Homeoprotein hhex-induced conversion of intestinal to ventral pancreatic precursors results in the formation of giant pancreata in Xenopus embryos

Abstract

Liver and ventral pancreas develop from neighboring territories within the endoderm of gastrulae. ventral pancreatic precursor 1 (vpp1) is a marker gene that is differentially expressed in a cell population within the dorsal endoderm in a pattern partially overlapping with that of hematopoietically expressed homeobox (hhex) during gastrulation. In tail bud embryos, vpp1 expression specifically demarcates two ventral pancreatic buds, whereas hhex expression is mainly restricted to the liver diverticulum. Ectopic expression of a critical dose of hhex led to a greatly enlarged vpp1-positive domain and, subsequently, to the formation of giant ventral pancreata, putatively by conversion of intestinal to ventral pancreatic precursor cells. Conversely, antisense morpholino oligonucleotide-mediated knockdown of hhex resulted in a down-regulation of vpp1 expression and a specific loss of the ventral pancreas. Furthermore, titration of hhex with a dexamethasone-inducible hhex-VP16GR fusion construct suggested that endogenous hhex activity during gastrulation is essential for the formation of ventral pancreatic progenitor cells. These observations suggest that, beyond its role in liver development, hhex controls specification of a vpp1-positive endodermal cell population during gastrulation that is required for the formation of the ventral pancreas.

The endoderm will give rise to the digestive and respiratory tracts with their associated organs, such as the liver, pancreas, intestine, stomach, and lungs. In mice, it has been proposed that the initial regionalization of the definitive endoderm occurs concurrently with its formation as it exits the primitive streak (1). In the endodermal germ layer of mouse and various other vertebrate species, including zebrafish and Xenopus, the first regional differences can be detected in the early gastrula stage via the highly restricted expression of the divergent homeobox gene hematopoietically expressed homeobox (Hhex) in the anterior-most, migrating mesendoderm (2–7). By the end of gastrulation, an initial anteroposterior pattern within the endoderm has been established, as reflected by the anterior expression of Hhex, Sox2, and Foxa2 and the posterior expression of the caudal type homeobox genes Cdx1, Cdx2, and Cdx4 (8).

The pancreas is one of the organs that are formed from the anterior endoderm. In amphibians and higher vertebrates, the pancreas initially emerges as one dorsal and two ventral pancreatic buds from the gut endoderm epithelium; later on, these primary anlagen fuse to form one organ (9). Consistent with their distant locations, the dorsal and ventral pancreatic buds are specified by different regulatory molecules, as revealed by gene targeting studies in mice (10). Furthermore, fate mapping studies in zebrafish, Xenopus, and mice have indicated that the ventral pancreatic and liver progenitor cells display a partially intermingled allocation in the ventral foregut endoderm of early neurula stage embryos (11–15). It remains to be defined whether and how these two different types of progenitor cells can be distinguished from each other during gastrula stages of development.

Formation of liver and ventral pancreas is under the control of bone morphogenetic protein and fibroblast growth factor pathway signals produced in the septum transversum mesenchyme and cardiac mesoderm, respectively (16). The homeodomain transcription factor Hhex serves key regulatory functions both in the context of liver and ventral pancreas development (17). In Hhex-null mouse embryos, the hepatic diverticulum is normally specified within the early ventral foregut endoderm (18, 19); however, proliferation of these cells in the liver diverticulum is inhibited, leading to morphogenetic and movement defects. Thus, the prospective ventral pancreatic endoderm domain just next to the liver diverticulum fails to move beyond the cardiogenic mesoderm domain that induces the liver but inhibits the pancreatic fate. As a consequence, the ventral pancreas is absent in Hhex-null mouse embryos (20). Further chimeric and conditional knockout studies have revealed that Hhex has additional roles in the context of hepatobiliary development (21, 22).

