking buffer for 2 h at RT, followed by conjugation with a secondary antibody linked to horseradish peroxidase (Amersham Pharmacia Biotech, Buckinghamshire, UK) in blocking buffer for 1 h at RT. Subsequently proteins were detected using the enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).

Immunohistochemistry For quantitative morphometric analysis of pancreatic cell types, an entire tissue block was sectioned (2 ␮m) for immunohistochemical stainings. Every seventh consecutive section was stained with the same primary antibody. Deparaffinized, rehydrated sections were incubated for 2 h at RT in 3% normal goat or rabbit serum (Zymed Laboratories, Inc., South San Francisco, CA) in PBS (pH 7.4) to block nonspecific binding sites. Sections were incubated overnight at 4 C with primary antibody, diluted in PBS containing 3% normal serum (guinea pig antiporcine insulin 1:500, rabbit antihuman glucagon 1:2000, rabbit antihuman somatostatin 1:2000, rabbit antihuman pancreatic polypeptide 1:2000, rabbit antihuman amylase 1:500; DAKO Corp. (Carpinteria, CA), monoclonal anti-pan cytokeratin (mixture) 1:500; Sigma, St. Louis, MO). After rinsing several times with PBS, sections were incubated for 30 min at RT with biotinylated-goat antirabbit IgG (biotinylated-rabbit antimouse IgG for cytokeratin) (Zymed Laboratories, Inc.), diluted in PBS, rinsed and incubated with peroxidase conjugated streptavidin (Zymed Laboratories, Inc.), diluted in PBS. The sections were finally developed with 3-amino-9-ethyl-carbazole substrate and rinsed with distilled water. Light counterstaining was performed with hematoxylin. After the immunohistochemical staining of every seventh section


with the same pancreas-specific antibody [insulin, glucagon, somatostatin, pancreatic polypeptide (PP), and amylase], the number of positively stained cells was manually counted under the light microscope. Cell proliferation was quantified by staining with Ki67 nuclear antigen. For double staining of hormone plus Ki67, the staining protocol was continued from the one described above. The slides were treated with 10 mm citrate buffer (pH 6.0) for 25 min in a microwave oven to reveal antigenic sites and rinsed in PBS. Double staining was performed using the Vectastain ABC kit (Vector Laboratories, Inc., Burlingame, CA). The sections were incubated for 1 h at RT in 3% goat serum in PBS (pH 7.4) to block nonspecific binding sites and incubated with rabbit polyclonal Ki67 antibody (NovoCastra, Newcastle, UK) for 1 h at RT, diluted 1:500 in PBS containing 3% goat serum. After rinsing with PBS, sections were incubated for 30 min in biotinylated antibody, rinsed, incubated for 30 min in Vectastain ABC-alkaline phosphatase reagent, rinsed, and developed with alkaline phosphatase substrate (Vector Blue). After color development, slides were rinsed in distilled water and mounted with Aquamount. Numbers of the following cell types were analyzed from double-stained sections: insulin (⫹ Ki67), glucagon (⫹ Ki67), somatostatin (⫹ Ki67), and PP (⫹ Ki67). Sections from paraformaldehyde-fixed paraffin-embedded pancreases of embryonic (E12.5 and E16.5), newborn and adult mice were used to study the expression pattern of erbB-4 receptor at different developmental stages. Deparaffinized, rehydrated slides were treated with 10 mm citrate buffer (pH 6.0) for 5 min in microwave-oven to reveal antigenic sites, rinsed in PBS and incubated for 2 h at RT in 3% normal goat serum. The slides were then incubated overnight at 4 C with primary antibody, diluted in PBS containing 3% normal goat serum [rabbit polyclonal antihuman erbB4 (1:50), Santa Cruz Biotechnology, Inc., Santa Cruz, CA]. The staining was then completed as described above.

RT-PCR Expression of EGF, TGF-␣, HB-EGF, BTC, and NRG-4 mRNAs in E13, newborn, and adult pancreases was studied by RT-PCR. Primers were intron-spanning to exclude nonspecific genomic DNA amplification. Primers used were rat glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) (5⬘-GTC TTC ACC ACC ATG GAG AAG GCT⬘ and 5⬘-TGT AGC CCA GGA TGC CCT TTA GTG⬘; EMBL/GenBank M17701, position 325– 854, fragment size 530), mouse EGF (5⬘TTG AAA TGG CCA ATC TGG ATG G⬘ and 5⬘TGA CAC CAT GAT TTC AGC CAC T; EMBL/GenBank J00380, position 2368 –2855, fragment size 488 bp), mouse TGF-␣ (5⬘GTT CTC AGG TCC AGC CAG TC and 5⬘GGT TCT CTC CTT CCA CCA GAT⬘; EMBL/GenBank U65016, position 2685– 3225, fragment size 541 bp), mouse HB-EGF (5⬘TCT GGA GCG GCT TCG GAG AG⬘ and 5⬘CAC GCC CAA CTT CAT TTC TC⬘; 5⬘primer exon 2 EMBL/GenBank L36024 and 3⬘primer exon 5 EMBL/GenBank L36027 corresponding to human HB-EGF cDNA EMBL/GenBank M60278, position 336 – 873, fragment size 538 bp), mouse BTC (5⬘CAC AGC ACA GTT GAT GGA CC⬘ and 5⬘CCG TTA AGC AAT ATT GGT CTC 3⬘; EMBL/GenBank L08394, position 100 – 649, fragment size 550 bp) and NRG-4 (5⬘CTC ACT CTT ACC ATC GCG GC⬘ and 5⬘CAG CCT TAT CTA TAC TGC TGA C⬘, position 328 –710, fragment size 383 bp (25). E13, newborn and adult pancreases were dissected free from their surrounding tissue and total RNA was isolated using GenElute total RNA isolation kit (Sigma). RNA samples were treated with deoxyribonuclease I [41 ␮l ribonuclease-free water, 5 ␮l 10 ⫻ buffer (500 mm NaCl; 10 mm dithiothreitol; 100 mm Tris-HCl, pH 7.5), 2 ␮l RNasin (40 U/␮l, Promega Corp., Madison, WI), and 3 ␮l RQ ribonuclease-free deoxyribonuclease (1 U/␮l, Promega Corp.)] for 50 min at 37 C. Thereafter, the samples were phenol/chloroform extracted and precipitated after addition of 1 vol 7.5 m NH4OAc and 2.5 vol ethanol. Total RNA was precipitated by centrifugation and washed once with 75% EtOH. After precipitation with ethanol the samples were dissolved in 15 ␮l ribonuclease-free water. Total RNA (2 ␮g) was reverse transcribed after addition of 10 ␮l of master mix (4 ␮l 5⫻ reverse transcription buffer, 2.5 ␮l 2.5 mm deoxy-NTP, 1 ␮l 500 ␮g/ml oligo(deoxythymidine)-primer, 0.5 ␮l RNasin 40 U/␮l, 0.5 ␮l Moloney murine leukemia virus-reverse transcriptase, 1.5 ␮l water) (Promega Corp.), at 37 C for 90 min. The reaction was inactivated at 95 C for 5 min. One microliter of reverse transcription reaction mixture in 13.3 ␮l of water was added to PCR master mixture [2.5 ␮l 10⫻ PCR buffer containing 15 mm MgCl2 (Perkin-Elmer, Foster City, CA), 2 ␮l deoxy-NTP (2.5 mm stock), 5 ␮l dimethylsulfoxide (50%

