Beng H Chong, Level 2 Pitney Building, St George Hospital, Belgrave Street, Kogarah, NSW 2217, Australia. Tel.: +612 91132010; fax: +612 91133998. E-mail: email@example.com
To cite this article: Liang SX, Khachigian LM, Ahmadi Z, Yang M, Liu S, Chong BH. In vitro and in vivo proliferation, differentiation and migration of cardiac endothelial progenitor cells (SCA1+/CD31+ side-population cells). J Thromb Haemost 2011; 9: 1628–37.
Summary. Background: Side-population (SP) cells are a select population identified by a capacity to efflux Hoechst dye and are enriched for stem/progenitor cell activity. Previous studies suggested that cardiac SP (CSP) cells could be divided into SCA1+/CD31− and SCA1+/CD31+ CSP cells. SCA1+/CD31− CSP cells have been shown to be cardiac stem/progenitor cells. However, SCA1+/CD31+ CSP cells have not been fully characterized. Objective: The aim of the present study was to characterize SCA1+/CD31+ CSP cells in the adult mouse heart, and investigate their abilities to proliferate, differentiate, vascularize and migrate in vitro and in vivo. Results: Using fluorescence-activated cell sorting (FACS), RT-PCR, and assays of cell proliferation, differentiation and migration, and a murine model of myocardial infarction (MI), we showed that SCA1+/CD31+ CSP cells express stem cell and endothelial-specific genes, and reside in the blood vessels. These cells were able to proliferate, differentiate, migrate and vascularize in vitro and in vivo. After MI, SDF-1α and CXCR4 were up-regulated in the damaged myocardium and on SCA1+/CD31+ CSP cells, respectively. Our results further showed that SDF-1α induced migration of these cells in vitro. Importantly, we found that SCA1+/CD31+ CSP cells could migrate into the ischemic region from the non-ischemic area within the myocardium and form a vascular tube-like structure after MI. Conclusions: Based on the gene expression profile, localization of SCA1+/CD31+ CSP cells, and their ability to proliferate, differentiate, migrate and vascularize in vitro and in vivo, we postulate that SCA1+/CD31+ CSP cells may represent endothelial progenitor cells in the mouse heart.
In recent years, different populations of cardiac stem or progenitor cells have been reported to reside within the adult heart. These cells have been identified and isolated using different methods, based on (i) expression of stem cell-associated markers such as c-kit and SCA1 , (ii) a specific culture system to obtain cardiospheres , (iii) expression of Isl-1, a homeodomain transcription factor , and (iv) ability to efflux Hoechst 33342 dye [4–6] giving rise to side-population (SP) cells that appear ‘off to the side’ of the main cell population (MP) as a result of their weak Hoechst staining. SP cells, enriched for stem and progenitor cells, have been identified in various mammalian tissues including bone marrow, liver, skeletal muscle, brain, heart and lung [7–9].
Cardiac SP (CSP) cells are capable of differentiation into cardiomyocytes, endothelial and smooth muscle cells [4,8,10]. CSP cells can be further divided into SCA1+/CD31− and SCAl+/CD31+ cells [5,11]. SCA1+/CD31− but not SCA1+/CD31+ CSP cells are able to differentiate into cardiomyocytes . Although SCA1+/CD31− CSP cells have been extensively studied, but not SCA1+/CD31+ CSP cells, the nature of which has remained unclear.
Cell migration is an important feature of stem cells. We have shown that SCA1+/CD31− CSP cells are able to migrate from healthy myocardium into the myocardial infarct (MI) region after acute MI, and that the SDF-1α/CXCR4 pathway plays an important role in this process . However, the role of the SDF-1α/CXCR4 pathway in SCA1+/CD31+ CSP cells is unknown.
In the present study, we characterized SCA1+/CD31+ CSP cells in the adult mouse heart, and investigated their abilities to proliferate, differentiate, vascularize and migrate in vitro and in vivo.
Materials and methods
MaterialsC57Bl/6J mice (Animal Resources Centre, Perth, Australia), collagenase B, dispase II (Roche Diagnostics, Mannheim, Germany), fetal calf serum, HEPES (Invitrogen, Carlsbad, CA, USA), PKH26, verapamil, Hoechst 33342, anti-mouse α-actinin and streptavidin–Cy3 (Sigma-Aldrich, St Louis, MO, USA) were obtained as indicated.
