Insulin-Expressing Colonies Developed From Murine Embryonic Stem ...

Original Article Insulin-Expressing Colonies Developed From Murine Embryonic Stem Cell–Derived Progenitors Hsun Teresa Ku,1 Jing Chai,1 Yoon-Jung Kim,1 Peter White,2 Sheetal Purohit-Ghelani,1 Klaus H. Kaestner,2 and Jonathan S. Bromberg3

Previous studies describe a unique culture method for the commitment of murine embryonic stem cells to early endocrine pancreata. In this report, early pancreatic-like ␤-cell progenitors were enriched and a colony assay devised to allow these progenitors to differentiate into insulin-expressing colonies in vitro. An embryonic stem cell line with enhanced green fluorescent protein (EGFP) inserted into one allele of neurogenin 3 (Ngn3), a marker for pancreatic endocrine progenitors, was differentiated. During the late stage of culture, 20 –30% of cells were Ngn3-EGFPⴙ. Gene expression profiling using the PancChip microarray platform demonstrated that Ngn3-EGFPⴙ cells differentially express endocrine-related genes. A novel semisolid culture method was developed to support the formation of individual insulin/C-peptide– expressing colonies from dissociated single cells. Approximately 0.1– 0.6% of Ngn3-EGFPⴙ cells gave rise to insulin-expressing colonies, a three- to fivefold enrichment of ␤-cell–like progenitors, or insulin-expressing colony-forming units (ICFUs), compared with nonsorted cells. All of the single colonies expressed insulin II, while 69% coexpressed insulin I and 44% coexpressed glucagon. Some single colonies expressed insulin I, insulin II, and Pdx-1 (pancreatic duodenal homeobox–1), but not glucagon. In other colonies, glucagon expression overlapped with C-peptide II in double immunostaining analysis, suggesting heterogeneity among the ICFUs and their resulting colonies. Together, these results demonstrate that progenitors that have the potential to give rise to insulin-expressing cells can be derived from murine embryonic stem cells. Diabetes 56:921–929, 2007

From the 1Departments of Gene and Cell Medicine and Surgery, Mount Sinai School of Medicine, New York, New York; the 2Institute for Diabetes, Obesity, and Metabolism and Department of Genetics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; and the 3Recanati/Miller Transplantation Institute, Mount Sinai School of Medicine, New York, New York. Address correspondence and reprint requests to H. Teresa Ku, Box 1496, Mount Sinai School of Medicine, New York, NY 10029-6574. E-mail: hsun. [email protected]. Received for publication 10 April 2006 and accepted in revised form 22 December 2006. Additional information for this article can be found in an online appendix at http://dx.doi.org/10.2337/db06-0468. bHLH, basic helix-loop-helix; CM, conditioned media; EGFP, enhanced green fluorescent protein; E, embryonic day; ICFU, insulin-expressing colonyforming unit; NEA, nicotinamide, exendin-4, and activin ␤B; Ngn3, neurogenin 3; Pdx-1, pancreatic duodenal homeobox–1; VEGF, vascular endothelial growth factor. DOI: 10.2337/db06-0468 © 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

DIABETES, VOL. 56, APRIL 2007

T

he pancreas is comprised of exocrine and endocrine tissues and develops from the foregut endoderm of the embryo. Many transcription factors, including Pdx-1 (pancreatic duodenal homeobox–1) (1,2), basic helix-loop-helix (bHLH) transcription factor PTF1a (3), neurogenin 3 (Ngn3) (4), homeodomain proteins Nkx2.2 (5) and Nkx6.1 (6), paired domain transcription factors Pax4 (7), Pax6 (8), bHLH transcription factor NeuroD (9), and forkhead homeobox A2 (Foxa2) (10), have been found to be important for the morphogenesis of pancreatic buds and the formation of exocrine and endocrine cells. Pdx-1, a homeodomaincontaining transcription factor, is one of the earliest genes expressed in the pancreatic lineage. Pdx1-lacZ knock-in mice stain positive for ␤-galactosidase as early as embryonic day (E) 8.5 in the region that forms the pancreatic buds (1,2). Mice that lack Pdx-1 develop only a small pancreatic rudiment, demonstrating the importance of this gene in pancreatic morphogenesis (2,11). Ngn3, a bHLH transcription factor, is expressed later (⬃E9.5) during development, and mice deficient for Ngn3 develop diabetes due to the lack of all endocrine lineages (4). Using genetic cell-lineage tracing that labels specific populations of cells at a given time, it was confirmed that early (⬍E12.5) Pdx-1⫹ progenitors give rise to all pancreatic lineages, including duct, exocrine, and endocrine cells (12). In contrast, cells that express Ngn3 differentiate only into endocrine, but not exocrine, lineages (12). Collectively, these studies support the view that sequential activation of transcription factors is accompanied by progressive restriction in the lineage potential of pancreatic progenitors. Although several transcriptional regulators and signaling pathways have been shown to play an important role in pancreatic development over the past decade, two major problems have prevented the isolation of pancreatic progenitors. One is the lack of suitable cell surface markers that can be used for enrichment of pancreatic progenitors by cell sorting. The other is the absence of a permissive culture method in which single pancreatic progenitors can differentiate according to their lineage potential. Colony assays have played important roles in deciphering the growth-factor requirements, interactions, and mechanisms governing the lineage commitment of hematopoietic stem cells (13). In the field of pancreatic progenitor cell biology, pancreatic organ cultures have been devised (14,15); however, such assays do not address biological questions at the single-cell level, and cultivation and differentiation of 921

