Epidermal Growth Factor Increases Undifferentiated Pancreatic Embryonic Cells In Vitro A Balance Between Proliferation and Differentiation Corentin Cras-Me´neur, Lynda Elghazi, Paul Czernichow, and Raphael Scharfmann

During embryonic life, the development of a proper implicated in pancreatic development (1), the control of mass of mature pancreatic tissue is thought to require pancreatic cell growth and differentiation by growth fac- the proliferation of precursor cells, followed by their tors remains poorly understood. The goal of the present differentiation into endocrine or acinar cells. We inves- work was to determine whether by perturbing the prolif- tigated whether perturbing the proliferation of pre- eration of pancreatic precursor cells, the final mass of cursor cells in vitro could modify the final mass of endocrine cells that develop could be modified, with the endocrine tissue that develops. For that purpose, we objective of defining new strategies for increasing the final used activators or inhibitors of signals mediated by ␤-cell mass. receptor tyrosine kinases. We demonstrated that when It has been proposed that ligands of receptor tyrosine embryonic day 13.5 rat pancreatic epithelium is cul- tured in the presence of PD98059, an inhibitor of the kinases that are implicated in the control of cell prolifer- mitogen-activated protein (MAP) kinase, epithelial cell ation and differentiation in a large number of organs (2) proliferation is decreased, whereas endocrine cell dif- could also control pancreatic development (3,4). However, ferentiation is activated. On the other hand, in the the exact function of such ligands of receptor tyrosine presence of epidermal growth factor (EGF), an activa- kinases during pancreatic development is not well estab- tor of the MAP kinase pathway, the mass of tissue that lished. It has recently been shown that in mice deficient in develops is increased, whereas the absolute number of epidermal growth factor receptor (EGFR), a member of endocrine cells that develops is decreased. Under this the receptor tyrosine kinase family, pancreatic cell devel- last condition, a large number of epithelial cells prolif- opment was abnormal (5). However, the way by which erate but remain undifferentiated. In a second step, signals mediated by EGFR control the final ␤-cell mass when EGF is removed from the pool of immature pan- creatic epithelial cells, the cells differentiate en masse that develops was not elucidated. Theoretically, ligands into insulin-expressing cells. The total number of insu- of EGFR could control the proliferation of precursor lin-expressing cells that develop can thus be increased cells, their ability to differentiate into mature cells, the by first activating the proliferation of immature epithe- level of proliferation of such mature cells, and, finally, cell lial cells with growth factors, thus allowing an increase survival. in the pool of precursor cells, and next allowing their Recently we developed and characterized an in vitro differentiation into endocrine cells by removing the model of pancreatic development (6,7) that allowed for growth factor. This strategy suggests a possible tissue testing the effects of specific growth factors on the devel- ␤ engineering approach to expanding -cells. Diabetes 50: opment of the endocrine pancreas (8,9). In the present 1571–1579, 2001 study, we used this experimental system to further study the implication of cell proliferation in the control of endocrine cell differentiation. Our data indicate that epi- roliferation of precursor cells and their differen- dermal growth factor (EGF), a ligand of EGFR, acts as a tiation into mature cells represent two crucial growth factor for embryonic pancreatic epithelial cells. processes for the proper development of an When epithelial cell proliferation occurs, endocrine cell Porgan. The pancreatic gland develops from the differentiation is repressed, while putative endocrine pre- gut endodermal epithelium. It is thought that such epithe- cursor cells remain present. Such precursor cells keep the lial precursor cells will first proliferate and then differen- ability to differentiate en masse in insulin-expressing cells tiate into endocrine cells forming the islets of Langerhans when EGF is removed. Taken together, these experiments and exocrine pancreatic tissue. Although progress has indicate that the mass of undifferentiated pancreatic epi- recently been made concerning the transcription factors thelial cells and the balance between epithelial cell prolif- eration and differentiation into insulin-expressing cells can be controlled by signals mediated by receptor tyrosine From the INSERM U457, Hospital R. Debre´, Paris, France. Address correspondence and reprint requests to Raphael Scharfman, PhD, kinases such as EGFR. Such a strategy can thus be used to INSERM U457, Hospital R. Debre´, 48, Boulevard Se´rurier, 75019 Paris, France. increase the final mass of ␤-cells developed. E-mail: [email protected]. Received for publication 19 September 2000 and accepted in revised form 30 March 2001. RESEARCH DESIGN AND METHODS BrdU, bromo-deoxy-uridine; e, embryonic day; EGF, epidermal growth factor; EGFR, EGF receptor; FCS, fetal calf serum; HBSS, Hanks’ balanced Dissections. Pregnant Wistar rats were purchased from Janvier breeding salt solution; MAP, mitogen-activated protein; PBS, phosphate-buffered saline. center (CERJ, Le Genet, France). The morning the vaginal plug was discovered

