Establishing clonal cell lines with endothelial-like potential from CD9(hi), SSEA-1(-) cells in embryonic stem cell-derived embryoid bodies

Abstract

Background: Differentiation of embryonic stem cells (ESCs) into specific cell types with minimal risk of teratoma formation could be efficiently directed by first reducing the differentiation potential of ESCs through the generation of clonal, self-renewing lineage-restricted stem cell lines. Efforts to isolate these stem cells are, however, mired in an impasse where the lack of purified lineage-restricted stem cells has hindered the identification of defining markers for these rare stem cells and, in turn, their isolation.

Methodology/principal findings: We describe here a method for the isolation of clonal lineage-restricted cell lines with endothelial potential from ESCs through a combination of empirical and rational evidence-based methods. Using an empirical protocol that we have previously developed to generate embryo-derived RoSH lines with endothelial potential, we first generated E-RoSH lines from mouse ESC-derived embryoid bodies (EBs). Despite originating from different mouse strains, RoSH and E- RoSH lines have similar gene expression profiles (r(2) = 0.93) while that between E-RoSH and ESCs was 0.83. In silico gene expression analysis predicted that like RoSH cells, E-RoSH cells have an increased propensity to differentiate into vasculature. Unlike their parental ESCs, E-RoSH cells did not form teratomas and differentiate efficiently into endothelial-like cells in vivo and in vitro. Gene expression and FACS analysis revealed that RoSH and E-RoSH cells are CD9(hi), SSEA-1(-) while ESCs are CD9(lo), SSEA-1(+). Isolation of CD9(hi), SSEA-1(-) cells that constituted 1%-10% of EB-derived cultures generated an E-RoSH-like culture with an identical E-RoSH-like gene expression profile (r(2) = 0.95) and a propensity to differentiate into endothelial-like cells.

Conclusions: By combining empirical and rational evidence-based methods, we identified definitive selectable surface antigens for the isolation and propagation of lineage-restricted stem cells with endothelial-like potential from mouse ESCs.

Figure 1. Derivation of E-RoSH cell lines.

(a) ESCs were plated singly on methycellulose based media to form EBs. At day 3–6, EBs were harvested, dissociated by collagenase and cultured as a monolayer on gelatinized feeder plate. RoSH-like colonies with adherent fibroblast-like cells and ring-like structures were selected and propagated on gelatinized plates to generate E-RoSH 1, 2, 3… Each of the cultures were then plated at a low density of 10–100 cells per 10 cm plate and single RoSH like colonies were picked to established sublines, E-RoSH 2.1, 2.2, 2.3. .. etc; b) A putative RoSH-like colony consisting of adherent short fibroblast-like cells with characteristic ring-like cells (inset) expanding over time; c) Alkaline phosphatase staining of E-RoSH2.1 and its parental E14 ES cells.

Figure 2. Characterisation of E-RoSH cell lines.

a) Cellular morphology of E-RoSH2.1 and RoSH2 cells in sub-confluent cultures; b,c) Pairwise comparison of global gene expression between E14 ESCs and E-RoSH2.1/E-RoSH3.2, and RoSH2 and E-RoSH2.1/E-RoSH3.2. Global gene expression analysis of E14 ESCs, E-RoSH, RoSH were performed by hybridizing total RNA from two biological samples each of E14 ESCs, E-RoSH2.1, E-RoSH3.2, and RoSH2 with Illumina BeadArray containing about 24,000 unique features; d) Biological processes in which 1115 highly expressed genes in ESCs have statistically significant higher frequencies (p<0.05). The biological processes were (left to right): neuron differentiation, neuron development, nervous system development, neuron maturation, nerve ensheathment, cellular nerve ensheathment, ionic insulation of neurons by glial cells, myelination, transmission of nerve impulse regulation of action potential, neurophysiological process, neuron morphogenesis during differentiation, neurite morphogenesis, axonogenesis, cell development, development, system development, cell maturation, cell differentiation, cellular morphogenesis, regulation of gene expression, epigenetic Imprinting, gametogenesis, morphogenesis, sexual reproduction, cell-cell signalling, cell communication, metabolism, cell organization and biogenesis, regulation of biological process, macromolecule catabolism, carbohydrate catabolism, cellular carbohydrate catabolism, monosaccharide catabolism, hexose catabolism, glucose catabolism, glycolysis, cellular macromolecule catabolism, cellular catabolism, alcohol catabolism, carbohydrate metabolism, cellular carbohydrate metabolism, monosaccharide metabolism, hexose metabolism, glucose metabolism, main pathways of carbohydrate metabolism alcohol metabolism, generation of precursor metabolites and energy, energy derivation by oxidation of organic compounds; e) Biological processes in which 1263 highly expressed genes in ESCs have statistically significant higher frequencies (p<0.05). The biological processes were (left to right): proteolysis, protein metabolism, cellular protein metabolism, cellular macromolecule metabolism, macromolecule metabolism, memory and vasculature development; f,g, h, i) Relative gene expression analysis by quantitative RT-PCR analysis. The expression level was normalized against that of ESCs and expressed as a logarithmic function; j) Western blot analysis for pluripotency-associated gene products in cell extracts of ESCs, EBs and E-RoSH2.1 cells.

