Subpopulations of Human Embryonic Stem Cells With Distinct Tissue-Specific Fates Can Be Selected From Pluripotent Cultures

Abstract

Directed differentiation of human embryonic stem cells (hESCs) has generated much interest in the field of regenerative medicine. While subpopulations of hESCs within pluripotent cultures have been identified based on expression of specific surface antigens, their significance and fates are not well understood. To determine whether such subpopulations indicate specific tissue fates or represent stochastic antigen distributions within proliferating cultures, we isolated CD133⁺ or CD135⁺ hESCs from proliferating cultures constitutively expressing enhanced green fluorescent protein (GFP), and co-cultured these with unselected GFP⁻ hESCs. After passage in culture, GFP⁺ hESCs reanalyzed for the persistence of CD133 or CD135 expression, as well as other surface antigens (Tra-1-60, SSEA-4, FGFR-1), demonstrated that these two subpopulations continued to express CD133 or CD135 over serial passage, and that CD133⁺ hESCs were enriched for SSEA-4 expression as well. Upon differentiation in vitro, CD133⁺GFP⁺ hESCs gave rise solely to ectoderm, as detected by expression of nestin. Tissues representing endoderm (α-fetoprotein⁺) and mesoderm (smooth muscle actin⁺) were not seen among GFP⁺ tissues. In contrast, selection against CD133 gave rise almost exclusively to mesoderm and endoderm. In contrast, CD135⁺GFP⁺ hESCs gave rise to tissues representing all three embryonic germ layers, and were virtually indistinguishable from CD135⁻-derived tissues. Similar results were obtained by in vivo differentiation in teratomas. These data establish that subpopulations of proliferating hESCs whose tissue fate is predetermined exist, and challenge the notion that all cells within proliferating hESC cultures are truly “pluripotent.” This co-culture approach also will enable identification of other distinct hESC subpopulations, and selection for these should prove valuable in generating tissue-specific reagents for cell-based therapy.

Introduction

Human embryonic stem cells (hESCs) have an unlimited capacity for self-renewal and the ability to terminally differentiate into cell types in vitro and in vivo derived from all three embryonic germ layers. An important step toward advancing their use in cell-based therapies for human disease, however, is their directed differentiation into specific lineages. To date, most hESC lines have been characterized by their expression of cell surface antigens [1]. These studies have identified a battery of glycolipids and glycoproteins that are found on a high percentage of undifferentiated hESCs, including for example the stage-specific antigens, SSEA-3 and SSEA-4, and the keratin sulfate-related antigens, Tra-1-60 and Tra-1-81, among others [2]. These antigens are commonly used to assess the pluripotency of hESCs, for within days upon the induction of differentiation their expression dramatically decreases [3].

Other surface antigens, such as CD133, CD135, FGFR, CD117, SSEA-1, and CD130, have been shown to be expressed on a subfraction of proliferating hESCs in culture across various derived lines [2], suggesting that these may constitute functional subpopulations within “pluripotent” hESC cultures. The concept that such “private” markers may confer lineage specificity within the culture is supported by the observation that some of these markers also are expressed on tissue-specific progenitors giving rise to hematopoietic, neural, and endothelial derivatives [2,4]. In addition, one previous report demonstrated developmental differences among undifferentiated hESCs expressing GCTM-2 [5]. Alternatively, in some cases the expression of such surface markers may be a stochastic process where within an asynchronously dividing population of hESCs, such markers are expressed in a cell cycle phase-dependent or other such manner.

To test the significance of two of these markers, CD133 and CD135, we performed fate mapping experiments in culture using a novel fluorescence-tagged co-culture system. We establish that subpopulations of hESCs possessing a predetermined lineage fate indeed exist within proliferating cultures, suggesting that the current paradigm of proliferating cultures as consisting of uniformly “pluripotent” cells should be reconsidered. We also have developed a selection and co-culture method to identify, isolate, and expand these hESCs subpopulations, which will provide a valuable tool for identifying and generating tissue-specific reagents for cell-based therapy.

Materials and Methods

hESC culture and differentiation

The H9 hESC line was maintained on irradiated mouse embryonic fibroblast feeder cells in a medium comprised of Knockout Dulbecco’s Modified Eagle Medium (DMEM) (Invitrogen, Carlsbad, CA) supplemented with 20% Knockout Serum Replacement (Invitrogen), 2 mM glutamine, 0.1 mM nonessential amino acids, 0.1 mM β-mercaptoethanol, and 15 ng/mL recombinant human FGF-basic (R&D Systems, Minneapolis, MN). All cultures were routinely screened for mycoplasma, and normal karyotype was monitored by array comparative genomic hybridization using published methods [6]. Differentiation was initiated by human embryoid body (hEB) formation in suspension as previously described [7]. Briefly, colonies of hESCs were dissociated into small clusters by exposure to Collagenase IV (Sigma-Aldrich, St. Louis, MO), then allowed to differentiate in a medium comprised of Knockout DMEM (Invitrogen) supplemented with 20% Defined Fetal Bovine Serum (Hyclone), 2 mM glutamine, 0.1 mM nonessential amino acids, and 0.1 mM β-mercaptoethanol. After 1 week in suspension, hEBs were attached to gelatin-coated 12-well culture plates and allowed to differentiate for an additional 14 days. For co-culture experiments, proliferating enhanced green fluorescent protein (GFP⁺) hESCs were sorted by FACS on the basis of CD133 or CD135 expression as described below, and ≥5 × 10⁵ sorted hESCs were then added directly to a plate containing proliferating wild-type (GFP⁻) colonies that had been passaged 24 h prior. The co-cultures were incubated overnight to allow the selected GFP⁺ cells time to integrate into the GFP⁻ colonies before changing the medium. The co-cultures were then passaged 48 h after addition of sorted cells, and then expanded and differentiated as described earlier. All differentiation experiments were performed at least three times in triplicate.

