- Top of page
- EXPERIMENTAL PROCEDURES
Here, we report a detailed fate map of the zebrafish pancreas at the early gastrula stage of development (6 hours postfertilization; hpf). We show that, at this stage, both pancreas and liver progenitors are symmetrically localized in two broad domains relative to the dorsal organizer. We demonstrate that the dorsal and ventral pancreatic buds can derive from common progenitor pools at 6 hpf, but often derive from independent populations. Endocrine vs. exocrine pancreas show a similar pattern of progenitors, consistent with descriptions of the dorsal bud being strictly endocrine and the ventral bud primarily exocrine. In general, we find that endocrine/dorsal bud progenitors are located more dorsally than the exocrine pancreas/ventral bud progenitors. Later in gastrulation (10 hpf), pancreas progenitors have migrated to bilateral domains at the equator of the embryo. Our fate map will assist with design and interpretation of future experiments to understand early pancreas development. Developmental Dynamics 236:1558–1569, 2007. © 2007 Wiley-Liss, Inc.
- Top of page
- EXPERIMENTAL PROCEDURES
The zebrafish is emerging as a useful model for the study of vertebrate organogenesis, combining powerful genetics with accessible, optically clear embryos that allow cell fates and movements to be traced. In recent years, several studies have revealed that the zebrafish pancreas shares many properties with the mammalian pancreas. Thus, the zebrafish pancreas consists of both an endocrine and an exocrine component (Slack,1995; Biemar et al.,2001), and many molecular pathways acting during early pancreatic development are evolutionarily conserved (e.g., Biemar et al.,2001; Gu et al.,2004; Stafford et al.,2004,2006; Molotkov et al.,2005). However, to understand how pancreatic progenitors become separated from the previously undifferentiated cells of the endoderm germ layer, one must first determine when and where these cells originate. In this study, we have produced a fate map of the zebrafish endocrine and exocrine pancreas that will assist future studies of pancreas development.
As in other species, the zebrafish pancreas develops from distinct dorsal and ventral buds associated with the gut tube (Field et al.,2003, Wallace and Pack,2003; Fig. 1). A transgenic line expressing green fluorescent protein in the postpharyngeal endoderm (gut-GFP, Field et al.,2003) has allowed visualization of the morphogenesis of the two zebrafish pancreatic buds. The dorsal bud is located slightly posterior to the ventral bud, and we hereafter refer to them as the posterodorsal and anteroventral buds, respectively. The posterodorsal bud is the first to form, becoming morphologically apparent at 24 hours postfertilization (hpf); immunohistochemistry shows this bud is composed exclusively of endocrine pancreatic cell types. By 40 hpf, the anteroventral bud has formed on the ventral side of the gut tube; this second bud comprises primarily exocrine cells, although there are a few endocrine cells present. By 52 hpf, the two buds have merged (Field et al.,2003; schematized in Fig. 1) to form a single pancreatic islet of endocrine cells, surrounded by exocrine tissue, with the entire structure localized to the right side of the ventral midline.
Figure 1. Development of the pancreas in zebrafish and general fate map methodology. A: pdx1 is expressed bilaterally adjacent to the notochord of the anterior trunk at 14 hours postfertilization (hpf; based on Biemar et al.,2001). B–D: Morphogenesis of the pancreatic buds in zebrafish (based on Field et al.,2003). B: 24 hpf. The posterodorsal bud is visible. C: At 40 hpf. Both posterodorsal and anteroventral buds are visible, but have not fused. D: At 52 hpf. The buds have fused. E: One marginal blastomere was injected at the 512- to 2000-cell stage. F: Location of injected cells that gave rise to endoderm. Red circles were locations for the 40 hpf analysis, and blue squares were for the 72 hpf analysis. G–J: Embryos were imaged at various time points using brightfield and fluorescence, and photos were merged. G: At 6 hpf. H: At 10 hpf. I: At 24 hpf. J: Dorsal view of 72 hpf embryo showing rhodamine dextran-labeled cells in the exocrine and endocrine pancreas. Anti-Islet1 (blue) labels the endocrine pancreas and anti–green fluorescent protein (GFP; green) labels the postpharyngeal endoderm in gut-GFP embryos. DB, posterodorsal pancreatic bud; EP, exocrine pancreas; I, pancreatic islet; IB, intestinal bulb; L, liver; N, notochord; P, pancreas; S, embryonic shield; VB, anteroventral pancreatic bud.