COLONY ASSAY FOR ␤-CELL–LIKE PROGENITORS

dissociated pancreatic progenitors into any lineage have so far not been possible. In this report, we identify a colony assay method for pancreatic-like ␤-cell progenitors isolated from murine embryonic stem cell– derived populations. RESEARCH DESIGN AND METHODS Growth and differentiation of embryonic stem cells. The murine embryonic stem lines R1, TL1, and Ngn3-EGFP (16) were maintained in an undifferentiated state as described (17). Differentiation protocols were as described (17), with minor modifications. Embryoid bodies grown in the presence of a selected batch of FCS (Gemini BioProducts, West Sacramento, CA) for 6 days were used in this study. Whole embryoid bodies were washed, counted, and plated to “tertiary culture” in six-well tissue culture plates or T-175 flasks (Becton Dickinson Falcon, San Jose, CA). The basic tertiary culture medium contained Dulbecco’s modifed Eagle’s medium/F-12 (1:1) (Mediatech, Herndon, VA) and 15% knockout serum replacement (Invitrogen, Carlsbad, CA). From day 0 to day 2 of tertiary culture, 5% serum was added to basic media to hasten the attachment of embryoid bodies. On day 7 of the tertiary culture, 10 mmol/l nicotinamide (Sigma, St. Louis, MO), 10 ng/ml human recombinant activin-␤B (R&D Systems, Minneapolis, MN), and 0.1 nmol/l exendin-4 (Sigma) were added. Dissociation and reaggregation experiments. For dissociation on day 0 of tertiary culture, single-cell suspensions from a total of 90 embryoid bodies treated with trypsin-EDTA were plated in three wells of a six-well plate. For dissociation on days 3, 7, and 10 of tertiary culture, a total of 90 nondissociated embryoid bodies were first plated in three wells and on designated days cells were trypsinized, washed, and then incubated with 2 mg/ml collagenase D (Roche Diagnostics, Indianapolis, IN) and 10 ␮g/ml DNase I (Calbiochem, San Diego, CA) for 30 min to yield single-cell suspensions. Cells were allowed to reaggregate in 24-well ultralow protein-binding plates (Costar; Corning, Acton, MA) at a concentration of 2–3 ⫻ 105 cells/ml in either basic tertiary media (days 0 and 3 of tertiary culture) or tertiary media supplemented with a combination of nicotinamide, exendin-4, and activin ␤B (NEA) (days 7 and 10 of tertiary culture) for 24 h. Cultures were maintained according to the schedule depicted in supplemental Fig. 1 (which can be found in an online appendix available at http://dx.doi.org/10.2337/db06 – 0468), and all cells were collected on day 14 of tertiary culture. Expression of insulin I was analyzed by real-time RT-PCR. Cell sorting and colony culture for progenitors. Cells from day 10 of tertiary culture were dissociated into a single-cell suspension as described above. Cells were washed with PBS containing 0.1% BSA and passed through a 20-␮m mesh before placing in a sorter (FACSVantage SE with FACSDiVa Option; Becton Dickinson, San Jose, CA). Cells were gated on forward and side scatter, and doublets were eliminated. Ngn3-EGFP⫹ cells were gated according to the FL1/EGFP and FL3/autofluorescence window so that ⬎99% of TL1 cells are negative as displayed in Fig. 1B. Cells were placed in colony culture at a density of 2.5 ⫻ 104 cells/well/0.5 ml. For Matrigel-based colony culture, one part of cold Matrigel was mixed with two parts of cold tertiary culture media and then added at 100 ␮l/well to eight-well chamber slides (Becton Dickinson Biosciences Discovery Labware, Bedford, MA) as droplets. After incubation at 37°C for 30 min to solidify the gels, cells in 400 ␮l/well basic tertiary media containing 50% tertiary conditioned media (CM), 5% FCS, 1 ng/ml vascular endothelial growth factor (VEGF)-A (R&D), 10 mmol/l nicotinamide, 0.1 nmol/l exendin-4, and 10 ng/ml human recombinant activin-␤B were added on top of the gels. The tertiary-CM was obtained by adding the basic tertiary media containing NEA on day 7 of the tertiary culture and collecting the medium on day 10. For methylcellulosebased colony culture, 1 ml cold culture mixture contained 1% 1,500 centipoise (high-viscosity) methylcellulose (Sinetsu Chemical, Tokyo, Japan), 5% Matrigel, 50% tertiary-CM, 5% FCS, 1 ng/ml VEGF-A, 10 mmol/l nicotinamide, 0.1 nmol/l exendin-4, and 10 ng/ml human recombinant activin ␤B. The culture mixture was placed in either eight-well chamber slides or 24-well ultralow protein-binding plates and incubated in a humidified 5% CO2 atmosphere. Colonies were scored on an inverted microscope after 8 days of culture. RNA analysis. Methods for procurement of cells, extraction of total RNA, and real-time RT-PCR analysis were as previously described (17). Total RNA from single colonies was extracted using the RNeasy Micro Kit (Qiagen, Valencia, CA). Immunocytochemistry. Colonies were hand-picked, spun onto slides using Cytospin 3 (Thermo Shandon, Pittsburgh, PA), immediately fixed in 4% paraformaldehyde/0.15% picric acid in PBS, pH 7.5, for 30 min, and then washed with PBS twice. Cell membranes were permeablized with 0.1% Triton-X 100 (Pierce, Rockford, IL) in PBS at 20°C for 10 min. Protein Block Serum-Free (Dako, Carpinteria, CA) solution was added to inhibit nonspecific staining of the cells. Goat anti-rat C-peptide II serum (1:800 dilution) was 922