Murine myocardial infarction model and in vivo cell migration study
MI was induced by permanent ligation of the left anterior descending coronary artery (LADCA) ligation in C57Bl/6J mice (female, 8 to 10 weeks old) as we previously described . Briefly, mice were anesthetized with 0.1 mL/10 g body weight of a cocktail consisting of 1 part Hypnorm, 1 part Midazolam and 2 parts distilled water, and ventilated with a standard rodent ventilator (Harvard Ventilator, Harvard Apparutus, Boston, MA, USA). The thoracic cavity was opened by incision between the second and the third intercostal space, a rodent rib spreader introduced to allow for visualization of the heart and MI induced. To test CSP cell migration from the non-ischemic to the ischemic area of the heart, SCA1+/CD31+ or SCA1+/CD31− CSP cells (approximately 150 000) labeled with PKH26, a red fluorescence dye, were injected into the non-ischemic right ventricle wall. The cell injection was administered into the right ventricle with the needle directed towards the right auricle to ensure that the cells were delivered away from the ischemic area . To investigate whether SDF-1α is able to initiate SCA1+/CD31+ CSP cells migration without MI, PKH26-labeled SCA1+/CD31+ CSP cells (approximately 150 000) were injected into the right ventricle wall as described above, and SDF-1 (500 ng in 50 μL phosphate-buffered saline [PBS]) was injected into the left ventricle wall but without left anterior descending coronary artery (LADCA) ligation. The heart was harvested 14 days post-injection and sectioned. In other experiments, hearts of mice with LADCA ligation were harvested 3 days post-MI to study SDF-1α and CXCR4 expression in the ischemic myocardium and CSP cells, respectively. In the control experiments, sham-operated mice underwent identical procedures without LADCA ligation. The studies described in the present study were approved by the institutional Animal Care and Ethics Committee.
SCA1+/CD31+ CSP cell isolation, fluorescence-activated cell sorting and analysis, and magnetic depletion of CD45+ cells
CSP cell isolation was performed as previously described [5,9]. Briefly, minced cardiac tissue was digested with 0.1% collagenase B, 2.4 U mL−1 dispase II and 2.5 mm L−1 CaCl2 for 45 min. Cell suspensions were incubated with Hoechst 33342 (5 μg mL−1) for 90 min at a concentration of 106 nucleated cells mL−1 in the presence or absence of verapamil (50 mm) . Cells were subsequently labeled with monoclonal rat anti-mouse antibodies, which included anti-SCA1-FITC, anti-CD31-PE, anti-c-kit-APC, anti-CD45-APC, anti-VEGFR2-APC, anti-CD34-APC and anti-CXCR4-APC (BD, San Diego, CA, USA), and polyclonal rabbit anti-mouse von Willebrand factor (VWF) detected with donkey anti-rabbit-APC (Chemicon, Temeula, CA, USA). For all studies, cells stained with an isotype/control antibody were employed as negative controls, and to establish gating parameters for positive cells. All stains were performed at 4 °C for 30 min in the dark.
Fluorescence-activated cell sorting (FACS) and analysis were performed using a FACStar Plus flow cytometer (BD). The Hoechst dye was excited at 350 nm. SP cells were identified as previously described . Data were analyzed using BD FACSDiva™ software v4.12 (BD).
After removal of contaminating CD45+ cardiac cells using biotin-conjugated anti-CD45 antibody and streptavidin-conjugated microbeads (Miltenyi Biotech, Bergisch Glabach, Germany), SCA1+/CD31+ CSP and MP cells were sorted and analyzed by FACS as previously described .
Methylcellulouse assay (colony-forming unit assay)
SCA1+/CD31+ CSP and MP cells (7000 cells mL−1) were plated in Methocult GF M3534 media according to the manufacturer’s instructions (StemCell Technologies, Vancouver, Canada) as previously described [4,5,10]. Cell colonies consisting of more than 25 cells were scored after 14 day in culture. In order to determine whether these cells are endothelial cells, methocult media was cut into very small pieces and incubated with Dulbecco’s modiﬁed Eagle’s medium (DMEM) at 37 °C with shaking for 30 min. After methocult media was dissolved in DMEM, the cells were collected and subsequently labeled with anti-CD31-PE, and analyzed by FACS.
