Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2007 Sep 6.
Published in final edited form as: Dev Biol. 2007 Mar 2;305(2):508–521. doi: 10.1016/j.ydbio.2007.02.033

TGF-β isoform signaling regulates secondary transition and mesenchymal-induced endocrine development in the embryonic mouse pancreas

Sidhartha S Tulachan 1, Eri Tei 2, Mark Hembree 2, Christopher Crisera 3, Krishna Prasadan 1, Masayuki Koizumi 1, Sohail Shah 1, Ping Guo 1, Erwin Bottinger 4, George K Gittes 1,*
PMCID: PMC1968155  NIHMSID: NIHMS23261  PMID: 17418116

Abstract

Transforming growth factor-beta (TGF-β) superfamily signaling has been implicated in many developmental processes, including pancreatic development. Previous studies are conflicting with regard to an exact role for TGF-β signaling in various aspects of pancreatic organogenesis. Here we have investigated the role of TGF-β isoform signaling in embryonic pancreas differentiation and lineage selection. The TGF-β isoform receptors (RI, RII and ALK1) were localized mainly to both the pancreatic epithelium and mesenchyme at early stages of development, but then with increasing age localized to the pancreatic islets and ducts. To determine the specific role of TGF-β isoforms, we functionally inactivated TGF-β signaling at different points in the signaling cascade. Disruption of TGF-β signaling at the receptor level using mice overexpressing the dominant-negative TGF-β type II receptor showed an increase in endocrine precursors and proliferating endocrine cells, with an abnormal accumulation of endocrine cells around the developing ducts of mid-late stage embryonic pancreas. This pattern suggested that TGF-β isoform signaling may suppress the origination of secondary transition endocrine cells from the ducts. Secondly, TGF-β isoform ligand inhibition with neutralizing antibody in pancreatic organ culture also led to an increase in the number of endocrine-positive cells. Thirdly, hybrid mix-and-match in vitro recombinations of transgenic pancreatic mesenchyme and wild-type epithelium also led to increased endocrine cell differentiation, but with different patterns depending on the directionality of the epithelial-mesenchymal signaling. Together these results suggest that TGF-β signaling is important for restraining the growth and differentiation of pancreatic epithelial cells, particularly away from the endocrine lineage. Inhibition of TGF-β signaling in the embryonic period may thus allow pancreatic epithelial cells to progress towards the endocrine lineage unchecked, particularly as part of the secondary transition of pancreatic endocrine cell development. TGF-β RII in the ducts and islets may normally serve to downregulate the production of beta cells from embryonic ducts.

Keywords: Pancreas, endocrine and exocrine differentiation, TGF-β, TGF-β receptors, dominant-negative TGF-β type II receptor, TGF-β pan-neutralizing antibody, in-vitro culture

Introduction

The early stages of pancreatic development entail a primary lineage selection by the pancreatic epithelium between endocrine and exocrine lineages (Gittes et al., 1996; Miralles et al., 1998b). This selection appears to be determined by stimuli from the associated pancreatic mesenchyme (Golosow and Grobstein, 1962; Rutter et al., 1978). Later in gestation, starting at E13.5–14.5 in mice, a secondary transition occurs in which a very different pathway appears to give rise to new endocrine cells, presumably from ducts (Pictet and Rutter, 1972). Specific control points for this transition are not well-delineated. The embryonic pancreatic milieu, including the pancreatic mesenchyme, consists of numerous soluble, membrane bound, extracellular and matrix-bound factors that may control pancreatic development and differentiation. Among such factors, many of the transforming growth factor-beta (TGF-β) superfamily members have been suggested to play an important role (Bottinger et al., 1997; Kim et al., 2000; Lee et al., 1995; Miralles et al., 1998a; Rawdon and Andrew, 1998; Ritvos et al., 1995; Sanvito et al., 1994; Yamaoka et al., 1998).

The TGF-β superfamily consists of three main subfamilies: TGF-β isoforms (numbered 1–3 in mammals), BMP’s and activins. All are multifunctional proteins that regulate cell proliferation, lineage determination, differentiation, and extracellular matrix synthesis. Canonical TGF-β superfamily signaling entails binding a type I and type II transmembrane serine/threonine kinase receptor. Upon binding ligands, type II receptors recruit and phosphorylate type I receptors, which in turn activate downstream smads, that mediate TGF-β-regulated gene expression (Massague, 1998; Shi and Massague, 2003).

Several studies have suggested a role for the TGF-β isoform subfamily in the regulation of pancreatic organogenesis, but with conflicting results. A clear and specific role for TGF-β isoform signaling in pancreatic development remains elusive. Much of the previous work focused on late gestational effects, or in adult transgenics (Lee et al., 1995; Miralles et al., 1998a; Sanvito et al., 1994; Sanvito et al., 1995), adult islets (Shalev et al., 2002) or engineering of embryonic stem cells (Skoudy et al., 2004). However, Sanvito et al. reported a potential pro-endocrine effect of the TGF-β1 present in Matrigel on early (E11.5) embryonic mouse pancreas development. Here a prolonged in-vitro culture with a progressive loss of acinar tissue suggested that there may have been significant acinar autolysis with relative endocrine sparing, which could have been perceived as a pro-endocrine effect. In addition, these early studies were not able to delve into the complexity of potential TGF-β isoform signaling pathways due to the relative lack of knowledge in the field at that time. Thus, the role of TGF-β isoform signaling in endocrine versus exocrine lineage selection, as well as in the amplification of those lineages, remains unclear. We previously demonstrated expression of TGF-β1, -β2, and - β3 starting from the early embryonic pancreatic epithelium through to late gestation endocrine and exocrine pancreas (Crisera et al., 1999). In addition, manipulations of TGF-β isoform signaling have been shown to affect adult pancreatic differentiation (Bottinger et al., 1997).

In this study, we have used several approaches to investigate the specific role of TGF-β isoform signaling in early pancreatic organogenesis. First, we studied the endogenous expression of TGF-β isoform receptors during pancreas development, and correlated expression with lineage selection. Second, we studied the effect of inhibiting TGF-β isoform signaling in the developing pancreas using a transgenic mouse expressing a dominant-negative TGF-β type II receptor (DNTβRII). The adult phenotype of these transgenic mice has been previously reported (Bottinger et al., 1997). We also inhibited TGF-β signaling in vitro, using either ligand inhibition with TGF-β isoform pan-neutralizing antibody, or else receptor inhibition with the recombination of epithelia or mesenchyme from early DNTβRII transgenic pancreatic rudiments with wild-type counterparts.

Our results show that the endogenous TGF-beta type II receptor (TGF-β RII) and type I receptor (TGF-β RI, or ALK5) are localized differentially to the epithelium and mesenchyme at early stages, and then localize to pro-endocrine/endocrine structures and ducts at later stages of pancreatic development. Blocking TGF-β signaling in the embryonic pancreas, either with the dominant-negative receptor transgenic mice, or with TGF-β-specific pan-neutralizing antibody, led to an increased number of endocrine cells. Interestingly, when TGF-β RII function was blocked in only the mesenchyme, by combining transgenic pancreatic mesenchyme with age-matched wild-type pancreatic epithelia, the greatest increase in the number of endocrine cells was seen. These data indicate that absence of TGF-β isoform signaling in the early embryonic period, both in pancreatic epithelium and in mesenchyme, may direct pancreatic epithelial cells towards the endocrine lineage, leading to growth and differentiation of endocrine cells. Importantly, signaling through endogenous TGF-β receptors in the ducts and islets may normally serve to downregulate proliferation and differentiation of new endocrine cells, via either neogenesis from ducts or proliferation of existing insulin-positive cells, in what is considered the secondary transition of pancreatic endocrine development.

