Expression of Coxsackievirus and Adenovirus Receptor Separates Hematopoietic and Cardiac Progenitor Cells in Fetal Liver Kinase 1-Expressing Mesoderm

Distinct progenitor types in fetal liver kinase 1-positive (Flk1⁺) cells were separated by the expression of coxsackievirus and adenovirus receptor (CAR), a tight junction component molecule. A novel population with the potential to differentiate into hematopoietic cells, endothelial cells, and cardiomyocytes was identified. Our findings indicate that CAR would be a novel and prominent marker for separating pluripotent stem cell- and embryo-derived Flk1⁺ mesodermal cells with distinct differentiation potentials.

Abstract

In developing embryos or in vitro differentiation cultures using pluripotent stem cells (PSCs), such as embryonic stem cells and induced pluripotent stem cells, fetal liver kinase 1 (Flk1)-expressing mesodermal cells are thought to be a heterogeneous population that includes hematopoietic progenitors, endothelial progenitors, and cardiac progenitors. However, information on cell surface markers for separating these progenitors in Flk1⁺ cells is currently limited. In the present study, we show that distinct types of progenitor cells in Flk1⁺ cells could be separated according to the expression of coxsackievirus and adenovirus receptor (CAR, also known as CXADR), a tight junction component molecule. We found that mouse and human PSC- and mouse embryo-derived Flk1⁺ cells could be subdivided into Flk1⁺CAR⁺ cells and Flk1⁺CAR⁻ cells. The progenitor cells with cardiac potential were almost entirely restricted to Flk1⁺CAR⁺ cells, and Flk1⁺CAR⁻ cells efficiently differentiated into hematopoietic cells. Endothelial differentiation potential was observed in both populations. Furthermore, from the expression of CAR, Flk1, and platelet-derived growth factor receptor-α (PDGFRα), Flk1⁺ cells could be separated into three populations (Flk1⁺PDGFRα⁻CAR⁻ cells, Flk1⁺PDGFRα⁻CAR⁺ cells, and Flk1⁺PDGFRα⁺CAR⁺ cells). Flk1⁺PDGFRα⁺ cells and Flk1⁺PDGFRα⁻ cells have been reported as cardiac and hematopoietic progenitor cells, respectively. We identified a novel population (Flk1⁺PDGFRα⁻CAR⁺ cells) with the potential to differentiate into not only hematopoietic cells and endothelial cells but also cardiomyocytes. Our findings indicate that CAR would be a novel and prominent marker for separating PSC- and embryo-derived Flk1⁺ mesodermal cells with distinct differentiation potentials.

Significance

Flk1-expressing (Flk⁺) mesodermal cells are assumed to be a heterogeneous population that includes hematopoietic progenitors and cardiac progenitors. However, information on cell surface markers for separating the Flk1⁺ cell subsets is currently limited. This study shows that hematopoietic lineage cells and cardiac lineage cells of pluripotent stem cell- and mouse embryo-derived Flk1⁺ cells could be separated based on the expression of coxsackievirus and adenovirus receptor (CAR), a tight junction component molecule. The results indicate that CAR is a novel and prominent cell surface marker for separating the Flk1⁺ cell subsets, and these differentiation methods based on the expression levels of CAR are expected to be instrumental in basic sciences and clinical studies.

Introduction

The Flk1 gene (also called kinase insert domain receptor [KDR]), which encodes a receptor for vascular endothelial growth factor (VEGF), is known to be expressed in lateral mesodermal cells in the developing embryo, and fetal liver kinase 1 (Flk1)-expressing mesodermal cells give rise to hematopoietic cells and endothelial cells [1–3]. A lineage-tracing analysis has demonstrated that Flk1⁺ cells can differentiate, not only into hematopoietic cells and endothelial cells, but also into cardiomyocytes and skeletal muscle cells [1, 4]. Just as in the case of in vivo development, in vitro differentiation studies have shown that Flk1⁺ mesodermal cells can be differentiated from pluripotent stem cells (PSCs), including embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs), and PSC-derived Flk1⁺ cells produce hematopoietic cells, endothelial cells, and cardiomyocytes [4–8]. Therefore, Flk1 is widely used as a mesodermal cell marker. However, the described findings also suggest that Flk1⁺ cells could be heterogeneous populations containing hematopoietic and cardiac progenitor cells. Cells expressing high levels of Flk1 were reported to be hematopoietic progenitors, and cells expressing low levels of Flk1 were cardiac progenitors [4]. Although the establishment of methods for the separation of distinct mesodermal progenitor cells in Flk1⁺ cells will enable us to selectively generate the lineage-committed cells from PSCs, it is difficult to segregate the different cell types in Flk1⁺ cells according to the expression levels of Flk1 alone. Thus, the identification of other markers would be required to isolate these progenitors.

Coxsackievirus and adenovirus receptor (CAR, also known as CXADR), a tight junction component molecule, was originally identified as a cell surface receptor for coxsackie B viruses and C-type adenoviruses [9, 10]. It has been shown that the expression levels of CAR are low in hematopoietic cells, but CAR is highly expressed in the developing brain and the heart [10–14]. CAR-deficient mice showed embryonic lethality at embryonic days 11.5–13.5 owing to cardiomyocyte dysfunction and heart failure, indicating the essential role of CAR in heart development [14–16]. Moreover, we have previously showed that undifferentiated mouse and human PSCs and PSC-derived embryoid bodies (EBs) expressed CAR [17–20]. However, it remains to be determined which type of cells can express CAR in early embryos and in in vitro differentiation cultures.

In the present study, we examined CAR expression in the mesodermal differentiation process from PSCs and tested whether mesodermal progenitor cells could be separated from PSC-derived heterogeneous cells on the basis of CAR expression. We found two types of populations, one that has high levels of CAR expression (CAR⁺ cells), and one that has a low or negative level of CAR expression (CAR^low/− cells, hereafter referred to as CAR⁻ cells), in PSC- and embryo-derived Flk1⁺ cells. In addition, hematopoietic lineage cells and cardiac lineage cells in the Flk1⁺ cells were enriched in the CAR⁻ and CAR⁺ cells, respectively. Although previous studies reported that Flk1⁺ cells lacking platelet-derived growth factor receptor-α (PDGFRα) (Flk1⁺PDGFRα⁻ cells) displayed hemoangiogenic differentiation tropism [8, 21, 22], our analysis revealed that CAR expression identified novel populations that differentiate, not only into hematopoietic cells and endothelial cells, but also into cardiomyocytes in Flk1⁺PDGFRα⁻ cells. Our data indicate that the tight junction molecule CAR would be a valuable marker for separating PSC- and embryo-derived Flk1⁺ mesodermal cells with distinct differentiation potentials.