Download figure to PowerPoint
Molecular markers reveal that pancreatic development is well under way at stages that precede bud morphogenesis. The earliest molecular marker of the zebrafish pancreas is the transcription factor Pdx1 (Biemar et al.,2001). Expression of Pdx1 in pancreatic progenitors is conserved in zebrafish, Xenopus, mouse, and chick; in all species Pdx1 labels both exocrine and endocrine pancreas progenitor cells, as well as part of the duodenum (Offield et al.,1996; reviewed in Edlund,2002). The zebrafish pdx1 gene is first expressed as early as the 10-somite stage (14 hpf), in bilateral patches of endoderm on either side of the midline (Biemar et al.,2001; schematized in Fig. 1A). Other pancreas cell type-specific markers begin to be expressed a short time later. The islet-1 gene is first expressed in the endoderm by the 12-somite stage (15 hpf), specifically in cells of the endocrine pancreas (Biemar et al.,2001), and at later stages in the mesenchyme surrounding the pancreas and liver (Ahlgren et al.,1997). Expression of insulin, a marker of differentiated β-cells, can also be detected from 15 hpf (Biemar et al.,2001; Kim et al.,2005). Like pdx1, insulin expression is initially detected in cells located on both sides of the midline; however, by 24 hpf, the insulin-expressing β-cells are localized to the posterodorsal bud together with other endocrine pancreas-specific cell types (Biemar et al.,2001; Field et al.,2003). In contrast, markers of the exocrine pancreas progenitors, mnr2 and ptf1a, cannot be detected until 24 hpf and 32 hpf, respectively (Wendik et al.,2004; Zecchin et al.,2004; Lin et al.,2004). In summary, expression of cell type-specific molecular markers reveals a differentiation program in which a subset of endodermal cells become specified as pancreatic progenitors and go on to differentiate into distinct endocrine and exocrine pancreatic cell types.
Fate maps can provide useful information about the location of progenitor cells at developmental stages that precede expression of cell type-specific molecular markers. Previous work has shown that fate maps for all three germ layers of the zebrafish are established by early gastrulation (approximately 6 hpf; Kimmel et al.,1990; Woo and Fraser,1995; Shih and Fraser,1995; Melby et al.,1996; Warga and Nüsslein-Volhard,1999; Vogeli et al.,2006). Fate maps are generated by labeling cells at early developmental stages and then determining what region or regions of the embryo the labeled cells contribute to at later developmental stages. Fate maps thus provide a probabilistic representation of where cells within a particular region are likely to be located at a later developmental stage, but do not provide any information on the state of commitment of the labeled cells. By contrast, tests of commitment require cell transplantation: cells can be termed committed if, when transplanted to an ectopic location, their differentiation program continues unperturbed. Experiments of this type have revealed that the majority of zebrafish endoderm cells are committed to an endodermal fate at 5 hpf, just as gastrulation begins (David and Rosa,2001). These committed endoderm cells express the Sox17 transcription factor, a general endodermal marker, which is essential for endoderm development (David and Rosa,2001; Reiter et al.,2001; Aoki et al.,2002). No experiments thus far have addressed when endodermal derivatives become committed to their specific fates.
Previously, a fate map of the zebrafish endoderm was generated by labeling cells in the blastula before gastrulation begins, establishing the location of their progeny in the late blastula at 5 hpf, before formation of germ layers, and again in the gastrula at 6 hpf, once the dorsal axis can be unambiguously identified, and then observing which endodermal organs these cells contributed to at 5 days post fertilization (Warga and Nüsslein-Volhard,1999). This fate map showed that just before the involution movements of gastrulation begin, all the endodermal precursor cells are located within four cell diameters of the blastoderm margin. Once gastrulation commences, these endoderm progenitor cells are the earliest cell population to internalize. The fate map further showed that the dorsoventral position of endoderm precursors in the early gastrula correlates approximately with their anteroposterior position in the 5 day old gut tube. For example, cells close to the dorsal organizer become pharyngeal endoderm, an anterior derivative, whereas cells from the ventral-most part of the gastrula become intestinal endoderm, a posterior derivative. This study also provided some limited evidence suggesting that progenitors of the asymmetric foregut organs, the pancreas and liver, may also be asymmetrically organized in the gastrula. Warga and Nüsslein-Volhard's (1999) data suggested that liver precursors derive from a dorsal region on the right side of the gastrula and a ventral region on the left, while the pancreas derives from dorsal regions on both sides, as well as a ventral region on the right side of the gastrula. These results raise the possibility that, already by the gastrula stage, some endodermal progenitors are asymmetrically localized.
