Sonic hedgehog Is Required Early in Pancreatic Islet Development

Transcription

1 Developmental Biology 244, (2002) doi: /dbio , available online at on Sonic hedgehog Is Required Early in Pancreatic Islet Development Philip J. diiorio,*,1,2 Jennifer B. Moss,*,1 Jennifer L. Sbrogna, Rolf O. Karlstrom, and Larry G. Moss*,,3 *Department of Medicine and Department of Physiology, Tufts University School of Medicine, Boston, Massachusetts 02111; and Department of Biology, Morrill Science Center, University of Massachusetts, Amherst, Massachusetts Pancreatic organogenesis relies on a complex interplay of cell-autonomous and extracellular signals. We demonstrate that the morphogen sonic hedgehog (Shh) is required for pancreatic development in zebrafish. Genetic mutants of Shh and its signaling pathway establish this dependence as specific to endocrine, but not exocrine, pancreas. Using cyclopamine to inhibit hedgehog signaling, we show that transient Shh signaling is necessary during gastrulation for subsequent differentiation of endoderm into islet tissue. A second hedgehog-dependent activity occurring later in development was also identified and may be analogous to the known action of Shh in gut endoderm to direct localization of pancreatic development. The early action of Shh may be part of a more general process allowing neuroendocrine cells to originate in nonneuroectodermally derived tissues Elsevier Science (USA) Key Words: pancreas; islet; zebrafish; hedgehog; insulin. INTRODUCTION The exocrine and endocrine cells of the pancreas arise from common precursors present in the early gut endoderm (Le Douarin, 1988; Percival and Slack, 1999). Studies in chick and mouse have begun to uncover a complex program dependent on numerous transcription (Edlund, 2001) and signaling factors (Kim and Hebrok, 2001) required for the formation of the pancreas and differentiation of its component cell types. FGF-mediated interactions with adjacent mesoderm direct early anteroposterior (A-P) patterning of the endoderm (Wells and Melton, 2000). Later, communication from the notochord and local mesenchyme dictate and refine subsequent organ-specific commitment (Kim et al., 1997a). The pancreatic diverticulum evaginates from the posterior foregut endoderm at a characteristic position, within a broader domain defined by the expression of a 1 These authors contributed equally to this work. 2 Present address: Diabetes Division, University of Massachusetts Medical School, Worcester, MA To whom correspondence should be addressed at Department of Physiology, Tufts University School of Medicine, 136 Harrison Avenue, Boston, MA Fax: (617) larry. moss@tufts.edu. homeodomain transcription factor, Pdx-1 (Idx-1, Ipf-1) (Ohlsson et al., 1993; Miller et al., 1994). The genetic ablation of Pdx-1 results in pancreatic agenesis, although early endocrine precursor cells do form (Ahlgren et al., 1996; Offield et al., 1996). While much is now known regarding later steps in pancreatic development, the early processes by which gut endoderm acquires competence to respond to later specific signals are not well understood. Sonic hedgehog (Shh) is a member of a family of secreted morphogens that has been demonstrated to play an important role in pancreatic development. The dorsal pancreatic bud forms within a gap in the endodermal expression of shh, within the larger domain of Pdx-1 expression (Bitgood and McMahon, 1995; Apelqvist et al., 1997). Pancreatic development in the adjacent shh-expressing endoderm may operate through the suppression of essential inductive signals, such as activin, from the overlying notochord (Kim et al., 1997a; Hebrok et al., 1998). For example, overexpression of Shh within the Pdx-1-expressing endoderm blocks pancreas formation, although differentiated exocrine and endocrine cells form and are mislocalized to the intestinal wall (Apelqvist et al., 1997). Inhibition of Shh signaling by the alkaloid cyclopamine in 10-somite chick embryos causes anterior and posterior expansion of pancreatic endoderm and a relative hypertro /02 $ Elsevier Science (USA) All rights reserved. 75

2 76 diiorio et al. FIG. 1. Disruption of Shh signaling alters pancreas development. (A) Homozygous 48-hpf INS-GFP transgenic zebrafish. Untreated embryos (left) contain a compact multicellular cluster of GFP-expressing cells. Addition of 50 M cyclopamine at 4 hpf (right) blocks formation of GFP islets. Some isolated, single GFP-expressing cells remain (arrows). (B) 48 hpf WT zebrafish in situ hybridized as whole-mount embryos with a proinsulin riboprobe depict normal clusters of insulin-expressing cells in the pancreas budding from the right side of the endoderm (left). 48 hpf Syu t4 / embryos have lost all insulin expression (middle) or exhibit expression (right) in isolated single cells (arrowheads). (C) Early insulin expression in a 24-hpf WT (left) and syu embryo (right). (D) Hybridization of WT embryos with a proglucagon probe (left) reveals cells in the perimeter of the islet as well as in an isolated cell within the gut at 48 hpf (arrow). No mrna for this neuroendocrine hormone is seen in syu embryos (right). (B D) Views are dorsal, with anterior to the left. phy of islet endocrine cell types (Kim et al., 1997b). In mice, two hedgehog (hh) genes, indian (ihh) and shh, are expressed in gut endoderm. Due to lethality before pancreas development, shh / ihh / mice are uninformative (Hebrok et al., 2000). However, the viable shh / ihh / mouse develops an annular pancreas and generates a relative excess of pancre-