Materials and Methods

Antibodies

The antibodies (Abs) used for flow cytometry and immunostaining are listed in supplemental online Tables 1 and 2. Purified monoclonal rat anti-mouse CAR Ab was kindly provided by Dr. Toshio Imai (KAN Research Institute, Inc., Hyogo, Japan, http://www.kan-research.co.jp/english/).

Cell Cultures

The mouse ESC line, BRC6 (Riken Bioresource Center, Tsukuba, Japan, http://www.en.brc.riken.jp), and the mouse iPSC line, 38C2 (a kind gift from Dr. S. Yamanaka, Kyoto University, Kyoto, Japan) [23], were cultured in leukemia inhibitory factor-containing ESC medium (EMD Millipore, Billerica, MA, http://www.emdmillipore.com) on mitomycin C (MMC)-treated mouse embryonic fibroblasts (MEFs; EMD Millipore). The human ESC line, KhES-3 (provided by Dr. N. Nakatsuji, Kyoto University, Kyoto, Japan) [24], and the human iPSC line, 201B7 (provided by Dr. S. Yamanaka, Kyoto University) [25], were routinely maintained on MMC-inactivated MEFs with ReproStem medium (ReproCELL, Yokohama, Japan, http://www.reprocell.com) supplemented with 5 ng/ml human fibroblast growth factor 2 (FGF2; Katayama Chemical, Osaka, Japan, http://www.katayamakagaku.co.jp/en/). KhES-3 cells were used following the Guidelines for Derivation and Utilization of Human Embryonic Stem Cells of the Ministry of Education, Culture, Sports, Science and Technology of Japan, after approval by the institutional ethical review board at the National Institute of Biomedical Innovation (Osaka, Japan). OP9 stromal cells (Riken Bioresource Center) were cultured with OP9 medium containing α-minimum essential medium (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) supplemented with 20% fetal bovine serum (FBS), 2 mM l-glutamine, nonessential amino acid (NEAA), and penicillin/streptomycin (all from Life Technologies, Carlsbad, CA, http://www.lifetechnologies.com).

In Vitro Differentiation of Mouse ESCs and iPSCs

EB differentiation of mouse ESCs and iPSCs was performed as reported previously [18, 26]. In brief, to initiate differentiation, mouse ESCs and iPSCs were trypsinized and plated on a culture dish for 30–60 minutes to remove the MEFs. To form EBs, the cells were then suspended in differentiation medium (Dulbecco’s modified Eagle’s medium; Wako Chemical, Osaka, Japan, http://www.wako-chem.co.jp/english) supplemented with 15% FBS, NEAA, penicillin/streptomycin, 2 mM l-glutamine, and 100 μM 2-mercaptoethanol (2-ME; Nacalai Tesque Inc., Kyoto, Japan, http://www.nacalai.co.jp/en) and plated on a round-bottom Lipidure-coated 96-well plate (Thermo Fisher Scientific, Yokohama, Japan, http://www.thermofisher.co.jp/) at 3 × 10³ cells (ESCs) or 1 × 10³ cells (iPSCs) per well. We mainly used 7-day-cultured EBs, except as indicated, because the proportion of Flk1-expressing cells in EBs increased to a peak on day 7 and decreased over the next 2 days under our culture conditions, as reported previously [26]. To evaluate the differentiation potential of day 7 EB-derived cells, the EBs were harvested and subjected to fluorescence-activated cell sorting (FACS). For hematopoietic differentiation, FACS-sorted cells were cultured on OP9 cells, with OP9 medium supplemented with 50 ng/ml mouse stem cell factor (SCF; PeproTech, Rocky Hill, NJ, http://www.peprotech.com), 50 ng/ml human Flt3-ligand (Flt3-L; PeproTech), 10 ng/ml mouse thrombopoietin (TPO; PeproTech), 5 ng/ml mouse interleukin (IL)-3 (R&D Systems, Minneapolis, MN, http://www.rndsystems.com), 5 ng/ml human IL-6 (PeproTech), and 50 μM 2-ME. To induce endothelial cells and cardiomyocytes, FACS-sorted cells were cultured on OP9 stromal cells, with OP9 medium supplemented with 50 μM 2-ME [6].

In Vitro Differentiation of Human ESCs and iPSCs

Mesodermal differentiation of human ESCs and iPSCs was performed according to the reported procedures [7], with slight modifications. In brief, to induce differentiation, the medium of human ESCs and iPSCs was initially changed to a serum-free StemPro34 medium, including supplements (Life Technologies). After 2 hours of incubation, human ESCs and iPSCs were harvested using 0.1 mg/ml of Dispase (Roche, Indianapolis, IN, http://www.roche.com) and incubated on a gelatin-coated dish to remove the MEFs. To form EBs, the cells were then cultured on 100-mm Petri dishes in basic medium (StemPro34 medium containing 50 μg/ml ascorbic acid [Sigma-Aldrich] and 450 μM monothioglycerol [MTG, Sigma-Aldrich]) supplemented with 10 μM Y27632 [Wako Chemical] and 2 ng/ml human bone morphogenetic protein 4 [BMP4, R&D Systems]). The next day, the EBs were collected and cultured for 3 days in basic medium supplemented with 10 ng/ml human BMP4, 6 ng/ml human Activin A (R&D Systems), and 5 ng/ml human FGF2 to induce mesodermal cells. On day 4, the culture medium for EBs was changed to basic medium supplemented with 10 ng/ml human VEGF (PeproTech) and 150 ng/ml human Dickkopf1 (R&D Systems) and were cultured for 3 days. Next, the day 7 EB-derived cells were sorted, and the differentiation potentials of the sorted cells were estimated. For hematopoietic differentiation, FACS-sorted cells were cultured on OP9 cells, with OP9 medium supplemented with 100 ng/ml mouse SCF, 100 ng/ml human Flt3-L, 10 ng/ml mouse TPO, 10 ng/ml human IL-3 (PeproTech), 10 ng/ml human IL-6, 50 μg/ml ascorbic acid, and 450 μM MTG. For cardiac differentiation, the sorted cells were cultured in a 96-well plate at 0.5–1 × 10⁵ cells per well in basic medium containing 10 ng/ml human VEGF, 10 ng/ml human FGF2, and 10 μM Y27632 for an additional 7 days.