Flow cytometry and FACS

Proliferating hESC colonies were scraped from six-well plates in cold phosphate-buffered saline (PBS). After centrifugation, cell pellets were briefly exposed to 0.05% trypsin-EDTA for 5 min at 37°C. Cells were then diluted into cold blocking buffer comprised of PBS, 20% horse serum, and 0.5 mM EDTA, and incubated on ice in blocking buffer with the appropriate antibodies. Conjugated primary antibodies used were APC-conjugated mouse antihuman CD133 (Miltenyi Biotec 130–090-826) at 1:10 dilution, APC-conjugated mouse antihuman CD135 (Caltag Labs MHCD13505) at 1:10, and PE-conjugated mouse antihuman SSEA-4 (R&D Systems FAB1435P) at 1:40. Unconjugated primary antibodies used included rabbit anti-FGFR-1 (Sigma F5421) with goat anti-rabbit-APC-Cy7 (Santa Cruz Biotechnology sc-3847) secondary antibody at 1:200 and mouse anti-Tra-1-60 (Chemicon MAB4360) with goat anti-mouse-APC (Santa Cruz Biotechnology sc-3818) secondary antibody at 1:200. After washing with blocking buffer, hESCs were analyzed on an LSR II flow cytometer (BD Biosciences, San Jose, CA) or sorted on a FACSAria (BD Biosciences) using standard filter sets as previously described [8,9]. Cells were co-stained with 1 µg/mL Hoechst 33248 for DNA content detection and 1 µg/mL propidium iodide to distinguish between live and dead cells.

Immunocytochemistry

The lineage fate of differentiated, co-cultured hESCs was determined by staining 21-day-old, adherent differentiating hEBs attached to coverslips in 12-well culture plates. Coverslips were fixed with 4% paraformaldehyde, then permeabilized with 50% methanol/50% PBS, 100% methanol, 50% methanol/50% PBS/0.1% Triton X-100, and finally PBS/0.1% Triton X-100. hEBs were incubated with blocking buffer (PBS/10% horse serum/1% BSA/0.1% Triton X-100), then with primary antibody (1–5 µg/mL) in blocking buffer. Primary antibodies used were mouse antihuman α-fetoprotein (Sigma A8452), mouse antihuman nestin (R&D Systems MAB1259), and mouse antihuman smooth muscle actin (R&D Systems MAB1420). Coverslips were washed with blocking buffer, incubated with 1:500 dilution of goat anti-mouse Alexa Fluor 594 (Invitrogen A20185), washed with PBS/1% Triton X-100, and then analyzed by confocal microscopy using a Zeiss LSM510 META with NLO at 63× magnification. Quantitation was performed by scoring the number of nuclei in α-fetoprotein-, smooth muscle actin-, or nestin-reactive GFP⁺ cells/total number of nuclei in GFP⁺ cells per ≥100 fields. Each experiment was performed at least three times in triplicate.

To localize CD133⁺ and CD135⁺ hESCs within proliferating and differentiating colonies of GFP⁺ cells, cells grown on coverslips were fixed with 4% paraformaldehyde, incubated with blocking buffer (PBS/150 mM sodium acetate, pH 7/0.1% nonfat dried milk), and stained with antibody against CD133 or CD135 (10 µg/mL) in blocking buffer. Coverslips were then washed with blocking buffer, stained with antibodies against Oct4 (10 µg/mL) and βIII-tubulin (5 µg/mL) in blocking buffer made 0.1% Triton X-100, counterstained with DAPI, and analyzed by confocal microscopy. For these analyses, monoclonal antibodies against CD133 (Miltenyi Biotec 130–092-395) or CD135 (Caltag MHCD13500), Oct4 (Santa Cruz Biotechnology sc-5279), and βIII-tubulin (R&D Systems MAB1195) were labeled with Alexa Fluor Monoclonal Antibody Labeling Kits (647 nm, 647 nm, 546 nm, and 594 nm, respectively; Molecular Probes Inc., Eugene, OR).

Teratomas

To form teratomas, 5 × 10⁵ co-cultured hESCs were mixed with an equal volume of 1 mg/mL Phaseolus vulgaris lectin (PHA-P L1668; Sigma), pelleted, and incubated in growth medium overnight at 37°C, 5% CO2 in a 0.4 µm MILLICELL (Millipore). At least two cell pellets were grafted under each kidney capsule of 8-week-old male CB17 SCID-Beige mice (three mice per co-culture) using standard techniques [10]. Transplanted cells formed teratomas in the recipients and were analyzed 10 weeks after grafting. Teratomas were fixed in 10% buffered formalin, embedded in paraffin, and 5 µm sections were stained with purified polyclonal rabbit anti-GFP (Molecular Probes A11120) at 1:1500 and biotinylated goat anti-rabbit IgG (Vector BA-1000). Slides were developed using the VECTASTAIN Elite ABC kit (Vector) and counterstained with hematoxylin and eosin to identify tissue structures.

Statistics

For quantitative studies involving continuous variables, the covariance of the means was compared using Student’s t-test. For tests with an output in discrete variables, and for comparison between groups, Chi-square analysis was used. A value of P < 0.05 was considered significant. All analyses were performed using SPSS v.16 (SPSS, Inc.) for Macintosh.

Results

Fate mapping of human embryonic stem cell subpopulations can be accomplished using a fluorescence-tagged co-culture system

Since the first analyses of hESC lines, it has been shown that within proliferating, presumably pluripotent cultures, there exist subpopulations of cells distinguished by the expression of specific surface markers [2]. In contrast to markers of pluripotency such as SSEA-3, Tra-1-60, Tra-1-81, and CD9, which are expressed on >80% of proliferating hESCs from various lines [2], the significance of these “private” markers, such as CD133 and CD135, and the subpopulations they identify is not known. To investigate the relevance of these subpopulations, we initially isolated these specific hESCs by fluorescence-activated cell sorting (FACS), and attempted to regrow these cells using standard culture techniques on mouse embryonic feeder cell layers. This resulted in cell death or very slow cell growth, suggesting that isolating such subpopulations away from the mixed culture milieu was detrimental to their continued expansion and subsequent differentiation.