3 Shh Promotes Zebrafish Pancreatic Islet Development 77 FIG. 2. Nature of the islet defect in the zebrafish mutant syu t4. The morphology of WT (A) vs syu embryos (B) was compared. Paraffin transverse cut sections (5 m), stained with hematoxylin and eosin. (A) WT islet (arrow). (B) The islet is selectively missing in syu embryos (arrow). (C) Carboxypeptidase A expression (arrow) in 72-hpf syu embryos embedded in paraffin, sectioned, and counterstained with Vector Green. (D I) Whole-mount in situ hybridization analysis of Wt and syu / embryos. Views are dorsal with anterior to the left, except (I), which is a lateral view. (D) Carboxypeptidase A expression in 48-hpf WT and syu / embryos. (E) Expression 72 hpf of transferrin (F) Pdx-1 expression in 48-hpf WT pancreas (arrow) and duodenum (arrowhead). (G) Isl-1 expression in 48-hpf syu / embryos. Arrows denote loss of expression. Foregut expression (arrowheads) is retained. (H) Loss of Beta2/NeuroD expression within the islet at 48 hpf (arrow) in syu / embryos. (I) Nkx2.2 (nk2.2) expression in WT 24-hpf embryos in early gut endoderm where the pancreatic diverticulum will form (arrow). FIG. 3. Insulin expression in hh signal transduction mutants. Insulin expression in 48-hpf WT (A), homozygous yot ty17a / (B) and smu b577 / embryos (C). Arrows denote the presence of insulin-labeled cells in the mutants (B, C). atic endocrine tissue (Hebrok et al., 2000). Thus, the established role of Shh signaling, when evaluated after the formation of the gut endoderm, appears to be inhibitory or restrictive to pancreas cell development. We used the transparent zebrafish embryo to address the possible role of Shh signaling at earlier stages of development. The zebrafish pancreas forms as a single dorsal evagination within a domain of Pdx-1-expressing endoderm located adjacent to the A-P boundary of somites 3 and 4 (Argenton et al., 1999). Islet precursor cells first express insulin at the 12-somite stage, h postfertilization (hpf) (Biemar et al., 2001). Clusters of insulin-positive cells subsequently increase in number, fuse into a single group by 24 hpf (30-somite stage), and condense into an islet surrounded by exocrine pancreatic tissue by 48 hpf (hatching stage) (Biemar et al., 2001). We generated a zebrafish transgenic line, in which insulin expression is represented by green fluorescent protein (GFP), and discovered that an inhibitor of hh signaling, cyclopamine, blocks formation of the pancreatic islet. A requirement for Shh in the pancreas was confirmed by analysis of zebrafish mutants that disrupt hh signaling. This novel requirement for shh early in embryogenesis may provide a clue to a long unanswered question of how islet endocrine cells, with their striking neuronal character, originate within the endoderm. MATERIALS AND METHODS GFP Transgenic Zebrafish Production and Imaging A 960-bp zebrafish insulin promoter was generated by PCR based on a published sequence (Milewski et al., 1998) and was cloned into a Tc3 expression vector (Raz et al., 1998) along with egfp (Clontech) to produce INS-GFP founders. Approximately 5 nl containing 20 ng/ml cesium-purified plasmid DNA and 50 ng/ml of capped Tc3A transposase mrna (mmessage mmachine; Ambion) was injected into one- to two-cell zebrafish embryos, generating singlecopy insertions. Embryos used in this study are generation 5 INS-GFP homozygotes. For fluorescence imaging, we used a Leica

4 78 diiorio et al. MZFI III dissecting microscope, a Hamamatsu Orca CCD camera, and Openlab software (Improvision) on a Macintosh G3 computer. Bright-field microscopy was recorded with a Zeiss Axiophot. Photoshop v.6 was used to prepare figures. Zebrafish Strains, Embryo Maintenance, and Cyclopamine Treatment Wildtype (WT) embryos were derived from the Oregon strain AB. Stages were assessed according to Kimmel et al. (1995) and are as follows: 4 hpf, blastula; 6 and 8 hpf, gastrula; 10 hpf, gastrulation complete; 12 hpf, 6 somites; 16 hpf, 14 somites; and 20 hpf, 24 somites. Syu t4 (Schauerte et al., 1998) is a deletion mutation in the shh gene, and Yot ty17a (Karlstrom et al., 1999) is a mutation in the gli-2 gene. Smu b577 (Barresi et al., 2000), a mutation in the smo gene, was a gift from S. Devoto. These homozygous lethal strains were maintained as heterozygotes and identified by their curved body axis. Homozygous INS-GFP embryos were grown at 28 C on agarosetreated dishes in embryo media and imaged (Westerfield, 1995). For cyclopamine treatment, chorions were removed by using pronase (Westerfield, 1995) and were grown on agarose-covered dishes. A 10 mm stock of cyclopamine (gift of W. Gaffield) was prepared in 95% ethanol and diluted to 50 M in embryo media. Histology and Whole-Mount in Situ Hybridization Five-micrometer paraffin sections were stained with hematoxylin and eosin. Whole-mount in situ hybridization was performed essentially as described (Oxtoby and Jowett, 1993) at 65 C, except for insulin (60 C). At 72 hpf, syu embryos were hybridized with carboxypeptidase A probe, embedded in paraffin, sectioned, and counterstained with vector green to indicate the presence of exocrine pancreas in the mutant. Insulin-hybridized mutant embryos were also sectioned (data not shown). Hybridization Probes Total RNA was prepared by using the Trizol method from adult pancreatic tissue, identifiable within the gut of INS-GFP zebrafish under the fluorescent dissecting microscope. Following reverse transcription using poly(dt) oligos and Superscript II RNA