In this study, we report on the identification of a marker gene, ventral pancreatic precursor1 (vpp1), that specifically demarcates a population of endodermal cells that we suggest are precursors of the ventral pancreas as early as gastrulation. In hhex knockdown Xenopus embryos, we found an early reduction of vpp1-positive cells and later a corresponding loss of ventral pancreas at tail bud and tadpole stages of development. Furthermore, overexpression of a low dose of hhex was sufficient to convert intestinal precursor cells into vpp1-positive cells that correlated with the formation of giant ventral pancreata in these tadpoles. Thus, use of the vpp1 marker gene has allowed us to characterize a population of cells in the anterior endoderm that is responsive to levels of hhex activity and is likely to represent the precursor cells that give rise to the ventral pancreas, suggesting an active role of hhex in ventral pancreas specification.

Results

Vpp1 Expression Identifies a Distinct Population of Endodermal Cells in the Ventral Pancreas-Forming Region of Gastrula Stage Xenopus Embryos.

In the context of a genome-wide study for regional-specific gene expression in Xenopus gastrulae (23), we identified a novel dorsal endoderm-specific marker gene which, by virtue of its primary structure, belongs to the Ig protein superfamily. We were unable to identify homologous genes in other vertebrate species in the GenBank database. Based on its specific expression in the ventral pancreatic buds at tail bud stages (Fig. 1A, 19 and 21), we named this gene ventral pancreatic precursor 1 (vpp1). Two alternative splicing isoforms were detected with the shorter version being predominantly expressed (Fig. S1). The biological function of vpp1 remains unknown. Microinjection of either vpp1 mRNA or antisense morpholino oligos did not reveal any obvious phenotypic effects (Fig. S2). Nevertheless, vpp1 proved to be a valuable tool to study endoderm regionalization.

Fig. 1.

vpp1 and hhex display partially overlapping expression in gastrulating dorsal endoderm and later demarcate ventral pancreatic buds and liver anlagen, respectively. (A) Comparison of vpp1 and hhex expression as revealed by whole mount in situ hybridization analysis with bisected (stages 11–18) or whole (stages 24–33) embryos. (1, 2, 5, 6, 9, 10, 13, and 14) Dorsal is toward the right; (17–22) lateral view, head toward the left. Images 3, 7, 11, and 15 are cartoons drawn representing both the single staining (images 1, 2, 5, 6, 9, 10, 13, and 14) and double staining whole mount in situ hybridization data (Fig. S3B), illustrating vpp1 (purple) and hhex (turquoise) expression. Boxed regions are magnified in 4, 8, 12, and 16, respectively. (B) RT-PCR analysis compares the temporal expression profile of vpp1 and hhex during Xenopus embryogenesis. Ornithine decarboxylase (odc) was used as the RNA loading control.

At the onset of gastrulation, vpp1 mRNA is first detected in a subset of dorsal endodermal cells that is partially overlapping with the territory of hhex-positive cells (Fig. 1A, 1–4). Hhex is first expressed in the leading edge of invaginating endomesodermal cells during gastrulation and, during later stages, demarcates the liver anlage (3–5). In full agreement with these earlier studies, our whole mount in situ hybridization data further indicated that, in early gastrulae, hhex exhibited a graded expression pattern with the highest levels in the dorsal-most endomesoderm (Fig. 1A, 2); it appears that vpp1-positive cells are located in the subblastoporal endodermal region of relatively low-level hhex expression but not in the territory of high-level hhex expression (Fig. 1A, 3 and 4). RT-PCR analysis revealed that zygotic expression of hhex was initiated earlier (stage 8.5) than that of vpp1 (stage 10) (Fig. 1B).

As gastrulation proceeds, the archenteron separates the vpp1-positive and hhex-positive cells, with hhex-expressing cells lining the most anterior wall and the anterior archenteron roof and vpp1-expressing cells lining the anterior archenteron floor (Fig. 1A, 5–12). In consequence, during neurulation (stage 14–22), two separate cell populations can be distinguished in the anterior ventral foregut endoderm: a more anterio-dorsal domain expressing hhex and a directly adjacent postero-ventral group of cells expressing vpp1 (Fig. 1A, 13–16). The overlapping expression domain that is relatively broad at stage 11 (Fig. 1A, 1–4) became smaller but persisted throughout early development (Fig. 1A and Fig. S3B).