Flow Cytometry and Cell Sorting

Mouse ESC- or iPSC-derived EB cells were incubated with purified rat anti-mouse CAR Abs on ice for 30 minutes before staining with Dylight649-conjugated secondary Abs. The cells were then sorted using FACSAria (BD Biosciences, San Diego, CA, http://www.bdbiosciences.com). In the analysis of the expression of CAR and other surface molecules, such as Flk1 and PDGFRα, the EB cells were initially stained with anti-CAR Abs, as mentioned. They were then washed 3 times with staining buffer (2% FBS/phosphate-buffered saline [PBS]) and subsequently stained with Abs for other molecules. When biotin-conjugated Abs were used, the cells were further incubated with phycoerythrin- or brilliant violet 421-conjugated streptavidin. The stained cells were sorted using a FACSAria (BD Biosciences) or SH800 (Sony, Tokyo, Japan, http://www.sony.net) cell sorter. Similarly, human ESC- or iPSC-derived EB cells were stained and sorted using purified mouse anti-human CAR Abs, brilliant violet 421-conjugated anti-mouse IgG Abs, and other surface markers. The purity of the sorted cells was always greater than 90%. FACS-sorted cells were cultured for 7 days, as described, and then harvested and stained with monoclonal Abs on ice for 30 minutes. For analysis of human ESC- or iPSC-derived cells cocultured on OP9 cells, human cells that express TRA-1-85 were gated [27]. In all flow cytometric analyses, dead cells were excluded by 7-aminoactinomycin D (eBioscience Inc., San Diego, CA, http://www.ebioscience.com) or Hoechst33342 (Life Technologies) staining. Analysis was performed on an LSRFortessa flow cytometer using FACSDiva software (BD Biosciences) and FlowJo software (FlowJo LLC, Ashland, OR, http://www.flowjo.com/).

Immunostaining

The cells were fixed with methanol on ice for 10 minutes or with 4% paraformaldehyde at room temperature for 15 minutes and then blocked by 2% bovine serum albumin/PBS. After blocking, the cells were incubated with primary Ab at 4°C overnight. For immunohistochemistry, the cells were reacted with horseradish peroxidase (HRP)-conjugated anti-mouse IgG (1:500, Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com) at room temperature for 1 hour. HRP activity was detected by ImmPACT DAB SUBSTRATE (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). For immunofluorescent staining, Alexa Fluor 488- or Alexa Fluor 594-conjugated secondary Abs (1:1,000; Life Technologies) were used. Nuclei were visualized using 4′6-diamidino-2-phenylindole (Sigma-Aldrich).

Colony Assay

FACS-sorted cells were directly cultured in Methocult M3434 medium (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell.com) for 10 days. The number of individual colonies was counted using microscopy.

Reverse Transcription-Polymerase Chain Reaction and Microarray Analysis

Total RNA was extracted using ISOGEN (Nippon Gene, Tokyo, Japan, http://www.nippongene.com) or RNAiso Plus (TaKaRa, Shiga, Japan, http://www.takara-bio.com). cDNA was synthesized from DNase I-treated total RNA with a Superscript VILO cDNA synthesis kit (Life Technologies) followed by conventional polymerase chain reaction (PCR) using ExTaq DNA polymerase (TaKaRa). Quantitative real-time reverse transcription-PCR was performed with Fast SYBR Green Master Mix using an ABI StepOne Plus (Life Technologies). Relative quantification was performed against a standard curve, and the values were normalized against the input determined for the housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase. The sequences of the primers used in the present study are listed in supplemental online Table 3. Microarray analysis was performed using a 3D-Gene Mouse Oligo chip 24K (Toray, Tokyo, Japan, http://www.toray.co.jp/). The gene expression levels were analyzed after a global normalization procedure. The Gene Expression Omnibus (GEO) accession number for the microarray analysis is GSE58170.

Embryo Experiments

All experiments were conducted according to the institutional ethical guidelines for animal experimentation at the National Institute of Biomedical Innovation. The morning of finding the vaginal plug was considered embryonic day 0.5 (E0.5). Pregnant ICR (CD1) mice were purchased from Japan SLC (Shizuoka, Japan, http://www.jslc.co.jp). Green fluorescent protein-transgenic (GFP-TG) male mice [28] were crossed with non-TG female mice to obtain GFP-expressing embryos. Embryos at stage E8.5 were removed from the decidua and Reichert’s membrane and dissociated in a cell dissociation buffer (Life Technologies) at 37°C for 15 minutes followed by gentle pipetting. A single-cell suspension was used for the experiments.

Limiting Dilution Assay

The sorted cells were seeded on confluent OP9 cells in 96-well plates containing 8 cell-dose groups: 10, 20, 40, 80, 160, 320, 640, or 1,280 sorted cells per well. After culturing with hematopoietic cytokines for 7 days, the number of wells with or without hematopoietic cells was counted using microscopy. Data analysis was performed using ELDA (extreme limiting data analysis; Walter+Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia, http://bioinf.wehi.edu.au/software/elda/).

Statistical Analysis

Statistical analysis was performed using an unpaired two-tailed Student’s t test.

Results

Hematopoietic Progenitor Cells Are Enriched in CAR⁻ Cells Derived From Mouse ESCs

We examined the expression patterns of CAR on undifferentiated mouse ESCs and ESC-derived EBs. Undifferentiated mouse ESCs uniformly expressed CAR (Fig. 1A, day 0), as shown previously [18]. Similarly, more than 90% of EBs cultured for 5 days expressed CAR (Fig. 1A, day 5). However, two types of populations, one that expressed high levels of CAR (CAR⁺ cells) and one that expressed low or negative levels of CAR (CAR^low/− cells; hereafter CAR⁻ cells), were observed in mouse ESC-derived EB cells after additional culture (Fig. 1A, day 7). The appearance of CAR⁻ cells in EBs suggests the presence of immature hematopoietic cells, because hematopoietic stem/progenitor cells express low levels of CAR [11, 13]. To verify this, the hematopoietic differentiation potentials of CAR⁺ cells and CAR⁻ cells were examined after isolation of each type of cell. The colony-forming cell assay showed that CAR⁻ cells derived from mouse ESCs and iPSCs generated significantly greater numbers of hematopoietic colonies, including mixed hematopoietic colonies, compared with the CAR⁺ cells (supplemental online Fig. 1A, 1B). Hematopoietic transcription factor genes, such as Scl, Runx1, and Gata-1, were also highly expressed in CAR⁻ cells (supplemental online Fig. 1C). Furthermore, CAR⁻ cells efficiently generated hematopoietic cells on OP9 stromal cells in the presence of hematopoietic cytokines (supplemental online Fig. 1D). These data clearly indicate the enrichment of hematopoietic progenitor cells within the population of CAR⁻ cells in EBs.