Endocrinology, November 2002, 143(11):4437– 4446 4439

stock), 1 ␮l of 5⬘and 3⬘ primers mix (both 10 ␮m in stock) and 0.25 ␮l AmpliTaq Gold (5 U/␮l, Perkin-Elmer)]. The samples were heated to 95 C for 5 min to activate the polymerase and then cycled for 35 times: denaturation 45 sec at 94 C, annealing 45 sec at 58 C for BTC, NRG-4, GAPDH, and 63 C for EGF, TGF-␣ and HB-EGF, elongation 45 sec at 72 C on GenAmp PCR System 9600 apparatus (Perkin-Elmer). The PCR samples were electrophoresed through 1.5% agarose gel and visualized with ethidium bromide staining. To confirm specificity of the PCR, all generated fragments were purified from low melting point agarose gel by phenol/chloroform extraction, precipitated after addition of 1 vol 7.5 m NH4OAc and 2.5 vol ethanol and sequenced in a sequencing core facility by routine methods.

Statistics Results are presented as the mean ⫾ sem. Significance of observed differences between experimental groups were tested either by one-way ANOVA followed by Fischer’s protected least significant difference test, or in case of skewed data, by nonparametric methods (Kruskal Wallis test and Mann Whitney U test) using Statview 5.0 software (SAS Institute, Cary, NC).

Results EGF family ligands are expressed in the embryonic pancreas

RT-PCR was performed to determine whether EGF, TGF-␣, HB-EGF, BTC, and NRG-4 are expressed in the developing pancreas. Total RNA from E13, newborn, and adult mouse pancreas was isolated and subjected to RT-PCR with specific EGF family primers. mRNAs for all the studied ligands were present already at E13, and their expression was detectable also in newborn and adult mouse pancreas (Fig. 1). RT-PCR with EGF primers gave two equally strong bands at E13, a weaker and a stronger band at newborn and only one band at adult stage. Sequence analysis of these two bands shows that the heavier band (488 bp) corresponded to the pre-pro area, whereas the lighter band (420 bp) had a deletion between base pairs 2525 and 2591. Whether this represents a splice variant or an inactive protein was not studied further. Specificity of all the PCR products was confirmed by sequence analysis.

FIG. 1. EGF-R ligands are expressed in developing pancreas. RTPCR was used to study the endogenous pancreatic expression of EGF family growth factors. GAPDH was used as a control gene. All the mRNAs for the studied ligands were expressed already at E13, and their expression was detectable also in newborn and adult mouse pancreas. A 100-bp ladder was used as a size marker.

4440


Ligands of erbB-1 and -4 favor differentiation of ␤-, ␦-, and PP-cells instead of ␣-cells

The whole-mount organ culture model mimics in vivo development because it allows three-dimensional branching morphogenesis to occur (32). E12.5 dorsal and ventral pancreatic buds together with stomach and duodenal loop were

FIG. 2. Pancreatic development proceeds normally in organ culture. Pancreas from E12.5 mouse embryo and a pancreatic explant grown for 5 d in control medium in the absence of any exogenous growth factors were stained for several pancreatic antibodies to demonstrate the cell differentiation during 5-d explant culture. In E12.5 pancreas, glucagon-positive cells are already abundant (A), whereas very few insulin-positive cells are seen (B). After 5 d of organ culture, all the endocrine cell types have developed [glucagon (C), insulin (D), somatostatin (E), PP (F)], as well as exocrine [amylase (G)] and also ductal cells [cytokeratin (H)].


