Human umbilical vein endothelium-derived cells retain potential to differentiate into smooth muscle-like cells.

Mouse embryonic stem-derived cells were recently shown to differentiate into endothelial and smooth muscle cells. In the present study, we investigated whether human umbilical vein endothelium-derived cells retain the potential to differentiate into smooth muscle cells. Examination of biochemical markers, including basic calponin, SM22alpha, prostaglandin E synthase, von Willebrand factor, and PECAM-1, as well as cell contractility, showed that whereas endothelium-derived cells cultured with fibroblast growth factor can be characterized as endothelial cells, when deprived of fibroblast growth factor, a significant fraction differentiates into smooth muscle-like cells. Reapplication of fibroblast growth factor reversed this differentiation. Activin A was up-regulated in fibroblast growth factor-deprived, endothelium-derived cells; moreover, the inhibitory effects of exogenous follistatin and overexpressed Smad7 on smooth muscle-like differentiation confirmed that the differentiation was driven by activin A signaling. These findings indicate that when deprived of fibroblast growth factor, human umbilical vein endothelium-derived cells are capable of differentiating into smooth muscle-like cells through activin A-induced, Smad-dependent signaling, and that maintenance of the endothelial cell phenotype and differentiation into smooth muscle-like cells are reciprocally controlled by fibroblast growth factor-1 and activin A.

were plated on collagen type I gels and subjected to contraction analysis.
Immunocytochemical Study-Cells were double-labeled with polyclonal anti-vWF and monoclonal anti-calponin primary antibodies and then secondarily labeled with Texas Red-conjugated donkey anti-rabbit IgG and fluorescein isothiocyanate-conjugated sheep anti-mouse IgG, respectively. The signals were then detected using a confocal laser scanning microscope (Carl Zeiss, Jena, Germany).
Semiquantitative Analysis of mRNA Expression-Semiquantitative RT-PCR/Southern blot analysis was carried out as described previously (21). Briefly, total RNA was isolated from cells using Isogen (Nippon Gene), after which a 1-g sample was reverse-transcribed using M-MLV reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Using 4% of the reverse-transcribed mix, cDNA fragments of test genes were amplified within the linear range by PCR using AmpliTaq Gold (PerkinElmer Applied Biosynthesis). The specific primers used are listed in Table I. Aliquots of the PCR products were resolved on 1.0% agarose gels and transferred to Hybond Nϩ membranes (Amersham Biosciences). The membranes were then hybridized with the corresponding cDNA probe labeled with digoxygenin-labeled dUTP (DIG-High Prime, Roche Molecular Biochemicals). The signals were visualized, incorporated, and quantitated using NIH Image software (version 1.62).
Immunoblot Analysis-The cells were lysed in lysis buffer (1% Triton ® X-100, 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1ϫ protease inhibitor mixture, Complete TM (EDTA-free); Roche Molecular Biochemicals), and the extracted proteins (60 g) were resolved by SDS-PAGE and electroblotted to polyvinylidene fluoride membranes (Millipore Corp., Bedford MA). The proteins of interest were then detected using appropriate antibodies and an ECL Western blotting detection system (Amersham Biosciences).
Cell Contraction Assay-HUVE-DCs were first cultured with or without FGF-1 plus heparin for 3 weeks on type I collagen-coated plastic culture dishes, after which they were relaxed by incubation with 10 mM 8-bromo-cGMP for 20 min at 37°C and embedded in collagen gels in 24-well plates (final concentration of 8-bromo-cGMP in the gel was 1 mM) (2.4 ϫ 10 5 cells in 0.75 ml gel/well). The collagen gels were prepared using Cellmatrix type I-P (0.3% porcine type I collagen; Nittagelatin Corp., Osaka, Japan) following the manufacturer's instructions. Briefly, 10ϫ M199 medium cell suspension in 1ϫ M199 medium, bicarbonate buffer, fetal bovine serum, and Cellmatrix type I-P solution were combined at a ratio of 1:1:1:1.5:5.5, poured into wells, and allowed to polymerize. The resultant SMC-embedded gels were covered with 0.5 ml of M199 containing 15% fetal bovine serum. After incubating the gels for 17 h, by which time they had detached from the well bottoms, ET-1 (250 nM) was added to the culture once every 30 min. In some cases, ET-receptor (ET-R) antagonists (type A, BQ-123 (30 M); type B, BQ-788 (10 M)) were added to the culture for 30 min before application of ET-1. For quantification, each gel piece was transferred into a 1.5-ml micro-tube containing 500 l of culture medium and then centrifuged, and its volume was measured. For morphological observation, relaxed differentiated cells were seeded onto precasted type I collagen gel (400 l in 24-well plates) at a density of 4 ϫ 10 4 cells/well, cultured for 17 h, and then treated with ET-1 and photographed under a phase-contrast microscope.