Most vertebrate fate maps that include endoderm have investigated broad general divisions of its derivatives from anterior to posterior along the gut (reviewed in Fukuda and Kikuchi,2005). However, a fate map of the mouse liver and ventral foregut has been established (Tremblay and Zaret,2005), and a fate map of the tadpole gut has revealed that the progenitors of Xenopus dorsal and ventral pancreatic buds are segregated as early as the neurula stage (stage 14; Chalmers and Slack,2000). The former zebrafish fate map (Warga and Nüsslein-Volhard,1999) did not supply extensive information on where zebrafish pancreatic progenitors originate within the undifferentiated early endoderm, nor did it address whether endocrine and exocrine pancreas cell types, or posterodorsal and anteroventral pancreatic buds, derive from the same or separate pools of endodermal progenitors. To address these questions, we have reinvestigated where progenitors of the pancreas originate from, and how these cells move, to produce the differentiated cell types. Only with this knowledge in hand can we hope to properly understand the inductive interactions critical to pancreas cell type specification.
In this study, we establish a more complete early gastrula fate map of the zebrafish pancreas. We compare the locations of pancreas and liver progenitors, of posterodorsal and anteroventral pancreatic bud progenitors, and of endocrine and exocrine pancreatic cell progenitors and find that liver and pancreas progenitors do not show strict asymmetric localization at 6 hpf as previously suggested. Our extended fate map also reveals that progenitors of the posterodorsal and anteroventral pancreatic buds are located in different yet overlapping domains. Endocrine and exocrine pancreatic progenitors are located in similar domains to the posterodorsal and anteroventral pancreatic buds, respectively, consistent with previous observations that the posterodorsal bud is exclusively endocrine and the anteroventral bud primarily exocrine. Our results suggest that endocrine progenitors derive from a population of gastrula stage cells that is, on average, located more dorsally than exocrine progenitors. We also determine the location of pancreatic progenitors in the late gastrula (10 hpf), as our previous studies demonstrated that this is an important developmental stage when retinoic acid signals are received by the endoderm to specify pancreas fate (Stafford and Prince,2002; Stafford et al.,2006). We find that pancreas progenitors move significant distances during gastrulation to converge bilaterally on either side of the midline at the anterior of the trunk by 10 hpf. In summary, our studies indicate that, at the onset of gastrulation, there are already biases in the fates of pancreatic precursors, with respect to bud and associated histotype, dependent on where they originate in the gastrula. Moreover, by the end of gastrulation, pancreatic precursors are found to populate a unique equatorial latitude of the endodermal layer.
- Top of page
- EXPERIMENTAL PROCEDURES
This study expands on a previous fate map of the endoderm generated by one of us (R.M.W.) and Nüsslein-Volhard (1999), by specifically addressing where progenitors of the pancreas originate in the gastrula of the zebrafish embryo. The earlier study showed that a cell labeled at the blastula margin of the 1,000-cell stage embryo has a high probability of giving rise to endoderm. Following a similar approach, we labeled single marginal blastomeres of gut-GFP embryos at the 512- to 2,000-cell stage with 70-kD lysinated rhodamine dextran (Fig. 1E) and localized the labeled progeny relative to the embryonic shield (dorsal organizer) of the early gastrula to determine the clone's position relative to dorsoventral (Fig. 1G). To simplify analysis and data presentation, we divided the embryo into quadrants: left-dorsal (1 degree–90 degrees), left-ventral (91 degrees–180 degrees), right-ventral (181 degrees–270 degrees), and right-dorsal (271 degrees–360 degrees). In each experimental embryo, we detected 2–16 rhodamine dextran-labeled cells at 6 hpf. The positions shown in the polar plots depicting our fate map data (e.g., Fig. 1F) indicate the mean location along the blastoderm margin of each clone of labeled cells.