To address this limitation without the use of reagents that might alter subpopulation dynamics [11], we developed a method utilizing fluorescence tagging to monitor specific subpopulations of hESCs in a co-culture system (Fig. 1A). We used an H9-derived hESC line that constitutively expresses GFP from a ubiquitin C promoter, and previously has been shown to proliferate and differentiate into all three embryonic germ layers similar to the parent wild-type H9 line [6]. The GFP⁺ hESCs were first sorted based on surface marker expression by FACS (Fig. 1B). The freshly sorted, GFP⁺ marker⁺ cells were then added to dispersed colonies of proliferating wild-type H9 (GFP⁻) hESCs and allowed to reform chimeric (ie, GFP⁺ marker⁺/GFP⁻) colonies under standard hESC growth conditions (Fig. 1C). We observed that the chimeric colonies exhibited proliferation rates and colony size and morphology that were indistinguishable from wild-type colonies (data not shown), and readily formed chimeric hEBs in suspension (Fig. 1D). These results demonstrated that fluorescence-tagged co-culture provides a system for expanding and differentiating hESC subpopulations in vitro that would allow for subsequent determination of lineage fate.

FIG. 1.

A fluorescence-tagged co-culture system for determining fate of human embryonic stem cell (hESC) subpopulations. (A) Overview of experimental approach. Tag hESCs with green fluorescent protein (GFP) refers to the establishment of hESCs stably expressing a UbiC-GFP cassette. (B) CD133⁺ cells were isolated by FACS from a proliferating culture of GFP-expressing hESCs (blue arrow). Cntrl Ab, isotype control antibody; CD133, APC-conjugated CD133 antibody. (C) GFP⁺/CD133⁺ cells were recultured with wild-type hESCs (GFP⁻) to produce “chimeric” colonies (top). Phase-contrast image shows feeder cells as well as the presence of GFP⁻ wild-type hESCs within chimeric colonies (bottom). Bar = 100 µm. (D) Expanded colonies were differentiated into “chimeric” human embryoid bodies (hEBs) (top). Phase-contrast image shows wild-type hEBs as well as the presence of GFP⁻ wild-type cells (red arrows) within chimeric hEBs (bottom). Bar = 100 µm.

Proliferating subpopulations of CD133⁺ human embryonic stem cells give rise to ectodermal derivatives, while CD135⁺ subpopulations contribute to all three embryonic germ layers

Among the various markers that have been identified on the surface of subpopulations within hESC cultures from various lines, several have consistently been reported across different hESC lines, including CD133, CD135, CD117, FGFR-1, SSEA-1, and CD130 [2]. For the purpose of proving our fluorescence-tagged co-culture system, we chose to focus on the hESC surface antigens CD133 and CD135. Although the biological role of CD133 is not well understood, the epitope has served as a useful marker for the isolation of progenitor cells from various lineages [12]. CD135, also known as FLT3 (fms-like tyrosine kinase-3), is a receptor tyrosine kinase known to be expressed on hematopoietic progenitors [13].

GFP⁺ hESCs were positively and negatively selected for either CD133 or CD135 by FACS. In our H9-derived, GFP-expressing hESC line [6], we routinely observed CD133 expression on ∼60% of cells and CD135 expression on ∼30% of cells (data not shown), which were within the range of percentages reported by others for various hESC lines [12,13]. The sorted hESCs were expanded in co-culture with wild-type (GFP⁻) H9 cells as described earlier (Fig. 1), then allowed to differentiate in suspension to form hEBs. These were plated and allowed to differentiate for a total of 21 days. To determine the fate of CD133⁺ and CD135⁺ cells, CD133⁺GFP⁺ versus CD133⁻GFP⁺ and CD135⁺GFP⁺ versus CD135⁻GFP⁺ chimeric hEBs were stained with lineage-specific antibodies (Figs. 2 and 3). α-Fetoprotein was used to detect endoderm, nestin was used to detect ectoderm, and smooth muscle actin was used to detect mesoderm. Confocal microscopy demonstrated that both CD135⁺GFP⁺ and CD135⁻GFP⁺ differentiated into all three embryonic germ layers, with a distribution of ∼25% endoderm, 40% ectoderm, and 35% mesoderm (Fig. 2). These distributions were virtually identical to those obtained for wild-type H9 hESCs (data not shown) suggesting that although CD135 may be found on cells of mesodermal origin [13], its presence on proliferating hESCs does not predict a specific fate.

FIG. 2.

Proliferating CD135⁺ human embryonic stem cells (hESCs) differentiate in vitro into tissues representing all three embryonic germ layers. (A) Human embryoid body (hEB) derived from CD135⁺GFP⁺ and CD135⁻GFP⁺ hESCs were differentiated in adherent cultures for 21 days and stained with antibodies against α-fetoprotein, smooth muscle actin, and nestin (red), and nuclei were counterstained with DAPI (blue). Representative data are shown. Co-localization of GFP (green) with lineage markers, determined by confocal microscopy, demonstrated that CD135⁺GFP⁺ as well as CD135⁻GFP⁺ hESCs ultimately gave rise to all three embryonic germ layers. Note the heterogeneous size range among smooth muscle actin-expressing cells, which occurred in both CD135⁺/GFP⁺ and CD135⁻/GFP⁺ cultures, compared to the homogeneous size of α-fetoprotein- and nestin-expressing cells. Bar = 10 µm. (B) Percentage of nuclei in cells co-expressing GFP and specific antibody/total number of nuclei in GFP⁺ cells in ≥100 fields. Data shown are mean ± SEM. There were no significant differences between cultures differentiated from CD135⁻GFP⁺ compared to CD135⁺GFP⁺ hESCs.