DIABETES, VOL. 50, JULY 2001 1571 GROWTH FACTORS AND ␤-CELL EXPANSION was designated as embryonic day (e) 0.5. Animals had free access to food (MAP) kinase pathway, on cell proliferation and differen- pellets and water. Pregnant rats at 13.5 days of gestation were killed by tiation. Epithelial rudiments were cultured for 3 days with injection of a lethal dose of pentobarbital (Sanofi, Libourne, France). The embryos were harvested, and the dorsal pancreatic buds were dissected as or without PD98059. As shown in Fig. 1A and B,inthe described (10). The pancreatic epithelium was separated from its surrounding presence of PD98059, the size of the rudiments and the mesenchyme, as described previously (6) with minor modifications. Briefly, number of BrdU-positive cells were decreased when com- the stomach, pancreas, and a small portion of the intestine were dissected pared with controls grown for the same period in the together and incubated with 0.5 mg/ml of collagenase A (Boehringer-Mann- heim, Mannheim, Germany) at 37°C for 30 min. They were then washed absence of PD98059. Moreover, the number of BrdU- several times with Hanks’ balanced salt solution (HBSS; Gibco) at 4°C. The positive cells per surface unit was also decreased (P Ͻ epithelium was then mechanically depleted from the surrounding mesen- 0.05) in the presence of PD98059. In fact, while the size of chyme using needles on a 0.25% Agar, 25% HBSS, 75% RPMI (Gibco) gel in a the tissue increased during the 3-day culture period in Petri dish. Organ culture. Pancreatic epithelia were embedded into 500 ␮l of collagen control conditions (Fig. 1B, top left panel; compare col- gel (10% RPMI 10ϫ [Sigma-Aldrich], 80% type I rat tail collagen [2.5 mg/ml; umns 1 and 2), no increase in terms of size of the tissue Sigma-Aldrich], and 10% sodium bicarbonate in NaOH 0.1 mol/l) into four-well was observed during a 3-day culture period in the presence plates (Nunc), as previously described (6,11). Once the gel had polymerized, of PD98059 when compared with uncultured rudiments 500 ␮l of RPMI 1640 (Gibco) containing penicillin (100 U/ml), streptomycin (100 ␮g/ml), HEPES (10 mmol/l), L-glutamin (2 mmol/l), and nonessential (Fig. 1B, top left panel; compare columns 1 and 3). amino acid (1ϫ; Gibco) were added. The medium was supplemented with 1% At the same time, the absolute mass of endocrine cells heat-inactivated fetal calf serum (FCS) (Hyclone). At this concentration of that differentiated in the presence of PD98059 was in- FCS, the development of the endocrine tissue was identical to the one creased when compared with the ones that developed in obtained with 10% FCS. Cultures were maintained at 37°C in a humidified the absence of PD98059 (Fig. 1A and B, bottom panel). In atmosphere of 95% O2/5% CO2. The medium was changed every 2 days, and 50 ng/ml human recombinant EGF (Sigma-Aldrich (diluted in phosphate-buffered fact, while the endocrine cell mass was increased by saline 0.1% bovine serum albumin or its carrier) were added every day. This 1.8-fold during a 3-day culture period in the absence of daily addition and concentration of EGF was chosen because in preliminary PD98059, it was increased 3.3-fold during a 3-day culture experiments it gave the strongest effect on cell differentiation. PD98059 (Calbiochem) was diluted in DMSO and used at 50 ␮mol/l. Comparable DMSO period in the presence of PD98059 when compared with concentrations were made in culture medium. To label cells in the S phase, uncultured rudiments (Fig. 1B, bottom panel). bromo-deoxy-uridine (BrdU; 10 ␮mol/l) was added to the culture medium. At EGF increases epithelial cell mass and represses the indicated times, the pancreatic epithelia were photographed and fixed for endocrine cell differentiation. When e13.5 pancreatic immunohistochemistry, as described below. Immunohistochemistry. Pancreatic rudiments were photographed and fixed epithelial buds were grown in culture in a collagen gel in in formalin 3.7% for 1 h, preembedded in an agarose gel (4% of type VII low the absence of EGF over 7 days, the size of the tissue gellingϪtemperature agarose [Sigma] in Tris-buffered saline), and embedded increased 1.5-fold. On the other hand, in the presence of in paraffin. Consecutive sections 4 ␮m thick were collected on gelatinized EGF, the size of the