dissected and cultured for 5 d with either EGF, HB-EGF, TGF-␣, BTC, or NRG-4 or with plain medium (control group), as described above. At the beginning of culture several glucagon-positive cells were detected (Fig. 2A), but only very few insulin-positive cells were seen (Fig. 2B). Somatostatin, PP, and amylase-positive cells were undetectable at



this stage. After 5 d of culture all endocrine cell types (Fig. 2, C–H) and also exocrine (amylase positive) and cytokeratin expressing ductal cells were present. The effects of growth factors on the in vitro development of endocrine and exocrine cell populations are shown in Tables 1 and 2. Addition of EGF family ligands (EGF, HB-EGF, BTC, or NRG-4) to the cultures did not affect the ratio of exocrine (amylase) vs. endocrine (insulin ⫹ glucagon ⫹ somatostatin ⫹ PP) cells (approximately 1.4:1) (Table 2). However, significant effects were observed within the endocrine cell populations. Of the EGF-R ligands tested (BTC, HB-EGF, EGF, and TGF-␣), only BTC induced a highly significant 2-fold increase in the number of ␤-cells in the explants, whereas also EGF and HB-EGF increased the proportion of ␤-cells out of all endocrine cells (Table 1). BTC, EGF, and NRG-4 reduced the proportion of glucagon expressing ␣-cells (Table 1). NRG-4 was the only factor significantly increasing both the number and proportion of somatostatin (␦) cells (Table 1). Also, this occurred at the expense of glucagon-producing ␣-cells. As a result of these changes, highly significant differences were observed in the ␤/␣- and ␦/␣-cell ratios depending on the growth factor stimuli. For example, BTC induced a 2.4-fold increase in the ␤/␣-cell ratio, and NRG-4 a 7-fold increase in the ␦/␣-cell ratio (Table 2). Also, the PP/␣-cell ratio was increased 4-fold by NRG-4 (Table 2). These calculations may be relevant in the context of cell lineage determination regulated by the growth factors. Importantly, the growth factors only affected the numbers of the various endocrine cell populations. The numbers of exocrine (amylase positive) cells, or the exocrine/endocrine ratio were not affected (Table 2). EGF family growth factors do not induce endocrine cell proliferation in pancreatic explant cultures

Some explants were double-stained with Ki67 and insulin, glucagon, somatostatin, or PP to study the proliferation of

endocrine cells. We have previously demonstrated that in the absence of functional EGF-R, proliferation of ␤-cells is reduced. However, none of the exogenously added EGF family growth factors affected the proliferation of endocrine cells in our ex vivo model. This indicates that the increase in ␤-cell number induced by BTC was not due to increased ␤-cell proliferation (Fig. 3 showing insulin-Ki67 and glucagon-Ki67 double-stainings and ␤-cell proliferation index in nonstimulated and in BTC-stimulated explants). Proliferation of ␦-cells was also unaffected by NRG-4 (no proliferating somatostatin-producing cells detected). The labeling index of other endocrine cells was also unaffected by stimulation with the different EGF family growth factors (not shown). Developmental expression of erbB-4 in the pancreas

As both BTC and NRG-4 can activate erbB-4, we next wanted to study the expression pattern of erbB-4 in the developing mouse pancreas. For this purpose, immunohistochemistry with a specific anti-erbB-4 antibody was performed on E12.5, E16.5, newborn, and adult mouse pancreas. As illustrated in Fig. 4, erbB-4 expression is seen throughout the developing pancreas at E12.5. At E16.5, the expression becomes mainly limited to the ductal epithelium, and a similar staining pattern is seen in the newborn pancreas. As reported previously (13), erbB-4 immunoreactivity in the adult pancreas is detectable both in some of the ductal epithelial cells and in the glucagon-containing ␣-cells at the periphery of islets. BTC-stimulated ␤-cell differentiation requires EGF-R but not erbB-4 signaling

In our next set of experiments, function of both BTC and NRG-4 as well as their receptors was inhibited to further define the importance of these factors in islet cell delineation

TABLE 1. Effects of the erbB ligands on pancreatic cell differentiation based on quantitative morphometric analysis of cell types in E12.5 pancreatic explants cultured for 5 d in serum-free medium with or without the erbB ligands Insulin

Control (n ⫽ 30) EGF (n ⫽ 8) TGF-␣ (n ⫽ 5) BTC (n ⫽ 28) HB-EGF (n ⫽ 8) NRG-4 (n ⫽ 11)

Glucagon

Somatostatin

PP

No. of cells

% of endocrine cells

No. of cells


No. of cells


No. of cells


368 ⫾ 51 381 ⫾ 98 592 ⫾ 72 683 ⫾ 76c 389 ⫾ 85 320 ⫾ 84

44.8 ⫾ 2.8 60.3 ⫾ 3.8a 49.3 ⫾ 3.3 56.8 ⫾ 2.3c 53.4 ⫾ 2.6b 39.9 ⫾ 3.4

262 ⫾ 30 180 ⫾ 52 339 ⫾ 48 278 ⫾ 39 245 ⫾ 33 154 ⫾ 41

35.4 ⫾ 2.4 25.1 ⫾ 2.8b 28.3 ⫾ 2.8 21.5 ⫾ 1.7c 37.7 ⫾ 3.9 19.2 ⫾ 3.6a

77 ⫾ 14 70 ⫾ 24 125 ⫾ 34 129 ⫾ 17b 45 ⫾ 16 156 ⫾ 22a

9.4 ⫾ 1.1 8.8 ⫾ 1.2 9.6 ⫾ 1.6 11.1 ⫾ 1.2 5.4 ⫾ 1.2 24.8 ⫾ 4.5c

93 ⫾ 19 51 ⫾ 23 180 ⫾ 72 125 ⫾ 17 24 ⫾ 8 149 ⫾ 57

10.3 ⫾ 1.3 5.8 ⫾ 1.6 12.8 ⫾ 4.6 10.6 ⫾ 1.3 3.5 ⫾ 1.2b 16.1 ⫾ 3.2

Concentration of all growth factors was 20 ng/ml, except for NRG-4 (1 ng/ml). Results are based on pooled data from two to six separate experiments (n ⫽ number of explants, a P ⬍ 0.01, b P ⬍ 0.05, c P ⬍ 0.001). TABLE 2. The ratio of exocrine to endocrine cells and the ratio of insulin, somatostatin, and PP cells to glucagon cells

