(A) 3000 of control or HES1-ER⁺ ES cells (H1) were forced to aggregate in 96-well plates in the presence of BMP4 (10 ng/ml), VEGF (5 ng/ml), Flt-3 ligand (5 ng/ml), SCF (20 ng/ml), IL-6 (5 ng/ml), and 200 nM of 4HT. 1.7 × 10⁴ dissociated EB12 cells were plated in methylcellulose supplemented with IL-3, GM-CSF, and Epo for hematopoietic colony assay. (B–C) Whole EBs (day 12) were transferred onto OP9 cells in the absence of 4HT. 7 days later, cells were harvested for FACS analysis (C). Human cells were gated based on the positive GFP and Tra-1-85 staining shown in (B).

(A–C) Gene expression of trophectoderm markers in day 7 EBs (H1). Control EBs (n=3); DNMAML⁺ EBs (n=4). (D) TROMA-I staining in day 14 EBs (H9). (E) Day 12 EBs (H9) were dissociated into single cells and stained with either TROMA-1 or β-hCG antibody for FACS analysis. (F–H) 10⁵ of hES cells were plated in the presence of BMP4 (50 ng/ml) for 10 days. The supernatents were collected for hCG measurement. The cell lines examined were cells transduced with parental, ICN1-ER (E), HES1-ER, ΔBHES1-ER (F), and DNMAML lentivector (G). (E) 200 nM of 4HT was added during the 10-day culture to induce functional ER fusion protein expression in H1 cells. Cells were also treated with either DMSO (empty bar) or 10 μM of GSI-2 (filled bar). 3 independent experiments were performed and data were grouped here.

(A–B) H9 ES cells were induced to differentiate into EBs in the presence of 20% FBS for 6 days. Continuous treatment of GSI-18 (20μM) or DMSO was started on Day 0 (A) or Day 2 (B). The mRNA level of the Day 6 EB cells was measured by qRT-PCR and expressed relative to the mean of the teratoma (which was normalized to 1). * indicates Ct=0 (undetected) after 40 cycles. (C–E) Blocking Notch signaling preserves uncommitted hES cells during differentiation in vitro. (C) Day 6 whole EBs (H9) were stained for Oct4 antibody. (D) 1000 of dissociated day 12 EB cells were cultured on MEFs in 96-well plate for 7 days. Colonies were further stained for AP. The number of colonies derived from control EB cells were normalized to 1. (E) Day 12 EBs (H1) were cultured on plain OP9 cells for 6 days. Cell mixtures were then isolated and co-stained with anti-Tra-1-60 (ES cell marker) and Tra-1-85 (human cell marker) antibodies.

Control or DNMAML transduced hES cells were injected intramuscularly into the SCID/Beige mice. Palpable teratoma was dissected into pieces for various assays. The induction of CBFRE-Luc activity by ICN1 was examined in dissociated DNMAML⁺ tumor cells. The DNMAML tumor cells that showed restored ICN1-induced CBF1 activity was designated as silenced DNMAML cells (DN-si). (A) qRT-PCR analysis of gene expression in teratoma cells (H9). The gene expression level of control teratoma cells was normalized as 1. (Control=3, DN⁺=4, DN-si=3). (B) Dissociated teratoma cells (H1) were immunostained for Tra-1-60 and Tra-1-85 antibodies and analyzed by FACS. Human cells were gated based on positive Tra-1-85 staining. (Control=5, DNMAML=6). (C) 10⁶ of teratoma cells (H1) were replated onto MEFs in 6-well plate, and cultured in ES media for 7–9 days. The survived cells were then stained for AP. (Control=8, DNMAML=9) (D) In secondary colony assay, teratoma-derived primary colonies were collected and replated (ratio of 1:1) onto new MEF cells. The cells were cultured for 7 more days and stained with AP. 3 control and 3 DNMAML teratoma samples were used in this assay. (* indicates no colonies formed). (E) Characterization of a tertiary colony derived from DNMAML-transduced teratoma cells.