tissue increased by 2.8-fold during a glass slides, deparaffinized, and microwaved for 12 min. Immunohistochem- istry was performed as previously described (6,8,9) using the following 7-day culture period. Thus, with EGF, after 7 days in antibodies: mouse anti-human insulin (Sigma; 1/2,500), mouse anti-human culture, a 1.9-fold increase in size (P ϭ 0.0040) was seen in glucagon (Sigma; 1/2,000), rabbit anti-glucagon (Diasorin; 1/2,000), rabbit EGF-treated epithelial buds compared with controls buds anti-human amylase (Sigma; 1/300), mouse anti-human pan-cytokeratin (Sig- grown in the same conditions (Fig. 2). Immunohistochem- ma; 1/100), mouse anti-BrdU (Amersham; 1/4), and mouse anti-porcine vimen- tin (Dako; clone V9, 1/100). istry analysis was next performed to define the effect of The fluorescent secondary antibodies obtained from Jackson Immunore- EGF treatment on the absolute mass of endocrine and aci- search Laboratories were fluorescein anti-rabbit antibodies (1/200), fluores- nar tissue that developed. EGF treatment did not modify cein anti-mouse antibodies (1/150), Texas Red anti-mouse antibodies (1/200), the absolute amylase-expressing cell mass (Fig. 3A and data and Texas Red anti-rabbit antibodies (1/200) In some experiments, adjacent sections were stained for endocrine plus not shown). On the other hand, in the presence of EGF, the mesenchymal markers (insulin/glucagon plus vimentin) and acinar plus absolute mass of insulin- and glucagon-expressing cells epithelial markers (amylase and cytokeratin, respectively). False colors were that developed was strongly decreased when compared generated using Graphic Converter 3.9.1 (Lemke Software), and the images with controls (2.96-fold decrease for insulin, P Ͻ 0.005; from adjacent sections were overlaid using Photoshop 5.5 (Adobe). Ͻ Confocal microscopy was performed using a CS4D confocal microscope 5.23-fold decrease for glucagon, P 0.005) (Fig. 3). (Leica). Sections 4 ␮m thick were analyzed every 0.5 ␮m. Recomposed color A large number of cells that stained negative for images were generated using GraphicConverter 4.0.1 (Lemke Software). endocrine/exocrine markers remained present after Images obtained for the different fluorochromes were then overlaid using 7 days in culture with EGF. After 7 days of culture in the Photoshop 5.5 (Adobe). Surface quantification and statistical analysis. For each pancreatic presence of EGF, a large number of cells stained negative epithelium that was analyzed, all sections (50Ϫ100 per rudiment) were for endocrine and acinar markers when compared with numbered using a Hamamatsu C5810 cooled tri-CCD camera. All images were rudiments grown in the absence of EGF (Fig. 4A and B). taken at the same magnification. On every image, the surfaces of the different As shown in Fig. 4B, such negative cells expressed high stainings were quantified with IPLab (version 3.2.4, Scananalytics). The levels of cytokeratins and were more numerous in EGF- surfaces were then added up per rudiment. When a section was missing (Ͻ2% of the total section number), we took the mean value of the staining in the treated buds than in buds grown in the absence of EGF adjacent sections into account. To quantify the total size of the rudiments, the (compare Fig. 4A and B). In fact, 4.65 times more cells surface of all the sections of the rudiment prepared for immunohistochemistry (P Ͻ 0.01) stained negative for endocrine and acinar were measured, and areas of all sections of each rudiment were added. markers in cultures performed in the presence of EGF Statistical analysis was performed using StatView 5.0 (SAS Institute). Because we could not assume that the data followed a binomial law, we used than in the absence of EGF. On the other hand, both in nonparametric Mann-Whitney U tests to compare the different groups. EGF-treated and control cultures, very few cells expressed vimentin, a marker of mesenchymal cells, indicating that RESULTS EGF has no effect on the few remaining pancreatic mes- Endocrine cell development is activated when cell enchymal cells (Fig. 4A and B). proliferation is repressed. We first tested the effects of To define whether, in the presence of EGF, cells ex- PD98059, an inhibitor of the mitogen-activated protein pressing high levels of cytokeratins proliferate, a 6-h pulse