Control (n ⫽ 30) EGF (n ⫽ 8) TGF-␣ (n ⫽ 5) BTC (n ⫽ 28) HB-EGF (n ⫽ 8) NRG-4 (n ⫽ 11)

Amylase (no. of cells)

Exocrine/endocrine

Insulin/glucagon

Somatostatin/glucagon

PP/glucagon

1080 ⫾ 269 915 ⫾ 171 N.D. 1500 ⫾ 243 1055 ⫾ 343 794 ⫾ 196

1.4 ⫾ 0.2 1.7 ⫾ 0.4 N.D. 1.2 ⫾ 0.2 1.5 ⫾ 0.2 2.1 ⫾ 0.7

1.5 ⫾ 0.2 2.7 ⫾ 0.4a 1.8 ⫾ 0.3 3.6 ⫾ 0.5b 1.6 ⫾ 0.2 3.0 ⫾ 0.6a

0.3 ⫾ 0.1 0.4 ⫾ 0.1 0.4 ⫾ 0.1 0.8 ⫾ 0.2c 0.2 ⫾ 0.04 2.3 ⫾ 0.7b

0.4 ⫾ 0.1 0.3 ⫾ 0.1 0.5 ⫾ 0.2 0.7 ⫾ 0.1 0.1 ⫾ 0.04 1.6 ⫾ 0.6c

Results are based on pooled data from two to six separate experiments (n ⫽ number of explants, a P ⬍ 0.05, b P ⬍ 0.001, c P ⬍ 0.01). N.D., Not done.

4442



BTC-induced stimulation. These results demonstrate that BTC-induced ␤-cell differentiation does not involve erbB-4 signaling. Finally, a neutralizing anti-BTC antibody did not induce a significant response in the development of islet cells in these cultures. Ability of this BTC neutralizing antibody to block the binding of BTC to its receptor was tested using the A431 cell line, which expresses a high level of EGF-R (see Materials and Methods). In this model, the antibody blocked the BTC-induced tyrosine phosphorylation of the EGF-R in a dose-dependent manner (not shown). Nonspecific goat IgG, which was used as a negative control, did not have any effect on cell populations of the explants (Fig. 5). This result indicates that BTC is not absolutely required for ␤-cell development, and that other EGF-R ligands may substitute for its effect. The ␦-cell lineage requires an erbB-4 ligand for its development

Because NRG-4, which is known to only bind to erbB-4, strongly stimulated ␦-cell development, we applied IgB4 to embryonic explants to examine the essential role of erbB-4binding ligands for the ␦-cell lineage (33). As shown in Fig. 5B, the addition of soluble IgB4 resulted in a greater than 50% reduction in the delineation of ␦-cells. Moreover, IgB4 totally inhibited the strong stimulatory effect of NRG-4 on the ␦/␣cell ratio, demonstrating that NRG-4 and possibly other erbB-4-binding ligands play a major role in the delineation of somatostatin-expressing cells. FIG. 3. Betacellulin does not induce ␤-cell proliferation in an explant culture. Sections from pancreatic explants grown for 5 d in control medium in the absence of any exogenous growth factors were stained for insulin plus Ki67 (A) and glucagon plus Ki67 (B) to study the proliferation of endocrine cells at the end of the culture period. As shown in C, betacellulin did not significantly induce the proportion of proliferating ␤-cells. Other EGF family growth factors were also unable to induce proliferation after a 5-d culture period (not shown) (n ⫽ number of explants; magnification 1000-fold; thick arrow, Ki67positive nucleus; thin arrow, insulin or glucagon-positive cell; *, double-positive cell).

in the model used here. To assess the importance of EGF-R, we studied pancreatic explants from the EGF-R (⫺/⫺) mice. Signaling through erbB4 was blocked indirectly, using soluble erbB-receptors. Finally, we used specific blocking antibodies to directly block the action of these ligands. We have previously shown that ␤-cell differentiation is retarded in the EGF-R (⫺/⫺) mice (12). Figure 5A shows that the ␤/␣-cell ratio is about 40% lower in the EGF-R-deficient explants as compared with wild-type tissue. The figure also illustrates that the stimulatory effect of BTC is totally abolished in the EGF-R (⫺/⫺) explants, indicating that EGF-Rmediated signaling is required for the effect. Blocking of erbB-4 was achieved with a bivalent soluble receptor (IgB4) harboring the extracellular domain of erbB-4 receptor fused to the Fc component of a human IgG (33). IgB4 binds to erbB ligands that would otherwise bind to endogenous receptors, thus acting as a general inhibitor of multiple erbB-4-binding ligands. As shown in Fig. 5A, IgB4 did not have a major effect on the ␤/␣-cell ratio and, more importantly, did not block the

NRG-4 directs pancreatic islet cell development toward the ␦-cell lineage

Finally, an inhibitory antibody to NRG-4 was employed to test the specific essential role of this ligand in pancreatic development, at least within the context of this explant model. NRG-4 inhibitory antibody was generated by injecting rabbits with a refolded peptide encoding the EGF domain of mouse NRG-4 as has been described (25). The EGF domain is the essential ligand component required for receptor binding, and therefore an antibody binding to this domain will result in a complex that inhibits ligand-receptor binding. Radiolabeled NRG-4 bound specifically to the two tested NRG-4 antisera but did not bind preimmune serum, an irrelevant antisera or IgB1. Conversely, as expected, the radiolabeled EGF, which was used as a control, did not bind any of the tested antisera and only bound to a soluble receptor of erbB-1 (IgB1) (not shown). As shown in Fig. 5B and Table 3, the NRG-4 blocking antibody induced a greater than 50% reduction in the development of somatostatin-expressing ␦-cells (P ⬍ 0.02), whereas the development of other endocrine cell types was not affected. This result supports all previous results presented, and indicates that NRG-4, an erbB-4 ligand, plays a major role in pushing the ␦-cell lineage in the differentiating explants. As NRG-4 administration to these cultures was shown not to alter the mitogenic capacity of islet cells, our results indicate that NRG-4 targets a common precursor toward the ␦-cell lineage in lieu of the ␣-, ␤-, or PP-cell lineage.