Materials and methods

Transgenic animals and genotyping

All the animal experiments were performed in accordance with guidelines established by the animal care facility of University of Pittsburgh and University of Missouri Kansas City. Transgenic mice expressing a dominant-negative TGF-βRII (DNTβRII), under the control of constitutive metallothionine (MT-1) promoter, were generous gifts from Dr. Lalage Wakefield and Erwin Bottinger. These mice were bred and embryos were harvested on gestational day E11.5, E12.5, E14.5, E16.5 and E18.5. Non-pancreatic tissues from the embryos were used for genotyping with the Extract-N-Amp PCR mix (Sigma, St. Louis, MO) kit and PCR probe that is specific for the transgene. Pancreata were isolated by microdisscetion from the transgenic embryos as well as from the CD1 littermate controls. When indicated, the activity of the MT-1 promoter was enhanced by maintaining the pregnant animals on drinking water containing 25 mM ZnSO4.

Organ culture

Neutralizing antibody experiment: Dorsal pancreatic rudiments of E11.5 CD1 mouse embryos were grown in hanging drop organ culture containing TGF-β isoform pan-neutralizing antibody (Upstate biotechnology). Non-immune rabbit serum was used as a control. A drop, consisting of 50μl of control media or media containing neutralizing antibody 80μg/ml was placed on a 35mm Petri dish. Pancreas rudiment was placed inside the drop and inverted so that the drop hangs. The bottom of the Petri dish was filled partially with media to keep the environment moist. The drop was replaced every day and the cultures were maintained for 6 days at 37ºC and 5% CO2.

Mix-and-match experiment: Time dated CD1 and DNTβRII mice were sacrificed at E11.5. Pancreatic epithelium and mesenchyme were separated by microdissection as described previously, and a series of mix-and-match hanging drop cultures were carried out between wild-type CD1 and transgenic DNTβRII pancreatic epithelium and mesenchyme. 10−5 M zinc chloride was added to the media to activate the MT-1 promoter. The cultures were maintained for 6 days.

Histology and Immunohistochemistry

Harvested tissues were fixed in 4% paraformaldehyde, cryoprotected by incubating in 30% sucrose overnight, embedded in Tissue Tek OCT compound and frozen in Liquid N2. 6–8 μm sections were cut by cryostat and haematoxylin and eosin (H and E) staining was done by standard protocol. For immunohistology, optimal dilutions and controls were used for each antibody used. DNTβRII was detected after antigen retrieval with sodium citrate (10mM, pH 3.0) by polyclonal rabbit anti-human TβRII 1:100 (residues 1–28, Upstate biotechnology, VA). Insulin guinea pig anti-swine 1:400 (Dako, Carpinteria, CA), glucagon mouse monoclonal 1:500 (Sigma, St. Louis, MO ), amylase rabbit anti-human 1:400 (Sigma, St. Louis, MO), endogenous TGF-β RII rabbit polyclonal anti-human 1:75 (Santa Cruz biotech, CA), TGF-β RI (ALK5) rabbit polyclonal anti-human 1:75 (Santa Cruz biotech, CA), ALK1 rabbit polyclonal anti- human 1:40 (Santa Cruz Biotech, CA), PDX-1 rabbit polyclonal 1:1400 (generous gift from Prof. Chris Wright, Vanderbilt University Medical School, Nashville, TN), CD-31 rat monoclonal anti-mouse 1:50 (BD Pharmingen, CA), neuroD1 rabbit polyclonal 1:100 (Santa Cruz Biotech, CA), cleaved caspase-3 rabbit monoclonal cleaved 1:100 (Cell signaling Tech, MA), neurogenin-3 mouse monoclonal 1:3000 (Beta Cell Biology Consortium, TN) and DBA FITC conjugated 1:100 (Vector Laboratories, CA) . Primary antibodies were incubated for 1 hour at room temperature or at 4ºC overnight. Biotinylated Vectastain ABC kit or AMCA/CY3/FITC fluorescent conjugated donkey secondary antibodies were used for 1 hour at room temperature. Immunoperoxidase was detected by DAB kit (Dako, Carpintaria, CA) or AEC (Sigma, St. Louis, MO) and fluorescently labeled samples were imaged using a fluorescent microscope.

Bromo-deoxyuridine (BrdU) incorporation

Pregnant mice were injected with BrdU (Sigma, St. Louis, MO) 200mg/Kg intraperitoneally two hours prior to harvesting the embryos. The pancreata were fixed, sectioned and slides treated with 2N HCl for 15 minutes, then 0.5% Pepsin in 0.1N HCl for 15 minutes at room temperature. To determine the proportion of proliferating endocrine cells (sum of insulin and glucagon positive cells), we used high power images (X400) taken from sections that were stained with insulin or glucagon and BrdU (Dako, Carpintaria, CA) and DAPI nuclear counterstaining. Cells positive for BrdU and endocrine cells were counted in at least six sections obtained from four independent mice.

Quantitative analysis

Serial 8μm sections were cut and collected in glass slides. One out of twelve (for in-vivo experiments) or seven (for in vitro experiments) consecutive sections was analyzed by immunohistochemistry and microscopic images were captured using 100X and 200X magnification. The results of each experiment were obtained by quantifying the positive staining area for the given antigen in at least 4 pancreatic rudiments. Data are presented as percentage of positive area per gross total area of the pancreas using Image-Pro Plus (Media Cybernetics, Silver Spring, MD). Statistical analysis was performed using student t-test and P value <0.05 was considered as statistically significant.

RNA extraction and Semi-quantitative Reverse transcriptase-PCR

Total RNA was isolated from whole pancreas of E11.5, E12.5, E14.5, E16.5 and E18.5 transgenic or wild-type mice using Qiagen RNeasy Mini kit, followed by Promega RQ1 DNase I treatment. First strand cDNA was synthesized by using sensiscript Reverse Transcriptase kit (Promega). Real-time semi-quantitative PCR reactions were carried out with SyberGreen PCR Master Mix in iCycler iQ apparatus (Bio-Rad, Hercules, CA). A mouse β-tubulin primer set was used to normalize the amount of total cDNA for each gene expression. Various concentrations of wild-type or transgenic cDNA (107, 106,105, 104, 103, 102 molecules/μl) were used for the calibration of standard curve for each primer set. In addition, the PCR products were separated and visualized on the agarose gel for confirmation. At least five different samples were analyzed at each sequential age to quantify the target gene. Primers were designed by primer3_www.cgi v 0.2 software. Oligos sequences used are listed as forward then reverse, 5’ to 3’: β-Tubulin 5’CCTTTTGGCCAGATCTTCAG3’ and 5’AACCAACTCAGCTCCCTCTG3’ amplify a product of 102 bp, Transgene (DNTβRII) 5’GCATCAGAAGAGGCCATCAA3’ and 5’ACTCACCCTGAAGTTCTCAG3’ amplify a product of 224 bp, TGF-β RII 5’CAATGCTGTGGGAGAAGTGA3’ and 5’CTCACACACGATCTGGATGC3’ amplify a product of 163 bp, TGF-β RI (ALK5) 5’ATTGCTGGTCCAGTCTGCTT3’ and 5’CCTGATCCAGACCCTGATGT3’ amplify a product of 188 bp, ALK1 5’TGACCTCAAGAGTCGCAATG3’ and 5’GTTGTTGCCGATATCCAGGT3’ amplify a product of 115 bp, TGF-β1 5’TGGAGCAACATGTGGAACTC3’ and 5’CGTCAAAAGACAGCCACTCA3’ amplify a product of 108 bp, TGF-β2 5’TGGCTTCACCACAAAGACAG3’ and 5’CCTCGAGCTCCTTCGCTTTTA3’ amplify a product of 115 bp, TGF-β3 5’GATGAGCACATAGCCAAGCA3’and 5’ATTGGGCTGAAAGGTGTGAC3’ amplify a product of 177 bp, neuroD1 5’CTCTGGAGCCCTTCTTTGAA3’ and 5’TGCAGGGTAGTGCATGGTAA3’ amplify a product of 161 bp, neurogenin-3 5’TTCGCCCACAACTACATCTG3’ and 5’CAGGGAATTCCTCCAATGAG3’ amplify a product of 199 bp.