FIG 1. Expression of Ngn3-EGFP in embryonic stem– derived cells. Ngn3-EGFP embryonic stem cells were differentiated to embryoid bodies in the presence of serum followed by plating into tertiary culture. Cells were dissociated into single cells and analyzed by flow cytometry for EGFP expression. Parental TL1 embryonic stem cells were used for negative gating controls. A: Kinetics of Ngn3-EGFP expression. B: Representative flow cytometric analysis is shown on day 10 of tertiary culture. purchased from Linco Research (St. Charles, MO). Rabbit anti-mouse Cpeptide I serum (1:200 dilution) was obtained from the Beta Cell Biology Consortium. Rabbit anti-human glucagon serum (1:75 dilution) was purchased from Dako. Primary antibodies were applied to samples and incubated at room temperature for 1 h. Cy3 or Cy5 conjugated donkey anti-rabbit or anti-goat secondary antibodies were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA). Dilution of the antibodies was made according to the manufacturer’s recommendations. Samples were washed with PBS twice before secondary antibodies were applied and incubated at room temperature for 1 h. Samples were washed again with PBS twice and observed by fluorescence microscopy using a Leica DMRA2 microscope. Images were captured as grayscale pictures and colors painted subsequently using Openlab software (Improvision; Quincy, Boston, MA). Flow cytometric analysis. Cells from embryoid bodies, tertiary culture, or semisolid culture were made into a single-cell suspension as described above. Cells were either directly analyzed by using a FACSCalibur (Becton Dickinson) for EGFP expression or fixed in 4% paraformaldehyde/0.15% picric acid and stained with anti–C-peptide I antibody as described for immunocytochemistry. Data were analyzed by CellQuest (Becton Dickinson) software. All events were gated with forward and side scatter profiles. Age-matched TL1 embryonic stem– derived cells or cells stained without primary antibody were used as negative gating controls for EGFP or C-peptide I expression, respectively. Statistical analysis. Student’s t test was used to determine statistical significance. Microarray expression profiling and data analysis. The mouse PancChip version 6.0 13K cDNA microarray (http://www.cbil.upenn.edu/EPConDB) (18) was used in this study. Four different biological replicates of mouse Ngn3EGFP⫹ and Ngn3-EGFP⫺ populations were analyzed. RNA samples were isolated as described (17) and amplified using Message Amp kit (Ambion, Austin, TX). Methods for quality control of the samples, reverse transcription, fluorescent labeling, and data acquiring were as described (19). For statistical analysis, the genes were identified as differentially expressed using the patterns from the Gene Expression package (PaGE, version 5.0) as described previously (20).

RESULTS

Ngn3-EGFP– expressing cells are present in tertiary culture. Previously, we described a culture system that allows murine embryonic stem cells to commit to early DIABETES, VOL. 56, APRIL 2007

H.T. KU AND ASSOCIATES

endocrine pancreas lineages in vitro (17) (supplemental Fig. 1). Since the progenitors for both the first (21) and second (22) waves of ␤-cells express Ngn3 (4,12,23), pancreatic endocrine progenitors were further characterized in this system by examining the expression of Ngn3 during embryonic stem cell differentiation. An embryonic stem line designated as Ngn3-EGFP, in which the EGFP coding sequence was inserted into one allele of the Ngn3 locus, was used (16). Haplodeficiency of Ngn3 does not result in developmental abnormality (4,16). Insulin I message expression levels in later stage tertiary cultures, analyzed by real-time RT-PCR, were similar among Ngn3EGFP, TL1 (parental line to Ngn3-EGFP), and R1 embryonic stem lines (Supplemental Fig. 2), demonstrating that the differentiation protocol (17) is applicable to a variety of murine embryonic stem lines. To test the extent of endocrine progenitor formation in the later-stage tertiary culture, the kinetics of Ngn3-EGFP expression during embryonic stem cell differentiation were examined by fluorescent flow cytometry (Fig. 1). Parental TL1 cells were used as a negative control for gating. Consistent with a previous report (24), Ngn3 was expressed in ⬃18% of undifferentiated embryonic stem cells (Fig. 1A). The percentage of Ngn3-EGFP⫹ cells decreased during the first 2 days of differentiation and increased to 4 –9% by day 6 of secondary embryoid body culture (Fig. 1A). The embryoid bodies were then placed into tertiary culture, and the percentage of Ngn3-EGFP⫹ cells increased to 20 –30% within 48 h. This level of expression was maintained throughout the 16 days of tertiary culture (Fig. 1A). Generation of insulin I– expressing cells after dissociation on day 10 of tertiary culture. Using trypsin to dissociate day 6 embryoid bodies into smaller clusters abrogates the development of insulin I⫹ cells in tertiary culture (17), suggesting that cell-to-cell interactions established in day 6 embryoid bodies are required to support ␤-cell–like formation. Since colony assays require plating of dissociated single cells, the stage at which ␤-cell–like progenitors could become independent of cell-to-cell interactions was assessed. On days 0, 3, 7, or 10 of the tertiary culture, embryoid body– derived populations were dissociated into single cells (Fig. 2). Direct plating of dissociated single cells into tissue culture plates did not result in significant expression of insulin I at later time points (not shown). We allowed the single cells to reaggregate at 2–3 ⫻ 105 cells/ml for 24 h before placing them back into the tertiary culture. Cells were harvested on day 14 of the tertiary culture and analyzed for insulin expression by RT-PCR. Dissociation and reaggregation on day 10 of tertiary culture permitted similar expression levels of insulin I mRNA compared with the control nondissociated cells (Fig. 2). In contrast, dissociation and reaggregation of day 6 embryoid bodies (or day 0 tertiary cells) resulted in undetectable levels of insulin I (17), and experiments performed on day 3 or day 7 of the tertiary culture led to reduced insulin expression. These results suggest that at least some progenitors to insulin I– expressing cells may commit by day 10 of tertiary culture and require minimal cell-to-cell interactions established during the earlier phase of the culture. Ngn3-EGFPⴙ cells isolated on day 10 of tertiary culture express pancreatic endocrine genes. Results from the dissociation and reaggregation studies (Fig. 2) suggest that some cells in day 10 of the tertiary culture may represent a relatively mature phase of progenitor DIABETES, VOL. 56, APRIL 2007