FIG. 4. Developmental pattern of pancreatic erbB-4 expression. ErbB-4 expression pattern was studied at different developmental stages by immunohistochemistry. At E12.5 (A) immunoreactive erbB-4 (red color) was seen throughout the developing pancreatic epithelium, whereas at E16.5 (B) as well as in the newborn (C) pancreas it was predominantly ductal. In the adult (D) pancreas, expression of erbB-4 was detected both in some of the ductal epithelial cells as well as in glucagonexpressing ␣-cells in the periphery of islets. Glucagon staining shown in E.

Discussion

Our results demonstrate that EGF family ligands can regulate the lineage determination of pancreatic endocrine cells at least ex vivo. We show that BTC favors ␤-cell differentiation. This effect of BTC is abolished in the EGF-R (⫺/⫺) pancreases, indicating that BTC activity is mediated through the EGF-R. We also show that a ligand of erbB-4, NRG-4, is required for the development of somatostatin producing ␦-cells. In addition, we show a ductal expression pattern of erbB-4, a receptor for BTC and NRG-4, in the developing pancreas. These data indicate that both the NRG/EGF family of growth factors and erbB protein tyrosine kinase receptors might have a role in pancreas development. The first appearance of pancreatic differentiation in the mouse is at E9.5 when the dorsal pancreatic bud arises from gut endoderm. Soon, also the ventral bud arises. These two buds then start to grow, and finally fuse at E16 –E17 to form a single organ. Pancreatic endocrine cells arise from common multipotent precursor cells that express the transcription factor pancreatic duodenal homeobox-1. In the developing buds, the endocrine cells start to differentiate before the exocrine cells. The first appearance of endocrine cells in the

mouse is at E9.5, whereas amylase becomes detectable by immunostaining from about E14.5 (1). Coexpression of different hormones by a single cell has also been demonstrated at early stages of development (34), though more recent studies show that endocrine cell types do develop independently and not via cells coexpressing endocrine hormones (35). Although much has been learned of islet cell differentiation at the level of transcription factors, the role of extracellular soluble factors remain in large part open. Here we demonstrate that it is possible to modulate the lineage determination of endocrine cells by growth factors. It appears that in the absence of EGF-R signaling, differentiation mainly follows a default path toward ␣-cells and PP cells. EGF-R ligands, most notably BTC, shift this balance toward the ␤-cell phenotype (Fig. 6). It is interesting that HB-EGF, which also activates the EGF-R, is clearly less potent in this activity. Because BTC also activates erbB-4 (36), it was of interest to study the role of this receptor in the ␤-cell differentiation pathway. ErbB-4 expression is high in the sites where islet differentiation is most active, the embryonic pancreatic ducts. However, our results clearly indicate that erbB-4 signaling is not required for this particular effect of

4444


FIG. 5. Effect of EGF-R ligands on endocrine cell lineages. A, Insulin to glucagon ratio was induced 2.4-fold (P ⬍ 0.001) by addition of BTC into culture medium of E12.5 pancreatic explants. This effect was abolished in the EGF-R (⫺/⫺) pancreata, but IgB4 (inhibitor of erbB-4 binding ligands) did not block the BTC-induced stimulation in ␤-cell increase. B, Somatostatin to glucagon ratio was induced 7-fold (P ⬍ 0.001) by addition of NRG-4. IgB4 had an inhibitory effect, and also blocked the action of NRG-4. Anti-NRG-4 antibody also had a significant inhibitory effect (n ⫽ number of explants; *, P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001; nonspecific goat IgG was used as a negative control).

BTC because the effect of BTC was not inhibited by a soluble erbB-4 construct. This result is validated by the observation that the same inhibitory strategy effectively blocked the specific effect of NRG-4, which only can bind to erbB-4 (25). These findings are consistent with experiments on the pancreatic pluripotent tumor cell line, AR42J, showing that BTCstimulated differentiation toward a ␤-cell phenotype is associated with binding of BTC only to EGF-R but not to erbB-4 (37). Naturally, these data do not exclude the possibility that erbB-4 could be activated after heterodimerization with erbB-1. We have also identified a role for NRG-4 in the differentiation of endocrine islet cells. Instead of ␤-cells, exogenous NRG-4 strongly favors the differentiation of the somatostatin producing ␦-cells. Blocking experiments with a specific NRG-4 antibody clearly demonstrated that NRG-4 is an essential factor for ␦-cell differentiation in the cultured explants