With rapid elongation of the embryos and with the formation of the liver diverticulum during the transition from neurula to tail bud stages of development, vpp1 expression became specific for the ventral pancreas, whereas hhex expression was then restricted to the liver. A small number of cells coexpressing both vpp1 and hhex remained to be detected (Fig. 1A, 17–22 and Fig. S3C). Vpp1 transcripts became hardly detectable at or beyond stage 35 in both whole mount in situ hybridization and RT-PCR analysis (Figs. 1B and 2N).

Fig. 2.

The expression of vpp1 is regulated by hhex. Overexpression of hhex-induced ectopic expression of vpp1 (A, D, G, and J), and hhex knockdown led to the down-regulation of vpp1 expression (C, F, I, and L) compared with controls (B, E, H, and K). (A–I) Bisected embryos: A–C, dorsal toward the right; D–F, anterior toward the left; G–I, anterior toward the top; each image displays twin halves dissected from the same embryo. (J–O) Lateral view, head toward the left. Phenotype shown are as follows: A, 34/34; C, 47/48; D, 28/32; F, 30/30; G, 25/30; I, 20/20; J, 15/15; L, 28/28; M, 26/28; O, 31/31.

In summary, these observations suggest the existence of distinct liver- and ventral pancreas-forming endodermal cell populations that can be distinguished as early as gastrulation. Analysis of vpp1 and hhex gene expression facilitated the following corresponding morphogenetic processes.

Ectopic Expression of hhex Results in an Expansion of the Endodermal vpp1-Positive Cell Population.

To test a possible influence of hhex on the formation and behavior of vpp1-expressing cells, we injected four-cell stage Xenopus embryos with increasing doses of hhex mRNA (12.5, 25, and 50 pg per embryo) into either the two dorsal or the two ventral blastomeres from the vegetal pole. The expression of vpp1 was then examined by whole mount in situ hybridization. We observed that dorsal injection with a critical dose of hhex (25 pg of mRNA per embryo) resulted in the induction of massive ectopic vpp1 expression in the dorsal endoderm (Fig. 2 A, D, G, and J). A lower dose (12.5 pg) had no obvious effect on vpp1 expression, whereas a higher dose (50 pg) caused severe retardation of gastrulation. Similar to the temporal expression profile of endogenous vpp1, the induced ectopic vpp1 expression persisted up to stage 34 and was then down-regulated (Fig. 2). Conversely, a dramatic reduction of vpp1 expression was observed in hhex MO-injected embryos (hhex morphants) (Fig. 2 C, F, I, and L), and the very little vpp1 signals retained in hhex morphants did not overlap with hhex expression (Fig. S4); hhex MO has been demonstrated to specifically block hhex translation (24, 25). Thus, hhex is required for vpp1 expression in Xenopus embryos.

To analyze whether the hhex-induced supernumerary vpp1-positive cells (Fig. 3 A–C) are the result of cell fate conversion or of increased proliferation of the endogenous vpp1-expressing cells, we first analyzed effects of ectopic hhex on the expression of darmin, an intestine specific marker (26). We found that darmin expression was significantly reduced in hhex-injected embryos at neurula stage (Fig. 3 D–F). Next, we analyzed effects of hhex injection on cell proliferation. Whole mount immunostaining for phospho-histone H3 revealed no increase in cell proliferation upon hhex overexpression (Fig. 3 J–L). Finally, activation of the endogenous hhex gene was not observed (Fig. 3 G–I), ruling out the possibility that ectopic hhex protein expression activates an autoregulatory loop, thereby leading to persistent hhex gene activation in these embryos. Taken together, these observations are compatible with the idea that a low dose of hhex is sufficient to convert intestine-forming endodermal cells into vpp1-positive cells.

Fig. 3.