FIG. 3.

Proliferating CD133⁺ human embryonic stem cells (hESCs) differentiate in vitro into neuroectoderm. (A) Human embryoid body (hEB) derived from CD133⁺GFP⁺ and CD133⁻GFP⁺ hESCs were differentiated in adherent cultures for 21 days and stained with antibodies against α-fetoprotein, smooth muscle actin, and nestin (red), and nuclei were counterstained with DAPI (blue). Representative data are shown. Co-localization of GFP (green) with lineage markers, determined by confocal microscopy, demonstrated that CD133⁺GFP⁺ hESCs gave rise almost exclusively to ectoderm (ie, co-localization of nestin [red] with GFP [green]), while CD133⁻GFP⁺ hESCs differentiated into endoderm and mesoderm (ie, co-localization of α-fetoprotein or smooth muscle actin [red], but not nestin, with GFP [green]). Bar = 10 µm. (B) Percentage of nuclei in cells co-expressing GFP and specific antibody/total number of nuclei in GFP⁺ cells in ≥100 fields. Data shown are mean ± SEM. Cultures differentiated from CD133⁺GFP⁺ hESCs resulted in a significantly greater percentage of nestin-expressing cells, compared to cultures from CD133⁻GFP⁺ hESCs. Conversely, cultures derived from CD133⁻GFP⁺ hESCs differentiated into significantly more α-fetoprotein- and smooth muscle actin-expressing cells, compared to cultures from CD133⁺GFP⁺ hESCs. **P < 0.0001.

In contrast, a specific lineage profile was obtained for differentiated CD133⁺GFP⁺ and CD133⁻GFP⁺ hESCs (Fig. 3). CD133⁺GFP⁺ hESCs demonstrated a threefold increase in the number of differentiated cells associated with expression of the ectodermal marker, nestin, compared to CD135⁺GFP⁺ or CD135⁻GFP⁺ cells (0.95 ± 0.04 vs. 0.38 ± 0.08 or 0.41 ± 0.09; P < 0.0001). Consistent with this finding, CD133⁻GFP⁺ cells demonstrated a 3-fold decrease in the number of differentiated hESCs associated with nestin expression compared to CD135⁺GFP⁺ or CD135⁻GFP⁺ cells (0.12 ± 0.05 vs. 0.38 ± 0.08 or 0.41 ± 0.09; P < 0.001). α-Fetoprotein expression was not found among CD133⁺GFP⁺-differentiated hESCs (Fig. 3). In contrast, differentiated CD133⁻GFP⁺ cells demonstrated a 2-fold enrichment in expression of this endodermal marker compared to CD135⁺GFP⁺ or CD135⁻GFP⁺ cells (0.41 ± 0.09 vs. 0.26 ± 0.08 or 0.24 ± 0.08; P < 0.01).

Expression of smooth muscle actin among CD133⁺GFP⁺-differentiated hESCs was very similar to that of α-fetoprotein (Fig. 3). Only 5% of differentiated CD133⁺GFP⁺ hESCs expressed this mesodermal marker, compared with its expression in 47% of CD133⁻GFP⁺ cells (0.05 ± 0.04 vs. 0.47 ± 0.08; P < 0.0001). In fact, the expression of smooth muscle actin in differentiated CD133⁻GFP⁺ cultures constituted a 1.3-fold enrichment compared to its expression in CD135⁺GFP⁺ or CD135⁻GFP⁺ cells (0.47 ± 0.08 vs. 0.36 ± 0.08 or 0.35 ± 0.05; P < 0.01).

To explore the stringency of the developmental restriction observed in the CD133⁺ subpopulation, we performed similar co-culture experiments with CD133⁺GFP⁺, CD133⁻GFP⁺, CD135⁺GFP⁺, and CD135⁻GFP⁺ cells grown under conditions previously described for directing hESC differentiation toward neuroectoderm or primitive endoderm [14,15]. We observed no significant differences in the ability of CD135⁺GFP⁺ or CD135⁻GFP⁺ cells to respond to directed differentiation (Supplementary Fig. 1; Supplementary materials are available online at http://www.liebertpub.com), supporting the observation that CD135 does not define a hESC subpopulation with a specific lineage fate. CD133⁺GFP⁺ cells, however, preferentially differentiated into neuroectoderm regardless of differentiation conditions (Supplementary Fig. 1), suggesting that the lineage fate of this subpopulation is restricted.

It is known that low levels of spontaneous differentiation can occur within hESC cultures grown under proliferation conditions, and that the cells toward the periphery of proliferating colonies can express early markers of specific embryonic germ layers. To rule out the possibility that the CD133⁺ and/or CD135⁺ hESCs examined in these experiments represent such differentiating cells, we determined the spatial localization of CD133⁺ and CD135⁺ hESCs within the proliferating colonies from which they were sorted for co-culture experiments (Fig. 4). Cells expressing either cell surface marker were observed throughout the proliferating colonies examined, and not just at the periphery (Fig. 4A,C). In addition, CD133 and CD135 expression co-localized with Oct4, a nuclear transcription factor associated with pluripotency (Fig. 4A–D). Similarly, in differentiating colonies where the neuroectodermal marker, βIII-tubulin, was observed toward the colony periphery during early differentiation or throughout the colony later in differentiation, CD133⁺ and CD135⁺ expression co-localized with Oct4 in cells that did not express βIII-tubulin (Fig. 4B,D).

FIG. 4.