Results

Basic phenotype of DNTβRII transgenic mouse embryonic pancreas

To examine the embryonic phenotype of the dominant-negative type II TGF-β receptor (DNTβRII) transgenic mice, we first studied histology and immunohistochemistry for markers of pancreatic differentiation in sequential ages of embryonic pancreas (Fig 1 and supplementary 1). At E11.5 and E12.5 there were no appreciable differences grossly, histologically, nor immunohistochemically between the DNTβRII transgenic embryos and wild-type littermates, regardless of whether transgene expression was further enhanced with ZnSO4 (Supplementary 1). However, by E14.5, the early phase of the secondary transition, transgenic pancreata began to show abnormal expansion of the endocrine and potentially the pro-endocrine components (see below) of the epithelium, with an increase in both insulin and glucagon-positive cells (Fig 1Ab,c). At E16.5 the phenotype of the DNTβRII embryonic pancreas diverged markedly from wild-type littermates. The gestational timing of this phenotypic divergence coincides with a period of profound exocrine and endocrine differentiation and expansion in normal embryonic pancreas. The transgenic pancreas had prominent expansion of the cord region (central region that typically harbors the endocrine cells), with periductal accumulation of cells. The majority of the periductal cells were endocrine, expressing insulin and glucagon (Figs 1Ae,f and 2Ab,Bb). Many periductal cells stained for pro-endocrine markers neurogenin-3 and neuroD1 (see below and Fig 4). Quantitative analysis showed a significant increase in insulin-positive cells at this age, with some further augmentation seen in zinc-treated animals (Fig 1C), representing a form of dose-response of these cells to TGF-β isoform inhibition. The insulin-positive cells appear to be mature beta cells as evidenced by positive staining for PDX-1 (Fig 2Ab). At E18.5, continued cord region expansion was seen, as well as the onset of the acinar atypia (supplementary 4) described previously for the adult pancreas (Bottinger et al., 1997). In spite of their abnormal appearance, the atypical acini of the transgenic pancreas did express amylase (Fig 1Ah,i and 2Bd). E16.5 and E18.5 transgenic pancreata also showed a significant increase in the number of blood vessels, as shown by an increase in staining for the vascular endothelial marker CD-31 (Fig 1Bb,d). Some of the transgenic pancreata exhibited abnormally dilated blood vessels with focal areas of congestion. However, the endothelial lining of these abnormal blood vessels appear to be intact, suggesting that the clusters of red blood cells may not be from hemorrhage, but rather from stagnant blood in malformed vessels (supplementary 1). This abnormal vascular phenotype is reminiscent of a diffuse vasculopathic phenotype seen in several null mutant mouse strains, including TGF-β1 (Dickson et al., 1995), TGF-βRII (Oshima et al., 1996), endoglin (Bourdeau et al., 2001) and ALK1 null mutants (Larsson et al., 2001; Oh et al., 2000).

Figure 1.

Figure 1

Open in a new tab

Phenotype of DNTβRII transgenic mouse embryonic pancreas: A) Embryonic pancreata at various gestational time points were analyzed by immunostaining for markers of pancreatic differentiation. Amylase (red), insulin (blue) and glucagon (green) in wild- type CD1 (a,d,g), transgenic DNTβRII (b,e,h) and transgene induced DNTβRII+Zn (c,f,i). At E14.5 (a–c), transgenic pancreas showed increased expansion of the “cord region” (defined as the more central region of the pancreas where typically the endocrine cells are found). At E16.5 (d–f), the transgenic phenotype diverged most markedly from controls. The transgenic pancreas had prominent expansion of the cord region with an increase in endocrine cells, most notably in Zn-treated embryos. At E18.5 (g–i), continued cord region expansion was seen. B) Transgene-induced DNTβRII+Zn pancreata at E16.5 (b) and E18.5 (d) showed increased and abnormally dilated blood vessels, as shown by CD31 (red) staining. C) Quantitative analysis of E16.5 pancreas showed a significant increase in the amount of insulin-positive area, and in the amount of CD31 positive endothelium in transgenic pancreas compared to wild-type. *P value <0.05 and scale bar 60μm

Figure 2.

Figure 2

Open in a new tab

Transgenic phenotype: expansion of the cord region with periductal accumulation of endocrine cells. Immunohistochemical analysis of embryonic pancreas in A) with duct marker DBA (green), insulin (blue), PDX-1 (red) and in B) insulin (red) amylase (blue) and DBA (green) of wild-type CD1 (a,c) or transgene-induced DNTβRII+Zn (b,d) at E16.5 (a,b) and E18.5 (c,d) showed prominent expansion of cord region with increased accumulation of endocrine cells in periductal region. Scale bar 60μm

Figure 4.

Figure 4

Open in a new tab

Increased endocrine precursors in the embryonic pancreas of transgenic mice: immunohistochemical analysis showed increased expression of endocrine precursor (A) neurogenin-3 positive (red) and (C) neuroD1-positive (red) cells in the transgene-induced pancreas at E14.5 (a,b) and E16.5 (c,d). At E16.5, there were abundant neurogenin-3 positive cells in and around the DBA-positive ducts (green) of transgene-induced pancreas. Quantitative RT-PCR analysis of transgene-induced pancreas showed significant upregulation in (B) neurogenin-3 and (D) neuroD1 mRNA expression. *P value <0.05 and scale bar 60μm

Proliferation of endocrine and pro-endocrine cells in embryonic DNTβRII transgenic pancreas

To investigate the mechanism of increase in endocrine cells in mid-gestation embryonic transgenic pancreas, we examined the presence of pro-endocrine cells and the proliferative status of the endocrine cells at different embryonic time points. Proliferating cells were assessed by bromo-deoxyuridine incorporation, and then immunostaining with insulin or glucagon and anti-BrdU antibodies. In general, the control CD1 embryonic pancreas had few proliferating endocrine cells (Sander et al., 2000) (Fig 3Aa,c and 6H). However, the proliferative status of the endocrine cells in transgene-induced pancreas was significantly increased at E16.5 (Fig 3Ab,B,Cb,D and supplementary 2). Proliferating ductal structures, as evidenced by double staining for DBA and BrdU, were also more apparent in the transgenic pancreas at this age (Supplementary 2), consistent with an amplified secondary transition. But by E18.5, the proliferative nature of the insulin or glucagon-positive cells in transgenic pancreas subsided (Fig 3Ad,B,Cd,D). The pro-endocrine markers neurogenin-3 and neuroD1 were also more highly expressed in transgene-induced pancreas at E14.5 and E16.5 (Fig 4Ab,d and Cb,d). At E16.5, there were abundant neurogenin3-positive cells in and around the DBA-positive ducts (green) of transgene-induced pancreas compared to wild-type controls. Quantitative RT-PCR analysis of the pro-endocrine markers in transgene-induced pancreas showed statistically significant increase in mRNA expression of neurogenin-3 at E14.5 and E16.5 (Fig 4B). Similarly, mRNA expression of neuroD1 was significantly upregulated at E16.5 (Fig 4D). Since neuroD1 is a post-mitotic marker, this higher expression may portend the stoppage of proliferation by E18.5. In addition to the decrease in enhanced proliferation of the endocrine cells at this stage, there was also a concomitant increase in caspase-3 positive apoptotic endocrine cells in the transgene-induced pancreas (Supplementary 3).