Labeled cells were localized again at tail bud (10 hpf: Fig. 1H) and at 24 hpf (Fig. 1I), and their final locations and fate were determined either at 40 hpf, a stage immediately before fusion of posterodorsal and anteroventral pancreatic buds, or at 72 hpf, when marker analysis allows exocrine and endocrine pancreas fates to be distinguished (Fig. 1J). A total of 360 embryos were labeled, 125 of which had labeled cells in endodermal derivatives at later stages (40 hpf, 57 embryos; 72 hpf, 68 embryos; Fig. 1F). Consistent with the previous report (Warga and Nüsslein-Volhard,1999), not all of these 125 clones were germ layer restricted to the endoderm: some also contributed to mesodermal derivatives, and in almost all cases labeled blastomeres also contributed to the extraembryonic enveloping layer (data not shown). Statistical analyses confirmed that the locations of the endoderm progenitor populations were randomly distributed along the margin at 6 hpf (see Experimental Procedures section), again consistent with previous fate maps using this methodology (Warga and Nüsslein-Volhard,1999; Kimmel et al.,1990). As predicted by the Warga and Nüsslein-Volhard endoderm fate map (1999), we found a positive correlation between a labeled clone's dorsoventral location at 6 hpf and its later anteroposterior position in the differentiating gut (Tables 1–4). Because the liver and pancreas are a relatively small portion of the gut, only 57 of the 125 endodermal clones contributed labeled cells to liver and/or pancreatic tissues (Tables 3, 4). These 57 clones form the basis for the remainder of our analysis.
Table 1. Average Dorsoventral Positions of Clones Contributing to Specific Endodermally Derived Tissues at 40 hpfa
| ||Pharynx||Esophagus||Swim bladder||Posterodorsal bud||Anteroventral bud||Liver||Intestinal bulbb||Intestine|
Table 2. Average Dorsoventral Positions of Clones Contributing to Specific Endodermally Derived Tissues at 72 hpfa
| ||Pharynx||Esophagus||Endocrine pancreas||Swim bladder||Exocrine pancreas||Liver||Intestinal bulb*||Intestine|
Table 3. Relationships of Cellular Contributions to Different Endodermal Tissues Based on the 40 hpf Analysisa
| ||Pharynx||Esophagus||Swim bladder||Posterodorsal bud||Anteroventral bud||Intestinal bulb||Liver||Intestine|
Table 4. Relationships of Cellular Contributions to Different Endodermal Tissues Based on the 72 hpf Analysisa
| ||Pharynx||Esophagus||Endocrine pancreas||Swim bladder||Exocrine pancreas||Liver||Intestinal bulb||Intestine|
Fate Map of the Pancreas and Liver in Zebrafish
To determine the location of progenitors of the pancreas and the liver, we combined our analyses at 40 and 72 hpf (Fig. 2A). We found that progenitor cells contributing to either organ could derive from any of the four early gastrula stage quadrants (Fig. 2A). However, clones that contributed to the liver but not the pancreas (“liver-only” clones, n = 15; Fig. 2B,E) were more likely to lie within the ventral half of the gastrula (Fig. 2A). By contrast, clones that contributed to the pancreas but not the liver (“pancreas-only” clones, n = 19; Fig. 2C,F) tended to lie within the dorsal half of the gastrula (Fig. 2A). These liver-only progenitors were located significantly further from the dorsal organizer than the pancreas-only progenitors (Mann–Whitney U Test, P = 0.01). Consistent with these data the average location of all clones that contributed to the liver was more ventral than the average location of all clones that contributed to the pancreas (Tables 1, 2). Clones that contributed to both pancreas and liver (n = 23; Fig. 2D,G) tended to be located in more intermediate positions (Fig. 2A). We found a slight bias for both pancreas and liver clones to come from the left-dorsal quadrant, although this is largely explained by the increased number of endodermal clones labeled in that quadrant (Fig. 1F): 54% of the clones that gave rise to endoderm in the left-dorsal quadrant gave rise to pancreatic or liver tissues as compared to 39% for the left-ventral, 42% for the right-ventral, and 45% for the right-dorsal. These data do not support the hypothesis of a robust left/right asymmetrical distribution of progenitors. In summary, while a single labeled blastomere can contribute to both liver and pancreas, there is nevertheless bias for more dorsally located endoderm cells in the early gastrula to contribute to the pancreas, and more ventrally located endoderm cells to contribute to the liver.