CD133⁺ and CD135⁺ human embryonic stem cells (hESCs) express Oct4 but not βIII-tubulin in proliferating colonies in vitro. Colonies of proliferating undifferentiated GFP⁺ hESCs (A and C) typical of the ones used for sorting (see Fig. 1), as well as colonies of GFP⁺ hESCs grown in medium with defined fetal bovine serum and allowed to differentiate spontaneously (B and D), were stained with antibodies against Oct4 (right panels: yellow), βIII-tubulin (left panels: red), and CD133 (left panels: green; A,B) or CD135 (left panels: green; C,D). Typical results are shown. CD133⁺ (A) and CD135⁺ (C) cells were distributed throughout proliferating colonies and co-localized with Oct4. βIII-tubulin expression was not apparent in these colonies. In differentiating colonies, βIII-tubulin was expressed toward the periphery of colonies in early stages of differentiation (D) and throughout the culture at later stages (B). However, CD133⁺ (B) and CD135⁺ (D) cells were distributed within patches of Oct4-expressing cells, and were not observed to co-localize with βIII-tubulin. Bar = 100 µm.

To determine the fate of CD133⁺ and CD135⁺ cells in an independent system not subject to the influence of in vitro culture conditions, we evaluated the differentiation of these same co-cultures in a teratoma formation assay (Fig. 5). Teratomas formed from co-cultures of CD133⁺GFP⁺ and wild-type hESCs demonstrated GFP-expressing cells solely within tissues of neuroectodermal origin (Fig. 5G–L), while teratomas formed from co-cultures of CD135⁺GFP⁺ and wild-type hESCs showed GFP-expressing cells within tissues derived from all three embryonic germ layers (Fig. 5A–F). Together with the results of differentiation in vitro, these studies demonstrated that some hESC subpopulations, as defined by surface marker expression such as CD133, are predestined toward specific lineage fates, while others such as CD135⁺ hESCs are not.

FIG. 5.

Proliferating CD133⁺ human embryonic stem cells (hESCs) differentiate into neuroectoderm in vivo. Proliferating co-cultures of CD135⁺GFP⁺ or CD133⁺GFP⁺ with wild-type hESCs were used to form teratomas by renal capsule grafting. Teratomas were sectioned, stained with H&E to identify embryonic tissues, and counterstained with anti-GFP antibody (brown) to localize derivatives of CD135⁺GFP⁺ or CD133⁺GFP⁺ cells within tissues. (A–F) CD135⁺-derived cells, as determined by immunostaining with anti-GFP antibody, were observed in tissues arising from embryonic ectoderm (A and B), mesoderm (C and D), and endoderm (E and F). (G–L) In contrast, CD133⁺-derived cells were observed in tissues arising from embryonic ectoderm, specifically neural epithelium (H) and nascent neural tube-like structures (G), but not in tissues derived from mesoderm (I,J) or endoderm (K and L). (A) Primitive endodermal structure subjacent to nascent neural tube. (B) Primitive squamous epithelium. (C) Nascent renal tubules and glomeruli within bed of primitive renal epithelium. (D) Cartilage surrounded by capsule of condensed mesenchyme. (E) Glandular intestinal structure surrounded by smooth muscle. (F) Endothelial structure within loose mesenchyme. (G) Nascent neural tube structure. (H) Endodermal structures, characterized by surrounding clear cells, within a bed of primitive neural epithelial cells. (I) Nascent renal tubules and glomeruli surrounded by primitive renal epithelium and loose mesenchyme. (J) Two pools of cartilage surrounded by nascent fibrous capsule, a ring of smooth muscle, and a primitive intestinal structure. (K) Primitive endoderm in loose mesenchyme. Bar = 100 µm (all panels except E and J). Bar = 200 µm (panels E and J).

CD133⁺ and CD135⁺ subpopulations of proliferating human embryonic stem cells stably express defining surface markers during proliferation and expansion

Having demonstrated that proliferating hESCs contain subpopulations that may be destined for specific lineage fates, we wanted to determine whether these subpopulations retain these defining markers in proliferating cultures, or if the expression of these markers at any given time point represents a stochastic process. To accomplish this, we assayed sorted CD133⁺GFP⁺ versus CD133⁻GFP⁺ and CD135⁺GFP⁺ versus CD135⁻GFP⁺ hESCs for the presence of SSEA-4 and Tra-1-60, surface markers associated with pluripotency and expressed on the majority (>90%) of proliferating hESCs across individual hESC lines [2]. We also examined these cells for FGFR-1, which is expressed on a small percentage of cells within proliferating hESC cultures [2].

The sorted CD133⁺GFP⁺, CD133⁻GFP⁺, CD135⁺GFP⁺, and CD135⁻GFP⁺ hESCs were expanded for three passages as described earlier and compared to zero-passage, unsorted GFP⁺ hESCs by flow cytometry (Fig. 6). The zero-passage, unsorted cells provided a surface antigen profile of the hESCs prior to sorting and expansion by co-culture (Fig. 6A). We observed that CD133⁺ cells remained CD133⁺ with proliferation and expansion, suggesting that CD133 expression is a stable phenomenon. We also observed that the percentage of CD133⁺GFP⁺ versus CD133⁻GFP⁺ hESCs co-expressing Tra-1-60, as well as CD135 and FGFR-1, was similar to that of unsorted cells (Fig. 6B). In contrast, over 90% of the CD133⁺GFP⁺ hESCs were found to co-express SSEA-4, whereas SSEA-4 was only detected on 70% of CD133⁻GFP⁺ cells (χ² = 7; DF 1; P = 0.008). This suggested that SSEA-4 preferentially co-localizes with CD133 in proliferating hESC cultures.

FIG. 6.