Figure 1.

Evaluation of the hematopoietic differentiation potential of mouse embryonic stem cell (ESC)-derived Flk1⁺CAR⁺ cells and Flk1⁺CAR⁻ cells. (A): Undifferentiated mouse ESCs (day 0) and ESC-derived embryoid bodies (EBs) (days 5 and 7) were stained with anti-CAR antibodies and then subjected to flow cytometric analysis. Representative results from one of three independent experiments are shown. (B): Mouse ESC-derived EBs (day 7) were stained with anti-Flk1 and anti-CAR antibodies. (C): Total RNA was extracted from fluorescence-activated cell sorting-purified F⁺C⁺ cells and F⁺C⁻ cells. Next, quantitative reverse transcription-polymerase chain reaction analysis was performed. The results shown are the mean of three independent experiments with the indicated standard deviations (SDs). ∗, p < .05; ∗∗, p < .01. (D): Purified F⁺C⁺ cells and F⁺C⁻ cells were plated and cultured in methylcellulose-containing media with hematopoietic cytokines. Ten days later, the number of hematopoietic colonies was counted using light microscopy. The numbers of total colonies (left) and subdivided colonies (right) are shown. The data are expressed as the mean ± SD (n = 3); ∗, p < .05; ∗∗, p < .01. (E, F): F⁺C⁺ cells and F⁺C⁻ cells were cocultured with OP9 cells, and the number of hematopoietic cells was counted (E). The cells were stained further with each antibody and subsequently subjected to flow cytometric analysis (F). Representative results from one of three independent experiments are shown. ∗, p < .01. Abbreviations: Ab, antibody; BFU-E, burst forming unit-erythroid; CAR, coxsackievirus and adenovirus receptor; CFC, colony forming cell; CFU-G, colony forming unit-granulocyte; CFU-GM, colony forming unit-granulocyte macrophage; CFU-M, colony forming unit-macrophage; CFU-Mix, colony forming unit-mixed; F⁺C⁺, Flk1⁺CAR⁺; F⁺C⁻, Flk1⁺CAR⁻; Flk1, fetal liver kinase 1; FSC, forward scatter; mCAR, mouse CAR; NS, not significant; SSC, side scatter.

CAR Expression Separates Hematopoietic and Cardiac Lineage Cells of Mouse ESC-Derived Flk1⁺ Cells

It is well-known that various types of mesodermal cells, such as hematopoietic cells, endothelial cells, and cardiomyocytes, are generated from Flk1⁺ cells in both in vitro ESC differentiation cultures and mouse developing embryos [4, 7, 8]. Hematopoietic activities were largely enriched within CAR⁻ cells, as described; however, we observed similar expression levels of Flk1 mRNA between the CAR⁺ and CAR⁻ cells (supplemental online Fig. 1C). Accordingly, flow cytometric analysis of Flk1⁺ cells could lead to identification of two distinct populations (Flk1⁺CAR⁻ [F⁺C⁻] cells and Flk1⁺CAR⁺ [F⁺C⁺] cells) according to the Flk1 and CAR expression (Fig. 1B). Because CAR is expressed in the heart of embryos and is essential for embryonic heart development [12, 14–16], we expected that hematopoietic or cardiac lineage cells in Flk1⁺ cells could be separated according to the expression of CAR. To explore this, we initially examined the hematopoietic potentials of F⁺C⁺ cells and F⁺C⁻ cells in EBs. We found a significant elevation of the genes involved in hematopoiesis and the generation of a large number of hematopoietic colonies in the F⁺C⁻ cells (Fig. 1C, 1D). Flow cytometric analysis of F⁺C⁺ cells and F⁺C⁻ cells after culturing on OP9 cells revealed the generation of CD45⁺ cells, Ter119⁺ cells, and CD41⁺ cells from both F⁺C⁺ and F⁺C⁻ cells, although F⁺C⁻ cells showed a 10–20-fold higher production of hematopoietic cells compared with F⁺C⁺ cells (Fig. 1E, 1F). This indicates that F⁺C⁻ cells could have more potent hematopoietic differentiation potential than F⁺C⁺ cells.

In contrast to the expression patterns of hematopoiesis-related genes, isolated F⁺C⁺ cells showed significantly higher expression levels of genes associated with cardiomyocytes compared with F⁺C⁻ cells (Fig. 2A). We next tested the cardiac differentiation potential of F⁺C⁺ and F⁺C⁻ cells by culturing on OP9 stromal cells [6]. After 7 days of culture, elevated expression levels of cardiac genes were observed in F⁺C⁺ cells (Fig. 2B). F⁺C⁺ cells showed cardiac troponin T (cTNT) and α-actinin expression and sarcomere formation (Fig. 2C, 2D). F⁺C⁺ cells also expressed myosin light chain 2a (MLC2a), MLC2v, or hyperpolarization-activated cyclic nucleotide-gated channel 4 (Fig. 2E), indicating the generation of atrial cardiomyocytes, ventricular cardiomyoctes, or pacemaker cells, respectively. In contrast, neither self-beating cells nor cTNT-expressing cells were observed in F⁺C⁻ cell cultures. Taken together, these data indicate that hematopoietic or cardiac potential of Flk1⁺ cells could be separated by the expression levels of CAR.

Figure 2.

Comparison of the cardiac differentiation potential of mouse ESC-derived Flk1⁺CAR⁺ cells and Flk1⁺CAR⁻ cells. (A): Total RNA was isolated from F⁺C⁺ cells and F⁺C⁻ cells, and the expression levels of the genes associated with cardiogenesis were examined using quantitative reverse transcription-polymerase chain reaction analysis. The data are expressed as the mean ± SD (n = 3). ∗, p < .05; ∗∗, p < .01. (B): The expression level of cardiac genes was investigated after F⁺C⁺ cells and F⁺C⁻ cells were cocultured with OP9 cells. The data are expressed as the mean ± SD (n = 3). ∗, p < .05; ∗∗, p < .01. (C–E): The expression of cTNT (C, E), sarcomeric α-actinin (D), MLC2a (E, upper), MLC2v (E, middle), and HCN4 hyperpolarization-activated cyclic nucleotide-gated channel 4 (E, bottom) of F⁺C⁺ cell- and F⁺C⁻ cell-derived cells was investigated by immunostaining. The nuclei were stained with DAPI. Because F⁺C⁻ cells could hardly differentiate into cardiac cells, we only presented the data for the F⁺C⁺ cell-derived cells. (F): F⁺C⁺ cells and F⁺C⁻ cells, which were cocultured on OP9 cells for 7 days, were stained with an antibody against CD31 (upper). The capacity of Ac-LDL uptake of F⁺C⁺ cell- and F⁺C⁻ cell-derived cells was investigated using Alexa Fluor 488-conjugated Ac-LDL (bottom). Representative results are shown. Scale bars = 300 μm (C) and 100 μm (E, F). Abbreviations: Ac-LDL, acetylated low-density lipoprotein; CAR, coxsackievirus and adenovirus receptor; cTNT, cardiac troponin T; DAPI, 4′6-diamidino-2-phenylindole; ESC, embryonic stem cell; F⁺C⁺, Flk1⁺CAR⁺; F⁺C⁻, Flk1⁺CAR⁻; Flk1, fetal liver kinase 1; HCN4, hyperpolarization-activated cyclic nucleotide-gated channel 4; MLC2, myosin light chain 2; ND, not determined.