Figure 2. Fate map of liver and pancreatic progenitors at 6 hours postfertilization (hpf). A: Polar plot showing location at 6 hpf of progenitors of the liver and pancreas. The embryos from the 40 hpf and 72 hpf analyses were combined. Gray squares are pancreatic progenitors, black circles are liver progenitors, and white triangles indicate a mixed liver and pancreas population. B–G: All confocal images are individual slices from a Z-stack. B: Embryo with labeled cells in the liver bud at 40 hpf. C: Embryo with labeled cells in the pancreatic bud. D: Embryo with labeled cells in the liver and pancreatic buds at 40 hpf. E: Example of embryo with labeled cells only in the liver at 72 hpf. F: Embryo with labeled cells in the pancreas at 72 hpf. G: Embryo with rhodamine dextran-labeled cells in the liver and the pancreas at 72 hpf. Arrows point to labeled cells that colocalize with the structure of interest. Asterisks (*) point to rhodamine dextran-labeled cells that are not in the pancreas or liver. Dashed lines outline the anteroventral bud and solid lines outline the posterodorsal bud. L, liver; P, pancreas
Download figure to PowerPoint
Fate Map of the Posterodorsal and Anteroventral Pancreatic Buds
To determine whether progenitor cells for the posterodorsal and anteroventral buds derive from specific locations in the gastrula, we analyzed our labeled clones at a stage in development when both pancreatic buds are visible but not fused (40 hpf). We found that individual clones of cells contributed to the posterodorsal bud, anteroventral bud, or frequently both (Fig. 3B–D; Table 3). Table 3 also indicates that while many of the clones contributed to two or more tissue types, these contributions tended to be limited to tissues that are closely juxtaposed. In general, the majority of clones that contributed to the posterodorsal but not anteroventral bud originated from both dorsal quadrants (9/10 clones, Fig. 3A). By contrast, clones that contributed to the anteroventral but not posterodorsal bud (n = 5) tended to originate from more ventral locations. Clones that contributed only to the posterodorsal bud were significantly closer to the dorsal organizer than clones that contributed only to the anteroventral bud (Mann-Whitney U test, P = 0.029). Clones of cells that contributed to both buds (n = 5) tended to be located in intermediate positions between these domains (Fig. 3A). We conclude that, although a single labeled clone can contribute to both pancreatic buds, there is nevertheless some bias by early gastrula stage for more dorsally located endoderm cells to give rise to posterodorsal bud derivatives, and more ventrally located endoderm cells to give rise to anteroventral bud derivatives.
Figure 3. Fate map of the posterodorsal and anteroventral pancreatic buds. A: Polar plot of the locations of the posterodorsal and anteroventral pancreatic bud progenitors at 6 hours postfertilization (hpf). These points are the pancreas positive locations seen in Figure 2A. B–D: Confocal slices of representative 40 hpf embryos. B: Example of embryo with labeled cells in the anteroventral bud (purple circles in A). C: Embryo with rhodamine dextran-labeled cells in the posterodorsal bud (yellow squares in A). D: Embryo with labeled cells in both buds (blue triangles in A). Arrows point to labeled cells that colocalize with the structure of interest. Asterisks (*) point to rhodamine dextran-labeled cells that are not in the pancreas or liver. The dashed line encircles the location of the ventral bud, and the solid line encircles the location of the dorsal bud.
Download figure to PowerPoint
Fate Map of the Exocrine and Endocrine Pancreas
The pancreas is a composite organ consisting of both exocrine and endocrine cells. We tested whether progenitors of endocrine vs. exocrine cells derive from the same respective regions of the 6 hpf fate map as posterodorsal vs. anteroventral bud cells. For this study, we raised labeled embryos to 72 hpf, when the pancreatic buds have merged and enlarged. At this stage, we could easily distinguish endocrine cells from exocrine by monitoring endocrine-specific expression of islet1 (Fig. 4C–E). Data from 72 hpf, shown in Figure 2A, was re-analyzed by excluding clones that contributed to liver but not pancreas and asking what proportion of the remaining clones contributed to endocrine, exocrine, or combined endocrine and exocrine pancreatic cell types (Fig. 4A; Table 4).
Figure 4. Fate map of the endocrine and exocrine pancreas in zebrafish at 6 hours postfertilization (hpf). A: Polar plot of the locations of endocrine and exocrine pancreatic progenitors at mid-gastrulation. B: Graph of percentage exocrine cells at 72 hpf vs. dorsoventral location of the clone at 6 hpf. The numbers of exocrine and endocrine cells were counted, and the percentage exocrine is the number of exocrine cells divided by the total number of cells. C–E: Confocal slices of representative 72 hpf embryos. C: Embryo with rhodamine dextran-labeled cells in the exocrine pancreas. D: Embryo with labeled cells in the endocrine pancreas. E: Embryo with labeled cells in both the endocrine and exocrine pancreas. Blue cells are islet1 positive. Arrows point to labeled cells that colocalize with the structure of interest. Asterisks (*) point to rhodamine dextran-labeled cells that are not in the pancreas or liver. EP, exocrine pancreas; I, pancreatic islet; IB, intestinal bulb; L, liver.