(Table 2 and Fig. 6). The possibility remains that this in vitro system may lack some ligands that can substitute for NRG-4 in vivo. Furthermore, blocking of erbB-4 signaling with a soluble ligand binding Fc-fusion protein effectively inhibited the action of NRG-4. This confirms in a primary tissue the original notion that NRG-4 acts specifically through erbB-4. In this study, we have not specifically addressed the role of erbB ligands on the proliferation of primitive pancreatic epithelial cells. It was recently shown that EGF stimulates the proliferation of E13.5 rat pancreatic epithelial cells isolated from their mesenchymal contacts and that their endocrine differentiation is halted while the growth stimulus continues (38). Although the negative effect of EGF on endocrine differentiation appears to be in contrast with our findings, it must be noticed that the model systems used are fundamentally different. In our tissue cultures, the pancreatic epithelium maintains its close association with mesenchyme, which provides a multitude of signals affecting the growth and differentiation of the epithelial cells (4). Important signals are provided at least by members of the TGF-␤ (17, 39) and FGF (40) superfamilies. In this respect, the whole organ culture model can be considered more physiological as compared with the isolation of the embryonic epithelium from its natural environment. Furthermore, the in vitro effects obtained with various EGF family members must also be put into biological context by correlating them with the growth factors’ developmental expression pattern. EGF is not generally considered as the fetal ligand for erbBs as it is only weakly expressed during development (41), whereas TGF-␣ and BTC appear to be developmentally more important EGF-R ligands. This emphasizes the importance of our findings on BTC-specific lineage determination (22, 42, 43). In our experiments, different ligands of the same erbB receptors had clearly divergent effects on the endocrine differentiation. The pattern of receptor activation stimulated by EGF family ligands is very complex. Some ligands can bind to and activate signaling of more than one erbB kinase. Binding of an EGF family ligand to its receptor can also lead to activation of another erbB type through heterodimerization of the two kinases (44). Most of the EGF family ligands bind to and activate EGF-R. NRGs are so far the only exceptions. They bind to and activate erbB-3 and -4 (45). BTC has been shown to displace EGF from its binding site on the EGF-R. It can also bind to erbB-4 (36). There are no specific ligands for erbB-2, but it is a preferred heterodimerization partner for other erbBs (44). Furthermore, it has been shown that individual erbB receptors can discriminate between different EGF-like ligands even within a single receptor dimer (46). Thus, it is obvious that not only receptor heterodimerization but also heterogenetic phosphorylation of tyrosine residues within a homodimer can lead to differential activation of intracellular signaling pathways. It will be important to determine the intracellular signaling pathways responsible for the specific differentiation induced by BTC and NRG-4. In conclusion, our results emphasize the roles of both EGF-R- and erbB-4-mediated signaling in islet cell differentiation. Especially BTC and NRG-4 emerge as potential regulators of islet cell lineage determination. Understanding of regulatory mechanisms governing ␤-cell development has obvious implications for the development of new therapeutic



TABLE 3. Effects of anti-NRG antibody on islet cell differentiation

␤-Cells (% of all endocrine cells) ␣-Cells (% of all endocrine cells) ␦-Cells (% of all endocrine cells) PP-Cells (% of all endocrine cells)

Control (n ⫽ 8)

Anti-NRG-4 (n ⫽ 7)

P

46.7 ⫾ 2.9 37.3 ⫾ 3.5 8.3 ⫾ 1.1 7.7 ⫾ 1.3

40.6 ⫾ 2.6 45.6 ⫾ 2.4 4.1 ⫾ 1.0 9.7 ⫾ 2.3

0.44 0.05 0.02 0.85

Pooled data from two separate experiments of E12.5 pancreatic explants.

FIG. 6. Hypothesis for the role of erbB-signaling in lineage determination of developing pancreatic islet cells. Both EGF-R/erbB-1- and erbB-4-mediated signaling has a role in islet cell differentiation. EGF-R ligands EGF, and especially BTC, favor ␤-cell development, whereas ligand(s) for erbB-4 favor ␦- and PP-cell development. Instead, attenuation of EGF-R and erbB-4 signaling favors ␣-cell development.

strategies for insulin-deficient diabetes mellitus. In vitro differentiation of ␤-cells has been demonstrated from embryonic stem cells (47, 48) and pancreatic tissue-specific stem cells (49, 50). Once the optimal conditions for stem cell expansion and ␤-cell-specific differentiation have been defined, this approach could be used for the production of large quantities of ␤-cells for transplantation. Our findings on erbB ligand-dependent differentiation could thus be applied in such an approach. Acknowledgments The authors would like to thank Mrs. Ulla Kiiski for her valuable help in embryo dissections and Ms. Pa¨ ivi Kinnunen and Ms. Erika Wasenius for their excellent technical assistance. Received April 5, 2002. Accepted June 20, 2002. Address all correspondence and requests for reprints to: Mari-Anne Huotari, M.D., Biomedicum Helsinki, Room C503b, P.O. Box 63 (Haartmaninkatu 8), FIN-00014 Helsinki, Finland. E-mail: mari-anne. [email protected]. This study was supported by the Juvenile Diabetes Research Foundation International (Grant 1-1999-694), the European Community (Contract No. QLG1-CT-1999-00276), the Helsinki University Hospital Research Funds, the foundation for Diabetes Research in Finland, the Research and Science Foundation of Farmos, and the Finnish Medical Foundation.

References 1. Edlund H 1998 Transcribing pancreas. Diabetes 47:1817–1823 2. Sander M, German MS 1997 The ␤ cell transcription factors and development of the pancreas. J Mol Med 75:327–340