polymerase (Life Technologies), the cdna was used as a template to amplify pancreas-specific genes by using specific primers based on a published sequence. We generated PCR products for a 300-bp proinsulin (Milewski et al., 1998), a 650-bp proglucagon (Argenton et al., 1999), a 488-bp carboxypeptidase A, and a 540-bp Beta2/ NeuroD cdna (Liao et al., 1999) that were cloned into pcrii (Invitrogen). The inserts were sequenced and confirmed by using NCBI Blast ( Overexpression of Sonic Hedgehog mrna Capped shh mrna was synthesized from the expression plasmid pt7ts-shh by using the SP6 Message Machine (Ambion), and approximately 0.5 nl of 1 mg/ml solution was injected into yolks of embryos at the two- to four-cell stage. Injection of 0.5 nl of 1 mg/ml GFP mrna or -galactosidase ( -gal) mrna caused no detectable morphological abnormalities and served as a control for RNA toxicity. RESULTS Shh Is Required for Islet Formation An INS-GFP transgenic zebrafish line in which the zebrafish insulin promoter targets expression of GFP to the insulin-expressing cells of the islet was generated. GFP expression was detected at the 17-somite stage (18 hpf) in cells of the midline gut endoderm ventral to somite 4 (data not shown). By 48 hpf, the free-swimming zebrafish contains an islet (Biemar et al., 2001), represented by a discrete cluster of GFP-positive cells (Fig. 1A). Formation of these GFP-positive islets was blocked when 4-hpf transgenic embryos were exposed to cyclopamine (Fig. 1A), a potent inhibitor of hh signaling (Cooper et al., 1998; Incardona et al., 1998). The islet cells were either absent or detectable as one or a few isolated cells in the gut endoderm. Cyclopamine is a general inhibitor of hh action that interferes with smoothened, a plasma membrane protein that interacts with the hh receptor patched and is required for hh signal transduction (Taipale et al., 2000). To verify that the defects seen after cyclopamine treatment are indeed due to defects in Shh signaling, we examined insulin gene expression and islet development in sonic you (syu) embryos, which carry a loss-of-function mutation in the shh gene (Schauerte et al., 1998). As was seen after cyclopamine treatment, the insulin-expressing islet cell cluster fails to develop in syu embryos. When observed at 48 hpf, insulin expression is either completely eliminated (34%; 14/41 embryos) or reduced to one or more isolated single cells (66%; 27/41 embryos) (Fig. 1B). Disruption of normal islet development is also detected earlier at 24 hpf (Fig. 1C), indicating that the loss of insulin-expressing cells is not due to degeneration. Thus, we conclude that Shh signaling is a primary requirement for normal islet development. To determine whether islet cell types other than insulinexpressing cells are lost in syu / mutants, we analyzed the expression of glucagon, an endocrine hormone secreted by pancreatic cells. In 48-hpf WT embryos, glucagon expression was localized to cells which have acquired their characteristic location around the perimeter of the islet (Fig. 1D). No glucagon expression was observed in syu / (0/20) embryos, suggesting that the entire islet endocrine cell population was affected by the absence of sonic hedgehog. Blocking hh Signaling Specifically Disrupts Endocrine but Not Exocrine Islet Development Hematoxylin and eosin staining of paraffin sections shows that the 72-hpf WT pancreas consists of a single dorsal islet surrounded by basophilic exocrine tissue (Fig. 2A). In syu embryos, only basophilic tissue is seen in the expected location of the pancreas (Fig. 2B). This structure expresses carboxypeptidase A, a marker for exocrine pancreas (Fig. 2C), whereas insulin-expressing cells are not present (data not shown). Consistent with this observation, carboxypeptidase A expression in WT embryos encircles

5 Shh Promotes Zebrafish Pancreatic Islet Development 79 the islet and extends caudally along the gut to somite 9 (Fig. 2D, top). Carboxypeptidase A is also expressed in syu embryos (Fig. 2D, bottom), although expression is reduced in the anterior component of exocrine tissue, perhaps due to the absence of the normally underlying islet mass. The effects of Shh action on the liver, another organ derived from adjacent gut endoderm, were monitored by transferrin expression. No major effect resulting from the loss of Shh signaling on liver development was apparent in either WT or syu embryos at 72 hpf (Fig. 2E, top and bottom). Thus, the absence of Shh does not result in a generalized defect of gut endoderm patterning or differentiation, and the pancreatic defect is specific to the endocrine islet. We next evaluated the expression of transcription factors involved in pancreatic islet development in other organisms. Pdx-1 controls early pancreatic bud development and the ultimate morphogenesis of the pancreas but is not required for the emergence of early precursor cells in the gut endoderm (Ahlgren et al., 1998; Offield et al., 1996). In 48-hpf WT zebrafish, Pdx-1 is present in the duodenum and the entire pancreatic bud (Fig. 2F, top), which is predominately islet tissue at this stage (Biemar et al., 2001). In syu embryos, Pdx-1 expression is retained in the duodenum (Fig. 2F, bottom) but not in the bud (arrows). In addition, Isl-1 is required for the development in mouse of the entire dorsal, but not the ventral, pancreatic lobe as well as for all islet endocrine tissue (Ahlgren et al., 1997). Isl-1 is expressed in the pancreatic islet and in the gut immediately anterior to the