Primary cell culture
SCA1+/CD31+ CSP and MP cells (that represented mature endothelial cells) were cultured with EGF, VEGF, IGF-1, bFGF, fetal bovine serum (FBS), hydorcortisone, ascorbic acid, heparin and antimicrobial agents on fibronectin-coated eight-well chambers (10 000 cells/well) at 37 °C, 5% CO2 for 14 days as previously described .
The heart tissue was harvested 3 or 14 days post-MI and injection, frozen in optimal cutting temperature (OCT) (Tissue-Tek, Redding, CA, USA) and sectioned at 5 μm. Immunofluorescence staining was performed as previously described . Primary antibodies used included polyclonal rabbit anti-mouse VWF, monoclonal rat anti-mouse ABCG2 (Chemicon), polyclonal rabbit anti-mouse SDF-1α (eBioscience, San Diego, CA, USA), polyclonal rabbit anti-mouse SCA1 (R&D System, Minneatolis, MN, USA), monoclonal rat anti-mouse CD31-biotin (BD) and monoclonal anti-mouse α-actinin; and conjugated secondary antibodies used included goat anti-rabbit biotin (Vector-Lab, Burlingame, CA, USA), streptavidin-FITC (BD) or -Cy3, Goat anti-mouse Alexa fluor 488 (Invitrogen) and rabbit anti-rat FITC (Chemicon). Nuclei were counterstained with 4,6-diamidino-2-phenylindol dihydrochloride (DAPI) before evaluation by fluorescent microscopy. In the present study, isotype irrelevant antibodies were employed as negative controls. Slides were examined with an Olympus (Redwood, CA, USA) DP70 microscope and pictures taken with a Spot digital camera BX51.
Quantitative real-time RT-PCR
Total RNA from freshly isolated SCA1+/CD31+ CSP and MP, and cultured SCA1+/CD31+ CSP cells, non-infarcted and infarcted myocardium was extracted using a RNAeasy-Mini-Kit according to the manufacturer’s instruction (Qiagen, Hilden, Gemany). Complementary DNA was generated using a RT kit based on the manufacturer’s instruction (Promega, Madison, WI, USA). Quantitative real-time RT-PCR (qRT-PCR) was performed using a Rotor-gene 3000 (Corbett Research, Sydney, Australia), and a SYBER Green qPCR-SuperMix-UDG kit (Invitrogen) according to the manufacturers’ instructions. Primers for the qRT-PCR were obtained from Qiagen, including ABCG2 (product ID: QT00173138), VEGFR2 (product ID: QT00097020), VEGFR1 (product ID: QT00096292), Tie2 (product ID: QT00114576), SMA (product ID: QT0006746), VWF (product ID: QT00116795), CD133 (product ID: QT01065162), Nkx2.5 (product ID: QT0024810), SDF-1α (product ID: QT00161112) and β-actin (product ID: QT01136772). The thermal cycler conditions were as follows: Cycle 1 (95 °C for 3 min) X1, Cycle 2 (95 °C for 30 s, 57 °C for 30 s, 72 °C for 30 s) X45, Cycle 3 (72 °C for 4 min). All samples were performed in triplicate. Gene expression levels of ABCG2, VEGFR2, VEGFR1, Tie2, SMA, VWF, CD133, Nkx2.5 and SDF-1α were normalized to a housekeeper gene, β-actin. Amplification data were analyzed using Corbett Rotor-gene 3000 software (Corbett Research) as we previously described . Gene expression data are presented in relative units.