During cardiac differentiation, endothelial sheet-like structures were observed in both populations. These cells expressed CD31 and displayed uptake of acetylated low-density lipoprotein (Ac-LDL) (Fig. 2F), demonstrating that F⁺C⁺ cells and F⁺C⁻ cells could differentiate into functional endothelial cells.

Flk1⁺ Mesodermal Cells Can Be Separated Into Three Populations According to Flk1, PDGFRα, and CAR Expression

Expression of PDGFRα is known to distinguish the hematopoietic and cardiac potential in Flk1⁺ mesodermal cells [8, 22]; Flk1⁺PDGFRα⁻ (F⁺Pa⁻) cells or Flk1⁺PDGFR⁺ (F⁺Pa⁺) cells contain hemoangiogenic progenitor cells or cardiac progenitor cells, respectively [8, 22]. Considering the results of previous reports and our data described in the present study, we expected that the populations of F⁺C⁺ cells and F⁺C⁻ cells would be identical to those of F⁺Pa⁺ and F⁺Pa⁻ cells, respectively. However, we observed a subset of CAR⁺ cells in F⁺Pa⁻ cells. Therefore, Flk1⁺ cells could be divided into three populations (F⁺Pa⁻C⁻ cells, F⁺Pa⁻C⁺ cells, and F⁺Pa⁺C⁺ cells) according to the expression patterns of three surface markers (Fig. 3A). Taking these results, together with those of previous reports, it is reasonable to assume that F⁺Pa⁻C⁻ cells and F⁺Pa⁺C⁺ cells are differentiated into hematopoietic and cardiac cells, respectively, although the differentiation capacity of F⁺Pa⁻C⁺ cells is still unknown. Therefore, we examined the differentiation capacity of these three populations.

Figure 3.

Identification of the novel Flk1⁺ cell subsets derived from mouse embryonic stem cells (ESCs) based on coxsackievirus and adenovirus receptor (CAR) expression. (A): Mouse ESC-derived embryoid bodies, which were cultured for 7 days, were stained with anti-Flk1, anti-PDGFRα, and anti-CAR antibodies. Representative data from fluorescence-activated cell sorting (FACS) plots are shown. (B): FACS-sorted cells were cultured on OP9 cells with hematopoietic cytokines for 7 days. The cells were then collected and stained with each antibody. CD31^highCD45⁻ cells and CD31^lowCD45⁺ cells represent endothelial cells and hematopoietic cells, respectively. Representative results from one of three independent experiments are shown. (C): The hematopoietic differentiation potentials of FACS-sorted cells were estimated using a limiting dilution assay. The wells in which no hematopoietic cells were detected were defined as negative wells. (D): Flk1⁺ mesodermal subsets were sorted and cultured on OP9 stromal cells for 7 days. Then, the expression of α-actinin (Di–Diii), cTNT and MLC2a (Div, Dv), and cTNT and MLC2v (Dvi, Dvii) were investigated by immunostaining. Scale bars = 100 μm. (E): Endothelial differentiation potential of each population was examined by an analysis based on CD31 expression and Ac-LDL uptake. Scale bars = 100 μm. (F): Mouse ESC-derived F⁺Pa⁻C⁻ cells, F⁺Pa⁻C⁺ cells, and F⁺Pa⁺C⁺ cells were sorted by FACS and then subjected to microarray analysis. Representative clusters of the indicated lineage-specific genes are shown in the heat map. Red shading indicates increased expression and green shading, decreased expression. Abbreviations: Ac-LDL, acetylated low-density lipoprotein; cTNT, cardiac troponin T; F⁺Pa⁺C⁺, Flk1⁺PDGFRα⁺CAR⁺; F⁺Pa⁻C⁺, Flk1⁺PDGFRα⁻CAR⁺; F⁺Pa⁻C⁻, Flk1⁺PDGFRα⁻CAR⁻; FITC, fluorescein isothiocyanate; Flk1, fetal liver kinase 1; MLC2, myosin light chain 2; PDGFRα, platelet-derived growth factor receptor α.

Flow cytometric analysis showed that hematopoietic cells were differentiated from both F⁺Pa⁻C⁻ and F⁺Pa⁻C⁺ cells after culturing on OP9 cells, although F⁺Pa⁻C⁺ cells generated fewer hematopoietic cells than did F⁺Pa⁻C⁻ cells (Fig. 3B). Consistently, limiting dilution analysis also revealed a lower frequency of hematopoietic progenitor cells from F⁺Pa⁻C⁺ cells relative to that from F⁺Pa⁻C⁻ cells (Fig. 3C). As expected, no hematopoietic cells were differentiated from the F⁺Pa⁺C⁺ cell cultures (Fig. 3B, 3C).

We next evaluated the cardiac and endothelial differentiation potential of the three populations. Not only F⁺Pa⁺C⁺ cells, but also F⁺Pa⁻C⁺ cells expressed α-actinin (Fig. 3Di–3Diii) and cTNT, concomitant with MLC2a or MLC2v (Fig. 3Div–3Dvii). Cardiomyocytes expressing cTNT were more frequently observed in F⁺Pa⁺C⁺ cell-derived cells (Fig. 3Di, 3Dii; supplemental online Fig. 2), indicating that the cardiac differentiation potential of F⁺Pa⁺C⁺ cells was greater than that of F⁺Pa⁻C⁺ cells. In contrast, F⁺Pa⁻C⁻ cells did not show any cardiac potential. We also observed endothelial cells from all populations (Fig. 3E). Thus, these results showed that CAR⁺ cells were also present within F⁺Pa⁻ cells, which are reported as hemoangiogenic progenitor fractions. Moreover, newly identified F⁺Pa⁻C⁺ cell populations could differentiate into cardiac cells, as well as hematopoietic and endothelial cells.