Ectopic hhex converts intestine-forming cells into vpp1-expressing cells. Whole mount staining was performed with probes indicated on the left. (A–K) Bisected embryos, anterior toward the left. Phenotype is as follows: B, 28/32; E, 34/40; H, 21/21; K, 42/42. (L) Statistic analysis of phospho-histone H3-positive cells on the whole surface of bisected vegetal endoderm, as demarcated by the red dashed lines in J and K. As the distribution of phospho-histone H3-positive cells across the whole vegetal endoderm surface did not show obvious bias, for simplicity, we counted the whole area rather than only putative vpp1-positive area. Forty-two bisected samples were counted from either control or hhex-injected embryos. P value (0.156) was determined by the paired, two-tailed t test.

hhex-Induced vpp1-Positive Cells Correlate with the Generation of Ectopic Ventral Pancreatic Progenitor Cells in Xenopus Embryos.

To address the consequence of the hhex-induced conversion of intestine-forming cells to vpp1-positive cells on late foregut organogenesis, we first examined the expression of the pancreatic precursor cell marker genes XlHbox8 and Ptf1a/p48 and the earliest differentiation marker insulin in hhex-injected stage 36 embryos. External examination of hhex-injected embryos revealed an enlargement of the ventral foregut correlated with a slight reduction of the midgut (Fig. 4 B and D).

Fig. 4.

Ectopic hhex-induced ventral pancreatic progenitors undergo normal endocrine and exocrine differentiation. Whole mount staining was performed with molecular probes indicated on the left. (A–L) Lateral view, head toward the left. Phenotype is as follows: B, 18/20; D, 40/40; F, 21/21; H, 31/31; J, 19/19; L, 32/32. (M and N) Right side views of the same embryos in K and L, respectively, showing that in the control embryo, because of the unilateral location of pancreas, stomach, and duodenum on the left side of the embryo at this developmental stage, fibrinogen stains a continuous area between the intestine and the liver bud. In contrast, in hhex-injected embryos, the anterior, fibrinogen-positive area is interrupted by the ectopically formed ventral pancreatic cells. (O and P) Ventral view, head toward the top. Phenotype is as follows: P, 51/51. (Q and R) Pancreata isolated from elastase:GFP transgenic Xenopus tadpoles (stage 47) with or without hhex overexpression. They were kept in the same Petri dish and photographed together in the same visual field and magnification. The dashed white line was added for a better alignment. Phenotype is as follows: R, 32/32. (S–U) Whole mount in situ hybridization of insulin on isolated liver and annular pancreata. Red asterisks highlight liver lobes. U shows the other side of the same liver and pancreata in T. The darker structure above the red asterisk in U is the deformed gall bladder. Note that there are two liver lobes in control embryos, but only one liver lobe in hhex-injected embryos (100%, n = 11).

XlHbox8 expression demarcates dorsal and ventral pancreatic buds, and part of stomach and duodenum at tail bud stages of development (27, 28). The enlargement of the ventral foregut induced by ectopic hhex coincides with a specific expansion of ventral XlHbox8 expression domain (Fig. 4 C and D). This expansion was resistant to the treatment with BMS453, a retinoic acid antagonist that specifically inhibits dorsal pancreas formation with minor effect on the ventral pancreas (29), supporting the idea that these cells have a ventral pancreatic identity (Fig. 4 C–F). Concomitant with the expansion of XlHbox8 expression, the anterior expression domain of darmin was found to be severely reduced, whereas the more posterior domain remained unaltered in hhex-injected embryos (Fig. 4 A and B), which spatially correlated with the activation of vpp1 and the down-regulation of darmin expression at early neurula stages after hhex injection. Within the endoderm, expression of XPtf1a/p48 is specifically restricted to the dorsal and ventral pancreatic buds (28, 30). As for XlHbox8, the ventral expression domain of XPtf1a/p48 was found to be expanded in hhex-injected embryos, although to a lesser extent (Fig. 4 G and H). Insulin expression is first detected in the dorsal pancreas by stage 32 (31). Overexpression of hhex did not affect insulin expression (Fig. 4 I and J), providing further support for the notion that the effects observed upon hhex overexpression are specific to the ventral pancreatic anlage. Further, fibrinogen expression in the liver anlage remained unaffected in hhex-injected tail bud stage embryos (Fig. 4 K–N). Taken together, these findings suggest that the hhex-induced supernumerary vpp1-positive cells contribute to an expanded population of ventral pancreatic precursor cells.