FIG. 2. Dissociation and reaggregation of cells during tertiary culture leads to various levels of insulin I expression. On the designated day of tertiary culture, cells were dissociated, reaggregated for 24 h, and allowed to further differentiate until day 14 of tertiary culture. Cells were collected and analyzed by quantitative real-time RT-PCR for insulin I expression. *Expression levels are not different compared with the nondissociation (Non) controls at P > 0.05. **Expression levels are different compared with the nondissociation controls at P < 0.05.

development. To test whether the gene expression patterns of the Ngn3-EGFP⫹ cells at this stage were consistent with a pancreatic endocrine fate, gene profiling analysis was performed with the PancChip microarray platform (25) on sorted Ngn3-EGFP⫹ and Ngn3-EGFP⫺ cells collected on day 10 of tertiary culture. Four independent sorts were performed, expressed genes were directly compared between the Ngn3-EGFP⫹ and Ngn3-EGFP⫺ cells, and genes with a ⬎1.5-fold change were further analyzed. A total of 418 and 158 genes were upregulated in Ngn3-EGFP⫹ and Ngn3-EGFP⫺ cells, respectively (supplemental Table). A partial list of genes of interest is presented in Table 1. Ngn3-EGFP⫹ cells differentially expressed endocrine markers such as islet amyloid polypeptide and glucagon, along with enzymes involved in glycolysis, such as glucose phosphate isomerase. In contrast, Ngn3-EGFP⫺ cells differentially expressed genes for TABLE 1 Expression of genes in sorted Ngn3-EGFP⫹ and Ngn3-EGFP⫺ cells from day 10 tertiary culture Ngn3-EGFP⫹– enriched genes Glucagon Hexokinase 2 Glucose phosphate isomerase 1 IAPP GLUT-1 (Slc2a1) Insulin I Insulin II Ngn3

Fold enriched

Ngn3-EGFP⫺– enriched genes

Fold enriched

2.9 2.1

Hemoglobin-␤ Sox21

2.9 2.6

2.0 1.8 1.6 1.4 1.4 1.4

Amylase VEGF-A Sox2 HES1 FGF10

2.4 2.3 1.9 1.9 1.5

Data represent the mean from 4 independent sorts. Partial list of genes differentially expressed by Ngn3-EGFP⫹ or Ngn3-EGFP⫺ cells sorted from day 10 of the tertiary culture. All genes reach confidence limit ⬎80%. 923


FIG. 3. Sorting of Ngn3-EGFPⴙ cells enrich insulin-expressing cell-forming activities in a semisolid culture condition. Equal numbers (2.5 ⴛ 104 cells/well) of single-sorted day-10 tertiary Ngn3-EGFPⴙ, Ngn3-EGFPⴚ, or presort cells were cultured for 8 days in Matrigel-based semisolid media along with VEGF-A, a combination of NEA, tertiary CM, and serum. A: Photomicrographs of colonies developed in the upper layers of culture (left panel) or cells attached at the bottom of culture (right panel). Culture was initiated with Ngn3-EGFPⴙ cells. Bar length: 100 ␮m. B: Photomicrographs of cultures developed from designated populations. Bar length: 1,000 ␮m. C: The total number of colonies was counted. Data represent means ⴞ SD from quadruplicate wells. *Colony numbers are different from those in cultures initiated from Ngn3-EGFPⴚ or presort cells at P < 0.05. D: Total cells developed from the designated populations were isolated and analyzed by real-time RT-PCR for insulin I expression. *Expression levels are different from Ngn3-EGFPⴚ or presort derived cells at P < 0.05. E: Undifferentiated embryonic stem (ES) cells or cultures developed from Ngn3-EGFPⴙ population were made into a single-cell suspension, immunostained with C-peptide I antibody, and analyzed by flow cytometry. F: Total differentiated cells derived from designated populations were procured and gene expression analyzed. cDNA samples prepared from E18.5 embryonic pancreas were used as positive controls for PCR. Figure continues on next page.

exocrine (amylase), neuronal (Sox2 and Sox21), and hematopoietic cells (hemoglobin ␤). Consistent with the function of Ngn3, which induces lateral inhibition by Notch signaling in neighboring cells, the Ngn3-EGFP⫺ cells differentially expressed Hes1, a downstream signaling molecule for Notch (26,27). It should be noted that Ngn3 mRNA was increased only 1.4-fold in the Ngn3⫹ population (Table 1). This may be due to the fact that most Ngn3⫹ cells express low levels of Ngn3 (Fig. 1B). In addition, because the EGFP transcript is more stable than Ngn3 (28), some Ngn3-EGFP⫹ cells may have already committed to one of the endocrine lineages and downregulated the expression of Ngn3 (4). To confirm the differential expression of genes from PancChip analysis as well as other pancreatic related genes, quantitative real-time RTPCR was performed. The results showed that Ngn3, insulin I, hexokinase 2, glucokinase, Glut2, MafA, Pax4, Pax6, and Pdx-1 expression increased, while Hes1 expression decreased in the Ngn3-EGFP⫹ population compared with the Ngn3-EGFP⫺ cells (supplemental Fig. 3). Together, these findings demonstrate that Ngn3-EGFP⫹ cells isolated from day 10 of tertiary culture contain pancreatic-like endocrine or progenitor cells, whereas the Ngn3-EGFP⫺ population contains exocrine and hematopoietic cells. Generation of insulin-expressing colonies from Ngn3-EGFPⴙ cells isolated on day 10 of tertiary culture. As shown in Fig. 2, in order to support continuous development of insulin I– expressing cells from their progenitors, reaggregation of dissociated cells from day 10 of the tertiary culture is required. The cellular interactions required in the reaggregated cultures may contain two components: growth factors and adhesion molecules. We tested whether these cellular interactions could be re924