pancreatic bud in WT zebrafish embryos (Fig. 2G, top). However, syu embryos lack islet expression of Isl-1, whereas it is retained in the adjacent gut endoderm (Fig. 2G, bottom). Another factor known to contribute to normal islet development is Beta2/NeuroD, a bhlh transcription factor expressed in early islet precursors and mature cells (Naya et al., 1997) that is expressed in the zebrafish islet (Fig 2h, top, arrow) yet is absent in the pancreas of 48-hpf syu embryos (Fig. 2H, bottom). Finally, we examined the expression of the homeodomain transcription factor Nkx2.2 (nk2.2), which is necessary for the establishment of insulin-expressing islet cells in mice (Sussel et al., 1998). Nkx2.2 is expressed within the developing pancreatic endoderm in 24-hpf WT embryos but not in syu embryos (Fig. 2I). The failure to detect any Nkx2.2 and Beta2/NeuroD expression in the syu gut endoderm suggests that the loss of hh signaling may selectively inhibit formation of early pancreatic endocrine precursors. Effects of Other hh Signaling Pathway Mutations on Pancreatic Islet Development Because additional hh family members, such as tiggywinkle and echidna hh, are expressed in zebrafish (Ekker et al., 1995; Norris et al., 2000), we analyzed mutations in downstream components of hh signaling. You-too (yot) mutations disrupt zebrafish gli-2, one of a family of transcription factors required for mediating hh action in the nucleus (Karlstrom et al., 1999). Compared with WT embryos (Fig. 3A), yot embryos (Fig. 3B) contain anterior, mislocalized insulin-expressing cells (arrow) that are reduced in number, indicating a role for gli genes in interpreting hh signals in developing islet cells. In contrast to syu embryos, insulin expression is retained in all yot mutants examined (22/22 embryos). Compensation by other gli family members may be responsible for the reduced severity of the phenotype compared with syu mutants. Therefore, we evaluated a mutation in slow muscle omitted (smu) embryos that contain a loss-of-function mutation in Smoothened, an extracellular membrane protein required for all hh signal transduction (Varga et al., 2001). Smu embryos exhibit a more profound disruption of development affecting additional aspects of embryogenesis than syu, such as ventral cyclopia (Barresi et al., 2000). However, the defect in insulin expression seen in smu embryos (Fig. 3C) is similar to that seen in syu embryos (Fig. 1B) in that insulin expression is either absent or found in single, isolated cells. Because the observed effect on insulin expression in yot, smu, or cyclopamine-treated embryos is no more severe than that in syu embryos, the effects on islet development appear to depend primarily on the action of Shh. However, subtle contributions from other hh family members during normal development cannot be ruled out. Overexpression of shh Expands Pancreatic Islet Development The inhibition in islet development caused by decreased hh signaling led us to examine whether overexpression of Shh protein early in development would affect islet phenotype. Injection of shh mrna is consistently associated with a dramatic increase in the number of insulinexpressing cells (Fig. 4; 151/154 embryos). Associated with this increase, the shh-injected embryos contain significant numbers of ectopic insulin-positive cells (35/154 embryos) extending anteriorly as far as somite 1 (Fig. 4F). In WT and control-injected embryos, occasional insulin-positive cells could be identified in the gut endoderm just anterior to the developing islet but within the normal domain of insulin expression (Fig. 4E; 2/56 embryos). In addition, both pancreatic and duodenal expression of Pdx-1 were increased in shh-injected embryos (48/51 embryos) when viewed at 48 hpf (Figs. 4G and 4H). Separate Early and Late Actions of hh Signaling on Islet Development To evaluate the stage-specific requirement for Shh action on the pancreas, we treated INS-GFP or WT embryos with cyclopamine at a series of early time points. The interruption of hh signaling on islet development when observed at 48 hpf for GFP fluorescence or in situ insulin expression is comparable (Figs. 5A and 5E; Table 1). When cyclopamine is administered at the beginning of gastrulation (6 hpf; shield stage), insulin expression is ablated or reduced to single-cell loci (Figs. 4B and 4F; Table 1). This phenotype is similar to

6 80 diiorio et al. FIG. 4. Insulin and Pdx-1 expression after shh overexpression in WT embryos. Left: mrna control-injected embryos. Right; shh mrna-injected embryos. After mrna injection at the two- to four-cell stage, embryos developed for 24 or 48 hpf and were subjected to whole-mount in situ analysis using the insulin or Pdx-1 probe. (A, C, E) Control GFP mrna-injected embryos hybridized at 24 hpf with the insulin probe. Positive cells are centered ventral to somite 4 (4). (B, D, F) Shh mrna-injected embryos hybridized with the insulin probe. Note the anterior extension of labeled cells. (G) Control -gal mrna-injected embryo hybridized with the Pdx-1 probe. (H) Shh mrna-injected embryo hybridized with the Pdx-1 probe. (A, B, E, F) Ventral views. (C, D) Lateral views. (G, H) Dorsal views. that observed in syu / embryos. When administered at 8 hpf (late gastrulation), insulin expression becomes detectable in an increasing percentage of embryos, and multicellular clusters are occasionally detected (Figs. 5C and 5G; Table 1). Treatment at 12 hpf (6-somite stage) permits insulin expression in all embryos, most of which form