Download figure to PowerPoint
Consistent with the 40 hpf fate map data, we found that clones that gave rise to endocrine and/or exocrine pancreas derived from all four quadrants of the gastrula (n = 22; Fig. 4). The clones that contributed to the endocrine but not exocrine pancreas (endocrine-only; n = 3) were located in the dorsal half of the embryo at 6 hpf, while the majority of clones that contributed to the exocrine but not endocrine pancreas (exocrine-only clones; 5/6) were located in the ventral half of the embryo (Fig. 4A,B). The more dorsal location of endocrine-only clones relative to exocrine-only clones was statistically significant as determined by the Mann–Whitney U test (P = 0.024). Notably, we found no bias toward right or left side. We also found that clones that contributed to both exocrine and endocrine pancreas (mixed clones; n = 13) were scattered randomly around the margin (Rayleigh test for uniform circularity, P > 0.10; Fig. 4A). However, further examination revealed that mixed clones that gave rise to a low percentage of exocrine cells were located more dorsally, whereas mixed clones that gave rise to a high percentage of exocrine cells were located more ventrally (Fig. 4B). We conclude that, like the bias at 6 hpf for posterodorsal vs. anteroventral pancreatic bud, there is a correlated bias for endocrine vs. exocrine pancreatic cell types. In general, cells located more dorsally within the early gastrula have a higher probability of giving rise to endocrine cell types of the posterodorsal bud, whereas cells located more ventrally have a higher probability of giving rise to exocrine cell types of the anteroventral bud (compare Fig. 3A with 4A).
Location of Pancreatic Progenitors at Tail Bud
In addition to determining the location of the pancreas progenitors at 6 hpf, we asked where they are located at 10 hpf, when retinoic acid is required to specify their fate (Stafford and Prince,2002). At this developmental stage, the embryo completely encompasses the yolk cell, and expression of sox17 (Alexander and Stainier,1999) reveals that the endodermal cells are distributed extensively over the yolk cell in broad bilateral swaths extending dorsolaterally from the midline (Fig. 5A,B). Examining embryos in dorsal view at 10 hpf (tail bud stage), and then again at 40 hpf, allowed us to determine the likely position of the pancreatic progenitors at 10 hpf. We did this by superimposing all the individual cells from multiple clones that gave rise to pancreas (Fig. 5C; 20 embryos). For each clone, the labeled cells were pseudocolored with a unique color, revealing a region that included cells of every color (gray square in Fig. 5C). This region should represent the common domain where pancreatic progenitors are located at 10 hpf.
Figure 5. Location of cells at tail bud that potentially give rise to the pancreas. A: Lateral view of sox17 expression at tail bud. B: Dorsal view of sox17 expression at tail bud. n: notochord. C: Locations of all cells (excluding cells in the extreme anterior and posterior) from 20 different embryos that had one or more labeled cells in one or both of the pancreatic buds. Each color represents a different embryo. Embryos were mounted in dorsal view with the equator of the embryo (approximately the level of the first somite) closest to the viewer. The gray square marks an area that includes labeled cells from all embryos and marks the likely location of the pancreatic progenitors at the end of gastrulation (tail bud stage, 10 hours postfertilization [hpf]). The outline of the notochord is indicated by dashed lines. D–J: Representative confocal projections from time-lapse experiment between 95% epiboly and 14 somites (approximately 9.5 hpf–16 hpf). Images are approximately equally spaced in time. Red circle contains cells that later colocalize with pdx1–green fluorescent protein (GFP) positive cells. K: Location of rhodamine dextran-labeled cells (red) at 16 hpf with cells that are positive for GFP (green). The dashed yellow line is the future division between somites 1 and 2. The solid yellow line marks the boundary between somites 1 and 2. The solid white line marks the notochord.
Download figure to PowerPoint
To test the hypothesis that the equatorial domain indicated in Figure 5C does indeed include pancreatic progenitors, we performed time-lapse video microscopy of labeled clones in pdx1-GFP transgenic fish (Huang et al.,2001). Labeled endoderm cells were followed from 10 hpf until approximately 16 hpf, after onset of pdx1-GFP expression (Biemar et al.,2001) in the pancreatic domain. By tracing backward in time, we were able to determine that cells which contributed to the Pdx-GFP–labeled pancreas were located at approximately the level of the first somite when it formed (Fig. 5D–G). These cells then moved medially toward the notochord, and subsequently moved posteriorly to the level of the second somite (Fig. 5D–J). Later they could be seen in the pancreas (Fig. 5K). Thus, cells contributing to the pancreas are located in the endodermal layer at approximately the equatorial region of the 10 hpf embryo.