Figure 3.

Figure 3

Open in a new tab

Increased proliferation of endocrine cells in the embryonic pancreas of transgenic mice: proliferating endocrine cells in the embryonic pancreas were examined through BrdU incorporation, and then staining for BrdU (red/green) and (A) insulin (green) or (C) glucagon (red). At E16.5 and E18.5, the control wild-type CD1 endocrine cells showed low proliferation with few BrdU and Insulin/Glucagon-positive cells (a,c). However, at E16.5 the proliferative status of the insulin or glucagon cells in transgene-induced pancreas was significantly increased (arrowhead b). Quantitative analysis of proliferating endocrine cells in transgene-induced pancreas showed a significant increase in proliferation of insulin-positive (B) and glucagon-positive cells (D) at E16.5. However, the level of proliferation became similar to wild-type by E18.5. *P value <0.05 and scale bar 60μm

Figure 6.

Figure 6

Open in a new tab

Ontogeny of the transgene in embryonic DNTβRII pancreas by immunohistochemistry: the expression was compared between zinc treated and untreated embryonic pancreas. At E11.5 the transgene was absent unless induced by zinc. Addition of zinc at E11.5 augmented expression in the epithelium (arrows) and weakly in the mesenchyme (arrowheads). At E12.5, transgene expression began to appear in the epithelium even without zinc (arrows). However, with zinc, strong expression was seen not only in the epithelium at E12.5 (arrows), but also very clearly in the mesenchyme (arrowheads). At E14.5 only the cord or pro-endocrine region (arrowheads) expressed the transgene, either with or without zinc, but expression appeared much stronger with zinc-treatment, with some additional weak expression in acinar cells (arrows). At E16.5, however, the transgene was expressed weakly in the cord region and strongly in the acinar cells, with or without zinc treatment. At E18.5 essentially all of the transgene expression was in the acini, and the level of expression did not appear different between zinc-treated or untreated groups. Scale bar 60μm.

Ontogeny of TGF-β receptor isoforms in embryonic pancreas

In order to determine the role of endogenous TGF-βRII and potential type I receptors mediating TGF-β isoform signaling in pancreatic organogenesis, we first performed ontogeny studies of receptors important for TGF-β isoform signaling (Fig 5).

Figure 5.

Figure 5

Open in a new tab

Ontogeny of endogenous TGF-β isoform receptors: expression of TGF-βRII (A,D,G,J,M), and expression of the two potential type I receptor binding partners of TGF-βRII, i.e. TGF-βRI/ALK5 (B,E,H,K,N) and ALK1 (C,F,I,L,O) was examined in wild-type embryonic pancreases at various time points of gestation. The expression of TGF-βRII in epithelium (arrows) was stronger and more localized compared to mesenchyme (arrowhead) at E11.5 and E12.5. At E14.5 and E16.5, TGF-βRII appeared to be mainly expressed in the pro-endocrine cord region (arrowheads) and in the ducts (arrows), and then at E18.5 predominantly in ducts (arrows) and blood vessels (arrowheads). The expression pattern of TGF-βRI/ALK5 paralleled TGF-βRII. ALK1 expression showed a different pattern, with low levels in the epithelium (arrow) at E11.5 (C), low levels in the epithelium (arrow) and very low levels in mesenchyme (arrowhead) at both E12.5 (F) and E14.5 (G). At E16.5 and E18.5, ALK1 was mainly expressed in blood vessels (L,O,arrowheads). The overlapping expression of RII and ALK1 in vasculature at this stage may suggest a role for the TGF-β RII/ALK1 dimer in regulating blood vessel development. Scale bar 60 μm.

The endogenous type II (TGF-β RII) receptor was expressed strongly in the epithelium at E11.5 and E12.5, in the cord region (pro-endocrine central region around ducts) at E14.5, and in the ducts and blood vessels at E16.5 and E18.5. The type II receptor only expressed weakly, however, in the mesenchyme, at E11.5 and E12.5 (Fig 5 A,D). Similarly, receptor type I (TGF-β RI or ALK5) was expressed strongly in the epithelium at E11.5 and E12.5, and in the developing cord region at E14.5. At E16.5 it was expressed not only in the ducts and cord region, but also in the developing blood vessels. At E18.5 it was predominantly expressed in ducts. The mesenchyme appeared to have weak expression for TGF-β RI at E11.5 and E12.5 (Fig 5 B,E). Taken together, because of the parallel spatiotemporal expression pattern of receptor II and receptor I, these results support a role for TGF-β isoforms binding to the heterodimeric receptor TGF-β RI with RII. This ligand binding may occur in both the mesenchyme and epithelium early in gestation, and later only in the epithelium, especially in the cord region and ducts. Based on the findings in the mutant mice discussed above, this binding may act to prevent or regulate expansion and differentiation of precursor cells in the cord region to form new endocrine cells.

We also examined the ALK1 receptor (Fig 5C,F,I,L,O), which can also bind TGF-β isoforms, but unlike TGF-β RI activates the BMP-specific SMADs (Smad 1, 5 and 8), and can compete with the TGF-β RI for TGF-β isoform binding. We found ALK1 to be expressed at low levels in the epithelium at E11.5 (Fig 5C), and in the epithelium and weakly in the mesenchyme (possibly in the vasculogenic mesenchyme) at E12.5. At E14.5 it was expressed weakly in the cord region, similar to TGF-β RI, thus suggesting possible competition between ALK1 and ALK5 for binding TGF-β isoforms. Lastly, at E16.5 and E18.5 ALK1 was mainly expressed in the blood vessels and, to a lesser extent, in the cord region. The overlapping expression of RII and ALK1 in the vasculature at E16.5 and E18.5 suggests a role for a heterodimer of those two receptors in regulating blood vessel development.

Ontogeny of the dominant-negative transgene in embryonic DNTβRII pancreas

In order to inhibit endogenous TGF-β isoform signaling, the transgenic dominant-negative receptor must be expressed at a high level, and in the same cells that express the endogenous receptor. Thus, in order to better understand the phenotype in the embryonic transgenic mice, we examined the embryonic pancreas of DNTβRII mice for transgene expression by immunohistochemistry using anti-human TGF-β RII (1–28, Upstate biotechnology) antibody that only recognizes the human transgenic receptor, not the endogenous receptor. The expression level was compared between zinc-treated and untreated embryonic pancreas (Fig 6, non-transgenic tissues stained negative, data not shown).

At E11.5 the transgene was absent unless induced by zinc. Addition of zinc augmented some localized expression in the epithelium (Fig 6B). At E12.5, transgene expression began to appear in the epithelium without zinc. However, with the addition of zinc strong expression was seen in the epithelial cord (or pro-endocrine) region, and very clearly in the mesenchyme (Fig 6D). This early expression of the transgene could be inhibiting TGF-β isoform signaling, leading to the subsequent expansion of the cord region seen later in the transgenic embryos. At E14.5 there was only cord region/pro-endocrine expression of the transgene, either with or without zinc, but the expression again appeared much stronger with zinc. In addition, some expression in acinar cells was seen only in the zinc-treated group. This early acinar expression may be a harbinger of the acinar atypia seen late in gestation and postnatally in these animals. The expression pattern at E14.5 also supports the enhancement of endocrine differentiation that we see by E16.5, especially in the presence of zinc (Fig 1Af and C). At E16.5 the transgene was expressed weakly in the cord region even with zinc, and became quite strong in the acinar cells with or without zinc (Fig 6 G,H). This change in expression may explain both the decreasing endocrine phenotype seen at E18.5, as well as the early development of acinar atypia at E18.5. These data suggest that signaling through TGF-β RII is necessary for both endocrine and exocrine maturation at different phases of development. At E18.5 essentially all of the transgene expression was in the acini, and the level of expression was not significantly different between zinc-treated and untreated groups, which may explain the lack of an adult endocrine phenotype in these mice (Fig 6I,J).