To compare the contractile capacity of differentiated HUVE-DCs and HPASMCs, relaxed cells were seeded onto precasted type I collagen gel (150 l in 48-well plates) at a density of 2.8 ϫ 10 4 cells/well. After incubating for 6 h, the cells were treated with 500 nM ET-1 for 25 min and then with 1 M ET-1 for 5 min. Each gel was then carefully detached from the well wall and bottom, incubated for an additional 30 min in the same well, and then transferred onto a transparent plastic sheet. The degree of cell contraction was then determined by optically measuring the size of the cell sheets on the gel pieces.

Induction of SMC-like Cells from HUVE-DCs by Prolonged
Deprivation of FGF-We initially established a primary culture of HUVECs (HUVE-DCs) as described previously (15,16). The cells were cultured in M199 medium supplemented with 15% serum, FGF-1, and heparin in collagen type I-coated flasks and were split 1:4 once every 3-4 days. Cells were always used for experimentation before they had reached 16 population doublings. As shown in Fig. 1a (left), the majority of cells appeared compact and uniform, and when cultured until confluent, they assumed the "cobble stone" appearance typical of ECs (not shown). Interestingly, when we removed FGF-1 and heparin from the growth medium for 3 weeks, an enlarged, flat cell type emerged (Fig. 1a, right). Immunostaining the cells for EC-and SMC-specific markers, i.e. vWF (22)(23)(24)(25) and basic calponin (26 -29), respectively, revealed that all cells maintained in FGF-1-containing medium were positive for anti-vWF antibody (red) and negative for anti-basic calponin (Fig. 1b,  left), but a significant fraction of the FGF-1-deprived cells (9.4 Ϯ 3.0% in number; n ϭ 8) were positive for anti-basic calponin antibody (green), a marker of SMCs (Fig. 1b, right). We then examined the expression of several SMC-and ECspecific marker genes in FGF-deprived HUVE-DC cultures us- ing RT-PCR combined with Southern blot analysis (Fig. 1c). mRNA expression of all the SMC marker genes examined, i.e. basic calponin, SM22␣ (30,31), and prostaglandin E synthase (32), was clearly enhanced; indeed, expression of basic calponin and SM22␣ increased up to 50-fold during the 4-week period of FGF deprivation. Notably, expression of vWF mRNA was also increased in FGF-deprived cells after 1 week but had returned to its original level within 4 week. On the other hand, expression of another endothelial-specific gene, pecam-1 (12), was little changed. The above results suggest that HUVE-DC cultures are composed of a heterogeneous population with respect to their potential to differentiate into SMC-like cells. To minimize this heterogeneity and exclude the possibility that SMCs may have contaminated the primary HUVE-DC cultures, we carried out a limiting dilution of HUVE-DC cultures in FGF-1-containing growth medium. After establishing 30 single-cell-derived populations, they were deprived of FGF. We found that in all of the cloned cell populations, cells deprived of FGF were initially positive only for vWF, but after a period of FGF deprivation, some began to express basic calponin, confirming that SMClike cells can indeed differentiate from ECs (Fig. 2). We then classified the clones into two groups: slowly differentiating and rapidly differentiating. Among the slowly differentiating clones (e.g. CL1), which represented 6 of the 30 cultures, basic calponin-positive cells were not detected after 2 weeks of FGF deprivation (Fig. 2a, left) but did appear within 3 weeks (not shown). Among the rapidly differentiating clones (e.g. CL14), basic calponin-positive cells appeared comparatively quickly (Fig. 2a, right). After 2 weeks, approximately one-fifth of the cells (11-23%, as determined by immunostaining) expressed basic calponin ( Consistent with the immunohistochemistry, semiquantitative analysis using RT-PCR/Southern blot revealed that levels of basic calponin mRNA reached a maximum in CL14 cells after 2 weeks of FGF deprivation and in CL1 cells after 3 weeks of deprivation (Fig. 2c). Expression of another smooth muscle cell-specific marker, SM22␣, followed a similar time course (data not shown). These effects of FGF deprivation proved to be FIG. 1. HUVE-DCs express SMC markers after prolonged FGF-1 deprivation. In a and b, HUVE-DCs were