Cullin mediates degradation of RhoA through evolutionarily conserved BTB adaptors to control actin cytoskeleton structure and cell movement.

Cul3, a Cullin family scaffold protein, is thought to mediate the assembly of a large number of SCF (Skp1-Cullin1-F-box protein)-like ubiquitin ligase complexes through BTB domain substrate-recruiting adaptors. Cul3 controls early embryonic development in several genetic models through mechanisms not understood. Very few functional substrate/adaptor pairs for Cul3 ubiquitin ligases have been identified. Here, we show that Cul3 knockdown in human cells results in abnormal actin stress fibers and distorted cell morphology, owing to impaired ubiquitination and degradation of small GTPase RhoA. We identify a family of RhoA-binding BTB domain adaptors conserved from insects to mammals, designated BACURDs. BACURDs form ubiquitin ligase complexes, which selectively ubiquitinate RhoA, with Cul3. Dysfunction of the Cul3/BACURD complex decreases cell migration potential and impairs RhoA-mediated convergent extension movements during Xenopus gastrulation. Our studies reveal a previously unknown mechanism for controlling RhoA degradation and regulating RhoA function in various biological contexts, which involves a Cul3/BACURD ubiquitin ligase complex.

Retinoic acid enhances the generation of hematopoietic progenitors from human embryonic stem cell-derived hemato-vascular precursors.

Current induction schemes directing hematopoietic differentiation of human embryonic stem cells (hESCs) are not well defined to mimic the sequential stages of hematopoietic development in vivo. Here, we report a 3-stage method to direct differentiation of hESCs toward hematopoietic progenitors in chemically defined mediums. In the first 2 stages, we efficiently generated T-positive primitive streak/mesendoderm cells and kinase domain receptor-positive (KDR(+)) platelet-derived growth factor receptor α-negative (PDGFRα(-)) hemato-vascular precursors sequentially. In the third stage, we found that cells in a spontaneous differentiation condition mainly formed erythroid colonies. Addition of all-trans retinoic acid (RA) greatly enhanced generation of hematopoietic progenitors in this stage while suppressing erythroid development. The RA-treated cells highly expressed definitive hematopoietic genes, formed large numbers of multilineage and myeloid colonies, and gave rise to greater than 45% CD45(+) hematopoietic cells. When hematopoietic progenitors were selected with CD34 and C-Kit, greater than 95% CD45(+) hematopoietic cells could be generated. In addition, we found that endogenous RA signaling at the second stage was required for vascular endothelial growth factor/basic fibroblast growth factor-induced hemato-vascular specification, whereas exogenously applied RA efficiently induced KDR(-)PDGFRα(+) paraxial mesoderm cells. Our study suggests that RA signaling plays diverse roles in human mesoderm and hematopoietic development.

CD24: a novel surface marker for PDX1-positive pancreatic progenitors derived from human embryonic stem cells.

Human ESCs provide a promising cell resource for the treatment of type I diabetes mellitus. Although PDX1-positive pancreatic progenitors can be efficiently generated from human ESCs by stepwise induction, further in vitro differentiation into functional, mature beta cells is not efficient or reproducible. Purification of pancreatic progenitor cells could facilitate the identification of signals that regulate beta cell differentiation and maturation. Here, we report the identification of a novel surface marker for PDX1-positive pancreatic progenitors based on an in vitro human ESC differentiation system. By costaining PDX1 and a panel of cell surface antigens at the pancreatic progenitor stage of human ESC differentiation, we discovered a positive marker, CD24. CD24-positive cells coexpressed most of the key transcription factors of pancreatic progenitors, and the expression of important pancreatic genes was greatly enriched in CD24-positive cells compared with the CD24-negative cells. In addition, CD24-positive cells could differentiate into insulin-producing cells but CD24-negative cells could not. These results indicate that CD24 could be a surface marker for PDX1-positive pancreatic progenitors derived from human ESCs. Enrichment of pancreatic progenitors with this marker will facilitate the investigation of beta cell maturation during the human ESC differentiation.

Mouse ZAR1-like (XM_359149) colocalizes with mRNA processing components and its dominant-negative mutant caused two-cell-stage embryonic arrest.