normal-sized, multicellular clusters (Figs. 5D and 5H; Table 1). At later time points (16 hpf, 14 somites; and 20 hpf, 24 somites), islet development appears to be unaffected by cyclopamine administration (Table 1). These data reveal a requirement during gastrulation for hh signaling in pancreatic islet development. Another consequence of cyclopamine treatment is the formation of multiple loci of insulin-positive cells at ectopic sites anterior to the normal islet location at the somite 3/4 boundary (Fig. 5; Table 1). Importantly, treatment at 12 hpf (6 somites) leads to normal-sized islet clusters, but the percentage of embryos with islet cells shifted to anterior positions is similar to that caused by treatment during gastrulation. This suggests the presence of two separate actions of shh during pancreas development. An early Shh activity appears to be required for the differentiation of endocrine cells and formation of endocrine islets. Without this early shh activity, only single-cell loci

7 Shh Promotes Zebrafish Pancreatic Islet Development 81 FIG. 5. Shh signaling is required early for patterning the endocrine islet. Living fifth-generation homozygous INS-GFP or WT fixed embryos imaged at 48 hpf untreated (A, E) or treated with 50 M cyclopamine at 6 (B, F), 8 (C, G), or 12 (D, H) hpf. Representative islet phenotypes are shown in living embryos as fluorescent (A D) or as bright-field images after fixation and in situ hybridization with insulin probe (E H). (A) 48-hpf untreated INS-GFP embryos. (B) Exposure to cyclopamine at 6 hpf. (C) Treatment at 8 hpf. (D) Cyclopamine exposure at 12 hpf. (E) Endogenous insulin expression in control ethanol-treated embryos. (F H) Cyclopamine-treated embryos. Cyclopamine was added at 5 hpf (F); branchial arches (arrowhead), 8 (G) and 12 hpf (H). Abbreviations: s4, somite 4; s1, somite1; n, notochord; arrows, position of INS-GFP fluorescent cells; arrowheads, position of insulin-expressing cells. can arise, often at anterior locations. When Shh activity is interrupted after gastrulation (once endoderm formation is complete), normal islets are observed, indicating that Shh is involved in the correct localization of the pancreatic islet. DISCUSSION We report a previously undetected early requirement for Shh in the specification of the pancreas from gut endoderm. This early role for Shh is distinct from any later role of hh signaling in restricting pancreas to a discrete endodermal location. The early action of Shh during pancreatic organogenesis is specific to endocrine islet cells and does not directly affect the exocrine pancreas, duodenum, or liver. Our initial observation that cyclopamine treatment of presomitic zebrafish embryos does not cause an increase in the amount of insulin-expressing islet tissue was unexpected since previous results for cyclopamine-treated 10-somite chick embryos (Kim and Melton, 1998) defined a repressive activity of Shh on pancreas development. One explanation is that the observed Shh signaling in the pancreas is unique to zebrafish. Alternatively, the zebrafish system allows us to look both genetically and pharmacologically to uncover a requirement for hh signaling in vertebrate pancreas development not observable in mammals. We were able to confirm our results by using a direct examination in INS- GFP transgenic embryos to determine the precise timing of the cyclopamine effect based on a temporal analysis of inhibition. We found that the requirement for Shh in islet development is transient and appears to be exerted prior to the end of gastrulation, before notochord development and during endoderm formation. At this stage, the known sources of Shh are the embryonic shield, which acts as a dorsal organizer, and an axially extending mesoendoderm similar to the chick primitive streak (Krauss et al., 1993). Fate mapping in zebrafish embryos shows that pancreatic precursors originate in the blastoderm margin at 40% epiboly, prior to the onset of gastrulation (Warga and Nusslein-Volhard, 1999). After involution, these cells migrate dorsally past the shield and arrive at the midline mesoendoderm by the end of gastrulation. Thus, pancreatic precursors normally migrate in close proximity to Shhexpressing cells at a time when our data show that they are dependent on Shh for specification of the pancreas. Early Shh signaling may act to maintain or induce proliferation in a naive and receptive subset of potential endodermal pancreatic precursors (Rowitch et al., 1999). Alternatively, a transient Shh signal during gastrulation could endow a subset of endodermal cells with competence to respond to later signals that can trigger the generation of the pancreatic endocrine phenotype. This possibility is supported by the observation that the region-specific differentiation of neurons in the chick ventral telencephalon depends on Shh signaling during gastrulation (Gunhaga et al., 2000). Furthermore, shh appears to play a similar early role in the development of the zebrafish adenohypophysis/ pituitary gland (data not shown). Novel experimental strat-