Hhex-Induced Expansion of the Ventral Pancreas Is Associated with Normal Exocrine and Secondary Endocrine Differentiation.

To test whether the hhex-induced ventral pancreatic progenitor cells undergo further pancreatic differentiation, hhex-injected embryos were raised to late tadpole stages (stage 43–48) and analyzed for the expression of XPDIp, an exocrine-specific differentiation marker (32), and for insulin. Staining of stage 43 Xenopus embryos for XPDIp by whole mount in situ hybridization revealed a rather dramatic expansion of the exocrine pancreatic structure upon hhex overexpression (Fig. 4 O and P). The same effect can also be visualized by using stage 47 embryos in elastase:GFP transgenic frogs generated in our laboratory (33); the number of GFP-positive cells, reflecting activation of exocrine-specific elastase, was several times higher in the hhex-induced giant ventral pancreata than in the pancreas of control embryos (Fig. 4 Q and R).

We further asked whether the second wave of insulin expression takes place in the enlarged, hhex-induced pancreata. The number and distribution of the scattered insulin-expressing cells in the exocrine cell mass of hhex overexpressing embryos was similar to that seen in the control pancreas, but the mass of tissue was increased (Fig. 4 S–U and Fig. S5). Another endocrine differentiation marker, glucagon, showed similar expression in control and hhex-induced pancreata (Fig. S6). Annular pancreas was frequently observed upon hhex overexpression (Fig. 4 T and U). It would be interesting to determine whether human pathological annular pancreas resulted from abnormal ventral pancreas development is linked to ectopic HHEX activity. It should be pointed out that, although it is difficult to quantify the liver mass, instead of two liver lobes seen in control embryos, only one lobe of liver was observed in hhex-injected embryos (Fig. 4 S–U). The mechanisms involved in the blockage of liver lobing by hhex overexpression remain to be investigated.

Knockdown of hhex in Xenopus Embryos Leads to an Early Reduction of vpp1-Positive Cells and the Loss of Ventral Pancreas.

Although largely unaffected upon hhex overexpression, endogenous hhex expression appears to be slightly up-regulated in hhex morphants at all developmental stages analyzed (Fig. 5 A–H), which is reminiscent of the situation described for embryonic day (E)8.5 Hhex^−/− mouse embryos, where the initial formation of the liver diverticulum was not blocked (19). In further support of this notion in frog embryos, the early endodermal expression of foxa2 (34), which largely overlaps with hhex expression, remains unaltered upon hhex MO injection (Fig. 5 I and J). In contrast, the liver expression of differentiation marker genes, such as for1 (Fig. 5L) (35) and fibrinogen (Fig. 5N), became undetectable in hhex morphants. The identity of these hhex-expressing cells in late-stage hhex morphants remains to be defined. In line with the reduction of vpp1 expression in hhex morphants, the loss of XlHbox8 and XPtf1a/p48 expression was specifically restricted to the ventral pancreas (Fig. 5 O–R and Fig. S7). The complete loss of XlHbox8 expression in hhex morphants reported in a previous study (25) was not observed in our experiments. Coinjection of vpp1 mRNA with hhex MO did not rescue the loss of XPtf1a/p48 expression in the ventral pancreatic buds (Fig. S8). The liver and ventral pancreas defects observed here in Xenopus hhex morphants resemble the phenotype seen in Hhex knockout mice (18–20). We were unable to detect obvious alteration of either cell proliferation rate or apoptosis upon hhex knockdown (Fig. 5 S–W). Taken together, our data suggest that the early reduction of the cells marked by vpp1 likely accounts for the loss of ventral pancreas upon hhex knockdown in Xenopus embryos.