Next, we compared the gene expression profiles of the three populations using microarray analysis (Fig. 3F). Reflecting the data described in the previous paragraphs, F⁺Pa⁻C⁻ cells markedly expressed the genes associated with hematopoiesis and angiogenesis, but they expressed low levels of cardiomyocyte-related genes. In contrast, in F⁺Pa⁺C⁺ cells, the expression of genes involved in hematopoiesis and angiogenesis was downregulated, and elevated expression of cardiomyocyte-related genes was observed. The F⁺Pa⁻C⁺ cell populations expressed most of the genes associated with hematopoiesis, angiogenesis, and cardiomyocytes. Collectively, our data suggest that the three types of populations in Flk1⁺ cells derived from mouse ESCs would be distinct populations with respect to gene expression patterns and differentiation potentials.

Expression of Flk1, PDGFRα, and CAR Distinguishes Flk1⁺ Cell Populations in Mouse Embryos

We next examined whether CAR expression can distinguish Flk1⁺ cell populations in vivo. We detected F⁺C⁻ cells and F⁺C⁺ cells in developing mouse embryos at E8.5 by FACS analysis (Fig. 4A) and found that embryo-derived F⁺C⁻ cells and F⁺C⁺ cells showed the hematopoietic- and cardiac-differentiation tropism, respectively (Fig. 4B, 4C, top). Embryo-derived F⁺C⁺ cells and F⁺C⁻ cells also displayed endothelial potential (Fig. 4C, bottom). These data indicate that CAR expression could also distinguish the two populations in embryo-derived Flk1⁺ cells that show hematopoietic- and cardiac- differentiation tropism. Furthermore, just as with the case of mouse ESC-derived cells, embryo-derived Flk1⁺ cells could be separated into three populations (F⁺Pa⁻C⁻ cells, F⁺Pa⁻C⁺ cells, and F⁺Pa⁺C⁺ cells) according to the expression of Flk1, PDGFRα, and CAR (Fig. 4D). Note that F⁺Pa⁻C⁺ cells were actually present in the embryos, and embryo-derived F⁺Pa⁻C⁺ cells differentiated into hematopoietic cells, endothelial cells, and cardiac cells, although the hematopoietic differentiation potentials were extremely low (Fig. 4E, 4F). The differentiation potentials of embryo-derived F⁺Pa⁻C⁻ cells and F⁺Pa⁺C⁺ cells were also identical to those of the ESC-derived populations (Fig. 4E, 4F). Using GFP-expressing embryos, which were obtained by crossing male GFP-TG mice with female wild-type mice, we confirmed the absence of contamination of maternal cells in our analysis (supplemental online Fig. 3). Our findings therefore indicate that CAR expression could also identify novel populations in Flk1⁺ mesodermal cells in vivo.

Figure 4.

Separation of Flk1⁺ mesodermal subsets in developing embryos according to the expression of CAR. (A): After mouse embryos (E8.5) were isolated and dissociated into single cells as described in Materials and Methods, the embryo-derived cells were stained with antibodies against Flk1 and CAR. (B, C): Embryo-derived F⁺C⁺ cells and F⁺C⁻ cells were purified using fluorescence-activated cell sorting and then cultured on OP9 stromal cells, as described in Materials and Methods, before flow cytometric analysis (B) or immunostaining (C). (D) Mouse embryo (E8.5)-derived cells were stained with anti-Flk1, anti-PDGFRα, and anti-CAR antibodies and then subjected to flow cytometric analysis. (E, F): Each Flk1⁺ mesodermal subset was cultured on OP9 stromal cells for 7 days to determine the hematopoietic (E) and cardiac (F) differentiation potential. Scale bars = 100 μm (C, F). Abbreviations: CAR, coxsackievirus and adenovirus receptor; cTNT, cardiac troponin T; DAPI, 4′6-diamidino-2-phenylindole; E8.5, embryonic day 8.5; F⁺C⁺, Flk1⁺CAR⁺; F⁺C⁻, Flk1⁺CAR⁻; F⁺Pa⁺C⁺, Flk1⁺PDGFRα⁺CAR⁺; F⁺Pa⁻C⁺, Flk1⁺PDGFRα⁻CAR⁺; F⁺Pa⁻C⁻, Flk1⁺PDGFRα⁻CAR⁻; FITC, fluorescein isothiocyanate; Flk1, fetal liver kinase 1; MLC2, myosin light chain 2; PDGFRα, platelet-derived growth factor receptor α; SSC, side scatter.

CAR Expression Separates Hematopoietic and Cardiac Potential of Human PSC-Derived KDR⁺ Cells

The data described in the previous section led to the expectation that hematopoietic and cardiac progenitor cells in human mesodermal cells could probably be separated according to the CAR expression level. To test this possibility, human ESCs and iPSCs were differentiated into KDR⁺ mesodermal cells by EB formation under serum-free conditions (Fig. 5A). Under this condition, we observed an elevated expression of genes in association with hematopoiesis (Runx1 and Scl), vasculogenesis (Cd31 and vWF), and cardiogenesis (Isl1 and Tbx5) during EB culture. In contrast, the expression levels of undifferentiated marker genes (Oct-3/4 and Nanog) were downregulated (Fig. 5B). Although homogenous expression of CAR was observed in undifferentiated human ESCs, both CAR⁺ cells and CAR⁻ cells were observed in human ESC-derived KDR⁺ mesodermal cells (Fig. 5C). Importantly, CAR⁻ cells expressed high levels of KDR, and the cells expressing low levels of KDR expressed CAR (Fig. 5C, bottom). Human ESC-derived KDR⁺CAR⁻ (K⁺C⁻) cells highly expressed the genes for hematopoietic transcription factors (Fig. 5D, left), and efficiently produced CD45⁺ hematopoietic cells after coculture with OP9 cells (Fig. 5E). However, KDR⁺CAR⁺ (K⁺C⁺) cells displayed increased expression of the genes involved in cardiogenesis in contrast to the K⁺C⁻ cells (Fig. 5D, right), and K⁺C⁺ cells exhibited a potent capacity for differentiation into cTNT-expressing cardiomyocytes (Fig. 5F). These cardiomyocytes also coexpressed MLC2a but not MLC2v (Fig. 5F). We also found CD144- or CD31-expressing endothelial cells in both populations (Fig. 5E, 5G). Similar results were obtained when human iPSCs were used instead of human ESCs (supplemental online Fig. 4). Therefore, these results indicate that the hematopoietic or cardiac lineage in human PSC-derived KDR⁺ mesodermal cells could be successfully separated according to the expression levels of CAR.