8 82 diiorio et al. TABLE 1 Insulin Expression in Developing, Cyclopamine-Treated Embryos Number of embryos Cyclopamine addition (hpf) Pattern of insulin-gfp expression observed at 72 hpf GFP Loci/embryo No expression Single cell loci Small clusters WT clusters (16) 54 (2) (8) 25 (4) 43 (4) (3) 114 (13) (26) (58) 75 1 Vehicle only 121 (180) 121 Note. Alteration of pancreatic islet development by cyclopamine is dependent on the developmental stage of treatment. Homozygous INS-GFP embryos treated at six time points with 50 M cyclopamine were allowed to develop for 72 hpf and imaged or fixed for insulin in situ analysis. Small clusters contain at least 1 locus of 2 4 fluorescent cells. A WT cluster contains at least 1 locus of 5 cells. Untreated embryos have a single locus of 5 12 cells at 72 hpf (vehicle). Treatment at early stages (4 and 6 hpf) generates no WT clusters and results in anterior repositioning of GFP cells. At the 5- to 7-somite stage (12 hpf), the inhibitory effect of cyclopamine on islet development sharply declines. The number of independent GFP loci/embryo disappears by the 13- to 15-somite stage (20 hpf). Numbers in parentheses indicate embryos analyzed by insulin in situ hybridization. egies will be required to confirm whether Shh also directs early mammalian pancreatic development. Our time series of cyclopamine treatment also reveals a second hh activity once gastrulation is complete. This activity appears to function in the correct anatomical localization of the islet within the gut endoderm since cyclopamine treatment during somitogenesis stages results in the anterior mislocalization of normal-sized insulinexpressing cell clusters. Because this anteriorization is seen after formation of the primitive gut endoderm, it is probably not due to abnormal migration of pancreatic precursors. It is likely that the loss of hh signaling after the endoderm has formed affects inductive processes, perhaps via a release of inhibition anteriorly. This potentially restrictive effect of hh signaling occurring after gastrulation in zebrafish could be analogous to the previously described antagonistic role of Shh identified in chick (Kim and Melton, 1998) and mouse (Apelqvist et al., 1997; Hebrok et al., 2000). However, different mechanisms may be involved for A-P positioning in zebrafish because, unlike in mouse and chick, we cannot detect Shh expression in the gut endoderm prior to 24 hpf (30 somites; data not shown). Early overexpression of Shh also results in islet cells located anterior to somite 4, suggesting the override of an anterior limit of pancreas induction. The interrelationship between early and late hedgehog activity may prove to be complex. However, the generation of differential effects by hh signaling on islet development or function is not unexpected. For example, in differentiated cell lines, insulin gene expression is increased by hh-stimulated Idx-1 (Pdx-1) expression (Thomas et al., 2001). In our experiments, Shh overexpression also increases Pdx-1 expression within the gut endoderm and the pancreas (Figs. 4G and 4H). Because cyclopamine treatment after 24 hpf is not associated with a decrease in the number of insulin-expressing islet cells, hh signaling may not be required for later islet growth in the zebrafish (Table 1). Our overexpression experiments suggest that the early action of Shh to promote islet development could be mediated through Pdx-1. However, the presence of Pdx-1 as well as Isl-1, another potential target of Shh action, in their respective gut domains in the syu / embryos suggests that the islet defect may not be due to a primary failure of Shh to activate Pdx-1 or Isl-1 during development. Recent work (Grapin-Botton et al., 2001) indicates that overexpression of Pdx-1 in chick gut endoderm is not sufficient to induce the formation of islet endocrine cells. Alternatively, the known Shh target genes Nkx2.2 and Pax 6 may play a role in the development of the pancreas (Sussel et al., 1998; Sander et al., 1997; St-Onge et al., 1997) that can be compared with central nervous system development (Barth and Wilson, 1995). We observed that the early pancreatic defect seen in zebrafish mutants deficient in Shh signaling differs significantly from pancreatic phenotypes of mouse genetic knockouts of candidate target genes, including Pdx-1 (Ahlgren et al., 1996; Offield et al., 1996), Hb9 (Li et al., 1999; Harrison et al., 1999), Isl-1 (Ahlgren et al., 1998), NeuroD/Beta2 (Naya et al., 1997), Nkx2.2 (Sussel et al., 1998), Nkx6.1 (Sander et al., 2000), or Pax6 (St-Onge et al., 1997; Sander et al., 1997). Importantly, the selective failure of endocrine islet development we observed does resemble the phenotype of neurogenin 3 inactivation in mice (Gradwohl et al., 2000). This proneural transcription factor functions early in pancreatic development as part of a notch-based mechanism that directs pancreatic precursors to differentiate into endocrine instead of exocrine lineages (Apelqvist et al., 1999; Jensen et al., 2000). Additionally, ectopic overexpression of neurogenin 3 can induce islet cells in early chick gut

9 Shh Promotes Zebrafish Pancreatic Islet Development 83 endoderm (Grapin-Botton et al., 2001). Finally, it has been demonstrated that the expression of another proneural family member, neurogenin 1, is regulated by Shh in the developing zebrafish central nervous system (Blader et al., 1997). Unfortunately, the identity of a zebrafish neurogenin 3 homolog has not yet been established. Islet endocrine cells have long been recognized as structurally and functionally similar to neurons. These peptide hormone-secreting cells express many proteins characteristic of a neural phenotype such as ion channels and key components of a calcium-stimulated, granular secretion apparatus. The resemblance between neural cells and islet cells is further highlighted by the fact that many transcription factors involved in islet cell development are shared by developing