grown in medium 199 containing 15% fetal bovine serum with (ϩ) FGF-1 (5 ng/ml) and heparin (10 g/ml). For induction of differentiation, the FGF-1 and heparin were removed (Ϫ) from the medium, and the cells were cultured for 3 weeks. In a, cells were photographed under a phase-contrast microscope. In b, cells were immunostained for vWF (red) and basic calponin (green). The scale bar ϭ 50 m. In c, HUVE-DCs were cultured for the indicated periods in the absence of FGF-1 and heparin, after which expression of mRNAs encoding various SMC and EC markers was examined by semiquantitative RT-PCR/Southern blot analyses (left panels). The relative expression level of each mRNA was plotted after normalization with GAPDH expression (right panels). For SMC markers, the maximal level detected was defined as 100% expression; for EC markers, the initial level was defined as 100% expression. reversible. In the experiment shown in Fig. 2d, FGF-1 deprivation increased expression of basic calponin and SM22␣ mRNA by day 4. However, addition of FGF-1 at that time reversed the increase, almost eliminating expression of basic calponin and SM22␣ mRNA by day 8. A similar reversal of SMC-like-differentiation by FGF was also observed in parental HUVE-DCs (data not shown). In addition, as in the parental HUVE-DC cultures (Fig. 1c), expression of vWF mRNA increased in both FGF-deprived CL1 and CL14 cells (Fig. 2c).
To further investigate the characteristics of cells expressing SMC markers, we next examined their responsiveness to ET-1, which is produced by endothelium (33) and acts via a smooth muscle-specific receptor (ET-R type A) to induce contraction.
Although expression of ET-R type A mRNA was undetectable in HUVE-DCs cultured with FGF (Fig. 3a, week 0), it was abundantly expressed after 2-4 weeks of FGF deprivation (Fig.  3a). We observed a similar induction of ET-R type A in CL1 cells (data not shown). Interestingly, the level of ET-R type A expression in the rapidly differentiating CL14 cells was high even in the presence of FGF (data not shown) and did not change during the 4-week period of FGF deprivation.
To determine whether they were capable of contraction, SMC-like cells were induced by FGF-free medium for 3 weeks and then plated on collagen type I gels containing 1 mM 8-Br-cGMP, which relaxed the cells. After incubating the gels for 17 h, we examined the effects of ET-1 (250 nM). Whereas the In a, HUVE-DCs were cultured in the absence of FGF-1 and heparin and for the indicated periods, after which their expression of ET-receptor type A and ET-1 mRNA was examined by semiquantitative RT-PCR/ Southern blot analyses (left panels). The relative levels of each mRNA were plotted after normalization with GAPDH expression (right panels). For ET-R type A, the maximum level was defined as 100% expression; for ET-1, the initial level was defined as 100% expression. In b, after depriving cells cultured on collagen-coated plastic plates of FGF-1 and heparin for 3 weeks, they were plated on gels formed of type I collagen and incubated with (middle and right) or without (left) ET-1 (250 nM) for 30 min. The cells in the right panel were exposed to ET-R inhibitors (30 M of BQ123 and 10 M of BQ788) for 10 min prior to application of ET-1. Phase-contrast micrographs are shown. The scale bar ϭ 50 m. In c, HUVE-DCs cultured in growth medium containing FGF-1 (F) or in FGF-free medium for 3 weeks (D) were harvested and embedded into gels formed of type I collagen. Their capacity to contract was analyzed by comparing the relative volumes of gels in which the cells were (ϩ) or were not (Ϫ) treated with ET-1 (250 nM) for 3.5 h. *, p Ͻ 0.05; Student's t test. In d, HUVE-DCs cultured in FGF-free medium for 3 weeks were harvested, embedded into collagen gel, and analyzed for their capacity to contract as in panel c. ET-R inhibitors (30 M of BQ123 and 10 M of BQ788) were added to some cultures for 30 min prior to application of ET-1. *, p Ͻ 0.05; Student's t test. In e, HUVE-DCs and CL14 cells cultured in FGF-free medium for 3 weeks, or HPASMCs cultured in FGF-free medium, were plated on type I collagen gels. The cells were treated (ϩ) or not treated (Ϫ) with ET-1, after which cell contraction was analyzed by measuring the size of the cell areas on the gels. *, p Ͻ 0.05; Student's t test.