Migration assay (chemotaxis assay)
The in vitro migration of SCA1+/CD31+ CPS cells was assessed using transwells polycarbonate membrane (Millipore, Billerica, MA, USA) with 8 μm pores size as we previously described . Briefly, 50 000 cells in 50 μL serum-free DMEM were added to the upper suffer of the membrane in each chemotaxis chamber. SDF-1α (R&D Systems) at a concentration of 0, 50, 100 or 500 ng mL−1 was placed in the lower chamber. After 4 h incubation at 37 °C in 5% CO2 and 95% humidity, the upper surface of membranes was scraped free of cells and debris, and gently washed with PBS. Membranes were then fixed and stained using a DiffQuick cell fixation and staining kit (Fronine, Sydney, Australia). Cell migration was measured by counting the number of cells that had migrated through pores and adhered to the lower surface of the membrane in five adjacent high-power fields (40×). To block the SDF-1α signal, the cells were incubated with anti-CXCR4 antibody (10 μg mL−1) (eBioscience) for 30 min before being added to the membrane with SDF-1α (500 ng mL−1) in the lower chamber. Experiments were performed in triplicate. Cell numbers of experimental groups were expressed as a fold increase compared with control groups (no SDF-1α), respectively.
Data are presented as mean ± SD. Experimental data were compared using Student’s t-test. Results were considered statistically significant when P < 0.05.
Gene expression and localization of SCA1+/CD31+ CSP cells
Adult mouse hearts were enzymatically dissociated into a single-cell suspension and subjected to FACS analysis. The Hoechst 33342-low CSP cell population was detected and Hoechst efflux in the cells could be completely inhibited by the specific inhibitor, verapamil (Fig. 1A). CSP cells widely expressed SCA1 and CD31 (Fig. 1B), but lacked CD45, a hematopoietic marker . Based on the expression of SCA1 and CD31, CSP cells could be divided into SCA1+/CD31−, SCAl+/CD31+ and SCA1−/CD31− cells [5,11]. To further characterize SCAl+/CD31+ CSP cells in comparison with SCA1+/CD31− CSP cells, both populations were examined for the expression of stem/progenitor cells markers (SCA1 and c-kit), endothelial markers (VEGFR2, CD31 and VWF) and hematopoietic/myeloid markers (CD45 and CD34). CD31− CSP cells expressed only SCA1 and none of the other markers listed above, and all CD31+ CSP cells expressed SCA1. They also stained positively for VEGFR2 and weakly for VWF. Both cell populations were negative for CD45, CD34 and c-kit (Table 1).
Table 1. Comparison of surface markers expressed by SCA1+/CD31− and SCA1+/CD31+ CSP cells
SCA1+/CD31+ CSP (%)
SCA1+/CD31− CSP (%)
36 ± 1
5 ± 1
Using immunofluorescence staining, we found that ABCG2+/CD31+ cells (representing CD31+ SP cells) were restricted to the vasculature in the myocardium (Fig. 2A–D, Fig. S1A–D), and so were SCA1+/CD31+ cells (Fig. S2A–D). In contrast, ABCG2+/CD31− (representing CD31− CSP) cells were localized randomly in the heart mesenchyme (Fig. 2A–D).
To further characterize SCA1+/CD31+ CSP cells, the expression of genes specific for stem/progenitor and endothelial cells in SCA1+/CD31+ CSP cells were examined using qRT-PCR. SCA1+/CD31+ MP cells (mature endothelial cells) were used as the positive control. Like SCA1+/CD31+ MP cells, freshly isolated SCA1+/CD31+ CSP cells expressed endothelial-specific genes including VEGFR1, VEGFR 2, Tie2 and CD133. However, expression of ABCG2 was detected only in SCA1+/CD31+ CSP cells, and not in SCA1+/CD31+ MP cells (Fig. 2E). Recent studies showed that ABCG2, the ATP-binding cassette (ABC) transporter, is restricted to SP cells and is responsible for effluxing the Hoechst 33342 dye conferring on the cells the ‘SP’ phenotype . In addition, higher expression of VEGFR2 and CD133 in SCA1+/CD31+ CSP cells, compared with SCA1+/CD31+ MP cells, was observed (Fig. 2E). Interestingly, SMA (a smooth muscle cell marker) was detected in SCA1+/CD31+ CSP cells, but not in SCA1+/CD31+ MP cells. As expected, Nkx2.5, a marker of early cardiac commitment, was undetectable in SCA1+/CD31+ CSP cells (Fig. 2F). These results are consistent with SCA1+/CD31+ CSP cells being cardiac endothelial stem/progenitor cells and SCA1+/CD31+ MP cells being mature endothelial cells.