neurons. These include Hb9, Isl-1, Nkx2.2, Nkx6.1, and Pax6 that participate in the hh-dependent differentiation of ventral spinal cord motor neurons (Briscoe and Ericson, 2001). While embryonic islet cells were once believed to arise from neural crest cells migrating into the endoderm, it is now clear that both endocrine and exocrine pancreatic cells arise from the endoderm (Le Douarin, 1988; Percival and Slack, 1999). The expression of a neuronal phenotype within endodermal neuroendocrine cells may point to a common mechanism shared between islet cells and neurons. Our studies show that Shh activity may be a mechanistic link between these two differentiation events. Clarification of the role Shh plays in this process awaits the identification of the early endocrine-competent precursor cells along with the specific genes that are targeted by the shh signal. ACKNOWLEDGMENTS Syu and Smu zebrafish strains were kindly provided by A. Chandrasekhar and S. DeVoto, respectively. D. Melton, L. Zon, R. Moon, M. Pack, C. Burgtorf, and H. Lerach shared DNA reagents. We thank N. Roy, A. Goff, A. Klein, J. Weng, J. Dowd, and L. Etheridge for their technical expertise. These studies were supported by Juvenile Diabetes Foundation (to L.G.M.), NIH DK55370 (to L.G.M.), NIH F32DK09743 (to P.J.D.), NIH NS39994 (to R.O.K.), NIH GRASP Digestive Disease Center P30 DK34928, New England Medical Center Research Fund, and a generous gift from Mr. Edmund Lynch. Note added in proof. After submission of our manuscript, a brief communication by Roy et al. [Roy, S., Qiao, T., Wolff, C., and Ingham, P. (2001) Hedgehog signaling pathway is essential for pancreas specification in the zebrafish embryo. Curr. Biol. 11, ] also reported that pancreatic development in zebrafish is dependent on hh signaling. REFERENCES Ahlgren, U., Jonsson, J., and Edlund, H. (1996). The morphogenesis of the pancreatic mesenchyme is uncoupled from that of the pancreatic epithelium in IPF1/PDX1-deficient mice. Development 122, Ahlgren, U., Jonsson, J., Jonsson, L., Simu, K., and Edlund, H. (1998). -Cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the -cell phenotype and maturity onset diabetes. Genes Dev. 12, Ahlgren, U., Pfaff, S. L., Jessell, T. M., Edlund, T., and Edlund, H. (1997). Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature 385, Apelqvist, A., Ahlgren, U., and Edlund, H. (1997). Sonic hedgehog directs specialised mesoderm differentiation in the intestine and pancreas. Curr. Biol. 7, Apelqvist, A., Li, H., Sommer, L., Beatus, P., Anderson, D. J., Honjo, T., Hrabe de Angelis, M., Lendahl, U., and Edlund, H. (1999). Notch signalling controls pancreatic cell differentiation. Nature 400, Argenton, F., Zecchin, E., and Bortolussi, M. (1999). Early appearance of pancreatic hormone-expressing cells in the zebrafish embryo. Mech. Dev. 87, Barresi, M. J., Stickney, H. L., and Devoto, S. H. (2000). The zebrafish slow-muscle-omitted gene product is required for Hedgehog signal transduction and the development of slow muscle identity. Development 127, Barth, K. A., and Wilson, S. W. (1995). Expression of zebrafish nk2.2 is influenced by sonic hedgehog/vertebrate hedgehog-1 and demarcates a zone of neuronal differentiation in the embryonic forebrain. Development 121, Biemar, F., Argenton, F., Schmidtke, R., Epperlein, S., Peers, B., and Driever, W. (2001). Pancreas development in zebrafish: Early dispersed appearance of endocrine hormone expressing cells and their convergence to form the definitive islet. Dev. Biol. 230, Bitgood, M. J., and McMahon, A. P. (1995). Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev. Biol. 172, Blader, P., Fischer, N., Gradwohl, G., Guillemont, F., and Strahle, U. (1997). The activity of neurogenin1 is controlled by local cues in the zebrafish embryo. Development 124, Briscoe, J., and Ericson, J. (2001). Specification of neuronal fates in the ventral neural tube. Curr. Opin. Neurobiol. 11, Cooper, M. K., Porter, J. A., Young, K. E., and Beachy, P. A. (1998). Teratogen-mediated inhibition of target tissue response to Shh signaling. Science 280, Edlund, H. (2001). Developmental biology of the pancreas. Diabetes 50 (Suppl. 1), S5 S9. Ekker, S. C., McGrew, L. L., Lai, C. J., Lee, J. J., von Kessler, D. P., Moon, R. T., and Beachy, P. A. (1995). Distinct expression and shared activities of members of the hedgehog gene family of Xenopus laevis. Development 121, Gradwohl, G., Dierich, A., LeMeur, M., and Guillemot, F. (2000). neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas. Proc. Natl. Acad. Sci. USA 97, Grapin-Botton, A., Majithia, A. R., and Melton, D. A. (2001). Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes. Genes Dev. 15, Gunhaga, L., Jessell, T. M., and Edlund, T. (2000). Sonic hedgehog signaling at gastrula stages specifies ventral telencephalic cells in the chick embryo. Development 127, Harrison, K. A., Thaler, J., Pfaff, S. L., Gu, H., and Kehrl, J. H. (1999). Pancreas dorsal lobe agenesis and abnormal islets of Langerhans in Hlxb9-deficient mice. Nat. Genet. 23,