control cells appeared flat and relaxed in the collagen gels (Fig.  3b, left), ET-1 caused the cells to contract and become more rounded in shape (Fig. 3b, middle). Measurement of the gel volumes showed that ET-1 treatment reduced the volume of gels made up with FGF-deprived cells to 79.9% of control (Fig.  3c, D) but had no effect on gels made up with FGF-treated cells (Fig. 3c, F). Pretreating the cells with specific ET-R type A and type B antagonists (30 M BQ123 and 10 M BQ788, respectively) inhibited both the ET-1-evoked change in morphology (Fig. 3b, right) and the gel shrinkage (Fig. 3d). Thus, sensitivity to ET-1 and cell contractility correlated with the emergence of SMC-like cells in culture.
To further characterize the HUVE-DC-derived SMC-like cells, their contractility was compared with that of pure HPASMCs. Because of the limited availability of the cells, contractility was analyzed by comparing ET-1-evoked changes in cell area on collagen gels using smaller numbers of cells than were used for the gel volume assay shown in Fig. 3d. HUVE-DCs and CL14 cells cultured in FGF-free medium for 3 weeks or HPASMCs cultured in FGF-free medium were harvested and plated on type I collagen gels containing 0.5 mM 8-Br-cGMP. After incubating the cells for 6 h, the areas occupied by cells treated (ϩ) or not treated (Ϫ) with ET-1 were compared. We found that the area occupied by HPASMCs decreased to 50.4% of control upon exposure to ET-1, whereas the areas occupied by the differentiated HUVE-DCs and CL14 cells declined to 73.1 and 74.9% of control, respectively (Fig. 3e). (10) after stimulation with exogenous polypeptide growth factors belonging to the TGF-␤ family. Because both migrating and resting ECs reportedly express activin A, which inhibits endothelial cell proliferation (34), we examined expression of activin A and other genes related to activin A signaling (Fig. 4a). We determined that activin A expression in CL14 cells increased after FGF-1 deprivation and that reapplying FGF-1 to the cells reversed this up-regulation (Fig. 4a). The same results were obtained with parental cells (data not shown). Expression of mRNAs encoding activin-inducible inhibitors of activin signaling, i.e. follistatin (35)(36)(37) and Smad7 (18, 38 -42), was also reversibly increased after FGF-1 deprivation (Fig. 4a), indicating that endogenous activin A signaling is activated in FGF-deprived cells.

Activin A Signaling Is Activated in FGF-deprived HUVE-DCs and Is Responsible for the SMC-like Differentiation-Earlier studies suggested that SMC-like cells arise from atrioventricular (8) and dorsal aorta ECs
To examine the role of activin A signaling in the induction of SMC-like differentiation, we first attempted to block the signaling by adding excess exogenous follistatin to the cultures. HUVE-DCs were cultured with or without FGF-1 for 5 days in the presence or absence of follistatin (1 g/ml applied every 12 h) (Fig. 4b). Adding follistatin to the culture medium significantly diminished induction of basic calponin and SM22␣ expression (Fig. 4b), although it had no effect on vWF expression (Fig. 4b). This result strongly suggested that endogenous activin A signaling is responsible for inducing the SMC-like phenotype in FGF-deprived HUVE-DCs. Consistent with that in-

FIG. 4. Importance of activin A signaling in the differentiation of HUVE-DCs into SMC-like cells.
In a, CL14 cells were cultured for 4 days in the absence of FGF-1 and heparin and then for another 4 days (until day 8) in the same medium without (Ϫ) or with (ϩ) FGF-1 and heparin. Expression of activin A mRNA and its related genes was examined by semiquantitative RT-PCR/Southern blot analyses (left panels). Relative expression levels were plotted after normalization with GAPDH expression; the initial level was defined as 100% expression (right panel). In b, HUVE-DCs were cultured for 5 days with or without FGF-1. Follistatin (1 g/ml) was added to half of the cultures without FGF-1 once every 12 h throughout the culture period. The expression of mRNAs encoding SMC and EC markers was examined by semiquantitative RT-PCR/Southern blot analyses. In c, HUVE-DCs were cultured for 7 days with or without FGF-1, with (ϩ) or without (Ϫ) activin A (100 ng/ml). Expression of marker genes was examined as in panel b. In d, HUVE-DCs were cultured for 9 days without FGF-1 in the presence (left) or absence (right) of activin A (100 ng/ml). The cells were then double-immunostained for vWF (red) and basic calponin (green). The scale bar ϭ 50 m. In e, CL1 and CL14 cells were cultured for 4 days in the absence of FGF-1 and heparin and then for another 4 days (until day 8) in the same medium with (ϩ) or without (Ϫ) FGF-1 and heparin. Expression of activin A mRNA was examined by semiquantitative RT-PCR/Southern blot analyses (left panels). Relative expression levels were plotted after normalization with GAPDH expression (right panel). In f, CL1 and CL14 cells were cultured for 7 days with or without FGF-1 in the presence or absence of activin A (100 ng/ml), after which marker gene expression was examined as in panel b.