Figure 3. Differentiation of E-RoSH cells.

In vitro differentiation a) Morphology of E-RoSH2.1 cell culture two weeks after plating E-RoSH2.1 cells on matrigel coated plate, b) The patent E-RoSH2.1 derived tubular structures were labeled with CFDA, a cytoplasmic green fluorescent dye (Molecular Probe, Eugene, OR) and propidium iodide, and viewed by confocal microscopy (left panel). The tubular structures were incubated with acetylated red fluoresecent diI-labelled LDL (Molecular Probe, Eugene, OR) for 24 hours and counterstained with SYTOX Green™, a green fluorescent nuclear dye (Molecular Probe, Eugene, OR) before analysis by confocal microscopy (right panel). c–e) Immunoreactivity for vWF, Tie-2 and CD34 on sections of E-RoSH2.1 derived tubular structures. vWF immunoreactivity was visualized using HRP-based detection system. Brown precipitates indicate positive staining. The nuclei were stained with Mayer's hematoxylin. CD34 and Tie-2 immunoreactivities were detected using secondary antibodies conjugated with FITC. Nuclei were counterstained with PI. f) Flow cytometry analysis of E-RoSH2.1 cells for endothelial markers before (right panels) and 60 hours after (left panels) induction of differentiation by plating cells on matrigel. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies or with secondary antibody alone. g) Gene expression during in vitro differentiation of E-RoSH2.1 cells on matrigel as measured by quantitative RT-PCR analysis. Relative gene expression is normalized against that at time 0 and expressed as a logarithmic function. h) In vivo differentiation. 1×10⁵ E-RoSH2.1 cells labeled with Qdot® nanocrystals (655 nm emission) were injected into a ESC-derived teratoma that was induced in SCID mice. Three days later, the mice were euthanized and the tumors were removed. The tumors were fixed in 4% paraformaldehyde and cryosectioned at 20 µm thickness. The sections were assayed for Tie-2 immunoreactivity using rabbit anti-Tie-2 followed by FITC-conjugated goat anti-rabbit antibody, and counterstained with DAPI. i. A typical section of a capillary plexus in the teratoma as viewed by phase contrast microscopy (top left panel), stained with DAPI, a nuclear stain (top right panel), stained with using rabbit anti-Tie-2 (bottom left panel), and cells labelled with Qdot (bottom right panel). ii. Merging of the three fluorescent stains. Yellow fluorescence indicates co-localisation of Qdot and FITC-conjugated antibody.

Figure 4. Identifying selectable surface antigens for the isolation of putative RoSH-like cells from differentiating ESCs.

a) Confocal microscopy of E-RoSH2.1 cells (top) and E14 ESCs (bottom). The cells were counterstained with DAPI, a nuclear stain after immunostaining for anti-SSE4-1 antibody conjugated with FITC and anti-CD9 antibody conjugated with PE. b) FACS analysis of i) E-RoSH2.1 cells, ii) E14 ESCs, iii) murine embryonic fibroblast (MEF) and iv) 10∶1 mixture of E14 ESCs and E-RoSH2.1 cells. The cells were labelled with anti-CD9 antibody conjugated with PE and anti-SSE4-1 antibody conjugated with FITC, and analyzed on a FACS Aria using FACS Diva software (BD Biosciences Pharmingen, San Diego, CA).

Figure 5. Derivation of E-RoSHL1 cell line from CD9^hi, SSEA1^lo cell population in EB culture.

a) One week after plating 5 day old EBs on gelatinized culture plates, the cells were harvested, labelled with anti-CD9 antibody conjugated with PE and anti-SSE4-1 antibody conjugated with FITC, and sorted. CD9^hi, SSEA1^lo cells in the P1 quadrant were selected as putative RoSH cells and plated on gelationized feeder plate; b) Morphology of semi-confluent E-RoSHL1 and E-RoSH2.1. Bar represent 15 µm; c) Pairwise comparison of global gene expression between E-RoSHL1 and E-RoSH cells. Global gene expression analysis were performed by hybridizing total RNA from two biological samples each of E-RoSHL1, E-RoSH2.1 and E-RoSH3.2 with Illumina BeadArray containing about 24,000 unique features; d) Flow cytometry analysis of E-RoSHL1 cells for endothelial markers before (righ panels) and 60 hours after (left panels) induction of differentiation by plating cells on matrigel. Nonspecific fluorescence was determined by incubation of similar cell aliquots with isotype-matched mouse monoclonal antibodies or with secondary antibody alone; e) Morphology of E-RoSHL1 culture one week after induction of differentiation by plating on matrigel. Bar represents 50 µm.