Human embryonic stem cell (hESC) subpopulations, as determined by surface marker expression, persist over multiple passages in culture. (A) Unsorted, proliferating, GFP-expressing hESCs were analyzed by flow cytometry for the co-expression of CD133 with SSEA-4, Tra-1-60, FGFR-1, and CD135. Representative histograms are shown. For all histograms (A-C), percentage shown in upper right corner indicates percentage of cells positive for marker analyzed on the X-axis that are positive for the marker analyzed on the Y-axis (eg, 91% of CD133⁺ cells were SSEA-4⁺). (B) Unsorted, proliferating, GFP-expressing hESCs were analyzed by flow cytometry for the co-expression of CD135 with SSEA-4, Tra-1-60, FGFR-1, and CD133. Representative histograms are shown. (C) Following three passages of co-cultures of CD133- or CD135-sorted cells with wild-type GFP⁻ hESCs, GFP-expressing cells were reanalyzed by flow cytometry for the co-expression of CD133 or CD135 with SSEA-4, Tra-1-60, and FGFR-1. This demonstrated that sorted cells in co-culture retain their original profile of these cell surface markers with very little drift. (D) Graphic analysis of histograms shown in (C) demonstrate that there is little if any shift in surface marker expression as a result of co-culture, and that there is significant co-expression of SSEA-4 with CD133. % Ab⁺ Cells, percentage of CD133⁺, CD133⁻, CD135⁺, or CD135⁻ cells that stain positive with antibodies against the indicated surface markers (ie, FGFR-1, SSEA-4, Tra-1-60). *P = 0.008.

Similarly, we observed that CD135⁺ cells remained CD135⁺ with proliferation and expansion, suggesting that CD135 expression is a stable phenomenon as well. The percentage of CD135⁺GFP⁺ versus CD135⁻GFP⁺ hESCs co-expressing Tra-1-60, SSEA-4, and FGFR-1 also were similar to that of unsorted cells. These results obtained with CD133⁺GFP⁺ and CD135⁺GFP⁺ hESC subpopulations support the concept that stable subpopulations of hESCs, as defined by the expression of specific surface markers, exist within proliferating cultures, and that these proliferating hESC cultures are more heterogeneous and perhaps less “pluripotent” than previously assumed.

Discussion

Our goal was to analyze the functional significance of hESC subpopulations, as defined by the expression of specific surface antigens. A fluorescence-tagged co-culture system allowed us to both determine the lineage fate of differentiating hESC subpopulations and characterize the cell surface profile of proliferating hESC subpopulations in vitro.

CD135 (Flt3) is a tyrosine kinase receptor that has been associated with tissues derived from mesoderm, including hematopoietic stem cells, myelomonocytic progenitors, primitive B-cell progenitors, and thymocytes [16]. CD133 (prominin-1) is a member of a class of novel pentaspan membrane proteins identified in both humans and mice. Although classified as a marker of primitive hematopoietic and neural stem cells [17], tissue expression arrays have shown that human CD133 mRNA is strongly expressed in differentiated tissues derived from all three embryonic germ layers, including kidney, mammary gland, trachea, pancreas, digestive tract, and testis [18]. A population of CD133⁺ cells isolated from the adult human kidney was capable of both self-renewal and multilineage differentiation in vitro and in vivo [19]. CD133 is also expressed on endothelial progenitor cells, which play a role in angiogenesis and neovasculogenesis during both tumor growth and wound healing [20]. CD133⁺ subpopulations, isolated from developing mouse brain, were capable of clonal expansion to form neurospheres in vitro and differentiated into multiple lineages in vitro and in vivo following transplantation into neonatal mice [21]. Although CD133 is expressed on a broad range of adult tissues, the significance of its expression on undifferentiated hESCs has previously not been understood.

Both CD133⁺GFP⁺ and CD133⁻GFP⁺ selected cells displayed the ability to efficiently self-renew in a co-culture with H9 GFP⁻hESCs. Differences between these two subpopulations became evident upon differentiation in vitro. The vast majority (∼90%) of the CD133⁺GFP⁺ subpopulation followed an ectodermal lineage fate. This contrasted sharply with differentiated CD133⁻GFP⁺ hESCs, which predominantly stained for endodermal and mesodermal markers. The ability of CD133⁺GFP⁺ hESCs to follow an ectodermal lineage fate appears to be CD133-specific, since CD135⁺GFP⁺ hESCs differentiate into all three somatic lineages similar to wild-type H9 hESCs. These results suggest that a CD133⁺ subpopulation of proliferating hESCs are predisposed to follow an ectodermal lineage fate. CD133⁺ hESCs may indeed represent one of many subpopulations of progenitor cells within an undifferentiated hESC colony, which can self-renew and follow a preprogrammed, preferential lineage fate.

Previous studies have demonstrated the association of the CD133 surface antigen with embryonic neural stem cells in developing mouse embryos [22,23]. After culture in vitro, CD133⁻ cells isolated from embryonic mouse brains failed to form neurospheres, while CD133⁺ cells formed neurospheres, gave rise to astrocytes and neuronal cells, and expressed nestin [24]. Another study used an in vitro system to follow the expression CD133 on differentiating mouse embryonic stem cells (mESCs) [25]. CD133 expression was associated with proliferating mESCs, and its expression was maintained during the early stages of differentiation into ectodermal, endodermal, and mesodermal precursor cells. At the terminal stages of differentiation, however, expression of CD133 was observed only on cells co-expressing the neuroectodermal marker, nestin [25]. In human studies, CD133⁺ cells derived from human brain tumors have been shown to display the properties of self-renewal and recapitulation of the original tumor, suggesting that CD133 may mark cancer stem cells specific to neural tumors [26 –28]. Together, these studies provide further evidence to suggest a role for CD133 as a lineage-specific antigen for neural precursor cells in mice and humans.

We also wanted to determine whether the expression of a lineage-confirmed surface marker was maintained over time within proliferating cultures. To address this issue, we monitored the expression of a set of surface antigens during the expansion of selected hESCs in the fluorescence-tagged co-culture system. We discovered that both the CD133⁺GFP⁺ and CD135⁺GFP⁺ subpopulations continued to express surface antigens CD133, CD135, FGFR-1, SSEA-4, and Tra-1-60, following multiple passages. The amounts of these surface antigens remained constant as compared to the levels of the same antigens detected on zero-passage, unsorted GFP⁺ hESCs. These results support the concept that stable subpopulations of hESCs exist within proliferating cultures. More importantly, these subpopulations can be expanded without the loss of marker expression or lineage specificity.