Maternal effect genes and encoding proteins are necessary for nuclear reprogramming and zygotic genome activation. However, the mechanisms that mediate these functions are largely unknown. Here we identified XM_359149, a Zar1-like gene that is predominantly expressed in oocytes and zygotes, which we designated Zar1-like (Zar1l). ZAR1L-EGFP formed multiple cytoplasmic foci in late two-cell-stage embryos. Expression of the ZAR1L C-terminus induced two-cell-stage embryonic arrest, accompanied with abnormal methylation of histone H3K4me2/3 and H3K9me2/3, and marked down-regulation of a group of chromatin modification factors including Dppa2, Dppa4, and Piwil2. When ectopically expressed in somatic cells, ZAR1L colocalized with P-body components including EIF2C1(AGO1), EIF2C2(AGO2), DDX6 and LSM14A, and germline-specific chromatoid body components including PIWIL1, PIWIL2, and LIN28. ZAR1L colocalized with ZAR1 and interacted with human LIN28. Our data suggest that ZAR1L and ZAR1 may comprise a novel family of processing-body/chromatoid-body components that potentially function as RNA regulators in early embryos.

A mouse model of inducible liver injury caused by tet-on regulated urokinase for studies of hepatocyte transplantation.

Mouse models of liver injury provide useful tools for studying hepatocyte engraftment and proliferation. A representative model of liver injury is the albumin-urokinase (Alb-uPA) transgenic model, but neonatal lethality hampers its widespread application. To overcome this problem, we generated a transgenic mouse in which transcription of the reverse tetracycline transactivator was (rtTA) driven by the mouse albumin promoter, and backcrossed the rtTA mice onto severe combined immunodeficient (SCID)/bg mice to generate immunodeficient rtTA/SCID mice. We then produced recombinant adenoviruses Ad.TRE-uPA, in which the urokinase was located downstream of the tetracycline response element (TRE). The rtTA/SCID mouse hepatocytes were then infected with Ad.TRE-uPA to establish an inducible liver injury mouse model. In the presence of doxycycline, uPA was exclusively expressed in endogenous hepatocytes and caused extensive liver injury. Enhanced green fluorescent protein-labeled mouse hepatocytes selectively repopulated the rtTA/SCID mouse liver and replaced over 80% of the recipient liver mass after repeated administration of Ad.TRE-uPA. Compared with the original uPA mice, rtTA/SCID mice did not exhibit problems regarding breeding efficiency, and the time window for transplantation was flexible. In addition, we could control the extent of liver injury to facilitate transplantation surgery by regulating the dose of Ad.TRE-uPA. Our inducible mouse model will be convenient for studies of hepatocyte transplantation and hepatic regeneration, and this system will facilitate screening for potential genetic factors critical for engraftment and proliferation of hepatocytes in vivo.

Uroplakin IIIb, a urothelial differentiation marker, dimerizes with uroplakin Ib as an early step of urothelial plaque assembly.

Urothelial plaques consist of four major uroplakins (Ia, Ib, II, and III) that form two-dimensional crystals covering the apical surface of urothelium, and provide unique opportunities for studying membrane protein assembly. Here, we describe a novel 35-kD urothelial plaque-associated glycoprotein that is closely related to uroplakin III: they have a similar overall type 1 transmembrane topology; their amino acid sequences are 34% identical; they share an extracellular juxtamembrane stretch of 19 amino acids; their exit from the ER requires their forming a heterodimer with uroplakin Ib, but not with any other uroplakins; and UPIII-knockout leads to p35 up-regulation, possibly as a compensatory mechanism. Interestingly, p35 contains a stretch of 80 amino acid residues homologous to a hypothetical human DNA mismatch repair enzyme-related protein. Human p35 gene is mapped to chromosome 7q11.23 near the telomeric duplicated region of Williams-Beuren syndrome, a developmental disorder affecting multiple organs including the urinary tract. These results indicate that p35 (uroplakin IIIb) is a urothelial differentiation product structurally and functionally related to uroplakin III, and that p35-UPIb interaction in the ER is an important early step in urothelial plaque assembly.

A human endothelial cell feeder system that efficiently supports the undifferentiated growth of mouse embryonic stem cells.