1572 DIABETES, VOL. 50, JULY 2001 C. CRAS-ME´ NEUR AND ASSOCIATES

FIG. 1. Effect of the inhibitor of the MAP kinase pathway PD98059 on cell proliferation and differentiation. Pancreatic epithelia were grown in culture for 3 days in the absence or presence of PD98059 (50 ␮mol/l). A: Immunostaining for BrdU (red) and insulin plus glucagon (green) in controls (DMSO) and PD98059-treated rudiments. B: Quantification of the size and number of BrdU-positive cells and the cell surface area occupied by insulin/glucagon؊positive cells that developed during 3 days in the presence of PD98059 when compared with control rudiments grown for the same period in the absence of PD98059. The sizes of the epithelia and total surface of insulin/glucagon؊positive cells were also .determined before culture. Three rudiments were analyzed for each condition. Data are means ؎ SE. *P < 0.05

DIABETES, VOL. 50, JULY 2001 1573 GROWTH FACTORS AND ␤-CELL EXPANSION

FIG. 2. Effect of EGF on the size of pancreatic epithelia in culture. e13.5 pancreatic epithelial rudiments grown in culture for 7 days with or without EGF. A: Photographs of the epithelia before and after 7 days of culture. The pictures were taken at the same magnification. Left panel: an epithelium before culture; middle panel: an epithelium grown for 7 days in control conditions, with endocrine budding visible (arrows); right panel: an epithelium grown for 7 days with EGF. B: Quantification of the total size of the epithelial rudiments before and after 7 days in culture in the absence or in the presence of EGF. The surface of all sections prepared for immunohistochemistry were measured and areas of all sections .were added. Seven rudiments were analyzed for each condition. Mean values in ␮m2 ؎ SE are shown on the graph. **P < 0.005 of BrdU was performed on day 7 of the culture. As shown and b] and B [compare columns 1 and 2]). On the other in Fig. 4D and E, cells that stained positive for BrdU and hand, no major increase was seen in the absolute number cytokeratin could be detected in rudiments grown with of insulin-expressing cells that developed during 7 or 14 EGF, indicating that cells expressing high levels of cyto- days in the absence of EGF (Fig. 5A [compare panels c and keratins proliferated upon EGF treatment. On the other d] and B [compare columns 3 and 4]). hand, in the absence of EGF, the few cells expressing high Next, the fate of the cells that proliferated during the levels of cytokeratins did not proliferate (Fig. 4C). The few first week of culture with EGF was followed. For that pur- cells that did proliferate in the absence of EGF were pose, the epithelial buds were grown for 7 days with EGF. amylase-expressing cells (data not shown). At the end of the seventh day, a BrdU pulse was per- Cells expressing high levels of cytokeratins that formed. Then 6 h later, the epithelial buds were either developed in the presence of EGF can differentiate fixed for immunohistochemistry or washed, transferred to into insulin-expressing cells. We next asked whether a new gel, and grown for an additional 7-day period in cells expressing high levels of cytokeratins that prolifer- B ated in the presence of EGF had the ability to differentiate the absence of EGF. As shown in Fig. 6 , cells that stained into insulin-expressing cells. For that purpose, two sets of positive for both insulin and BrdU could be found after 7 experiments were performed. additional days in culture, whereas such double-positive First, e13.5 epithelial buds were grown in culture in the cells were not present at the end of the BrdU pulse at day presence of EGF. After 7 days, EGF was removed and the 7 (Fig. 6A). Only a fraction (ϳ50%) of BrdU-positive cells rudiments were kept in culture for an additional week. A were insulin positive after a pulse at day 7 and 7 additional 8.93-fold increase (P ϭ 0.0027) in the number of insulin- days in culture (Fig. 6B), suggesting that not all the poten- expressing cells was observed in epithelial buds treated tial stem cells that were proliferating with EGF did differ- for 7 days with EGF followed by a 7-day period in the entiate into insulin-expressing cells upon EGF removal. absence of EGF when compared with epithelial buds kept Taken together, these data strongly suggest that insulin- in culture for 7 days with EGF (Fig. 5A [compare panels a expressing cells did differentiate in the absence of EGF