Quantification of the endogenous and transgenic receptor mRNA

A significantly higher expression level of the transgene than of the endogenous receptor is necessary for effective inhibition. Therefore, quantification of the TGF-β receptors’ mRNA was performed at various gestational time points by semi-quantitative RT-PCR (Fig 7A–C). The endogenous TGF-β RII mRNA level gradually increased over gestation, consistent with a greater role late in gestation. On the other hand, low expression at early stages imply that its function would be more easily overridden by the dominant-negative transgene, thus making the younger embryonic pancreas potentially more likely to display a phenotype in the transgenic animals (Fig 7A). TGF-β RI mRNA (ALK5) showed the highest level of expression at E16.5, and then a sharp decline at E18.5, suggesting an important role early at E14.5 and E16.5 (Fig 7B). This temporal pattern implies that the strong phenotype seen in the TGF-β RII transgenic mice at mid-gestational time points may be due to inhibition of TGF-β RII binding to ALK5. The drop in ALK5 mRNA levels at E18.5 suggests that other ALKs, particularly ALK1, may instead dimerize with TGF-β RII to mediate other actions, such as blood vessel development. The ALK1 mRNA level rises suddenly late in gestation, consistent with a possible role in blood vessel formation and/or growth (Fig 7C), also supported by the apparent vascular phenotype seen at E16.5 and E18.5.

Figure 7.

Figure 7

Open in a new tab

Quantitative analysis: quantification of mRNA for endogenous TGF-β receptors (A–C), transgenic receptor (D) and TGF-β ligands (E–G) were performed at various gestational time points by semi-quantitative RT-PCR. The endogenous TGF-β RII mRNA level gradually increased over gestation (A). TGF-β RI/ALK5 mRNA showed its highest level of expression at E16.5, and then a sharp decline at E18.5 (B).The drop in ALK5 mRNA levels at E18.5 suggests that other ALKs, particularly ALK1, may instead dimerize with TGF-β RII to mediate actions such as blood vessel development. The mRNA level of ALK1 showed a sudden rise late in gestation, consistent with a possible role in blood vessel development and growth (C). Quantification of the transgenic receptor showed significant upregulation of expression in zinc-treated pancreas at E11.5, E12.5 and E14.5 compared to untreated embryos (D and inset). However, baseline expression of the transgene at E16.5 and E18.5 becomes very high even in the absence of zinc, probably due to zinc-independent expression of the transgene in acinar tissue. Addition of zinc at these later ages did not affect the expression level of the transgene. The relative level of TGF-β isoform ligand mRNA were similar between wild-type and transgenic mice through E16.5. However, the level of expression then differed selectively at E18.5. The level of TGF-β1 and TGF-β3 mRNA significantly increased in transgenic pancreas treated with zinc, whereas TGF-β2 mRNA was decreased. *P value <0.05

PCR quantification of the transgenic dominant-negative receptor showed significant upregulation of expression in zinc-treated pancreas at E11.5, E12.5 and E14.5 compared to untreated embryos (Fig 7D and inset). This pattern may also help to explain the stronger endocrine phenotype in zinc-treated pancreas at E16.5 compared to those without zinc treatment. However, the baseline expression of the transgene at E16.5 and E18.5 becomes very high, even in the absence of zinc. This baseline increase may be due to expression of the transgene in the acinar tissue alone at later stages of development, as was seen with the immunohistochemistry (Fig 6). Addition of zinc at these later ages did not seem to affect the expression level of the transgene. This change in expression may contribute to the primarily acinar phenotype seen in postnatal and adult transgenic mice.

Quantification of TGF-β ligands in embryonic transgenic pancreas

To compare the level of TGF-β ligands between wild-type and transgenic mice we used semi-quantitative mRNA expression analysis (Fig 7 E-G). The expression levels for the three ligands were fairly similar between wild-type and transgenic mice through E16.5. However, the level of expression then differed selectively at E18.5. The level of TGF-β1 and TGF-β3 were significantly increased in transgenic pancreas treated with zinc (Fig 7E,G and P value <0.05). On the other hand, the level of TGF-β2 was lower in transgenic pancreas when compared with wild-type (Fig 7F). These results are consistent with the previous study in adult DNTβRII transgenic pancreas (Bottinger et al., 1997). These changes in the level of TGF-β ligands, together with the changes in expression pattern of the transgene in transgenic pancreas at late gestation, may help to explain the lack of an endocrine phenotype late in gestation and postnatally.

In vitro mix-and-match experiments

In order to study the role of TGF-β RII signaling specifically in either the pancreatic mesenchyme or in the epithelium, we performed a series of in vitro recombination experiments between transgenic (TGF-β RII dominant negative) and wild-type E11.5 pancreatic epithelium and mesenchyme. Four different mix-and-match groups were made (Fig 8A). The recombination of wild-type epithelia and wild-type mesenchyme was considered as a control (Fig 8Ac,g,k). The other recombinations were between wild-type epithelium and transgenic mesenchyme, transgenic epithelium and wild-type mesenchyme, or transgenic epithelium and transgenic mesenchyme. 100μM ZnCl2 was supplemented in the media of all of the cultures to enhance transgene expression. Immunohistochemical staining (Fig 8A) and quantitative analysis (Fig 8B) showed that wild-type epithelium recombined with wild-type mesenchyme had the lowest amount of insulin-positive cells, and the greatest amount of amylase staining. When transgenic (TGF-βRII blocked) epithelium was recombined with either wild-type or transgenic mesenchyme, there was an augmentation in the number of insulin-positive cells. However, the greatest augmentation of insulin-positive area was seen when transgenic mesenchyme was recombined with wild-type epithelium (Fig 8Aa,e,i and B). PDX-1 staining suggested that most or all of the insulin-positive endocrine cells were mature β-cells (Fig 8Ai,j,k,l). These results indicate that TGF-β receptor II signaling is important in both epithelium and mesenchyme, and perhaps more so in the mesenchyme.

Figure 8.

Figure 8

Open in a new tab

In vitro mix and match experiment: four different mix-and-match in vitro recombinations between transgenic (DNTβRII) and wild-type (CD1) E11.5 pancreatic epithelium (e) and mesenchyme (m) were made. The recombination of wild-type mesenchyme and wild- type epithelium was considered as a control (c,g,k). 100μM ZnCl2 was supplemented in the media of all the cultures to enhance transgene expression. Immunohistochemical staining (A) and quantitative analysis (B) showed that control cultures had the lowest amount of insulin-positive cells, and the greatest amount of amylase staining. When transgenic epithelium was recombined with either wild-type or transgenic mesenchyme, there was an augmentation of insulin-positive cells. However, the greatest augmentation of insulin-positive area was seen with recombination of transgenic mesenchyme with wild-type epithelium (a,e,i and B). PDX-1 staining suggested that most or all of the insulin-positive endocrine cells were mature β-cells (i,j,k,l). Scale bar (a-d) 60μm and (e-l and a–d inset) 120μm

In vitro TGF-β pan-neutralizing antibody experiment

In order to inhibit TGF-β isoform signaling through a different mechanism, and at a different level (ligand rather than receptor), we used TGF-β ligand pan-neutralizing antibodies to determine whether endocrine upregulation would be reproducible in normal embryos through in vitro manipulation. Treatment with 80μg/ml neutralizing antibody led to a more than two-fold increase in insulin-positive cells (Fig 9Ag and B). Co-expression of PDX-1 in these insulin-positive cells again suggested that the cells were mature β-cells (Fig 9Ah). Rabbit non-immune serum control at a similar concentration had no effect (Fig 9Aa-d).