Feeder cells are commonly used to culture embryonic stem cells to maintain their undifferentiated and pluripotent status. Conventionally, mouse embryonic fibroblasts (MEFs), supplemented with leukemia inhibitory factor (LIF), are used as feeder cells to support the growth of mouse embryonic stem cells (mESCs) in culture. To prepare for fresh MEF feeder or for MEF-conditioned medium, sacrifice of mouse fetuses repeatedly is unavoidable in these tedious culture systems. Here we report the discovery of a human endothelial cell line (ECV-304 cell line) that efficiently supports growth of mESCs LIF-free conditions. mESCs that were successfully cultured for eight to 20 passages on ECV-304 feeders showed morphological characteristics similar to cells cultured in traditional feeder cell systems. These cells expressed the stem cell markers Oct3/4, Nanog, Sox2, and SSEA-1. Furthermore, cells cultured on the ECV-304 cell line were able to differentiate into three germ layers and were able to generate chimeric mice. Compared with traditional culture systems, there is no requirement for mouse fetuses and exogenous LIF does not need to be added to the culture system. As a stable cell line, the ECV-304 cell line efficiently replaces MEFs as an effective feeder system and allows the efficient expansion of mESCs.

Mitochondrial lipid droplet formation as a detoxification mechanism to sequester and degrade excessive urothelial membranes.

The apical surface of the terminally differentiated mammalian urothelial umbrella cell is mechanically stable and highly impermeable, in part due to its coverage by urothelial plaques consisting of 2D crystals of uroplakin particles. The mechanism for regulating the uroplakin/plaque level is unclear. We found that genetic ablation of the highly tissue-specific sorting nexin Snx31, which localizes to plaques lining the multivesicular bodies (MVBs) in urothelial umbrella cells, abolishes MVBs suggesting that Snx31 plays a role in stabilizing the MVB-associated plaques by allowing them to achieve a greater curvature. Strikingly, Snx31 ablation also induces a massive accumulation of uroplakin-containing mitochondria-derived lipid droplets (LDs), which mediate uroplakin degradation via autophagy/lipophagy, leading to the loss of apical and fusiform vesicle plaques. These results suggest that MVBs play an active role in suppressing the excessive/wasteful endocytic degradation of uroplakins. Failure of this suppression mechanism triggers the formation of mitochondrial LDs so that excessive uroplakin membranes can be sequestered and degraded. Because mitochondrial LD formation, which occurs at a low level in normal urothelium, can also be induced by disturbance in uroplakin polymerization due to individual uroplakin knockout and by arsenite, a bladder carcinogen, this pathway may represent an inducible, versatile urothelial detoxification mechanism.

Mature oocytes derived from purified mouse fetal germ cells.

BACKGROUND: Mouse fetal germ cells have been successfully purified from fetal gonads. However, there are no published reports describing a procedure for deriving mature oocytes from isolated fetal germ cells. The purpose of this present study is to explore whether purified fetal germ cells are able to differentiate into mature oocytes through an in vivo grafting procedure. METHODS AND RESULTS: First, intact 11.5 and 12.5 days post-coitum (dpc) female gonads with or without the attached mesonephros and the reaggregated female gonad cells were transplanted into the recipient mice. The results demonstrate both the gonad accompanied by mesonephroi and the innate gonad structure are not absolutely required for 11.5 dpc and 12.5 dpc oogonia to generate mature oocytes. Next, oogonia were purified from female gonads, aggregated with different ovarian somatic cells and transplanted into the recipient mice. Purified 12.5 dpc oogonia were able to generate mature oocytes by aggregating with 12.5 dpc ovarian somatic cells, but not with 16.5 dpc or 0 days postpartum ovarian somatic cells. We also tested 12.5 dpc male germ cells but they were unable to undergo oogenesis. CONCLUSIONS: Our study demonstrates that mature oocytes can be derived from purified fetal germ cells through an aggregation and transplantation procedure. It also suggests that the synchronized interactions between oogonia and gonadal somatic cells are important to ensure normal folliculogenesis.

Induction of oocyte-like cells from mouse embryonic stem cells by co-culture with ovarian granulosa cells.

In vitro derivation of oocytes from embryonic stem (ES) cells has the potential to be an important tool for studying oogenesis as well as advancing the field of therapeutic cloning by providing an alternative source of oocytes. Here, we demonstrate a novel, two-step method for inducing mouse ES cells to differentiate into oocyte-like cells using mouse ovarian granulosa cells. First, primordial germ cells (PGCs) were differentiated within the embryonic body (EB) cells around day 4 as defined by the expression of PGC-specific markers and were distinguished from undifferentiated ES cells. Second, day 4 EB cells were co-cultured with ovarian granulosa cells. After 10 days, these cells formed germ cell colonies as indicated by the expression of the two germ cell markers Mvh and SCP3. These cells also expressed the oocyte-specific genes Figalpha, GDF-9, and ZP1-3 but not any testis-specific genes by RT-PCR analysis. EB cultured alone or cultured in granulosa cell-conditioned medium did not express any of these oocyte-specific markers. In addition, EB co-cultured with Chinese hamster ovary (CHO) cells or cultured in CHO cell-conditioned medium did not express all of these oocyte-specific markers. Immunocytochemistry analysis using Mvh and GDF-9 antibodies confirmed that some Mvh and GDF-9 double-positive oocyte-like cells were generated within the germ cell colonies. Our results demonstrate that granulosa cells were effective in inducing the differentiation of ES cell-derived PGCs into oocyte-like cells through direct cell-to-cell contacts. Our method offers a novel in vitro system for studying oogenesis; in particular, for studying the interactions between PGCs and granulosa cells.