1574 DIABETES, VOL. 50, JULY 2001 C. CRAS-ME´ NEUR AND ASSOCIATES

FIG. 3. Immunohistological analysis of pancreatic epithelium grown in vitro during 7 days with or without EGF. Pancreatic epithelia were grown for 7 days without or with EGF. They were sectioned and consecutive sections were stained for amylase or insulin and glucagon. A: Super- imposition of staining from adjacent sections for amylase (green), insulin (red), and glucagon (blue, artificial color) in epithelia developed without (control) or with EGF. B: Absolute surface areas in ␮m2 occupied by insulin- (left) and glucagon-expressing (right) cells in rudiments grown for 7 days with or without EGF. Six to seven rudiments were analyzed for each condition. **P < 0.005. from precursor cells that were proliferating in the pres- It is well established that during embryonic life, endo- ence of EGF. crine and acinar cells derive from precursor cells located More insulin-expressing cells are generated when the in the pancreatic epithelium (12). Such precursor cells, proliferation of putative precursor cells is first in- which are not yet characterized, will proliferate and dif- duced by EGF. The insulin cell mass that developed ferentiate into mature cells. It has been shown in vivo that during 14 days in the absence of EGF was compared with when isolated embryonic pancreatic epithelium is grafted the one that developed when epithelial buds were cultured under the kidney capsule, it develops into endocrine during the first week in the presence of EGF followed by tissue, but acinar tissue does not develop. These experi- 1 week in the absence of EGF. As shown in Fig. 5A ments indicated that the development of the endocrine (compare panels b and d) and B (compare columns 2 and tissue can occur in the absence of signals from the 4), more insulin-expressing cells (1.78-fold increase; P Ͻ mesenchyme that surrounds the epithelium (13). It has 0.05) developed when the epithelial buds were cultured also been shown in vitro that when e12Ϫe13 pancreatic during the first week with EGF. This indicated that EGF epithelium is cultured in the absence of mesenchyme, increased the mass of precursor cells that then differenti- pancreatic endocrine cells develop properly (6,7). In fact, ated into insulin-expressing cells. in vitro, more insulin-expressing cells develop in the absence of mesenchyme, when epithelial cell proliferation DISCUSSION is low, than in the presence of mesenchyme, when prolif- We demonstrated that in vitro, EGF has the ability to eration is high, suggesting that signals from the mesen- expand the pool of embryonic pancreatic epithelial pre- chyme activate the proliferation of immature epithelial cursor cells. In the meantime, their differentiation into cells and repress their differentiation into endocrine tissue endocrine tissue is repressed. Once expanded and after (6). Taken together, these results strongly suggest that EGF is removed, these precursor cells differentiate and activation of the proliferation of immature embryonic form endocrine cells by a default pathway. epithelial cells represses their differentiation into endo-