Figure 9.

Figure 9

Open in a new tab

In vitro neutralizing antibody experiment: in order to inhibit TGF-β isoform signaling at the ligand level, TGF-β pan-neutralizing antibodies were used. Treatment with 80 μg/ml neutralizing antibody (e-h) showed a significant increase in insulin-positive cells (P value < 0.05) (B). PDX-1 (red) staining of these insulin-positive cells suggested that the cells were mature β-cells (d and h). Rabbit non-immune serum control at a similar concentration had no effect (a–d). *P value <0.05 and scale bar (b–d and f–h) 120μm

Discussion

TGF-β isoform signaling appears to have a complex influence on pancreatic lineage selection and differentiation. When TGF-β1 was expressed under the rat insulin promoter in adult mice, suppression of acinar growth was seen (Lee et al., 1995). Similarly, Sanvito et al. showed that exogenous TGF-β1 caused apoptosis of developing pancreatic acinar cells with what appeared to be a relative increase in endocrine elements (Sanvito et al., 1994). Based on this last study, the possibility was raised that TGF-β isoforms may have a pro-endocrine role in the developing pancreas. Data presented in that study were based on explant culture of E11.5/12.5 mouse pancreas treated with exogenous TGF-β1. Moreover, an initial five-day culture with TGF-β1 had no effect on endocrine/exocrine differentiation. When the culture was maintained for more than ten days in collagen I gel, the pancreas showed a sharp decrease in acinar cells, with a relative sparing of endocrine cells. In contrast, a study performed by Miralles et al. in E12.5 rat pancreas rudiment cultured for seven days showed no significant difference in the total number of endocrine/exocrine cells when treated with exogenous TGF-β1 at the same concentration as Sanvito et al. (Miralles et al., 1998a). These results suggest that the decrease in acinar and increase in endocrine differentiation perceived by Sanvito et al. may have been due to the prolonged 10-day culture period. This longer period may have allowed autolysis of acinar cells due to the release of acinar digestive enzymes.

Early embryonic lethal phenotypes of TGF-β receptor type II null mutants, as well as of double and triple null mutants for the TGF-β isoforms, make interpretation of the role of TGF-β isoform signaling at the genetic level in the pancreas difficult to determine. A dominant-negative form of TGF-β type II receptor was described in 1997 (Bottinger et al., 1997). While the embryonic phenotype was not investigated, the adult phenotype was specifically pancreatic, with acinar atypia and ductal hyperplasia, with occasional acinar tumor formation. Here we found that inhibition of TGF-β receptor type II signaling led to an enhanced number of periductal endocrine cells, with vascular ectasia and stasis, during mid-to-late gestation. In order to understand the mechanism of these effects, we needed to study the distribution and level of expression of both the transgenic dominant-negative receptor (identified through human-specific moieties in the receptor) and the endogenous receptor. Knowledge of the distribution and levels of expression are necessary for understanding the mechanisms of the dominant-negative effect since dominant-negative inhibition requires higher expression of the transgene specifically in cells that also express the endogenous TGF-β receptor type II. In addition, we could enhance transgene expression using zinc added to either the mother’s drinking water, or else to the culture medium for in vitro grown tissues.

The increased number of insulin-positive cells in embryonic pancreata with blocked TGF-β isoform signaling appears to be partially mediated through both epithelial and mesenchymal TGF-β receptor type II signaling. The endocrine enhancement was shown in vivo in the transgenic animals, as well as in vitro using hybrid transgenic/wild-type pancreatic epithelial/mesenchymal constructs, or else wild-type tissues treated with pan-neutralizing TGF-β isoform antibodies. Importantly, the endogenous receptor was seen to be expressed both in the epithelium and the mesenchyme at critical early times when lineage selection is occurring. Qualitatively, the endogenous receptor expression levels appeared to be lower in the mesenchyme, which may therefore be more easily overridden by the dominant-negative receptor (the transgenic dominant-negative receptor was also expressed in both epithelium and mesenchyme, though qualitatively lower in the mesenchyme), which could explain the greater endocrine enhancement with isolated transgenic mesenchyme combined with wild-type epithelium in the mix-and-match experiments.

TGF-β isoform signaling has been found to typically promote differentiation and growth arrest in epithelia, however, in stroma or mesenchyme it typically has the opposite effect, promoting growth. This differential action between epithelium and stroma (mesenchyme) may also be occurring in the early (E11.5–E12.5) embryonic pancreas. Thus, contrary to hypotheses presented from earlier work in the field, TGF-β signaling to the epithelium may actually directly suppress endocrine proliferation and enhance endocrine differentiation, whereas TGF-β signaling to the mesenchyme may promote mesenchymal proliferation. Since we and others have shown that mesenchyme is a pro-exocrine agent during the early critical times of pancreatic lineage selection, enhanced proliferation of mesenchyme (induced by TGF-β signaling) may thus favor exocrine differentiation over endocrine differentiation. In addition, since vasculogenesis arises from the mesenchyme, and blood vessel proximity to pancreatic endoderm has been shown to promote endocrine differentiation (Lammert et al., 2001), enhanced vasculogenesis due to ALK-1 signaling inhibition could be playing a role in enhanced endocrine differentiation as well.

The pro-endocrine effect produced by TGF-β signaling inhibition was most prominent at E16.5. This gestational time corresponds to the height of the secondary wave of endocrine differentiation in the developing mouse pancreas. TGF-β isoform signaling, which typically acts as a suppressor of epithelial proliferation, may act as a negative regulator of the secondary wave of insulin cell proliferation and differentiation from ducts. Our in vivo study suggested that the inhibition of TGF-β isoform signaling in these tissues leads to an increase in pro-endocrine as well as proliferating endocrine cells. The rapid cessation of proliferation between E16.5 and E18.5 could be due to an inappropriate commitment of cells to the endocrine lineage, perhaps with premature terminal differentiation. Alternatively, the cessation could be due to loss of TGF-β signaling inhibition due to changes in the expression of the endogenous and transgenic receptor that occurs during that gestational time window. Our ontogeny analysis suggested that the endogenous type II TGF-β receptor is expressed in the epithelium early in gestation, which at that time may be dimerizing with the ALK5 type I receptor to regulate lineage selection. Later in gestation both the TGF-β type II receptor and ALK5 type I receptor appeared to be most strongly expressed in ducts, possibly acting as a control mechanism for the generation of endocrine cells from ducts.

Bottinger et al. reported acinar atypia with some acinar cell tumors in adult DNTβRII transgenic mice, but with no obvious endocrine phenotype. We found normalization of the endocrine phenotype in late gestation. This normalization may be explained by changes in the expression pattern of the dominant-negative receptor, which becomes exclusively expressed in acinar cells. Likewise, changes in the level of TGF-β isoform ligands may override the dominant-negative effect in endocrine cells, leading to conditions favoring early apoptotic death of endocrine cells (supplementary 3).

Our findings of increased CD-31 staining and vascular ectasia in the late-gestation transgenic embryonic pancreas may be due to blocked signaling through the TGF-β type II receptor dimerized with the ALK1 type I receptor. ALK1 has been shown to compete with ALK5 for TGF-β receptor type II binding, particularly in endothelial progenitor cells (Goumans et al., 2003; Goumans et al., 2002), and hereditary hemorrhagic telangiectasia, often in the pancreas, has been associated with ALK1 mutations in animals and humans (Azuma, 2000; Chuang et al., 1977; Halpern et al., 1968; Torsney et al., 2003). In addition, we found that both TGF-β receptor type II and ALK1 type I receptor were strongly expressed in developing blood vessels during late phases of pancreatic development.