GATA factors induce mouse embryonic stem cell differentiation toward extraembryonic endoderm.

The GATA family of transcription factors are implicated in early embryonic development. There are six factors in this family in vertebrates. GATA4 and GATA6 have been demonstrated to induce mouse embryonic stem (mES) cells differentiation toward extraembryonic endoderm (ExE). We investigated the effect of GATA3 on the differentiation of mES cells both in the ES cell and in the embryoid body (EB) states. The results demonstrate that GATA3 overexpression can initiate the ES cell differentiation program toward ExE. Furthermore, overexpression of GATA1 and GATA2 in ES cells and EBs resulted in similar effects. We believe this finding can augment our understanding of mouse ES cell differentiation.

Mice transgenic for human angiotensin-converting enzyme 2 provide a model for SARS coronavirus infection.

To establish a small animal model of severe acute respiratory syndrome (SARS), we developed a mouse model of human severe acute respiratory syndrome coronavirus (SARS-CoV) infection by introducing the human gene for angiotensin-converting enzyme 2 (hACE2) (the cellular receptor of SARS-CoV), driven by the mouse ACE2 promoter, into the mouse genome. The hACE2 gene was expressed in lung, heart, kidney, and intestine. We also evaluated the responses of wild-type and transgenic mice to SARS-CoV inoculation. At days 3 and 7 postinoculation, SARS-CoV replicated more efficiently in the lungs of transgenic mice than in those of wild-type mice. In addition, transgenic mice had more severe pulmonary lesions, including interstitial hyperemia and hemorrhage, monocytic and lymphocytic infiltration, protein exudation, and alveolar epithelial cell proliferation and desquamation. Other pathologic changes, including vasculitis, degeneration, and necrosis, were found in the extrapulmonary organs of transgenic mice, and viral antigen was found in brain. Therefore, transgenic mice were more susceptible to SARS-CoV than were wild-type mice, and susceptibility was associated with severe pathologic changes that resembled human SARS infection. These mice will be valuable for testing potential vaccine and antiviral drug therapies and for furthering our understanding of SARS pathogenesis.

Failure of hepatocyte marker-expressing hematopoietic progenitor cells to efficiently convert into hepatocytes in vitro.

OBJECTIVE: Whether bone marrow (BM) hematopoietic stem/progenitor cells can directly differentiate into nonhematopoietic cells remains controversial. The aim of this study is to further investigate the potentiality of BM hematopoietic progenitor cells to convert into hepatocytes in vitro. MATERIALS AND METHODS: Different subsets of BM cells from C57/BL6 mice were isolated using markers of hematopoietic stem cells by magnetic cell sorting and by flow cytometry. These cells were induced to transdifferentiate to hepatocytes in vitro in the presence of various cytokines or of hepatocytes (or tissue) from damaged liver, which have been reported to stimulate the conversion. Hepatic gene markers in freshly isolated or cultured BM cells were determined by reverse transcriptase polymerase chain reaction and immunofluorescence. RESULTS: Freshly isolated hematopoietic progenitor cells (HPC) expressed a low level of messenger RNAs of hepatic cell-specific markers including albumin and alpha-fetoprotein (AFP), but did not significantly upregulate expression of these markers, even in the presence of cytokines or cocultured hepatocytes (or tissue). HPCs induced in vitro did not express the message of alpha-anti-trypsin-a mature hepatocyte marker. At protein level, the specific staining of AFP was not detected in the HPCs, either freshly isolated or in vitro induced. Albumin protein was detected in freshly isolated albumin mRNA-positive and -negative BM cell subpopulations. Albumin-stained BM cells disappeared after being induced for 5 days, but restained if mouse serum was supplemented in medium for a 24-hour extended culture, suggesting that albumin was absorbed by BM cells instead of de novo expression. CONCLUSIONS: HPCs expressed mRNAs of hepatic cell markers, but could not efficiently convert into hepatocytes in vitro under our experimental conditions. Our observation raises a cautionary note in determining whether in vitro transdifferentiation of BM cells to hepatocytes can actually take place.