DIABETES, VOL. 50, JULY 2001 1575 GROWTH FACTORS AND ␤-CELL EXPANSION

FIG. 4. Characterization of the cells developed in the presence of EGF that stain negative for endocrine and acinar markers. A and B: Pancreatic epithelia were grown for 7 days in the absence (A) or presence of EGF (B). Consecutive sections were analyzed by immunohistochemistry for amylase (green), cytokeratin (red), insulin plus glucagon (blue, artificial color), and vimentin (yellow, artificial color). The images were next superimposed. Note that in EGF-treated rudiments, nearly all cells that stained negative for endocrine and acinar markers expressed high levels of cytokeratin. Moreover, very few cells stained positive for vimentin, either in the presence or absence of EGF. C؊E: Pancreatic epithelia were grown for 7 days without (C) or with (D and E) EGF. Double-staining for cytokeratin (red) and BrdU (green). The picture presented in (E) was obtained after confocal microscopy. crine cells. Cell proliferation would thus act as a repressor that developed had decreased. These undifferentiated cells of endocrine cell differentiation. The data presented in this develop massively into insulin-expressing cells by a de- study further support this hypothesis. Indeed, when e13.5 fault pathway after EGF removal. It is important that sig- pancreatic epithelium is cultured in the presence of EGF, nificantly more insulin-expressing cells were present after epithelial cell mass increases, whereas endocrine cell dif- 14 days of culture when EGF was added during the first ferentiation is repressed. On the other hand, in the pres- week and removed during the second week, than after 14 ence of PD98059, an inhibitor of the MAP kinase pathway, days of culture in the absence of EGF. Thus, in vitro, the cell proliferation is decreased, but endocrine cell develop- development of a normal number of endocrine cells in the ment is activated. However, the possibility cannot be fully pancreas requires a proper proliferation of precursor cells excluded that the MAP kinase inhibitor or EGF acts direct- and their differentiation at the right time. Perturbation of ly on both cell proliferation and endocrine differentiation, one of these events decreases the development of the without any link between proliferation and differentiation. endocrine tissue. This is the case in vivo in mice with a In the present study, in control cultures performed in perturbed Notch signaling pathway in the pancreas, such the absence of EGF, most of the cells differentiated into as mice overexpressing neurogenin-3 under the control of endocrine or exocrine cells during a 7-day culture period. the PDX-1 promoter (14) or mice deficient in Hes1 (15). In Little additional cell differentiation occurred when the these mutant mice, an early and massive differentiation of epithelium was grown during an additional week in the the pool of precursor cells occurs. Such a pool is depleted absence of EGF to activate endocrine cell differentiation and the final mass of endocrine cells that will develop is (Fig. 5B, columns 3 and 4). Thus it seems that in these decreased. This could also be the case for mice deficient in conditions, the pool of precursor cells was strongly de- EGFR that have a small pancreas with an abnormal pleted. On the other hand, in the presence of EGF, a large development of the endocrine tissue (5). number of cells proliferated and remained undifferentiated In vitro, when pancreatic epithelium is grown in control after 7 days in culture, and the number of endocrine cells conditions, the majority of the cells differentiate into

1576 DIABETES, VOL. 50, JULY 2001 C. CRAS-ME´ NEUR AND ASSOCIATES

FIG. 5. Epithelial cells that developed during the first week of culture in the presence of EGF differentiated into insulin-expressing cells when EGF was removed. A: Pancreatic epithelia were grown for 7 days with (a and b) or without EGF (c and d). The rudiments were either fixed (a and c) or grown for an additional week in the absence of EGF (b and d). The rudiments were sectioned and analyzed by immunohistochemistry for insulin. B: Quantification of insulin-positive cell mass that developed in each condition. An 8.93-fold (P < 0.005) increase in the insulin-positive cell mass occurred during the second week of culture when EGF was removed at day 7. Moreover, at day 14, more insulin-expressing cells developed (P < 0.05) when the rudiments were grown during the first week with EGF and during the second week in the absence of EGF when compared with rudiments grown 2 weeks in the absence of EGF. Six to seven rudiments were analyzed in each condition. *P < 0.05; **P < 0.005; NS, not significantly different.