terpretation, addition of exogenous activin A to FGF-deprived cultures induced expression of basic calponin and SM22␣ mRNA (Fig. 4c), as well as calponin protein (Fig. 4d). Interestingly, when we simultaneously added FGF-1, the activity of exogenous activin A was suppressed (Fig. 4c).
The importance of activin A signaling in the induction of SMC-like differentiation of HUVE-DCs was examined further by comparing expression of activin A in CL1 and CL14 cells, which, respectively, show low and high potentials for SMC-like differentiation (Fig. 2). Following FGF-1 deprivation, activin A expression rapidly increased in CL14 cells, reaching 775% of control (Fig. 4e, right panel). By contrast, activin A expression in CL1 cells increased only modestly, reaching 152 and 176% of control after 4 and 8 days of FGF-1 deprivation, respectively (Fig. 4e, right panel). We also examined the effect of exogenously applied activin A on cell differentiation and found that, in the absence of FGF, exogenous activin A induced expression of basic calponin and SM22␣ mRNA in both CL1 and CL14 cells but to a much greater extent in the latter (Fig. 4f). Thus, activin A signaling appears to correlate with SMC-like differentiation of HUVE-DCs at the cell population level, suggesting that such differentiation is determined, at least in part, by the level of activin A expression following FGF deprivation.
We also attempted to block activin A signaling by overexpressing an inhibitory Smad in HUVE-DCs. Because HUVE-DCs are difficult to transfect using typical expression vectors, we infected the cells with AdCMV-Smad7, an adenoviral vector carrying mouse Smad7. As expected, the level of Smad7 expression was dependent on the vector dosage (Fig. 5a). Transfection with Smad7 significantly diminished positive Smad signaling in HUVE-DCs: it inhibited both the activin A-induced phosphorylation of Smad2 (Fig. 5b) and its nuclear translocation (38 -41) (Fig. 5c). In addition, control cells, but not Smad7 transfectants, underwent SMC-like differentiation when FGF-deprived and thus showed increased expression of basic calponin and SM22␣ mRNAs (Fig. 5d) and of basic calponin protein (green) (Fig. 5e). There was no difference in the level of vWF mRNA expression in control and Smad7-overexpressing cells.
Finally, because some TGF-␤ family members, i.e. TGF-␤1 and TGF-␤3, have been shown to induce expression of SMC markers in bovine aortic ECs (8) and mesenchymal cells (10), we examined the effects of neutralizing anti-TGF-␤1, -␤2,and -␤3 antibody on the expression of SMC marker genes. We found that the antibody did not significantly affect expression of these genes, nor did direct addition of TGF-␤1 (16 ng/ml) or BMP-4 (50 ng/ml) to the cultures (data not shown). DISCUSSION We have shown that, although HUVE-DCs cultured with FGF-1 constitute a population of ECs, when deprived of FGF-1, a significant fraction will differentiate into cells that fulfill many of the criteria used to define SMCs and that Smad-dependent activin A signaling is responsible for this differentiation. Because a more detailed comparison of these differentiated HUVE-DCs and SMCs obtained from the walls of the blood vessels is yet to be conducted, we are tentatively designating these cells as SMC-like. Still, this is the first study to clearly demonstrate that ECs obtained from a normal human vascular tube can differentiate into SMC-like cells identified not only by biochemical markers but also by their ability to contract in response to ET-1 and that SMC-like differentiation and dedifferentiation are reciprocally controlled by activin A and FGF-1.