3. Hellerstro¨m C, Swenne I 1985 Growth pattern of pancreatic islets. In: Volk BW, Wellmann KF, eds. The diabetic pancreas. 2nd ed. New York: Plenum Press; 53–79 4. Pictet R, Rutter WJ 1972 Development of the embryonic endocrine pancreas. In: Steiner D, Freinkel N, eds. Handbook of physiology. Section 7: Endocrinology, vol I. Baltimore: Williams and Wilkins; 25– 66 5. Gittes GK, Galante PE, Hanahan D, Rutter WJ, Debas HT 1996 Lineagespecific morphogenesis in the developing pancreas: role of mesenchymal factors. Development 122:439 – 447 6. Miralles F, Czernichow P, Scharfmann R 1998 Follistatin regulates the relative proportions of endocrine versus exocrine tissue during pancreatic development. Development 125:1017–1024 7. Miralles F, Czernichow P, Ozaki K, Itoh N, Scharfmann R 1999 Signaling through fibroblast growth factor receptor 2b plays a key role in the development of the exocrine pancreas. Proc Natl Acad Sci USA 96:6267– 6272 8. Hart AW, Baeza N, Apelqvist A, Edlund H 2000 Attenuation of FGF signalling in mouse ␤-cells leads to diabetes. Nature 408:864 – 868 9. Kraus MH, Issing W, Miki T, Popescu NC, Aaronson SA 1989 Isolation and characterization of ERBB3, a third member of the ERBB/epidermal growth factor receptor family: evidence for overexpression in a subset of human mammary tumors. Proc Natl Acad Sci USA 86:9193–9197 10. Plowman GD, Culouscou J-M, Whitney GS, Green JM, Carlton GW, Foy L, Neubauer MG, Shoyab M 1993 Ligand-specific activation of HER4/ p180erbB4, a fourth member of the epidermal growth factor receptor family. Proc Natl Acad Sci USA 90:1746 –1750 11. Miettinen PJ, Heikinheimo K 1992 Transforming growth factor-␣ (TGF-␣) and insulin gene expression in human fetal pancreas. Development 114:833– 840 12. Miettinen PJ, Huotari M-A, Koivisto T, Ustinov J, Palgi J, Rasilainen S, Lehtonen E, Keski-Oja J, Otonkoski T 2000 Impaired migration and delayed differentiation of pancreatic islet cells in mice lacking EGF-receptors. Development 127:2617–2627 13. Kritzik MR, Krahl T, Good A, Gu D, Fox H, Sarvetnick N 2000 Expression of ErbB receptors during pancreatic islet development and regrowth. J Endocrinol 165:67–77 14. Erickson SL, O’Shea KS, Ghaboosi N, Loverro L, Frantz G, Bauer M, Lu LH, Moore MW 1997 ErbB3 is required for normal cerebellar and cardiac development: a comparison with ErbB2- and heregulin-deficient mice. Development 124:4999 –5011 15. Gassmann M, Casagranda F, Orioli D, Simon H, Lai C, Klein R, Lemke G 1995 Aberrant neural and cardiac development in mice lacking the ErbB4 neuregulin receptor. Nature 378:390 –394 16. Lee K-F, Simon H, Chen H, Bates B, Hung M-C, Hauser C 1995 Requirement for neuregulin receptor erbB2 in neural and cardiac development. Nature 378:394 –398 17. Sanvito F, Herrera P-L, Huarte J, Nichols A, Montesano R, Orci L, Vassalli J-D 1994 TGF-␤1 influences the relative development of the exocrine and endocrine pancreas in vitro. Development 120:3451–3462 18. Arnush M, Gu D, Baugh C, Sawyer SP, Mroczkowski B, Krahl T, Sarvetnick N 1996 Growth factors in the regenerating pancreas of ␥-interferon transgenic mice. Lab Invest 74:985–990 19. Krakowski ML, Kritzik MR, Jones EM, Krahl T, Lee J, Arnush M, Gu D, Mroczkowski B, Sarvetnick N 1999 Transgenic expression of epidermal growth factor and keratinocyte growth factor in ␤-cells results in substantial morphological changes. J Endocrinol 162:167–175 20. Kaneto H, Miyagawa J, Kajimoto Y, Yamamoto K, Watada H, Umayahara Y, Hanafusa T, Matsuzawa Y, Yamasaki Y, Higashiyama S, Taniguchi N 1997 Expression of heparin-binding epidermal growth factor-like growth factor during pancreas development. A potential role of PDX-1 in transcriptional activation. J Biol Chem 272:29137–29143 21. Shing Y, Christofori G, Hanahan D, Ono Y, Sasada R, Igarashi K, Folkman J 1993 Betacellulin: a mitogen from pancreatic ␤ cell tumors. Science 259:1604 – 1607 22. Miyagawa J, Hanafusa O, Sasada R, Yamamoto K, Igarashi K, Yamamori K, Seno M, Tada H, Nammo T, Li M, Yamagata K, Nakajima H, Namba M, Kuwajima M, Matsuzawa Y 1999 Immunohistochemical localization of betacellulin, a new member of the Egf family, in normal human pancreas and islet tumor cells. Endocr J 46:755–764 23. Mashima H, Ohnishi H, Wakabayashi K, Mine T, Miyagawa J, Hanafusa T, Seno M, Yamada H, Kojima I 1996 Betacellulin and activin A coordinately convert amylase-secreting pancreatic AR42J cells into insulin-secreting cells. J Clin Invest 97:1647–1654