DIABETES, VOL. 50, JULY 2001 1577 GROWTH FACTORS AND ␤-CELL EXPANSION

FIG. 6. Insulin-expressing cells that devel- oped after EGF removal derived from cells that proliferated when EGF was present. Pancreatic epithelia were grown for 7 days with EGF and pulsed with BrdU (10 ␮mol/l) during the last 6 h. They were either fixed (A) or kept in culture without BrdU during an additional week in the absence of EGF (B) and finally fixed. They were sectioned and immunohistochemistry for insulin (green) and BrdU (red) was performed. Although insulin-expressing cells present at day 7 stained negative for BrdU, some insulin-ex- pressing cells found at day 14 stained posi- tive for BrdU (white arrows), indicating that such insulin-expressing cells differentiated during the second week of culture in the absence of EGF from cells that were prolif- erating when EGF was present. endocrine or acinar tissue after a 7-day culture period. On ers expressed on their surface. This type of marker will be the other hand, in the presence of EGF, a large number of necessary to purify multipotent pancreatic embryonic cells remain negative for endocrine and acinar markers. In cells, as is the case for other tissues such as hematopoietic the present study, we provided different arguments show- stem cells (24) or neural crest stem cells (25). ing that cells that stain negative for endocrine and acinar This study demonstrated that proliferation and differen- markers are precursor cells. First, such cells that stain tiation can be balanced and that EGF can control this negative for endocrine and acinar markers express a high equilibrium. Whether this is operational in vivo needs to be level of cytokeratin; however, during development, endo- demonstrated. However, EGF is certainly efficient in vitro, crine and acinar cells derive from the endodermal epithe- and could be used to increase the amount of differentiated lial cells that stain positive for cytokeratins (16,17). ␤-cells in vitro. Recently, because of its therapeutic poten- Second, the cells expressing high levels of cytokeratin can tial, the expansion of ␤-cells in vitro has been the focus of proliferate upon EGF treatment and next differentiate into a large number of studies. In the vast majority of such insulin-expressing cells. Moreover, when epithelial buds studies, specific growth factors have been tested for their were grown with EGF, pulsed with BrdU at the end of the ability to increase the proliferation of pre-existing ␤-cells. 7th day, and grown for an additional week in the absence The growth effect of factors such as hepatocyte growth of EGF, cells that stained positive for both insulin and factor, growth hormone and prolactin, EGF, and platelet- BrdU were found at the end of the culture. Such double- derived growth factor has been tested on mature ␤-cells positive cells were not present at the end of the BrdU from rodents and humans (26–28). In these studies, no pulse at day 7. Third, the expansion of such a pool of major increase in ␤-cell mass has been found using these putative precursor cells gives rise to a high amount of factors. We demonstrated in the present study that an insulin-expressing cells when differentiation occurs upon alternative strategy is to increase the proliferation of EGF removal. precursor cells using ligands of receptor tyrosine kinases We demonstrated here that the cells that express high such as EGF. Once EGF is removed, such amplified levels of cytokeratin and stain negative for endocrine and precursor cells differentiate en masse into endocrine cells. acinar markers do proliferate upon EGF treatment. Such an effect of EGF on cell proliferation resembles the one ACKNOWLEDGMENTS found when pancreatic duct cells from adult guinea pigs are treated with EGF (18). Our data also demonstrate that C.C.-M. was a recipient of a fellowship from the Fondation at the same time, differentiation into endocrine cells is de France. This work was supported by grants from the repressed. However, the expanded cells keep the capacity Juvenile Diabetes Foundation International. to differentiate after EGF removal. The effects of EGF We thank Dr. M. Kedinger for her guidance with the presented here resemble the ones found when EGF is establishment of the mesenchyme depletion procedure added to cultures of embryonic or adult precursor cells and for her useful advice. I. Le Nin is acknowledged for her derived from the nervous system. In this tissue, EGF has technical assistance and Arnaud Mailleux (Institute Curie, been found to act as a mitogen for precursor cells, Paris) is acknowledged for help with confocal microscopy. inducing their proliferation while repressing their differ- entiation. When EGF is removed, cell differentiation into REFERENCES mature cells such as neurons or glia occurs (19–23). 1. Edlund E: Transcribing pancreas. Diabetes 47:1817–1823, 1998 Ligands of receptor tyrosine kinases such as EGF should 2. Schlessinger J, Ullrich A: Growth factor signaling by receptor tyrosine kinases. Neuron 9:383–391, 1992 thus be useful in generating a large pool of precursor cells. 3. LeBras S, Czernichow P, Scharfmann R: A search for tyrosine kinase Such a pool is useful for studying in detail these expanded receptors expressed in the rat embryonic pancreas. Diabetologia 41:1474– pancreatic precursor cells by characterizing specific mark- 1481, 1998