We observed that the percentage of CD133⁺GFP⁺ versus CD133⁻GFP⁺ hESCs co-expressing the pluripotency antigen, Tra-1-60, as well as CD135 and FGFR-1, was similar to that of the unsorted cells. In contrast, over 90% of the CD133⁺GFP⁺ hESCs were found to co-express SSEA-4, whereas SSEA-4 was only detected on 70% of CD133⁻GFP⁺ cells. This suggested that SSEA-4 preferentially co-localizes with CD133 in proliferating hESC cultures. A previous study has also demonstrated that greater than 90% of CD133⁺ cells co-express SSEA-4 in a proliferating H9 hESC culture [2]. A recent report suggests a link between CD133 co-expression with SSEA-4 and the positive selection of human neural progenitor cells. SSEA-4 was found to be expressed during human central nervous system development in 6–9-week-old human embryos, and the selection of cells from embryonic forebrains, expressing both SSEA-4 and CD133, resulted in the enrichment for neural progenitor cells [29]. The SSEA4⁺/CD133⁺ subpopulation was expanded in a neurosphere culture, and the SSEA4⁺/CD133⁺ phenotype was retained after several passages [29].

Our data demonstrate that subpopulations characterized by the expression of distinct surface markers exist within proliferating hESC cultures in vitro. Importantly, this suggests that the concept of “pluripotency” within proliferating hESC cultures may not apply to all cells within the culture. We were able to efficiently expand these selected subpopulations using a fluorescence-tagged co-culture system, and demonstrate that marker expression of the selected hESCs was retained during expansion. The majority of CD133-selected hESCs, differentiated by embryoid body formation and by teratoma formation, followed a neuroectodermal lineage fate. This suggests that CD133 will be useful as a marker for the in vitro expansion of hESCs that will give rise specifically to neural progenitor cells for the potential treatment of neurodegenerative and neurotraumatic disorders.

Footnotes

Acknowledgments

This work was supported by a Comprehensive Research Grant from the California Institute for Regenerative Medicine (RC1-00104), a Public Health Service Grant (HL085377) from NHLBI, and a gift from the Pollin Foundation to H.S.B., and a Comprehensive Research Grant from the California Institute for Regenerative Medicine (RC1-00347) and funds from the Eli and Edythe Broad Center of Regeneration Medicine and Stem Cell Research at UCSF to A.D.L. F.W.K. was supported by a National Research Service Award (HL007544) from NHLBI. We thank Margaret Mayes and Philip Ursell for assistance with histology, and Patrick McQuillen and Meenakshi Gaur for helpful discussions.

Author Disclosure Statement

No competing financial interests exist.

References

Adewumi

, Aflatoonian

, Ahrlund-Richter

, Amit

, Andrews

, Beighton

, Bello

, Benvenisty

, Berry

, Bevan

, Blum

, Brooking

, Chen

, Choo

, Churchill

, Corbel

, Damjanov

, Draper

, Dvorak

, Emanuelsson

, Fleck

, Ford

, Gertow

, Gertsenstein

, Gokhale

, Hamilton

, Hampl

, Healy

, Hovatta

, Hyllner

, Imreh

, Itskovitz-Eldor

, Jackson

, Johnson

, Jones

, Kee

, King

, Knowles

, Lako

, Lebrin

, Mallon

, Manning

, Mayshar

, McKay

, Michalska

, Mikkola

, Mileikovsky

, Minger

, Moore

, Mummery

, Nagy

, Nakatsuji

, O'Brien

, Oh

, Olsson

, Otonkoski

, Park

, Passier

, Patel

, Pedersen

, Pera

, Piekarczyk

, Pera

, Reubinoff

, Robins

, Rossant

, Rugg-Gunn

, Schulz

, Semb

, Sherrer

, Siemen

, Stacey

, Stojkovic

, Suemori

, Szatkiewicz

, Turetsky

, Tuuri

, van den Brink

, Vintersten

, Vuoristo

, Ward

, Weaver

, Young

and Zhang

. (2007). Characterization of human embryonic stem cell lines by the International Stem Cell Initiative. Nat Biotechnol, 25:803–816.

Carpenter

, Rosler

, Fisk

, Brandenberger

, Ares

, Miura

, Lucero

and Rao

. (2004). Properties of four human embryonic stem cell lines maintained in a feeder-free culture system. Dev Dyn, 229:243–258.

Draper

, Pigott

, Thomson

and Andrews

. (2002). Surface antigens of human embryonic stem cells: changes upon differentiation in culture. J Anat, 200:249–258.

Hoffman

and Carpenter

. (2005). Characterization and culture of human embryonic stem cells. Nat Biotechnol, 23:699–708.

Peh

, Lang

, Pera

and Hawes

. (2008). CD133 expression by neural progenitors derived from human embryonic stem cells and its use for their prospective isolation. Stem Cells Dev [Epub 2008/07/25; COI 10.1089/scd.2008.0124].

Nicholas

, Gaur

, Wang

, Pera

and Leavitt

. (2007). A method for single-cell sorting and expansion of genetically modified human embryonic stem cells. Stem Cells Dev, 16:109–117.

Ivey

, Muth

, Arnold

, King

, Yeh

, Fish

, Hsiao

, Schwartz

, Conklin

, Bernstein

and Srivastava

. (2008). MicroRNA regulation of cell lineages in mouse and human embryonic stem cells. Cell Stem Cell, 2:219–229.

Epting

, Lopez

, Pedersen

, Brown

, Spitz

, Ursell

and Bernstein

. (2008). Stem cell antigen-1 regulates the tempo of muscle repair through effects on proliferation of alpha7 integrin-expressing myoblasts. Exp Cell Res, 314:1125–1135.