SCA1+/CD31+ CSP cells are capable of colony formation, differentiation and vascularization in vitro
We next examined the in vitro-colony forming potential of SCA1+/CD31+ CSP and MP cells isolated from adult murine hearts. SCA1+/CD31+ CSP and MP cells were isolated by FACS and plated on methylcellulose media and examined for colony formation after 14 days in culture, respectively. As shown in Fig. 3, SCA1+/CD31+ CSP cells produced significantly more colonies compared with SCA1+/CD31+ MP cells, indicating a substantially greater potential for colony formation. Expression of CD31 (98%) was detected on cells isolated from methocult media indicating that the colony-forming cells are endothelial in nature. To examine their differentiation capacity, SCA1+/CD31+ CSP and MP cells were sorted and cultured in endothelial growth conditions, respectively. After 14 days in culture, SCA1+/CD31+ CSP cells formed vascular tube-like networks (Fig. 4A). These cells stained uniformly with the endothelial marker VWF (Fig. 4B–D) suggesting they were endothelial or endothelial-like cells. In contrast, SCA1+/CD31+ MP cells attempted to form vascular tube-like structures in 3 days of culture. However, they underwent apoptosis on further culture and vascular tube-like structures were not seen after 14 days in culture (Fig. 4E). Next, freshly isolated and cultured SCA1+/CD31+ CSP cells were analyzed with qRT-PCR for expression of stem/progenitor cell- and endothelial cell-associated genes: ABCG2, VEGFR2, VWF and CD133, respectively. We found that expression of ABCG2 was only detected in freshly isolated SCA1+/CD31+ CSP cells and was undetectable in both cultured SCA1+/CD31+ CSP cells and the MP cells (Fig. 4F). Compared with freshly isolated MP cells and cultured SCA1+/CD31+ CSP cells, VEGFR2 and CD133 expression was significantly increased and VWF expression reduced in the freshly isolated SCA1+/CD31+ CSP cells (Fig. 4F), suggesting that SCA1+/CD31+ CSP cells were able to differentiate into mature endothelial cells in vitro. These results are again consistent with SCA1+/CD31+ CSP cells being progenitors of cardiac endothelial cells.
Transplanted SCA1+/CD31+ CSP cells form avascular tube-like structure in the ischemic myocardium after migration from non-ischemic myocardium
PKH26-labeled SCA1+/CD31+ CSP cells (purity: 95%) were isolated from adult mouse hearts and injected into the right ventricle wall of left anterior descending (LAD) artery-ligated heart. Two weeks post MI, PKH26-labeled transplanted cells were located in the infarcted region and were found to form microvascular tube-like structures (Fig. 5A–D). These cells expressed VWF (mature endothelial cell marker) suggesting that these are mature endothelial cells that have differentiated from the transplanted SCA1+/CD31+ CSP cells (Fig. 6A–D). The injection site, identified with the visible needle tract by the dye staining the myocardium of right ventricle wall (Fig. 5E–G), was situated at a considerable distance from the infarct region in the left ventricle. Our results suggested that SCA1+/CD31+ CSP cells migrated into the ischemic region from the injection site differentiated into endothelial cells and formed vascular tube-like structures after MI. In the control experiment (no LADCA ligation), PKH26-labeled SCA1+/CD31+ CSP cells remained in the right ventricle after cell injection and did not migrate into the left ventricle even when SDF-1α (500 ng) was injected into the myocardium of the left ventricle (data not shown).
Up-regulation of the expression of CXCR4 on SCA1+/CD31+ CSP cells in post-ischemic hearts
We hypothesized that the SDF-1α/CXCR4 pathway plays an important role in migration of SCA1+/CD31+ CSP cells to the ischemic region of MI heart. We also postulated that expression of CXCR4 on SCA1+/CD31+ CSP cells is up-regulated after acute MI. To ensure that SCA1+/CD31+ CSP cells are devoid of contaminating CD45+ SP cells (bone marrow-derived cells), CD45+ cells were removed by magnetic bead separation as previously described , producing a purity of CD45− CSP cells of > 97% in post-MI heart cell suspension (data not shown). Based on multiple colour FACS analysis, we found that CXCR4 expression was substantially increased on SCA1+/CD31+ CSP cells after acute MI. Compared with the non-ischemic heart (Fig. 7A–C), there was greater than a five-fold increase in CXCR4 expression on SCA1+/CD31+ CSP cells in Day 3 post-MI heart (Fig. 7D–G). Notably, as with non-CD45 depleted CSP cells, the efflux of Hoechst 33342 dye in CD45-depleted CSP cells was blocked by verapamil (50 mm), confirming that CD45 depletion did not change the SP phenotype (Fig. 7D).