1578 DIABETES, VOL. 50, JULY 2001 C. CRAS-ME´ NEUR AND ASSOCIATES

4. LeBras S, Miralles F, Basmaciogullari A, Czernichow P, Scharfmann R: 16. Slack J: Developmental biology of the pancreas. Development 121:1569– Fibroblast growth factor 2 promotes pancreatic epithelial cell proliferation 1580, 1995 via functional fibroblast growth factor receptors during embryonic life. 17. Bouwens L: Cytokeratins and cell differentiation in the pancreas. J Pathol Diabetes 47:1236–1242, 1998 184:234–239, 1998 5. Miettinen PJ, Huotari M, Koivisto T, Ustinov J, Palgi J, Rasilainen S, 18. Verme TB, Hootman SR: Regulation of pancreatic duct epithelial growth in Lehtonen E, Keski-Oja J, Otonkoski T: Impaired migration and delayed vitro. Am J Physiol 258:G833ϪG840, 1990 differentiation of pancreatic islet cells in mice lacking EGF-receptors. 19. Reynolds B, Weiss S: Generation of neurons and astrocytes from isolated Development 127:2617–2627, 2000 cells of the adult mammalian central nervous system. Science 255:1707– 6. Miralles F, Czernichow P, Scharfmann R: Follistatin regulates the relative 1710, 1992 proportions of endocrine versus exocrine tissue during pancreatic devel- 20. Reynolds BA, Weiss S: Clonal and population analyses demonstrate that an opment. Development 125:1017–1024, 1998 EGF-responsive mammalian embryonic CNS precursor is a stem cell. Dev 7. Miralles F, Serup P, Cluzeaud F, Vandewalle A, Czernichow P, Scharfmann Biol 175:1–13, 1996 R: Characterization of beta cells developed in vitro from rat embryonic 21. Weiss S, Reynolds BA, Vescovi AL, Morshead C, Craig CG, van der Kooy D: pancreatic epithelium. Dev Dyn 214:116–126, 1999 Is there a neural stem cell in the mammalian forebrain? Trends Neurosci 8. Miralles F, Battelino T, Czernichow P, Scharfmann R: TGF-beta plays a key 19:387–393, 1996 role in morphogenesis of the pancreatic islets of Langerhans by controlling 22. Gritti A, Frolichsthal-Schoeller P, Galli R, Parati EA, Cova L, Pagano SF, the activity of the matrix metalloproteinase MMP-2. J Cell Biol 143:827– Bjornson CR, Vescovi AL: Epidermal and fibroblast growth factors behave 836, 1998 as mitogenic regulators for a single multipotent stem cell-like population 9. Miralles F, Czernichow P, Ozaki K, Itoh N, Scharfmann R: Signaling from the subventricular region of the adult mouse forebrain. J Neurosci through fibroblast growth factor receptor 2b plays a key role in the 19:3287–3297, 1999 development of the exocrine pancreas. Proc Natl Acad SciUSA 23. Tropepe V, Sibilia M, Ciruna BG, Rossant J, Wagner EF, van der Kooy D: 96:6267–6272, 1999 Distinct neural stem cells proliferate in response to EGF and FGF in the 10. Gittes G, Galante P: A culture system for the study of pancreatic organo- genesis. J Tiss Cult Meth 15:23–28, 1993 developing mouse telencephalon. Dev Biol 208:166–188, 1999 11. Montesano R, Mouron P, Amherdt M, Orci L: Collagen matrix promotes 24. Spangrude GJ, Heimfeld S, Weissman IL: Purification and characterization reorganization of pancreatic endocrine cell monolayers into islet-like of mouse hematopoietic stem cells. Science 241:58–62, 1988 organoids. J Cell Biol 97:935–939, 1983 25. Morrison SJ, White PM, Zock C, Anderson DJ: Prospective identification, 12. Pictet R, Rutter W: Development of the embryonic pancreas. In Handbook isolation by flow cytometry, and in vivo self-renewal of multipotent of Physiology. Vol 1. Steiner DF, Freinkel N, Eds. Baltimore, Williams & mammalian neural crest stem cells. Cell 96:737–749, 1999 Wilkins, 1972, p. 25–66 26. Brelje TC, Sorenson RL: Role of prolactin versus growth hormone on islet 13. Gittes G, Galante P, Hanahan D, Rutter W, Debas H: Lineage specific B-cell proliferation in vitro: implications for pregnancy. Endocrinology morphogenesis in the developing pancreas: role of mesenchymal factors. 128:45–57, 1991 Development 122:439–447, 1996 27. Hayek A, Beattie GM, Cirulli V, Lopez AD, Ricordi C, Rubin JS: Growth 14. Apelqvist A, Li H, Sommer L, Beatus P, Anderson DJ, Honjo T, Hrabe de factor/matrix-induced proliferation of human adult beta-cells. Diabetes Angelis M, Lendahl U, Edlund H: Notch signalling controls pancreatic cell 44:1458–1460, 1995 differentiation. Nature 400:877–881, 1999 28. Lefebvre VH, Otonkoski T, Ustinov J, Huotari MA, Pipeleers DG, Bouwens 15. Jensen J, Pedersen EE, Galante P, Hald J, Heller RS, Ishibashi M, L: Culture of adult human islet preparations with hepatocyte growth factor Kageyama R, Guillemot F, Serup P, Madsen OD: Control of endodermal and 804G matrix is mitogenic for duct cells but not for ␤-cells. Diabetes endocrine development by Hes-1. Nat Genet 24:36–44, 2000 47:134–137, 1998

DIABETES, VOL. 50, JULY 2001 1579