Epting

, Lopez

, Shen

, Liu

, Bristow

and Bernstein

. (2004). Stem cell antigen-1 is necessary for cell-cycle withdrawal and myoblast differentiation in C2C12 cells. J Cell Sci, 117:6185–6195.

10.

Cunha

. (1972). Epithelio-mesenchymal interactions in primordial gland structures which become responsive to androgenic stimulation. Anat Rec, 172:179–195.

11.

Watanabe

, Ueno

, Kamiya

, Nishiyama

, Matsumura

, Wataya

, Takahashi

, Nishikawa

, Muguruma

and Sasai

. (2007). A ROCK inhibitor permits survival of dissociated human embryonic stem cells. Nat Biotechnol, 25:681–686.

12.

Shmelkov

, St Clair

, Lyden

and Rafii

. (2005). AC133/CD133/Prominin 1. Int J Biochem Cell Biol, 37:715–719.

13.

Drexler

and Quentmeier

. (2004). FLT3: receptor and ligand. Growth Factors, 22:71–73.

14.

D'Amour

, Agulnick

, Eliazer

, Kelly

, Kroon

and Baetge

. (2005). Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat Biotechnol, 23:1534–1541.

15.

Zhang

, Wernig

, Duncan

, Brustle

and Thomson

. (2001). In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nat Biotechnol, 19:1129–1133.

16.

Del Zotto

, Luchetti

and Zamai

. (2001). Cd135. J Biol Regul Homeost Agents, 15:103–106.

17.

Mizrak

, Brittan

and Alison

. (2008). CD133: molecule of the moment. J Pathol, 214:3–9.

18.

Florek

, Haase

, Marzesco

, Freund

, Ehninger

, Huttner

and Corbeil

. (2005). Prominin-1/CD133, a neural and hematopoietic stem cell marker, is expressed in adult human differentiated cells and certain types of kidney cancer. Cell Tissue Res, 319:15–26.

19.

Bussolati

, Bruno

, Grange

, Buttiglieri

, Deregibus

, Cantino

and Camussi

. (2005). Isolation of renal progenitor cells from adult human kidney. Am J Pathol, 166:545–555.

20.

Ribatti

. (2004). The involvement of endothelial progenitor cells in tumor angiogenesis. J Cell Mol Med, 8:294–300.

21.

Lee

, Kessler

, Read

, Kaiser

, Corbeil

, Huttner

, Johnson

and Wechsler-Reya

. (2005). Isolation of neural stem cells from the postnatal cerebellum. Nat Neurosci, 8:723–729.

22.

Weigmann

, Corbeil

, Hellwig

and Huttner

. (1997). Prominin, a novel microvilli-specific polytopic membrane protein of the apical surface of epithelial cells, is targeted to plasmalemmal protrusions of non-epithelial cells. Proc Natl Acad Sci USA, 94:12425–12430.

23.

Sawamoto

, Nakao

, Kakishita

, Ogawa

, Toyama

, Yamamoto

, Yamaguchi

, Mori

, Goldman

, Itakura

and Okano

. (2001). Generation of dopaminergic neurons in the adult brain from mesencephalic precursor cells labeled with a nestin-GFP transgene. J Neurosci, 21:3895–3903.

24.

Pfenninger

, Roschupkina

, Hertwig

, Kottwitz

, Englund

, Bengzon

, Jacobsen

and Nuber

. (2007). CD133 is not present on neurogenic astrocytes in the adult subventricular zone, but on embryonic neural stem cells, ependymal cells, and glioblastoma cells. Cancer Res, 67:5727–5736.

25.

Kania

, Corbeil

, Fuchs

, Tarasov

, Blyszczuk

, Huttner

, Boheler

and Wobus

. (2005). Somatic stem cell marker prominin-1/CD133 is expressed in embryonic stem cell-derived progenitors. Stem Cells, 23:791–804.

26.

Bao

, Wu

, McLendon

, Hao

, Shi

, Hjelmeland

, Dewhirst

, Bigner

and Rich

. (2006). Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature, 444:756–760.

27.

Hemmati

, Nakano

, Lazareff

, Masterman-Smith

, Geschwind

, Bronner-Fraser

and Kornblum

. (2003). Cancerous stem cells can arise from pediatric brain tumors. Proc Natl Acad Sci USA, 100:15178–15183.

28.

Singh

, Hawkins

, Clarke

, Squire

, Bayani

, Hide

, Henkelman

, Cusimano

and Dirks

. (2004). Identification of human brain tumour initiating cells. Nature, 432:396–401.

29.

Barraud

, Stott

, Mollgard

, Parmar

and Bjorklund

. (2007). In vitro characterization of a human neural progenitor cell coexpressing SSEA4 and CD133. J Neurosci Res, 85:250–259.

Subpopulations of Human Embryonic Stem Cells With Distinct Tissue-Specific Fates Can Be Selected From Pluripotent Cultures

Abstract

Introduction

Materials and Methods

hESC culture and differentiation

Flow cytometry and FACS

Immunocytochemistry

Teratomas

Statistics

Results

Fate mapping of human embryonic stem cell subpopulations can be accomplished using a fluorescence-tagged co-culture system

Proliferating subpopulations of CD133+ human embryonic stem cells give rise to ectodermal derivatives, while CD135+ subpopulations contribute to all three embryonic germ layers

CD133+ and CD135+ subpopulations of proliferating human embryonic stem cells stably express defining surface markers during proliferation and expansion

Discussion

Footnotes

Acknowledgments

Author Disclosure Statement

References

Proliferating subpopulations of CD133⁺ human embryonic stem cells give rise to ectodermal derivatives, while CD135⁺ subpopulations contribute to all three embryonic germ layers

CD133⁺ and CD135⁺ subpopulations of proliferating human embryonic stem cells stably express defining surface markers during proliferation and expansion