Up-regulation of expression of SDF-1α in ischemic myocardium
Recent studies have suggested that expression of SDF-1α (CXCR4 ligand) was up-regulated after acute MI [16,17]. Using immunohistochemical staining, expression of SDF-1α was easily detectable in ischemic myocardium (Fig. 8A–D) but was undetectable in the non-ischemic myocardium of the same heart (Fig. 8E–H). These results were confirmed by qRT-PCR data (Fig. 8I).
SDF-1α induces migration of SCA1+/CD31+ CSP cells via CXCR4 in vitro
We hypothesize that SDF-1α induces chemotaxis of SCA1+/CD31+ CSP cells. Utilizing a chemotaxis assay, a significant migratory response towards a SDF-1α gradient (50, 100 and 500 ng mL−1) was detected in this cell population when compared with the control (SDF-1α 0 ng mL−1) (Fig. 9). The chemotactic response to SDF-1α was abolished by blockage of CXCR4 with an anti-CXCR4 antibody (Fig. 9). This further confirmed the role of the SDF-1α–CXCR4 interaction in this migratory response.
Unlike SCA1+/CD31− CSP cells which have been extensively studied, little is known about SCA1+/CD31+ CSP cells. The present study provides new data contributing to our understanding of SCA1+/CD31+ CSP cells as cardiac endothelial progenitor cells, which play a critical role in cardiac repair and regeneration. This is the first detailed study of phenotype, localization, and the ability to form cell colony, differentiate, migrate and vascularize by SCA1+/CD31+ CSP cells in adult mouse heart.
Unlike SCA1+/CD31− CSP cells, SCA1+/CD31+ CSP and MP cells expressed CD31, suggesting that they are both cardiac endothelial cells. This was confirmed by their gene expression profiles. Importantly, however, SCA1+/CD31++ CSP cells have a unique expression of ABCG2 and high expression of VEGFR2 and CD133; this phenotype is consistent with these SP cells being endothelial stem/progenitor cells . Studies have shown that the ABC transporter, ABCG2, is restricted to the SP cells and is now used as a stem cell marker . On the other hand, expression of SMA, a smooth muscle cell marker in SCA1+/CD31+ CSP cells but not MP cells, suggests an ancestral relationship between endothelial and vascular smooth muscle cells. This possibility is consistent with previous studies showing that vascular smooth muscle cells are derived from an FLK1+ (VEGFR2)/TAL1-endothelial-like cell during vasculogenesis .
Like others, we found that CSP cells are highly enriched for cells expressing SCA1, but contrary to popular belief, the majority of SCA1+ cardiac cells are not SP cells [5,21]. In fact, our result shows that SCA1+/CD31+ CSP cells displayed a much higher colony-forming and cell-differentiating potential than SCA1+/CD31+ MP cells. This suggests that the capacity for colony-formation and cellular differentiation of SCA1+ cardiac cells can be attributable to SCA1+ SP cells, rather than SCA1+ cardiac cells as a whole. This finding is consistent with previous reports that SP cells isolated from different tissues have high colony-forming and cell-differentiating capability compared with non-SP cells [5,22].
Our results show that SCA1+/CD31+ CSP cells also have vascularization capability. The cells were able to differentiate into mature endothelial cells and form vascular tube-like structures in vitro. In contrast, SCA1+/CD31+ MP cells failed to form similar structures on prolonged culture (for 14 days). This is consistent with previous studies suggesting that mature endothelial cells are terminally differentiated cells with very little vascularization capability .