Fig. 5.

Liver and ventral pancreas defects in hhex morphants. Probes or methods for whole mount staining are indicated on the left. (A and B) Bisected embryos, dorsal toward top. (C–H) Lateral view, head toward the left. (I and J) Bisected embryos, anterior toward the left. (K–R) Lateral view, head toward the left. The white arrowhead (N) indicates the missing expression of fibrinogen in the liver anlage. (S–V) Bisected embryos, anterior toward the left. Phenotype is as follows: B, 19/20; D, 20/22; F, 24/25; H, 34/41; J, 25/27; L, 38/44; N, 19/20; P, 36/42; R, 16/72 (for detailed phenotype analysis, see Fig. S7); T, 22/23; V, 41/41. (W) Statistical analysis of phospho-histone H3-positive cells in the outlined areas in U and V. Thirty-four samples from 17 bisected embryos were counted from either control MO or hhex MO-injected embryos. P value (0.309) was determined by the paired, two-tailed t test.

hhex Activity Before the End of Gastrulation Is Essential for the Patterning of Ventral Pancreatic Precursors.

It has been reported that hhex functioned as a transcriptional repressor to define the anterior identity of Xenopus embryos (36). Given the partial coexpression of hhex and vpp1 from gastrulation until the formation of liver and ventral pancreatic buds, we made use of a dexamethasone-inducible hhex-VP16GR fusion protein to interfere with endogenous hhex activity, which allowed us to discern the critical time window for hhex functioning in the specification of the ventral pancreatic buds. In agreement with the previous study (36), hhex also acted as a transcriptional repressor in the context of ventral pancreas and liver development, because early activation of injected hhex-VP16GR phenocopied hhex morpholino knockdown effects on XlHbox8 and fibrinogen expression, albeit more severely (Fig. S9, comparing Fig. S9C with Fig. 5P). Serial activation of injected hhex-VP16GR with dexamethasone showed that inhibition of XlHbox8 by hhex-VP16 became insignificant when the chemical was added after gastrulation (Fig. S10), suggesting that the most critical period for endogenous hhex function in patterning the ventral pancreas is during gastrulation.

Discussion

Vpp1 is a molecular marker that shows relatively broad overlapping expression with hhex in Xenopus early and midgastrulae. Overexpression of hhex resulted in an expansion of the vpp1-positive cells as early as during gastrulation, and led to a vast expansion of the ventral pancreas during tadpole stages of development. We do not know why vpp1 does not spread through the entire hhex-positive domain in normal embryos, but it is possible that an inhibitor is present in the domain specified for liver that represses the expression of vpp1. In contrast, hhex knockdown led to a reduction of vpp1 expression and loss of ventral, but not dorsal pancreas. Interfering with hhex function using an inducible version of hhex-VP16 indicated that hhex activity during gastrulation was essential for the formation of ventral pancreatic buds. These observations expand our understanding of the function of hhex in frog ventral pancreas development and reveal a vpp1-positive population of endodermal cells during gastrulation with properties expected of precursor cells destined to form ventral pancreas.