4446


24. Huotari M-A, Palgi J, Otonkoski T 1998 Growth factor mediated proliferation and differentiation of insulin producing INS-1 and RINm5F cells: identification of betacellulin as a novel ␤-cell mitogen. Endocrinology 139:1494 –1499 25. Harari D, Tzahar E, Romano J, Shelly M, Pierce JH, Andrews GC, Yarden Y 1999 Neuregulin-4: a novel growth factor that acts through the ErbB-4 receptor tyrosine kinase. Oncogene 18:2681–2689 26. Chen X, Levkowitz G, Tzahar E, Karunagaran D, Lavi S, Ben-Baruch N, Leitner O, Ratzkin BJ, Bacus SS, Yarden Y 1996 An immunological approach reveals biological differences between the two NDF/heregulin receptors, ErbB-3 and ErbB-4. J Biol Chem 271:7620 –7629 27. Miettinen PJ, Berger JE, Meneses J, Phung Y, Pedersen RA, Werb Z, Derynck R 1995 Epithelial immaturity and multiorgan failure in mice lacking epidermal growth factor receptor. Nature 376:337–341 28. Siblia M, Wagner EF 1995 Strain-dependent epithelial defects in mice lacking the EGF receptor. Science 269:234 –238 29. Threadgill DW, Dlugosz AA, Hansen LA, Tennenbaum T, Lichti U, Yee D, La Mantia C, Mourton T, Herrup K, Harris RC, Barnard JA, Yuspa SH, Coffey RJ, Magnuson T 1995 Targeted disruption of mouse EGF receptor: effect of genetic background on mutant phenotype. Science 269:230 –234 30. Saxen L, Lehtonen E 1987 Embryonic kidney in organ culture. Differentiation 36:2–11 31. Giard DJ, Aaronson SA, Todaro GJ, Arnstein P, Kersey JH, Dosik H, Parks WP 1973 In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J Natl Cancer Inst 51:1417–1423 32. Ritvos O, Tuuri T, Eramaa M, Sainio K, Hilden K, Saxen L, Gilbert SF 1995 Activin disrupts epithelial branching morphogenesis in developing glandular organs of the mouse. Mech Dev 50:229 –245 33. Pinkas-Kramarski R, Eilam R, Alroy I, Levkowitz G, Lonai P, Yarden Y 1997 Differential expression of NDF/neuregulin receptors ErbB-3 and ErbB-4 and involvement in inhibition of neuronal differentiation. Oncogene 15:2803–2815 34. Teitelman G, Alpert S, Polak JM, Martinez A, 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 35. Jensen J, Heller RS, Funder-Nielsen T, Pedersen EE, Lindsell C, Weinmaster G, Madsen OD, Serup P 2000 Independent development of pancreatic ␣- and ␤-cells from neurogenin 3-expressing precursors: a role for the Notch pathway in repression of premature differentiation. Diabetes 49:163–176 36. Riese II DJ, Bermingham Y, van Raaij TM, Buckley S, Plowman GD, Stern DF 1996 Betacellulin activates the epidermal growth factor receptor and erbB-4, and induces cellular response patterns distinct from those stimulated by epidermal growth factor or neuregulin-␤. Oncogene 12:345–353


37. Ishiyama N, Kanzaki M, Seno M, Yamada H, Kobayashi I, Kojima I 1998 Studies on the betacellulin receptor in pancreatic AR42J cells. Diabetologia 41:623– 628 38. Cras-Meneur C, Elghazi L, Czernichow P, Scharfmann R 2001 Epidermal growth factor increases undifferentiated pancreatic embryonic cells in vitro: a balance between proliferation and differentiation. Diabetes 50:1571–1579 39. Yamaoka T, Idehara C, Yano M, Matsushita T, Yamada T, Ii S, Moritani M, Hata J, Sugino H, Noji S, Itakura M 1998 Hypoplasia of pancreatic islets in transgenic mice expressing activin receptor mutants. J Clin Invest 102:294 –301 40. Bhushan A, Itoh N, Kato S, Thiery JP, Czernichow P, Bellusci S, Scharfmann R 2001 Fgf10 is essential for maintaining the proliferative capacity of epithelial progenitor cells during early pancreatic organogenesis. Development 128: 5109 –5117 41. Miettinen PJ 1993 TGF-␣ and EGF expression in human fetal GI-tract. Pediatr Res 33:481– 486 42. Rappolee DA, Brenner CA, Schultz R, Mark D, Werb Z 1988 Developmental expression of PDGF, TGF-␣, and TGF-␤ genes in preimplantation mouse embryos. Science 241:1823–1825 43. Wilcox JN, Derynck R 1988 Developmental expression of transforming growth factor ␣ and ␤ in mouse fetus. Mol Cell Biol 8:3415–3422 44. Graus-Porta D, Beerli RR, Daly JM, Hynes NE 1997 ErbB-2, the preferred heterodimerization partner of all ErbB receptors, is a mediator of lateral signaling. EMBO J 16:1647–1655 45. Sliwkowski MX, Schaefer G, Akita RW, Lofgren JA, Fitzpatrick VD, Nuijens A, Fendly BM, Cerione RA, Vandlen RL, Carraway III KL 1994 Coexpression of erbB2 and erbB3 proteins reconstitutes a high affinity receptor for heregulin. J Biol Chem 269:14661–14665 46. Sweeney C, Lai C, Riese DJ, Diamonti AJ, Cantley LC, Carraway KL 2000 Ligand discrimination in signaling through an erbb4 receptor homodimer. J Biol Chem 275:19803–19807 47. Maedler K, Spinas GA, Lehmann R, Sergeev P, Weber M, Fontana A, Kaiser N, Donath MY 2001 Glucose induces ␤-cell apoptosis via upregulation of the Fas receptor in human islets. Diabetes 50:1683–1690 48. Soria B, Roche E, Berna G, Leon-Quinto T, Reig JA, Martin F 2000 Insulinsecreting cells derived from embryonic stem cells normalize glycemia in streptozotocin-induced diabetic mice. Diabetes 49:157–162 49. Bonner-Weir S, Taneja M, Weir GC, Tatarkiewicz K, Song KH, Sharma A, O’Neil JJ 2000 In vitro cultivation of human islets from expanded ductal tissue. Proc Natl Acad Sci USA 97:7999 – 8004 50. Ramiya VK, Maraist M, Arfors KE, Schatz DA, Peck AB, Cornelius JG 2000 Reversal of insulin-dependent diabetes using islets generated in vitro from pancreatic stem cells. Nat Med 6:278 –282