placed by a combination of VEGF-A (29), serum, CM from a similar stage of cells, a combination of the ␤-cell specification and differentiation factors NEA (17), and extracellular components, such as Matrigel (30). Dissociated day 10 tertiary culture cells were sorted into Ngn3-EGFP⫹ and Ngn3-EGFP⫺ populations and placed into the semisolid Matrigel-containing culture along with the aforementioned growth factors. As a control, single nonsorted (designated as “presort”) cells were cultured. After 8 days of culture, round colonies ⬃60 –100 ␮m in diameter appeared in the Matrigel layer (Fig. 3A; left panel). Individual cells in the colonies were small, dark, and not light reflective. These colonies were rarely found at the bottom of the culture wells, where fibroblasts and epitheloid cells attached (Fig. 3A; right panel). The number of these colonies was highest in cultures initiated with Ngn3-EGFP⫹ population, followed by those initiated with presort cells, and then from Ngn3-EGFP⫺ cells (Figs. 3B and C). A thin coating (⬍1 mm) of Matrigel, instead of a thick gel (2–3 mm), did not support the development of such colonies (not shown), suggesting the importance of a three-dimensional lattice for the formation of these colonies. Correlation of insulin I expression with the number of colonies generated from the three different populations was examined by real-time RT-PCR analysis. Consistent with the number of the colonies shown in Figs. 3B and C, insulin I was expressed at the highest levels in the Ngn3EGFP⫹– derived population and at the lowest levels in the Ngn3-EGFP⫺– derived population (Fig. 3D). To determine the percentage of insulin⫹ cells, all cells (attached as well as cells in the Matrigel) were dispersed into single cells and stained with antibody specific for C-peptide I. The DIABETES, VOL. 56, APRIL 2007


FIG 3—Continued

results showed that after 8 days of culture in Matrigel, ⬎6% of the total population derived from Ngn3-EGFP⫹ cells express C-peptide I (Fig. 3E). In contrast, ⬍1% of undifferentiated negative-control embryonic stem cells express C-peptide I. DIABETES, VOL. 56, APRIL 2007

To determine more complete gene expression profiles, multiplex RT-PCR was performed on total differentiated cells originating from Ngn3-EGFP⫹, Ngn3-EGFP⫺, and presort populations (Fig. 3F). Insulin II, glucagon, Sox17 (an endodermal marker [31]), hexokinase 2, Glucokinase, 925


FIG. 4. Gene expression analysis of individually picked colonies. A: A total of 16 different colonies, with the morphology similar to that shown in Fig. 3A (left panel), were handpicked and analyzed by conventional and quantitative RT-PCR. Undifferentiated embryonic stem cells were used as negative control. N.D., not done. B and C: Individual colonies were cytospun onto slides and stained with C-peptide II antibody or only with secondary antibody (Cy3 labeled). Original magnification: 400ⴛ. D: Positive control staining of adult murine islets. Original magnification, 100ⴛ. E: Double immunostaining of a colony for C-peptide II (Cy3 labeled) and glucagon (Cy5 labeled). Original magnification: 1,000ⴛ. Control staining with secondary antibodies showed negative results (not shown). Figure continues on next page. 926



FIG. 5. Linear relationship between the number of cells plated and the number of colonies generated. Graded numbers of sorted day 10 tertiary Ngn3-EGFPⴙ cells were plated per well. Insulin-expressing colonies were scored after 8 days of culture. Data represent means ⴞ SD from quadruplicate wells. Fig. 4 —Continued

MafA, and Glut2 were expressed at higher levels in Ngn3EGFP⫹-derived compared with Ngn3-EGFP⫺– derived cells. In contrast, expression of Hes1, Pdx-1, Nkx2.2, Pax6, Sox2, and ␤III tubulin was not different among the differentiated cells. Since Sox2 and ␤III tubulin are neuronalspecific markers (32,33), this result suggests that semisolid culture does not preferentially support neuronal cell expansion, although survival of Sox2- and ␤III tubulin– expressing cells is evident. Pdx-1 (11,34), Nkx2.2 (5,35), and Pax6 (36,37) are expressed by both pancreatic and neuronal cells, and thus their expression may represent the presence of pancreatic endocrine or neuronal cells derived from the Ngn3-EGFP⫹ or Ngn3-EGFP⫺ populations, respectively. Interestingly, the Ngn3 transcript was detected at a higher level in cells derived from the Ngn3EGFP⫺ population compared with that derived from Ngn3EGFP⫹ cells, indicating that some of the cells in the Ngn3-EGFP⫺ population isolated on day 10 of tertiary culture can subsequently give rise to Ngn3⫹ cells. Alternatively, the lower expression of Ngn3 in the Ngn3-EGFP⫹– derived population may indicate that the culture conditions drive progenitor cell differentiation rather than expansion, as Ngn3 expression is normally extinguished when endocrine progenitors are committed to lineagespecific cells during development (38). Amylase-expressing cells were not detected in the Ngn3-EGFP⫹– derived population. Together, these results further demonstrate the endocrine potential of the Ngn3-EGFP⫹ population, and the Matrigel-containing semisolid culture method is adequate to support endocrine-like cell development in vitro. Results shown in Fig. 3 could not prove whether individual colonies express insulin I or insulin II, and it is not clear whether they coexpress glucagon or other endocrine-related genes. Colonies were individually handpicked, and their gene expression was analyzed by quantitative RT-PCR. Undifferentiated embryonic stem DIABETES, VOL. 56, APRIL 2007