It is possible that the majority of stem and progenitor cells in the ischemic region may die by necrosis and apoptosis after MI. Stem and progenitor cells migrating from the surrounding non-ischemic areas, therefore, are needed for repair of damaged tissue and will play an important role in endothelial reconstruction. In our previous study, we have demonstrated that SCA1+/CD31− CSP cells are able to migrate to the ischemic area from the non-ischemic region within myocardium and differentiate into cardiomycyte- and endothelial-like cells after MI . In the present study, we found similarly that transplanted SCA1+/CD31+ CSP cells were able to migrate from the non-ischemic myocardium to the ischemic myocardium and form vascular tube-like structures after MI to further aid vascularization. The identity of the ‘channels’ through which SCA1+/CD31− CSP cells migrate, remains unclear. The cells could migrate along vessels in the micro-circulation or through the myocardial interstitium or both [21,24].
We found that the expression of CXCR4 was markedly increased on SCA1+/CD31+ CSP cells, and SDF-1α (CXCR4 ligand) become expressed and detectable in damaged myocardium by immuno-fluorescence staining and qRT-PCR after MI, but not in non-ischemic myocardium. Importantly, we have demonstrated that SDF-1α induced migration of SCA1+/CD31+ CSP cells in vitro, which was blocked by a specific antibody against CXCR4, suggesting that SDF-1α/CXCR4 pathway may play an important role in migration of SCA1+/CD31+ CSP cells after MI. As without MI, the CSP cells did not migrate to the left ventricle when it was injected with SDF-1α, other chemokines generated by tissue injury would appear to be required to act in concert with SDF-1α as SDF-1α alone is insufficient to induce CSP migration in vivo, which is consistent with the previous study .
As stem cells, SP cells reside in the niche of an organ and contribute to the long-term maintenance and repair of the tissue . Stem cell niches play an important role in regulating the proliferation and differentiation of stem cells . In our previous reports, we have shown that SP cells located in the blood vessel wall of a murine embryonic lung were able to differentiate into endothelial and smooth muscle cells [7,14]. Similarly, in the present study, we have found that ABCG2+/CD31+ (SCA1+/CD31+ CSP) cells were distributed within blood vessels. It seems likely that the stem cell niches in the microvascular environment are important for maintaining SCA1+/CD31+ CSP cell self-renewal and lineage fate. On the other hand, ABCG2+/CD31− (CD31− CSP) cells reside randomly in the mesenchyme of myocardium. In a recent study, Yoon et al. showed that CSP cells of a mouse heart were able to differentiate into endothelial cells . However, their CSP cells expressed both cardiac and endothelial specific genes and were likely to contain different CSP cell subpopulations. Without subpopulation characterization, it is unclear which subpopulation of their CSP cells has endothelial differentiation capability. Nevertheless, our studies and others have clearly demonstrated that CSP cells of a mouse heart contain at least two subpopulations of cells (SCA1+/CD31− and SCA1+/CD31+ cells), according to the expression of SCA1 and CD31 [5,11]. This notion of two distinct subpopulations of CSP cells is strongly supported by the different gene expression profile and location in the heart of SCA1+/CD31+ CSP cells that are distinct from those of SCA1+/CD31− CSP. Our data further indicate the lineage commitment of SCA1+/CD31+ CSP cells and may thus explain the inability of SCA1+/CD31+ CSP cells to differentiate into cardiomyocytes .
In conclusion, in the study we have isolated and characterized SCA1+/CD31+ CSP cells in the adult mouse heart. Based on their gene expression profile, localization, and colony-formation, differentiation, vascularization and migration capability in vitro and in vivo, we suggest that SCA1+/CD31+ CSP cells may serve as a source of cardiac endothelial cell progenitors in the adult mouse heart. Understanding and enhancing the processes of SCA1+/CD31+ CSP cell proliferation, differentiation and migration may hold enormous potential for development of novel therapeutic strategies for vascularization and vessel regeneration in the ischemic heart disease.
S.X. Liang: designed and performed research, and wrote the manuscript; L.M. Khachigian: analyzed data and wrote the manuscript; Z. Ahmadi and M. Yang: analyzed data and edited the manuscript; S. Liu: performed research; B.H. Chong: designed research, organized the study, analyzed the data and edited the manuscript.
This work was supported by the Australian Research Council (DP0558687 to B.C.). We are grateful to S. Wan and G. Mackenzie for their excellent technical work. We thank Dr N. Chapman for her critical review of the manuscript.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.