Baculovirus surface display of SARS coronavirus (SARS-CoV) spike protein and immunogenicity of the displayed protein in mice models.

The baculovirus surface display technique has provided an ideal tool to display foreign proteins with natural conformation, functions, and immunogenicity. In this work, we explored the application of this technique on SARS-associated coronavirus (SARS-CoV) spike (S) protein, and further analyzed the immunogenicity of displayed S protein. The entire ectodomain of S protein was fused between the gp64 signal peptide and the VSV-G membrane anchor and successfully displayed on the baculovirus surface. Subcutaneous injection with purified S-displayed baculoviruses without adjuvant elicited highly effective production of specific and neutralizing antibodies against S protein in mice. These results confirmed a successful surface display of S protein on baculoviruse, and suggested a potential role of S-displayed baculoviruses as a novel live virus-based vaccine candidate for SARS-CoV.

[RNA interference in three ES cell lines from different mouse strains].

RNA interference phenomenon in three different murine ES cell lines (MESPU13, B3, and R1) is reported. A vector(pdsGFP) was used that transcribed hairpin double-stranded RNA of GFP gene to transfect ES cells by using lipofectin. The transient transcription of dsRNA induced RNAi (RNA interference) in the ES cells. That is, the double-stranded RNA of GFP gene potently turned down the expression of the GFP gene. On the hand, the linearized plasmid pdsGFP-puro was electroporated into MESPU13 ES cells, and the expression level of GFP after puromycin screening was turned down obviously in about 30% ES cell clones; and in a few clones, the expression level of GFP was not observed under the fluorescence microscope and GFP mRNA was not detectable by RT-PCR. Further more, another vector (pdsOCT4) was constructed that transcribed double-stranded RNA of OCT-4 gene which is specifically expressed in ES cells. ES cell clones that stably integrated the vector were screened after the electrotransfection of the cells with the above construct. 51 random-selected clones were amplified and 48 of them were checked by semi-quantitative RT-PCR. In 11 of them the mRNA of OCT-4 was undetectable by RT-PCR. This means that RNAi can be used to study mammal and human gene's function in ES cell lines from different strain mice.

SNX31: a novel sorting nexin associated with the uroplakin-degrading multivesicular bodies in terminally differentiated urothelial cells.

Uroplakins (UP), a group of integral membrane proteins, are major urothelial differentiation products that form 2D crystals of 16-nm particles (urothelial plaques) covering the apical surface of mammalian bladder urothelium. They contribute to the urothelial barrier function and, one of them, UPIa, serves as the receptor for uropathogenic Escherichia coli. It is therefore important to understand the mechanism by which these surface-associated uroplakins are degraded. While it is known that endocytosed uroplakin plaques are targeted to and line the multivesicular bodies (MVBs), it is unclear how these rigid-looking plaques can go to the highly curved membranes of intraluminal vesicles (ILVs). From a cDNA subtraction library, we identified a highly urothelium-specific sorting nexin, SNX31. SNX31 is expressed, like uroplakins, in terminally differentiated urothelial umbrella cells where it is predominantly associated with MVBs. Apical membrane proteins including uroplakins that are surface biotin-tagged are endocytosed and targeted to the SNX31-positive MVBs. EM localization demonstrated that SNX31 and uroplakins are both associated not only with the limiting membranes of MVBs containing uroplakin plaques, but also with ILVs. SNX31 can bind, on one hand, the PtdIns3P-enriched lipids via its N-terminal PX-domain, and, on the other hand, it binds uroplakins as demonstrated by co-immunoprecipitation and proximity ligation assay, and by its reduced membrane association in uroplakin II-deficient urothelium. The fact that in urothelial umbrella cells MVBs are the only major intracellular organelles enriched in both PtdIns3P and uroplakins may explain SNX31's MVB-specificity in these cells. However, in MDCK and other cultured cells transfected SNX31 can bind to early endosomes possibly via lipids. These data support a model in which SNX31 mediates the endocytic degradation of uroplakins by disassembling/collapsing the MVB-associated uroplakin plaques, thus enabling the uroplakin-containing (but 'softened') membranes to bud and form the ILVs for lysosomal degradation and/or exosome formation.