cells were used as a negative control. Among 16 single colonies analyzed, all expressed insulin II (Fig. 4A). In contrast, some, but not all, expressed insulin I (11 of 16), glucagon (7 of 16), Pdx-1 (8 of 16), Glut2 (2 of 8), and glucokinase (5 of 8), suggesting heterogeneity among these insulin-expressing colonies. To confirm protein expression, immunostaining of C-peptide II and glucagon was performed (Fig. 4B). Consistent with gene-expression analysis, all of the individual colonies examined expressed C-peptide II (Fig. 4B), and some coexpressed glucagon (Fig. 4E). These results demonstrate that those morphologically distinctive colonies (Fig. 3A; left panel) consist of insulin-expressing cells. Characterization of culture conditions for insulinexpressing colony formation. To define more completely the culture requirements for the development of insulin-expressing colonies from their progenitors or ICFUs, several parameters of the culture methods were tested. As noted, a thin coating of Matrigel did not support colony formation, suggesting the importance of the threedimensional lattice for differentiation of ICFUs. In colony assays used for hematopoietic cells, methylcellulose provides the semisolid state and three-dimensional space for TABLE 2 Requirement of Matrigel addition to methylcellulose in insulinexpressing colony formation Matrigel None 1:20

No. of colonies per total cells 1.25 ⫻ 104 2.5 ⫻ 104 ND 28 ⫾ 6

0 60 ⫾ 9

Data are means ⫾ SD from quadruplicate samples. A total of 2.5 ⫻ 104 or 1.25 ⫻ 104 ngn3-EGFP⫹ cells were plated per well in methylcellulose-based semisolid media that contain CM; a combination of nicotinamide, exendin-4, and activin BB; VEGF-A; and FCS with (1:20) or without (none) Matrigel. The number of insulinexpressing colonies was scored after 8 days of culture. ND, not determined. 927


TABLE 3 Effects of culture components on insulin-expressing colony formation Culture components CM NEA VEGF-A FCS Mean ⫾ SD

No. of colonies ⫹ ⫹ ⫹ ⫹ 41 ⫾ 7

⫺ ⫹ ⫹ ⫹ 8 ⫾ 5*

⫹ ⫺ ⫹ ⫹ 10 ⫾ 3*

⫹ ⫹ ⫺ ⫹ 42 ⫾ 8

⫹ ⫹ ⫹ ⫺ 27 ⫾ 10

Data are means ⫾ SD from quadruplicate samples. A total of 2.5 ⫻ 104 ngn3-EGFP⫹ cells were plated in Matrigel-based semisolid media containing CM; a combination of nicotinamide, exendin-4, and activin ␤B; VEGF-A; and FCS. Individual components that were omitted are indicated. The number of insulin-expressing colonies was scored after 8 days of culture. *Colony numbers are different from those in cultures containing all growth factors at P ⬍ 0.05.

colony growth (39). However, insulin-expressing colonies did not develop with methylcellulose alone (Table 2). Addition of Matrigel to methylcellulose at 1:20 vol/vol restored the formation of insulin-expressing colonies (Table 2). These results indicate that a three-dimensional lattice alone is not sufficient for colony formation and that basement membrane components in Matrigel are required to support differentiation of the ICFUs. Next, we tested which of the soluble culture components were critical for colony formation. Individual components were omitted from cultures originating from sorted Ngn3-EGFP⫹ cells and colonies scored after 8 days of Matrigel-based culture. Omission of CM and the combination of NEA, but not VEGF-A, imposed significant barriers to colony formation (Table 3). Omission of serum from culture decreased the colony number and size (not shown) to some extent. Together, these results indicate that except for exogenous VEGF-A, all other culture components were necessary to support insulin-expressing colony formation. It should be noted that although exogenous VEGF-A was not required, this molecule was expressed by the Ngn3-EGFP⫺ population (Table 1) and thus may be present in CM collected from day 10 of the tertiary culture. Finally, a linear relationship between the number of cells plated and the resulting number of insulin-expressing colonies was observed (Fig. 5). This result suggests that colony formation is independent of cell density and is consistent with the hypothesis that some progenitors from day 10 of tertiary culture can differentiate autonomously and do not require direct cell-to-cell interaction. DISCUSSION

In this report, we extend our previous findings (17) to demonstrate that embryonic stem cell– derived tertiary cultures contain ICFUs, which, when plated into semisolid media, give rise to insulin-expressing colonies. The ICFUs can be enriched by sorting of cells that express Ngn3EGFP on day 10 of the tertiary culture. It should be noted that only 0.1– 0.6% of the total sorted Ngn3-EGFP⫹ cells gave rise to insulin-expressing colonies, demonstrating that the Ngn3-EGFP⫹ population is heterogeneous. This low frequency of ICFUs also correlates with the moderate increase of insulin-expressing cells from 2% in the previous study (17) to 6% (Fig. 3E) in the current report. Thus, additional markers need to be examined to further enrich ICFUs from the Ngn3-EGFP⫹ cells. PancChip analysis demonstrated that ⬃16 genes differentially expressed in the Ngn3-EGFP⫹ cells encode for cell surface proteins (supplemental Table). It will be interesting to test whether these cell surface proteins can be used to increase the frequency of ICFUs. Until ICFUs can be sufficiently en928