In conclusion, our results expand on the early reports of the role of TGF-β isoform signaling in pancreatic development, perhaps now with a better understanding of specific pathways involved. TGF-β isoforms may have variable functions throughout pancreatic development. During early-mid gestation, TGF-β isoforms appear to signal through a heterodimeric complex of type II and ALK5 receptors to control endocrine proliferation and differentiation. TGF-β isoform signaling between epithelium and mesencyhyme may play an important role in the endocrine versus exocrine lineage determination and the first wave of endocrine development. However, evidence in the literature suggests that this first wave may be vestigial and the secondary transition is more significant with respect to final beta cell mass. TGF-β isoform signaling appears to be critically important for suppressing the proliferation and differentiation of pro-endocrine cells in the ductal and periductal region of the mid-gestation pancreas.

Supplementary Material

01. Supplementary figure 1.

Histology of the embryonic pancreas from transgenic mice: H&E staining of wild-type CD1 (A,D,G,J,M), transgenic DNTβRII (B,E,H,K,N) and transgene-induced DNTβRII+Zn (C,F,I,L,O) pancreas at various gestational ages. At E11.5 and E12.5 no appreciable differences were seen between wild-type and transgenic pancreas. At E14.5 (G-I), transgenic pancreas showed increased expansion of the cord region (arrowheads in H,I). At E16.5 (J–L), the transgenic phenotype diverged most markedly from controls. The transgenic pancreas had prominent expansion of the cord region, with periductal accumulation of cells (arrowheads in K,N,O). At E18.5 (M–O), continued cord region expansion was seen with early acinar atypia. The phenotype was stronger in the transgene-induced pancreas, with abnormally dilated blood vessels (arrows in L,O). Scale bar 60 μm.

02. Supplementary figure 2.

Increased proliferation of endocrine and ductal structures in transgene-induced pancreas: At E16.5, transgene-induced DNTβRII+Zn pancreas showed higher numbers of BrdU-positive cells in DBA-positive ducts (arrowheads in C and D) and insulin-positive endocrine cells (arrows in C and D) compared to wild-type CD1 pancreas. C and D are higher magnification of A and B, respectively. Scale bar 60μm

03. Supplementary figure 3.

Apoptosis of endocrine cells during late gestation in transgene-induced pancreas: caspase-3 (red), a key enzyme involved in the terminal apoptotic cascade of cell death is activated more in and around the insulin (green) and glucagon (blue) positive cells in transgene-induced pancreas (B,D) compared to wild-type (A,C) at E16.5 (A,B) and E18.5 (C,D). This difference may suggest that an increase in apoptosis represents a key mechanism whereby DNTβRII transgenic pancreas loses the ability to sustain the enhanced endocrine phenotype. Scale bar 60μm

04. Supplementary figure 4.

Early acinar atypia during late gestation in the transgene-induced pancreas: simple histological analysis of E18.5 pancreas by H&E staining showed normal floral arrangement of acinar clusters in wild-type CD1 controls (A,B). However, transgene-induced pancreas (C,D) showed localized disruption of acinar architecture, with apparently lobule-level acinar degeneration (arrowhead C). The acinar dysmorphogenesis could be due to overproduction of enzymes. This finding is reminiscent of the acinar atypia seen in the adult transgenic pancreas (Bottinger et al, 1997). B and D are higher magnification of A and C, respectively. Scale bar 60μm

Acknowledgments

The authors would like to thank Prof. LM Wakefield for providing DNTβRII transgenic mice and Prof. CV Wright for providing PDX-1 antibody. This study was supported by grants to G.K.G. from NIH (1 R01 DK58400-01 and 1 R01 DK064952-01), JDRF1 (2-1999-636), tobacco fund (19831-066584), and CHP funds (19831-046903).

Abbreviations

TGF-β

Transforming growth factor-beta

TGF-β (RI, RII,RIII)

Transforming growth factor beta (type I receptor, type II receptor, type III receptor)