TGFß inhibition enhances the generation of hematopoietic progenitors from human ES cell-derived hemogenic endothelial cells using a stepwise strategy.

Embryonic hematopoiesis is a complex process. Elucidating the mechanism regulating hematopoietic differentiation from pluripotent stem cells would allow us to establish a strategy to efficiently generate hematopoietic cells. However, the mechanism governing the generation of hematopoietic progenitors from human embryonic stem cells (hESCs) remains unknown. Here, on the basis of the emergence of CD43(+) hematopoietic cells from hemogenic endothelial (HE) cells, we demonstrated that VEGF was essential and sufficient, and that bFGF was synergistic with VEGF to specify the HE cells and the subsequent transition into CD43(+) hematopoietic cells. Significantly, we identified TGFß as a novel signal to regulate hematopoietic development, as the TGFß inhibitor SB 431542 significantly promoted the transition from HE cells into CD43(+) hematopoietic progenitor cells (HPCs) during hESC differentiation. By defining these critical signaling factors during hematopoietic differentiation, we can efficiently generate HPCs from hESCs. Our strategy could offer an in vitro model to study early human hematopoietic development.

Holoclone forming cells from pancreatic cancer cells enrich tumor initiating cells and represent a novel model for study of cancer stem cells.

BACKGROUND: Pancreatic cancer is one of the direct causes of cancer-related death. High level of chemoresistance is one of the major obstacles of clinical treatment. In recent years, cancer stem cells have been widely identified and indicated as the origin of chemoresistance in multi-types of solid tumors. Increasing evidences suggest that cancer stem cells reside in the cells capable of forming holoclones continuously. However, in pancreatic cancer, holoclone-forming cells have not been characterized yet. Therefore, the goal of our present study was to indentify the holoclone-forming pancreatic cancer stem cells and develop an in vitro continuous colony formation system, which will greatly facilitate the study of pancreatic cancer stem cells. METHODOLOGY/PRINCIPAL FINDINGS: Pancreatic cancer cell line BxPC3 was submitted to monoclonal cultivation to generate colonies. Based on the morphologies, colonies were classified and analyzed for their capacities of secondary colony formation, long-term survival in vitro, tumor formation in vivo, and drug resistance. Flowcytometry and quantitative RT-PCR were performed to detect the expression level of cancer stem cells associated cell surface markers, regulatory genes and microRNAs in distinct types of colonies. Three types of colonies with distinct morphologies were identified and termed as holo-, mero-, and paraclones, in which only holoclones generated descendant colonies of all three types in further passages. Compared to mero- and paraclones, holoclones possessed higher capacities of long-term survival, tumor initiation, and chemoresistance. The preferential expression of cancer stem cells related marker (CXCR4), regulatory genes (BMI1, GLI1, and GLI2) and microRNAs (miR-214, miR-21, miR-221, miR-222 and miR-155) in holoclones were also highlighted. CONCLUSIONS/SIGNIFICANCE: Our results indicate that the pancreatic tumor-initiating cells with high level of chemoresistance were enriched in holoclones derived from BxPC3 cell line. Generation of holoclones can serve as a novel model for studying cancer stem cells, and attribute to developing new anti-cancer drugs.

Generation of homogeneous PDX1(+) pancreatic progenitors from human ES cell-derived endoderm cells.

One key step in producing insulin-secreting cells from human embryonic stem (hES) cells is the generation of pancreatic and duodenal homeobox gene 1 (PDX1)-expressing pancreatic progenitor cells. All-trans retinoic acid (RA) has important roles in pancreas development and is widely used to induce pancreatic differentiation of ES cells. When RA was added directly to the activin A-induced hES cells, <20% cells were positive for the pancreatic marker PDX1, whereas the other cells were mainly hepatic cells. We found that when the activin A-induced hES cells were replated and seeded at low cell densities, the addition of RA induced significant pancreatic differentiation and over 70% of cells in culture expressed PDX1. When the endodermal cells were isolated with the surface marker CXCR4 from the activin A-induced culture and further differentiated with RA, a homogeneous PDX1(+) cell population (over 95% pure) was generated. The PDX1(+) cells could further differentiate into cells that expressed pancreatic transcription factors and pancreatic endocrine or exocrine markers. We also found that RA inhibited the hepatic differentiation of endodermal cells that were seeded at low cell densities, and this inhibition may have been through the inhibition of Smad1/5/8 activity. Thus, we present a highly efficient and reproducible protocol for generating PDX1(+) pancreatic progenitor cells from hES cells.