It was notable that expression of activin A, follistatin, and Smad7 was all up-regulated in the HUVE-DCs after FGF-1 deprivation and that the up-regulation was reversed by administration of FGF-1. Since activin A reportedly induces expres-sion of Smad7 and follistatin mRNA (37,42), we suggest that activin A signaling would tend to suppress differentiation through a negative feedback mechanism involving follistatin and Smad7, although obviously this effect is not sufficient to completely inhibit activin A-induced differentiation of HUVE-DCs into SMC-like cells when FGF-1 was not present. The signaling pathway downstream of the FGF receptors involves tyrosine kinase-mediated activation of the FGF receptor and subsequent phosphorylation of various signaling molecules, including p38, phosphatidylinositol 3-kinase, and MEK (43)(44)(45)(46)(47)(48)(49). We are currently investigating which of these mediators are important for inhibition of differentiation of HUVE-DCs into SMC-like cells.
It is unclear whether it is appropriate to describe the induction of SMC-like cells from HUVE-DCs as transdifferentiation. Although HUVE-DCs maintained with FGF are widely studied The cells were then lysed, and expression of the target protein was analyzed by immunoblotting. AdCMV-LacZ was used in control experiments. b, activation of Smad2 by activin A. HUVE-DCs were treated with activin A (100 ng/ml) for the indicated times, after which phosphorylated (upper) and total (lower) Smad2 were detected by immunoblotting. In c, HUVE-DCs were transfected with AdCMV-Smad7 (panels 1 and 2) or AdCMV-LacZ (panels 3 and 4), after which they were left untreated (panels 1 and 3) or exposed to activin A (100 ng/ml) for 1 h (panels 2 and 4). The cells were then fixed, and subcellular localization of Smad2 was detected using anti-Smad2 monoclonal antibody and a fluorescently labeled secondary antibody. The scale bar ϭ 10 m. d, inhibition of SMC marker expression by overexpression of Smad7. HUVE-DCs were infected with AdCMV-Smad7 or AdCMV-LacZ and then cultured for 5 days in FGF-1-free medium containing activin A (100 ng/ml). Expression of mRNAs encoding SMC and EC markers was examined by semiquantitative RT-PCR/Southern blot analyses. In e, inhibition of SMC differentiation by overexpression of Smad7. HUVE-DCs were infected with AdCMV-Smad7 (panels 1 and 3) or AdCMV-LacZ (panels 2 and 4) and cultured in FGF-1-free medium containing activin A (100 ng/ml) for 7 days. Phase-contrast (1 and 2) and fluorescent (3 and 4) micrographs of the cells double-immunostained for anti-vWF (red) and basic calponin (green) are shown. The scale bar ϭ 50 m. as normal human endothelial cells, their differentiated characteristics become better defined when they are deprived of FGF for a short period; for example, FGF-deprived HUVECs readily form tube-like structures on extracellular matrix. 2 Indeed, in the present study, removal of FGF transiently enhanced expression of vWF in cells that were negative for SMC markers. Thus, HUVE-DCs cultured in the presence of FGF may be "immature" ECs with the capacity to differentiate into "mature" ECs or to transdifferentiate into SMC-like cells. In this regard, our finding that a small fraction of the cells simultaneously expressed vWF and basic calponin suggests that expression of endothelial-and smooth muscle-specific antigens are at least not mutually exclusive.
It would also be interesting to know whether ECs derived from human adult blood vessels harbor the same potential to differentiate into SMC-like cells as do HUVE-DCs. We therefore examined the potential of HPAECs to differentiate into SMC-like cells and found that, indeed, HPAECs cultured for 1-3 weeks in the absence of FGF-1 showed increased expression of mRNAs for prostaglandin E synthase, basic calponin, and SM22␣, although the increases were more modest than were seen in HUVE-DCs (data not shown). These results strongly suggest that ECs derived from adult human blood vessels also retain the potential to differentiate into SMC-like cells. To clarify this issue, however, more detailed investigation, including characterization of cell differentiation at the single cell level, will be essential.
Also necessary is a more detailed characterization of the SMC-like cells differentiated from HUVE-DCs. Although we found that differentiated HUVE-DCs were less responsive to ET-1 than HPASMCs, our findings may simply reflect the fact that only a small proportion of these cells were SMC-like; among rapidly differentiating HUVE-DCs, including CL14 cells, approximately one-fifth (11-23%) expressed basic calponin after 2 weeks of FGF deprivation. Alternatively, it is also possible that, at the single cell level, SMC-like cells derived from HUVE-DCs are not truly SMCs. A line of study addressing this question is currently under way.