10 84 diiorio et al. Hebrok, M., Kim, S. K., and Melton, D. A. (1998). Notochord repression of endodermal Sonic hedgehog permits pancreas development. Genes Dev. 12, Hebrok, M., Kim, S. K., St. Jacques, B., McMahon, A. P., and Melton, D. A. (2000). Regulation of pancreas development by hedgehog signaling. Development 127, Incardona, J. P., Gaffield, W., Kapur, R. P., and Roelink, H. (1998). The teratogenic Veratrum alkaloid cyclopamine inhibits sonic hedgehog signal transduction. Development 125, Jensen, J., Pedersen, E. E., Galante, P., Hald, J., Heller, R. S., Ishibashi, M., Kageyama, R., Guillemot, F., Serup, P., and Madsen, O. D. (2000). Control of endodermal endocrine development by Hes-1. Nat. Genet. 24, Karlstrom, R. O., Talbot, W. S., and Schier, A. F. (1999). Comparative synteny cloning of zebrafish you-too: Mutations in the Hedgehog target gli2 affect ventral forebrain patterning. Genes Dev. 13, Kim, S. K., and Hebrok, M. (2001). Intercellular signals regulating pancreas development and function. Genes Dev. 15, Kim, S. K., Hebrok, M., and Melton, D. A. (1997a). Notochord to endoderm signaling is required for pancreas development. Development 124, Kim, S. K., Hebrok, M., and Melton, D. A. (1997b). Pancreas development in the chick embryo. Cold Spring Harbor. Symp. Quant. Biol. 62, Kim, S. K., and Melton, D. A. (1998). Pancreas development is promoted by cyclopamine, a hedgehog signaling inhibitor. Proc. Natl. Acad. Sci. USA 95, Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling, T. F. (1995). Stages of embryonic development of the zebrafish. Dev. Dyn. 203, Krauss, S., Concordet, J. P., and Ingham, P. W. (1993). A functionally conserved homolog of the Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in zebrafish embryos. Cell 75, Le Douarin, N. (1988). On the origin of pancreatic endocrine cells. Cell 53, Li, H., Arber, S., Jessell, T. M., and Edlund, H. (1999). Selective agenesis of the dorsal pancreas in mice lacking homeobox gene Hlxb9. Nat. Genet. 23, Liao, J., He, J., Yan, T., Korzh, V., and Gong, Z. (1999). A class of neurod-related basic helix-loop-helix transcription factors expressed in developing central nervous system in zebrafish. DNA Cell Biol. 18, Milewski, W. M., Duguay, S. J., Chan, S. J., and Steiner, D. F. (1998). Conservation of PDX-1 structure, function, and expression in zebrafish. Endocrinology 139, Miller, C. P., McGehee Jr., R. E., and Habener, J. F. (1994). IDX-1: A new homeodomain transcription factor expressed in rat pancreatic islets and duodenum that transactivates the somatostatin gene. EMBO J. 13, Naya, F. J., Huang, H. P., Qiu, Y., Mutoh, H., DeMayo, F. J., Leiter, A. B., and Tsai, M. J. (1997). Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice. Genes Dev. 11, Norris, W., Neyt, C., Ingham, P. W., and Currie, P. D. (2000). Slow muscle induction by Hedgehog signalling in vitro. J. Cell Sci. 113, Offield, M. F., Jetton, T. L., Labosky, P. A., Ray, M., Stein, R. W., Magnuson, M. A., Hogan, B. L., and Wright, C. V. (1996). PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum. Development 122, Ohlsson, H., Karlsson, K., and Edlund, T. (1993). IPF1, a homeodomain-containing transactivator of the insulin gene. EMBO J. 12, Oxtoby, E., and Jowett, T. (1993). Cloning of the zebrafish krox-20 gene (krx-20) and its expression during hindbrain development. Nucleic Acids Res. 21, Percival, A. C., and Slack, J. M. (1999). Analysis of pancreatic development using a cell lineage label. Exp. Cell Res. 247, Raz, E., van Luenen, H. G., Schaerringer, B., Plasterk, R. H., and Driever, W. (1998). Transposition of the nematode Caenorhabditis elegans Tc3 element in the zebrafish Danio rerio. Curr. Biol. 8, Rowitch, D. H., S-Jacques, B., Lee, S. M., Flax, J. D., Snyder, E. Y., and McMahon, A. P. (1999). Sonic hedgehog regulates proliferation and inhibits differentiation of CNS precursor cells. J. Neurosci. 19, Sander, M., Neubuser, A., Kalamaras, J., Ee, H. C., Martin, G. R., and German, M. S. (1997). Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development. Genes Dev. 11, Sander, M., Sussel, L., Conners, J., Scheel, D., Kalamaras, J., Dela Cruz, F., Schwitzgebel, V., Hayes-Jordan, A., and German, M. (2000). Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of beta-cell formation in the pancreas. Development 127, Schauerte, H. E., van Eeden, F. J., Fricke, C., Odenthal, J., Strahle, U., and Haffter, P. (1998). Sonic hedgehog is not required for the induction of medial floor plate cells in the zebrafish. Development 125, St-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A., and Gruss, P. (1997). Pax6 is required for differentiation of glucagonproducing alpha-cells in mouse pancreas. Nature 387, Sussel, L., Kalamaras, J., Hartigan O Connor, D. J., Meneses, J. J., Pedersen, R. A., Rubenstein, J. L., and German, M. S. (1998). Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells. Development 125, Taipale, J., Chen, J. K., Cooper, M. K., Wang, B., Mann, R. K., Milenkovic, L., Scott, M. P., and Beachy, P. A. (2000). Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine. Nature 406, Thomas, M., Lee, J., Rastalsky, N., and Habener, J. (2001). Hedgehog signaling regulation of homeodomain protein islet duodenum homeobox-1 expression in pancreatic -cells. Endocrinology 142, Varga, Z. M., Amores, A., Lewis, K. E., Yan, Y. L., Postlethwait, J. H., Eisen, J. S., and Westerfield, M. (2001). Zebrafish smoothened functions in ventral neural tube specification and axon tract formation. Development 128, Warga, R. M., and Nusslein-Volhard, C. (1999). Origin and development of the zebrafish endoderm. Development 126, Wells, J. M., and Melton, D. A. (2000). Early mouse endoderm is patterned by soluble factors from adjacent germ layers. Development 127, Westerfield, M. (1995). The Zebrafish Book: Guide for the Laboratory Use of Zebrafish (Danio rerio). Univ. of Oregon Press, Eugene. Received for publication September 28, 2001 Accepted December 19, 2001 Published online February 25, 2002