DNTβRII/DβRII

Dominant-negative TGF beta type II receptor

E11.5

Embryonic day 11

RT-PCR

Reverse transcriptase polymerase chain reaction

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Azuma H. Genetic and molecular pathogenesis of hereditary hemorrhagic telangiectasia. J Med Invest. 2000;47:81–90. [PubMed] [Google Scholar]
  2. Bottinger EP, et al. Expression of a dominant-negative mutant TGF-beta type II receptor in transgenic mice reveals essential roles for TGF-beta in regulation of growth and differentiation in the exocrine pancreas. Embo J. 1997;16:2621–33. doi: 10.1093/emboj/16.10.2621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bourdeau A, et al. Potential role of modifier genes influencing transforming growth factor-beta1 levels in the development of vascular defects in endoglin heterozygous mice with hereditary hemorrhagic telangiectasia. Am J Pathol. 2001;158:2011–20. doi: 10.1016/s0002-9440(10)64673-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Chuang VP, et al. Angiography of pancreatic arteriovenuos malformation. AJR Am J Roentgenol. 1977;129:1015–8. doi: 10.2214/ajr.129.6.1015. [DOI] [PubMed] [Google Scholar]
  5. Crisera CA, et al. The ontogeny of TGF-beta1, -beta2, -beta3, and TGF-beta receptor-II expression in the pancreas: implications for regulation of growth and differentiation. J Pediatr Surg. 1999;34:689–93. doi: 10.1016/s0022-3468(99)90357-3. discussion 693–4. [DOI] [PubMed] [Google Scholar]
  6. Dickson MC, et al. Defective haematopoiesis and vasculogenesis in transforming growth factor-beta 1 knock out mice. Development. 1995;121:1845–54. doi: 10.1242/dev.121.6.1845. [DOI] [PubMed] [Google Scholar]
  7. Gittes GK, et al. Lineage-specific morphogenesis in the developing pancreas: role of mesenchymal factors. Development. 1996;122:439–47. doi: 10.1242/dev.122.2.439. [DOI] [PubMed] [Google Scholar]
  8. Golosow N, Grobstein C. Epitheliomesenchymal interaction in pancreatic morphogenesis. Dev Biol. 1962;4:242–55. doi: 10.1016/0012-1606(62)90042-8. [DOI] [PubMed] [Google Scholar]
  9. Goumans MJ, et al. Activin receptor-like kinase (ALK)1 is an antagonistic mediator of lateral TGFbeta/ALK5 signaling. Mol Cell. 2003;12:817–28. doi: 10.1016/s1097-2765(03)00386-1. [DOI] [PubMed] [Google Scholar]
  10. Goumans MJ, et al. Balancing the activation state of the endothelium via two distinct TGF-beta type I receptors. Embo J. 2002;21:1743–53. doi: 10.1093/emboj/21.7.1743. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Halpern M, et al. Hereditary hemorrhagic telangiectasia. An angiographic study of abdominal visceral angiodysplasias associated with gastrointestinal hemorrhage. Radiology. 1968;90:1143–9. doi: 10.1148/90.6.1143. [DOI] [PubMed] [Google Scholar]
  12. Kim SK, et al. Activin receptor patterning of foregut organogenesis. Genes Dev. 2000;14:1866–71. [PMC free article] [PubMed] [Google Scholar]
  13. Lammert E, et al. Induction of pancreatic differentiation by signals from blood vessels. Science. 2001;294:564–7. doi: 10.1126/science.1064344. [DOI] [PubMed] [Google Scholar]
  14. Larsson J, et al. Abnormal angiogenesis but intact hematopoietic potential in TGF-beta type I receptor-deficient mice. Embo J. 2001;20:1663–73. doi: 10.1093/emboj/20.7.1663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Lee MS, et al. Accumulation of extracellular matrix and developmental dysregulation in the pancreas by transgenic production of transforming growth factor-beta 1. Am J Pathol. 1995;147:42–52. [PMC free article] [PubMed] [Google Scholar]
  16. Massague J. TGF-beta signal transduction. Annu Rev Biochem. 1998;67:753–91. doi: 10.1146/annurev.biochem.67.1.753. [DOI] [PubMed] [Google Scholar]
  17. Miralles F, et al. TGF-beta plays a key role in morphogenesis of the pancreatic islets of Langerhans by controlling the activity of the matrix metalloproteinase MMP-2. J Cell Biol. 1998a;143:827–36. doi: 10.1083/jcb.143.3.827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Miralles F, et al. Follistatin regulates the relative proportions of endocrine versus exocrine tissue during pancreatic development. Development. 1998b;125:1017–24. doi: 10.1242/dev.125.6.1017. [DOI] [PubMed] [Google Scholar]
  19. Oh SP, et al. Activin receptor-like kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis. Proc Natl Acad Sci U S A. 2000;97:2626–31. doi: 10.1073/pnas.97.6.2626. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Oshima M, et al. TGF-beta receptor type II deficiency results in defects of yolk sac hematopoiesis and vasculogenesis. Dev Biol. 1996;179:297–302. doi: 10.1006/dbio.1996.0259. [DOI] [PubMed] [Google Scholar]
  21. Pictet RL, Rutter WJ. Handbook of Physiology. Williams and Wilkins; Washington DC: 1972. Development of the embryonic endocrine pancreas. [Google Scholar]
  22. Rawdon BB, Andrew A. Effects of tri-iodothyronine (T3), insulin, insulin-like growth factor I (IGF-I) and transforming growth factor beta1 (TGFbeta1) on the proportion of insulin cells in cultured embryonic chick pancreas. Anat Embryol (Berl) 1998;198:245–54. doi: 10.1007/s004290050181. [DOI] [PubMed] [Google Scholar]
  23. Ritvos O, et al. Activin disrupts epithelial branching morphogenesis in developing glandular organs of the mouse. Mech Dev. 1995;50:229–45. doi: 10.1016/0925-4773(94)00342-k. [DOI] [PubMed] [Google Scholar]
  24. Rutter WJ, et al. An analysis of pancreatic development: role of mesenchymal factor and other extracellular factors. Symp Soc Dev Biol. 1978:205–27. doi: 10.1016/b978-0-12-612981-6.50019-3. [DOI] [PubMed] [Google Scholar]
  25. Sander M, et al. Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of beta-cell formation in the pancreas. Development. 2000;127:5533–40. doi: 10.1242/dev.127.24.5533. [DOI] [PubMed] [Google Scholar]
  26. Sanvito F, et al. TGF-beta 1 influences the relative development of the exocrine and endocrine pancreas in vitro. Development. 1994;120:3451–62. doi: 10.1242/dev.120.12.3451. [DOI] [PubMed] [Google Scholar]
  27. Sanvito F, et al. TGF-beta 1 overexpression in murine pancreas induces chronic pancreatitis and, together with TNF-alpha, triggers insulin-dependent diabetes. Biochem Biophys Res Commun. 1995;217:1279–86. doi: 10.1006/bbrc.1995.2906. [DOI] [PubMed] [Google Scholar]
  28. Shalev A, et al. Oligonucleotide microarray analysis of intact human pancreatic islets: identification of glucose-responsive genes and a highly regulated TGFbeta signaling pathway. Endocrinology. 2002;143:3695–8. doi: 10.1210/en.2002-220564. [DOI] [PubMed] [Google Scholar]
  29. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cell membrane to the nucleus. Cell. 2003;113:685–700. doi: 10.1016/s0092-8674(03)00432-x. [DOI] [PubMed] [Google Scholar]
  30. Skoudy A, et al. Transforming growth factor (TGF)beta, fibroblast growth factor (FGF) and retinoid signalling pathways promote pancreatic exocrine gene expression in mouse embryonic stem cells. Biochem J. 2004;379:749–56. doi: 10.1042/BJ20031784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Torsney E, et al. Mouse model for hereditary hemorrhagic telangiectasia has a generalized vascular abnormality. Circulation. 2003;107:1653–7. doi: 10.1161/01.CIR.0000058170.92267.00. [DOI] [PubMed] [Google Scholar]
  32. Yamaoka T, et al. Hypoplasia of pancreatic islets in transgenic mice expressing activin receptor mutants. J Clin Invest. 1998;102:294–301. doi: 10.1172/JCI2769. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01. Supplementary figure 1.

Histology of the embryonic pancreas from transgenic mice: H&E staining of wild-type CD1 (A,D,G,J,M), transgenic DNTβRII (B,E,H,K,N) and transgene-induced DNTβRII+Zn (C,F,I,L,O) pancreas at various gestational ages. At E11.5 and E12.5 no appreciable differences were seen between wild-type and transgenic pancreas. At E14.5 (G-I), transgenic pancreas showed increased expansion of the cord region (arrowheads in H,I). At E16.5 (J–L), the transgenic phenotype diverged most markedly from controls. The transgenic pancreas had prominent expansion of the cord region, with periductal accumulation of cells (arrowheads in K,N,O). At E18.5 (M–O), continued cord region expansion was seen with early acinar atypia. The phenotype was stronger in the transgene-induced pancreas, with abnormally dilated blood vessels (arrows in L,O). Scale bar 60 μm.

NIHMS23261-supplement-01.pdf (360.4KB, pdf)
02. Supplementary figure 2.

Increased proliferation of endocrine and ductal structures in transgene-induced pancreas: At E16.5, transgene-induced DNTβRII+Zn pancreas showed higher numbers of BrdU-positive cells in DBA-positive ducts (arrowheads in C and D) and insulin-positive endocrine cells (arrows in C and D) compared to wild-type CD1 pancreas. C and D are higher magnification of A and B, respectively. Scale bar 60μm

NIHMS23261-supplement-02.pdf (138.4KB, pdf)
03. Supplementary figure 3.

Apoptosis of endocrine cells during late gestation in transgene-induced pancreas: caspase-3 (red), a key enzyme involved in the terminal apoptotic cascade of cell death is activated more in and around the insulin (green) and glucagon (blue) positive cells in transgene-induced pancreas (B,D) compared to wild-type (A,C) at E16.5 (A,B) and E18.5 (C,D). This difference may suggest that an increase in apoptosis represents a key mechanism whereby DNTβRII transgenic pancreas loses the ability to sustain the enhanced endocrine phenotype. Scale bar 60μm

NIHMS23261-supplement-03.pdf (139.8KB, pdf)
04. Supplementary figure 4.

Early acinar atypia during late gestation in the transgene-induced pancreas: simple histological analysis of E18.5 pancreas by H&E staining showed normal floral arrangement of acinar clusters in wild-type CD1 controls (A,B). However, transgene-induced pancreas (C,D) showed localized disruption of acinar architecture, with apparently lobule-level acinar degeneration (arrowhead C). The acinar dysmorphogenesis could be due to overproduction of enzymes. This finding is reminiscent of the acinar atypia seen in the adult transgenic pancreas (Bottinger et al, 1997). B and D are higher magnification of A and C, respectively. Scale bar 60μm

NIHMS23261-supplement-04.pdf (428.5KB, pdf)

RESOURCES