Figure 5.

Separation of the hematopoietic and cardiac mesoderm in KDR⁺ cells derived from human pluripotent stem cells (PSCs) according to CAR expression. (A): The experimental protocols for differentiation into hematopoietic cells and cardiac cells from human PSCs. (B): The expression of genes associated with the undifferentiated state (Oct-3/4 and Nanog), hematopoiesis (Runx1 and Scl), endothelial cells (CD31 and vWF), and cardiogenesis (Isl1 and Tbx5) in human embryonic stem cells (ESCs) (white bars, human ESC line, KhES-3) or induced pluripotent stem cells (iPSCs) (gray bars, human iPSC line, 201B7) was investigated using quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis. The data are expressed as the mean ± standard deviation (SD) (n = 3). (C): Human ESCs (KhES-3) were stained with an anti-human CAR antibody (upper, day 0). Human ESCs were cultured for 7 days as described in Materials and Methods and subsequently stained with anti-human CAR and anti-human KDR antibodies (below, day 7). The stained cells were subjected to flow cytometric analysis. (D): Total RNA was isolated from human ESC-derived K⁺C⁺ cells and K⁺C⁻ cells. Next, qRT-PCR analysis was performed. The data are expressed as the mean ± SD (n = 3). ∗, p < .05; ∗∗, p < .01. (E): K⁺C⁺ cells and K⁺C⁻ cells derived from human ESCs were sorted and cultured on OP9 cells with the hematopoietic cytokines. Next, fluorescence-activated cell sorting (FACS) analysis was performed. For this analysis, human cells that expressed TRA-1-85 were gated. (F, G): FACS-sorted cells were cultured on gelatin-coated 96-well plates in the presence of VEGF, FGF2, and Y27632. Next, the cells were stained with each antibody. The expression of cTNT, MLC2a, MLC2v, and CD31 in the cultures was detected using immunostaining. (F, Upper): Representative gross image of cTNT⁺ cell colonies in 96-well plates is shown in red. Scale bars = 100 μm (F, G). Abbreviations: BMP4, bone morphogenetic protein 4; cTNT, cardiac troponin T; D, day; EB, embryoid body; FGF2, fibroblast growth factor 2; Flt3-L, Fms-like tyrosine kinase ligand; FSC, forward scatter; IL, interleukin; K⁺C⁺, KDR⁺CAR⁺; K⁺C⁻, KDR⁺CAR⁻; KDR, kinase insert domain receptor; MLC2, myosin light chain 2; SCF, stem cell factor; TPO, thrombopoietin; VEGF, vascular endothelial growth factor.

Just as was the case with the mouse ESCs and mouse embryos, human ESC- and iPSC-derived KDR⁺ cells could be divided into three populations (K⁺Pa⁻C⁻ cells, K⁺Pa⁻C⁺ cells, and K⁺Pa⁺C⁺ cells) according to the expression of three surface markers (Fig. 6A, 6B). The differentiation potential of these 3 populations was largely identical to those of mouse ESCs and embryos (Fig. 6C, 6D). Collectively, our findings showed that CAR is a novel marker for distinguishing heterogeneous mesodermal cells derived from both mouse and human PSCs.

Figure 6.

Separation of hiPSC-derived KDR⁺ populations into three subsets according to the expression of KDR, PDGFRα, and CAR. (A, B): hiPSC (201B7)- and hESC (KhES-3)-derived embryoid body cells were stained with anti-human KDR, anti-human PDGFRα, and anti-human CAR antibodies before flow cytometric analysis. Representative fluorescence-activated cell sorting (FACS) plots are shown. (C): hPSC-derived K⁺Pa⁻C⁻ cells, K⁺Pa⁻C⁺ cells, and K⁺Pa⁺C⁺ cells were sorted and cultured on OP9 stromal cells in the presence of hematopoietic cytokines. FACS analysis was performed after staining with the indicated antibodies. Representative results from one of three independent experiments are shown. (D): Three populations derived from hPSCs were cultured on gelatin-coated 96-well plates with vascular endothelial growth factor, fibroblast growth factor 2, and Y27632. Next, α-actinin expression in the cells was examined using immunostaining. Representative results from one of three independent experiments are shown. Nuclei were counterstained with 4′6-diamidino-2-phenylindole. Scale bars = 100 μm. Abbreviations: CAR, coxsackievirus and adenovirus receptor; hESC, human embryonic stem cell; hiPSC, human induced pluripotent stem cell; K⁺Pa⁺C⁺, KDR⁺PDGFRα⁺CAR⁺; K⁺Pa⁻C⁺, KDR⁺PDGFRα⁻CAR⁺; K⁺Pa⁻C⁻, KDR⁺PDGFRα⁻CAR⁻; KDR, kinase insert domain receptor; MLC2, myosin light chain 2; PDGFRα, platelet-derived growth factor receptor α.

Discussion

Flk1⁺ mesodermal cells are known to give rise to various types of cells, including hematopoietic cells, endothelial cells, and cardiomyocytes. Flk1⁺ cells are therefore assumed to be heterogeneous populations containing hemoangiogenic and cardiac mesoderm. Fate mapping analysis revealed that the cells expressing high or low levels of Flk1 differentiated into hemoangiogenic cells or cardiomyocytes, respectively [4]. However, it is difficult to separate distinct progenitor fractions in Flk1⁺ mesodermal cells only by the expression levels of Flk1 alone. Thus, the identification of a cell surface marker would be necessary for isolation of these progenitors. We have shown that mouse and human PSC- and mouse embryo-derived Flk1⁺ cells can be subdivided into CAR⁺ and CAR⁻ cells (Figs. 1, 4, 5). Our analysis also revealed that the cells harboring cardiac differentiation potential were restricted to the F⁺C⁺ cell subsets, and the F⁺C⁻ cell subsets efficiently differentiated into hematopoietic cells (Figs. 1, 2, 4, 5). As illustrated in Figure 7, our findings have clearly demonstrated that CAR expression can separate hematopoietic and cardiac mesoderm in Flk1⁺ cells derived from PSC differentiation cultures and mouse developing embryos.

Figure 7.