riched, it will remain uncertain whether ICFUs are those that express Ngn3 or the immediate descendents of the Ngn3⫹ cells, as the EGFP transcript is more stable than Ngn3. The results demonstrate that individual insulin-expressing colonies display various gene expression profiles (Fig. 4A), suggesting that the ICFUs are heterogeneous in lineage potential or developmental kinetics. In the previous study, the presence of insulin⫹ glucagon⫹ cells was observed as early as day 11 of the tertiary culture (17). Thus, it may not be surprising that ⬃44% of the insulinexpressing colonies coexpressed glucagon (Fig. 4A). Since glucagon expression overlapped with C-peptide II in double immunostaining analysis (Fig. 4E), ICFUs isolated on day 10 of tertiary culture may contain the first wave-like endocrine progenitors. It remains to be determined whether some colonies, such as numbers 13 and 14 in Fig. 4A, that expressed insulin I, insulin II, and Pdx-1, but not glucagon, represent second wave-like pancreatic ␤-cells. Since neuronal cells express insulin II but not insulin I (40,41), it is also possible that some ICFUs are Ngn3expressing neuronal cells. For example, colony number 6 in Fig. 4A expressed only insulin II, but not insulin I, Pdx-1, or glucagon. A useful colony assay for progenitors is composed of two components. First, the culture media should be semisolid and thus limit the movements of individual progenitors, while allowing the differentiated cells derived from those progenitors to expand in space. Second, the culture components, i.e., growth factors and extracellular matrix, should be permissive for the progenitors of interest to differentiate according to their lineage potential (13). While the culture components used in this study contain many undefined factors, the insulin-expressing colony assay described in this report represents the first functional and quantitative tool by which the biology of individual ICFUs can be studied in vitro. ACKNOWLEDGMENTS

This work is supported in part by Juvenile Diabetes Research Foundation International Grant 1-2004-12 (to H.T.K.) and NIDDK UO1 DK-056947 (to K.H.K.). The authors thank Italas Georges and the Flow Cytometry Shared Research Facility at the Mount Sinai School of Medicine for sorting assistance. REFERENCES 1. Guz Y, Montminy MR, Stein R, Leonard J, Gamer LW, Wright CV, Teitelman G: Expression of murine STF-1, a putative insulin gene transcription factor, in ␤-cells of pancreas, duodenal epithelium and pancreatic exocrine and endocrine progenitors during ontogeny. Development 121:11–18, 1995 2. Offield MF, Jetton TL, Labosky PA, Ray M, Stein RW, Magnuson MA, Hogan DIABETES, VOL. 56, APRIL 2007


BL, Wright CV: PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum. Development 122:983–995, 1996 3. Krapp A, Knofler M, Ledermann B, Burki K, Berney C, Zoerkler N, Hagenbuchle O, Wellauer PK: The bHLH protein PTF1–p48 is essential for the formation of the exocrine and the correct spatial organization of the endocrine pancreas. Genes Dev 12:37523–763, 1998 4. Gradwohl G, Dierich A, LeMeur M, Guillemot F: neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas. Proc Natl Acad Sci U S A 97:1607–111., 2000 5. Sussel L, Kalamaras J, Hartigan-O’Connor DJ, Meneses JJ, Pedersen RA, Rubenstein JL, German MS: Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic ␤-cells. Development 125:22132–221, 1998 6. Sander M, Sussel L, Conners J, Scheel D, Kalamaras J, Dela Cruz F, Schwitzgebel V, Hayes-Jordan A, German M: Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of ␤-cell formation in the pancreas. Development 127:5533–540., 2000 7. Sosa-Pineda B, Chowdhury K, Torres M, Oliver G, Gruss P: The Pax4 gene is essential for differentiation of insulin-producing ␤-cells in the mammalian pancreas. Nature 386:399 – 402, 1997 8. St-Onge L, Sosa-Pineda B, Chowdhury K, Mansouri A, Gruss P: Pax6 is required for differentiation of glucagon-producing ␣-cells in mouse pancreas. Nature 387:406 – 409, 1997 9. Naya FJ, Huang HP, Qiu Y, Mutoh H, DeMayo FJ, Leiter AB, Tsai MJ: Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice. Genes Dev 11:23232– 334, 1997 10. Lee CS, Sund NJ, Behr R, Herrera PL, Kaestner KH: Foxa2 is required for the differentiation of pancreatic ␣-cells. Dev Biol 278:484 – 495, 2005 11. Ohlsson H, Karlsson K, Edlund T: IPF1, a homeodomain-containing transactivator of the insulin gene. EMBO J 12:4251– 429., 1993 12. Gu G, Dubauskaite J, Melton DA: Direct evidence for the pancreatic lineage: NGN3⫹ cells are islet progenitors and are distinct from duct progenitors. Development 129:24472– 457, 2002 13. Ogawa M: Differentiation and proliferation of hematopoietic stem cells. Blood 81:2844 –253, 1993 14. Gittes GK, Galante PE, Hanahan D, Rutter WJ, Debase HT: Lineage-specific morphogenesis in the developing pancreas: Role of mesenchymal factors. Development 122:439 – 447., 1996 15. Miralles F, Battelino T, Czernichow P, Scharfmann R: TGF-␤ plays a key role in morphogenesis of the pancreatic islets of langerhans by controlling the activity of the matrix metalloproteinase MMP-2. J Cell Biol 143:827– 836, 1998 16. Lee CS, Perreault N, Brestelli JE, Kaestner KH: Neurogenin three is essential for the proper specification of gastric enteroendocrine cells and the maintenance of gastric epithelial cell identity. Genes Dev 16:14881– 497, 2002 17. Ku HT, Zhang N, Kubo A, O’Connor R, Mao M, Keller G, Bromberg JS: Committing embryonic stem cells to early endocrine pancreas in vitro. Stem Cells 22:12051–217, 2004 18. Kaestner KH, Lee CS, Scearce LM, Brestelli JE, Arsenlis A, Le PP, Lantz KA, Crabtree J, Pizarro A, Mazzarelli J, Pinney D, Fischer S, Manduchi E, Stoeckert CJ Jr, Gradwohl G, Clifton SW, Brown JR, Inoue H, Cras-Meneur C, Permutt MA: Transcriptional program of the endocrine pancreas in mice and humans. Diabetes 52:16041– 610, 2003 19. White P, Brestelli JE, Kaestner KH, Greenbaum LE: Identification of transcriptional networks during liver regeneration. J Biol Chem 280: 37153–722, 2005 20. Manduchi E, Grant GR, McKenzie SE, Overton GC, Surrey S, Stoeckert CJ Jr: Generation of patterns from gene expression data by assigning confidence to differentially expressed genes. Bioinformatics 16:685– 698, 2000 21. Teitelman G, Alpert S, Polak JM, Martinez A, Hanahan D: Precursor cells of mouse endocrine pancreas coexpress insulin, glucagon and the neuronal proteins tyrosine hydroxylase and neuropeptide Y, but not pancreatic polypeptide. Development 118:10311– 039, 1993 22. Herrera PL: Adult insulin- and glucagon-producing cells differentiate from two independent cell lineages. Development 127:23172–322, 2000