(A–C) hES cells (H9) were cultured in the presence of 20% FBS to induce EB formation. (A) EBs of day 3 were shown. (B) The percent of viable cells was determined. Two independent experiments were performed, with duplicates per group in each experiment. (C) Gene expression in day 5 H9 EB cells. The expression level of each gene in teratoma was arbitrarily set as 1. (D–F) 1000 H1 cells were forced to aggregate after centrifugation in 96-well plate. 1–2 EBs were formed in each well after 12-day culture (D). (E) ES cells were differentiated in the presence or absence of 10 μM of GSI-2. 200 nM of 4HT were added into all cultures to induce the functional HES1 expression. Cystic and total day 12 EBs were documented. Three independent experiments were grouped to present here. (F) Enforced expression of HES1 partially restored the normal EB (day 12) differentiation blocked by GSI. The expression of each gene in DMSO treated control EB was normalized to 1. (* indicates Ct=0 after 40 cycles)

(A) H9 ES cells were overlaid on 3T3 cells in the presence of MEF CM for 4 days. On Day 4, 1/15 of the trypsinized cells were replated on MEFs in 96-well plate for quantitative colony formation (n=8 per group). A representative data set from 3 independent experiments is shown. (B–C) 10⁵ of GFP⁺ hES cells (H1) were plated in the 6-well plates pre-coated with 10 μg/ml of DLL1^ext-IgG or control human IgG proteins. On Day 8, the percentage of Oct4⁺ cells was measured by FACS (B). The harvested cells were further plated onto MEFs at a density of 2 × 10⁴ per well for colony forming assay (C) (n=8). Three independent experiments were performed. (D–E) H9 cells stably transduced with parental or ICNER lentivector were cultured on Matrigel in the absence or presence of 4HT (200 nM) for 6 days. 2000 of trypsinized ES cells were replated for colony forming assay (D). In an independent experiment, 8-day treated cells were stained for Oct4 antibody (E). (F–G) Control or HES1-ER⁺ hES cells (H1) were cultured on Matrigel in the presence or absence of 4HT (200 nM) for two passages. Then dissociated cells were plated in triplicates at a density of 10⁴ in 12-well of Matrigel coated plates. After 4 days of culture, colonies were stained for the AP (F) and counted (G). 3 independent experiments were performed.

(A) Undifferentiated ES cells (H1) were co-cultured with MEFs in the presence or absence of GSI-18 (10 μM) for 3 days. Dissociated teratoma-derived cells were cultured in the same manner. The cells were then transfected with CBFRE-Luc construct together with β-gal expression plasmid overnight and subjected for Luc assays. (B) Gene expression of representative Notch downstream target in H9 hES cells transduced with either DNMAML or the parental control lentivector. The relative level of each gene in the teratoma was defined as 100%. * and ^ indicate very low level of HES5 expression. (C) H1 hES cells were overlaid on 3T3 cells expressing DLL1 or JAG1 for 1 day before transfection. CBFRE-Luc with full-length Notch1 expression vector (pBOS-N1^FL) or parental vector was co-transfected into cells. (D) CBFRE-GFP was electroporated into H1 cells. Except for an aliquot of cells continuously cultured under the ES culture condition, remaining transfected cells were induced to form EB in the presence of 20% FBS. (E–F) EBs derived from H9 hES cells were collected daily for protein isolation. The expression of cleavage Notch1 (cN1) (E) and DLL1 protein (F) was detected by western blot. The proteins of mouse E14 germinal eminence were served as positive control for cleavage Notch1 (E). (G–H) cN1(G–H), Oct4 (G), and Nestin (H) expression at the single cell level (H1 hES cells) two days after EB formation. Similar results were obtained in H9 EB cells. Scale bar: 100 μm. (I) Self-renewal capacity in hES cells in which endogenous Notch activity was blocked by either DNMAML or GSI-18. hES cells were plated in 24-well Matrigel-coated plate at a density of 3 × 10⁴ per well and cultured in the presence or absence of GSI-18 (20μM) for 1 passage of 5 days. Cells were counted at the end of each passage. The frequency of colony-forming undifferentiated hES cells (AP⁺) per 1000 harvested cells were also assayed after plating them on MEFs in 96-well plates for 5–6 days. Relative levels of colony-forming (CF) abilities were calculated by multiplying frequency of AP⁺ colonies by the total cell number before colony assay. The level of CF ability of either DNMAML⁺ or GSI treated cells was expressed relative to the mean of the control ES cells, which was arbitrarily defined as 1. These experiments were repeated 3–6 times and the pooled results are plotted (n≥15). (J) The percentages of Oct4⁺Tra-1-60⁺ control or DNMAML⁺ H1 cells were determined by FACS analysis. A representative data set from 3 independent experiments was shown.