Differentiation potentials of three Flk1⁺ subsets. Mouse and human PSC- and mouse embryo-derived Flk1⁺ mesodermal cells could be separated into three populations (Flk1⁺, PDGFRα⁻, CAR⁻ cells [F⁺Pa⁻C⁻]; Flk1⁺, PDGFRα⁻, CAR⁺ cells [F⁺Pa⁻C⁺]; and Flk1⁺, PDGFRα⁺, CAR⁺ [F⁺Pa⁺C⁺] cells). F⁺Pa⁻C⁻ cells and F⁺Pa⁺C⁺ cells have the potential to differentiate into hematopoietic cells and cardiomyocytes, respectively. Newly identified F⁺Pa⁻C⁺ cells could differentiate into, not only hematopoietic cells and endothelial cells, but also cardiomyocytes. Abbreviations: CAR, coxsackievirus and adenovirus receptor; Flk1, fetal liver kinase 1; PDGFRα, platelet-derived growth factor receptor α.

To date, PDGFRα has been widely used as a well-known marker for the separation of distinct mesodermal subsets in Flk1⁺ cells; the F⁺Pa⁺ cells and F⁺Pa⁻ cells contained cardiac and hematopoietic progenitor cells, respectively [8, 22]. Our data showed that the differentiation potential of F⁺C⁺ cells and F⁺C⁻ cells was similar to that of F⁺P⁺ cells and F⁺Pa⁻ cells, respectively, indicating that CAR, as well as PDGFRα, would be a useful marker for segregation of different mesodermal subsets in Flk1⁺ cells. Mouse PSC-derived F⁺Pa⁻ cells could be further divided into F⁺Pa⁻C⁺ cells and F⁺Pa⁻C⁻ cells according to the CAR expression level (Fig. 3). In addition, F⁺Pa⁻C⁺ cell populations could differentiate not only into hematopoietic cells and endothelial cells but also into cardiomyocytes (Fig. 3). Remarkably, this population was actually present in mouse embryos and human PSC-derived cells (Figs. 4, 6). The cardiac differentiation potential of F⁺Pa⁻ cell subsets has not been reported until now. Although F⁺Pa⁻C⁺ cells differentiate into cardiomyocytes less efficiently than F⁺Pa⁺C⁺ cells, it will clearly be necessary to study the difference between the F⁺Pa⁻C⁺ cells and F⁺Pa⁺C⁺ cells as cardiac progenitors. It might be that F⁺Pa⁻C⁺ cells are more immature than F⁺Pa⁺C⁺ cells, because F⁺Pa⁻C⁺ cells expressed the genes associated with hematopoietic cells, endothelial cells, and cardiomyocytes (Fig. 3F). It is still unknown whether F⁺Pa⁻C⁺ cells are multipotent progenitor cells or a mixture of hemoangiogenic progenitors [2, 5] and cardiovascular progenitors [6, 29]. Additional investigation is needed to characterize the newly identified F⁺Pa⁻C⁺ cell subset.

Zhang et al. [30] showed that podocalyxin (podxl) expression separates the hematopoietic and cardiac potential of mouse ESC-derived Flk1⁺ mesoderm. Flk1⁺ cells expressing podxl showed definitive hematopoietic potential, and Flk1⁺ cells lacking podxl gave rise to cardiomyocytes [30]. Their results, combined with our data, would indicate that the characteristics of F⁺podxl⁺ cells and F⁺podxl⁻ cells resemble those of F⁺C⁻ cells and F⁺C⁺ cells, respectively. However, the differentiation potential of F⁺podxl⁺ cells and F⁺podxl⁻ cells in embryos has not been examined, and the appearance of F⁺podxl⁺ cells and F⁺podxl⁻ cells in human PSC-derived mesoderm was not documented in their study [30]. In contrast, we examined Flk1⁺ mesodermal subsets using mouse embryos and human PSCs and found that F⁺C⁺ cells and F⁺C⁻ cells were present in the embryos and human PSC-derived cells. In addition, we demonstrated that the differentiation potential of mouse embryo and human PSC-derived populations was largely identical to that of mouse PSC-derived cells, as described in Results. It would be of interest to examine the temporal relationship to CAR and Podxl in PSC- and embryo-derived Flk1⁺ cells.

Recent reports have shown that transcription factor Etv2 (also known as ER71) and Scl controlled the lineage specification into hematopoietic mesoderm versus cardiac mesoderm [21, 22, 31]. In addition, Flk1⁺Isl1⁺Nkx2.5⁺ cells have been reported to be cardiovascular progenitor cells in PSC- and developing embryo-derived cells [32]. However, because these molecules, with the exception of Flk1, functioned as transcription factors, the generation of transgenic ESCs or knock-in ESCs is required for isolation of mesodermal progenitor cells based on the expression of transcription factors. In contrast, our methods using CAR antibodies make it possible to easily collect the distinct subsets of Flk1⁺ cells without genetic manipulation. Therefore, CAR would be a valuable cell surface marker for separating hematopoietic and cardiac progenitor cells. In the case of using human PSCs, our sorting method using CAR expression would be expected to be instrumental in clinical applications, such as regenerative medicine.

Human PSC (hPSC)-derived functional cells are expected to be needed for application for drug discovery and regenerative medicine. For instance, hPSC-derived cardiomyocytes enable us to assess the cardiac toxicity of drugs in vitro and to create the cardiac cell sheets for regenerative medicine. In addition, hPSC-derived hematopoietic cells are considered an alternative cell source of adult hematopoietic cells for transplantation therapy. To realize these clinical research goals, the identification of surface markers on progenitor population is thought to be needed to purify the desired lineage. The present study clearly showed that CAR would be a useful marker for separating hematopoietic lineage and cardiac lineage in KDR⁺ mesodermal cells, thus indicating that our findings could contribute to both generate more efficient protocols for differentiation into hematopoietic or cardiac cells from hPSCs and advance clinical applications using hPSC-derived hematopoietic cells and cardiomyocytes. In particular, considering that a large number of cells are required for cell transplantation therapy and drug discovery, K⁺C⁺ cell subsets would be suitable for collecting cardiac lineage cells, because K⁺C⁺ cell subsets can also purify the cardiac potential in K⁺P⁻ cell subsets. Therefore, by combining K⁺C⁺ cell subsets with other techniques such as a cell sheet technology, a much greater number of cardiomyocyte sheets with high purity could probably be generated from human PSCs, thereby accelerating the research into cardiac regeneration.

Author Contributions

K. Tashiro: conception and design, financial support, collection and assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript; N.H., T.Y., and K. Takayama: collection and assembly of data, final approval of manuscript; A.O.: collection and assembly of data, data analysis and interpretation, final approval of manuscript; H.M.: conception and design, data analysis and interpretation, final approval of manuscript; K.K.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.

Disclosure of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.