Comparing vpp1 and hhex expression with the data from two previous fate mapping studies (12, 37), we suggest that vpp1-positive cells in gastrulae are the precursors of the ventral pancreas formed later in development. First, the dynamic expression of hhex and vpp1 in gastrulae exactly reflects the fate mapping study tracing the movement and destination of the suprablastoporal and subblastoporal endodermal cells during gastrulation (37), suggesting the consistency of the expression of these two genes in the two cell populations, respectively, through gastrulation. Second, in an effort to map 14 different regions of stage 14 embryos (early neurula) to organogenesis, Chalmers and Slack (12) encountered a complex origin of liver and ventral pancreas from distinct anterior territories. These authors appealed for finer grained fate mapping studies to establish the extent to which the ventral pancreatic rudiment extends laterally and, thus, how the origin of the ventral pancreas differs from that of the liver (12). The expression of vpp1 and hhex in stage 13/14 embryos provides a basis for explaining this complexity at a higher resolution, while leaving a definitive answer to future genetic labeling studies with vpp1 and hhex promoters. In mice, Hhex also displays expression in a small region of ventral pancreatic progenitors as seen by coexpression with Pdx1 (20, 38). Only vital dye staining-based fate mapping studies exist for the mouse ventral foregut endoderm destined to form liver and ventral pancreas (13–15), whereas genetic labeling studies revealing the fate of murine Hhex+/Pdx1+ cells are missing. Single-cell labeling experiments in the zebrafish show that a small number of bipotential cells are able to give rise to both liver and ventral pancreas (11). It remains to be defined whether vpp1+/hhex+ cells in Xenopus embryos represent such bipotential cells that give rise to both ventral pancreas and liver.

Several studies in mice have demonstrated that a number of events occurring after gastrulation are crucial for the stable fate decision of ventral pancreas versus liver (38–41). In this study, our finding of the early segregation of vpp1-positive and hhex-positive endodermal cells in Xenopus embryos during gastrulation supports the notion proposed in mice that the initial regionalization of the definitive endoderm occurs concurrently with its formation as it exits the primitive streak (1). Both loss-of-function and gain-of-function analyses of hhex demonstrate the correlation of the early vpp1 expression and the late specific ventral pancreatic phenotype, suggesting an earlier fate decision of ventral pancreas versus liver during gastrulation. Together with our previous finding that ectopic activation of a combined Pdx1 and Ptf1a/p48 during stages 13–27 can convert posterior endoderm into pancreatic progenitor cells that will give rise to giant pancreata (28), our present data support the concept that the size of the progenitor cell pools is important in setting the final organ size of the pancreas (42).

Because most of the early hhex-expressing cells are liver precursor cells, we reasoned that overexpression of hhex might be sufficient to induce formation of an enlarged liver. However, we were unable to find conditions that result in an increase in the number of liver precursor cells by means of hhex mRNA injection. Knockdown of hhex also had no obvious effect on the size of this cell population during early stages of development. Thus, it appears that the identity of the endogenous liver precursor cells might be under the control of a particular in vivo niche, which is not affected by either increase or decrease of hhex activity. Indeed, a gene regulatory network controlling hhex transcription in the anterior endoderm of the organizer has recently been reported, indicating a rather complex cross-talk between several signaling pathways (6).

It remains unclear how overexpression of hhex results in the robust level of ectopic vpp1 expression described here. Vegetal injection of hhex mRNA into the two ventral blastomeres of four-cell stage embryos had no effect on liver or ventral pancreas development. Similarly, we found that overexpression of hhex in endodermalized animal cap cells was not sufficient to activate vpp1 and pancreatic marker expression. Together, these data suggest that unidentified cofactor(s) exist(s) in the dorsal endodermal cells, which cooperate(s) with hhex in converting these cells to vpp1-expressing cells. It remains to be investigated whether hhex directly interacts with the vpp1 promoter.

Materials and Methods

Isolation of vpp1 cDNA.

Stage 10 Xenopus embryos were dissected into four explants: animal pole, dorsal marginal zone, ventral marginal zone, and vegetal pole. The global gene expression in these explants was analyzed by Affymetrix DNA microarray (23). Vpp1 was identified as it showed high expression in the explants of dorsal marginal zone.

Embryo Manipulation and Analysis.

Wild-type Xenopus laevis embryos were obtained by hormone-induced egg laying and in vitro fertilization. Whole mount in situ hybridization was performed as described (28). Whole mount immunostaining for phospho-histone H3 and TUNEL staining was done as described by Saka and Smith (43), and Hensey and Gautier (44), respectively. Vibratome sections (30 μm thickness) were prepared with a Leica VT1000S Vibratome. RT-PCR was carried out as described (29).