23. Mellitzer G, Martin M, Sidhoum-Jenny M, Orvain C, Barths J, Seymour PA, Sander M, Gradwohl G: Pancreatic islet progenitor cells in neurogenin 3-yellow fluorescent protein knock-add-on mice. Mol Endocrinol 18:27652– 776, 2004 24. Blyszczuk P, Czyz J, Kania G, Wagner M, Roll U, St-Onge L, Wobus AM: Expression of Pax4 in embryonic stem cells promotes differentiation of nestin-positive progenitor and insulin-producing cells. Proc Natl Acad Sci U S A 100:998 –1003, 2003 25. Scearce LM, Brestelli JE, McWeeney SK, Lee CS, Mazzarelli J, Pinney DF, Pizarro A, Stoeckert CJJ, Clifton SW, Permutt MA, Brown J, Melton DA, Kaestner KH: Functional genomics of the endocrine pancreas: the pancreas clone set and PancChip, new resources for diabetes research. Diabetes 51:1997–2004, 2002 26. Jensen J, Pedersen EE, Galante P, Hald J, Heller RS, Ishibashi M, Kageyama R, Guillemot F, Serup P, Madsen OD: Control of endodermal endocrine development by hes-1. Nat Genet 24:36 – 44, 2000 27. Lee JC, Smith SB, Watada H, Lin J, Scheel D, Wang J, Mirmira RG, German MS: Regulation of the pancreatic pro-endocrine gene neurogenin3. Diabetes 50:928 –936, 2001 28. Gu G, Wells JM, Dombkowski D, Preffer F, Aronow B, Melton DA: Global expression analysis of gene regulatory pathways during endocrine pancreatic development. Development 131:165–179, 2004 29. Lammert E, Cleaver O, Melton D: Induction of pancreatic differentiation by signals from blood vessels. Science 294:564 –567, 2001 30. Maldonado TS, Crisera CA, Kadison AS, Alkasab SL, Longaker MT, Gittes GK: Basement membrane exposure defines a critical window of competence for pancreatic duct differentiation from undifferentiated pancreatic precursor cells. Pancreas 21:93–96, 2000 31. Kanai-Azuma M, Kanai Y, Gad JM, Tajima Y, Taya C, Kurohmaru M, Sanai Y, Yonekawa H, Yazaki K, Tam PP, Hayashi Y: Depletion of definitive gut endoderm in Sox17-null mutant mice. Development 129:23672–379, 2002 32. Uwanogho D, Rex M, Cartwright EJ, Pearl G, Healy C, Scotting PJ, Sharpe PT: Embryonic expression of the chicken Sox2, Sox3 and Sox11 genes suggests an interactive role in neuronal development. Mech Dev 49:23–36, 1995 33. Asai DJ, Remolona NM: Tubulin isotype usage in vivo: a unique spatial distribution of the minor neuronal-specific ␤-tubulin isotype in pheochromocytoma cells. Dev Biol 132:398 – 409, 1989 34. Schwartz PT, Perez-Villamil B, Riverai A, Moratalla R, Vallejo M: Pancreatic homeodomain transcription factor IDX1/IPF1 expressed in developing brain regulates somatostatin gene transcription in embryonic neural cells. J Biol Chem 275:19106 –19114, 2000 35. Briscoe J, Sussel L, Serup P, Hartigan-O’Connor D, Jessell TM, Rubenstein JL, Ericson J: Homeobox gene Nkx2.2 and specification of neuronal identity by graded sonic hedgehog signaling. Nature 398:622– 627, 1999 36. Sander M, Neubuser A, Kalamaras J, Ee HC, Martin GR, German MS: Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development. Genes Dev 11:16621– 673, 1997 37. Ericson J, Rashbass P, Schedl A, Brenner-Morton S, Kawakami A, van Heyningen V, Jessell TM, Briscoe J: Pax6 controls progenitor cell identity and neuronal fate in response to graded shh signaling. Cell 90:169 –180, 1997 38. Jensen J, Heller RS, Funder-Nielsen T, Pedersen EE, Lindsell C, Weinmaster G, Madsen OD, Serup P: Independent development of pancreatic ␣- and ␤-cells from neurogenin3-expressing precursors: a role for the notch pathway in repression of premature differentiation. Diabetes 49:163–176, 2000 39. Nakahata T, Ogawa M: Identification in culture of a class of hemopoietic colony-forming units with extensive capability to self-renew and generate multipotential hemopoietic colonies. Proc Natl Acad Sci U S A 79:38433– 847, 1982 40. Devaskar SU, Singh BS, Carnaghi LR, Rajakumar PA, Giddings SJ: Insulin II gene expression in rat central nervous system. Regul Pept 48:55– 63, 1993 41. Devaskar SU, Giddings SJ, Rajakumar PA, Carnaghi LR, Menon RK, Zahm DS: Insulin gene expression and insulin synthesis in mammalian neuronal cells. J Biol Chem 269:84458 – 454, 1994

929