Time-Course Analysis on the Differentiation of Bone Marrow-Derived Progenitor Cells Into Smooth Muscle Cells During Neointima Formation
Objective—Bone marrow-derived progenitor cells have been implicated to contribute to neointima formation, but the time course and extent of their accumulation and differentiation into vascular cells and, most importantly, the long-term contribution of bone marrow-derived progenitor cells to the vascular lesion remain undefined.
Methods and Results—Wire-induced injury of the femoral artery was performed on chimeric C57BL/6 mice transplanted with bone marrow from transgenic mice expressing enhanced green fluorescence protein, and vessels were harvested at 3 days, 1, 2, 3, 4, 6, and 16 weeks after dilatation (n=8 animals per time point). Using high-resolution microscopy, we unexpectedly found that the expression of smooth muscle cell or endothelial cell markers in enhanced green fluorescence protein positive cells was a very rare event. Indeed, most of the enhanced green fluorescence protein positive cells that accumulated during the acute inflammatory response were identified as monocytes/macrophages, and their number declined at later time points. In contrast, a substantial fraction of highly proliferative stem cell antigen-1 and CD34+ but enhanced green fluorescence protein negative and thus locally derived cells were detected in the adventitia.
Conclusion—These data provide evidence that the differentiation of bone marrow-derived progenitor cells into smooth muscle cell or endothelial cell lineages seems to be an exceedingly rare event. Moreover, the contribution of bone marrow-derived cells to the cellular compartment of the neointima is limited to a transient period of the inflammatory response.
Smooth muscle cells (SMCs) play a decisive role in the pathogenesis of vascular diseases and its clinical manifestations. In addition to atherosclerosis, neointima (NI) formation is a major burden in vascular medicine and concerns patients after percutaneous coronary intervention, bypass operation, or graft vasculopathy. It had been widely accepted that intimal SMCs in proliferative vascular diseases are derived from resident medial SMCs or adventitial fibroblasts.1,2 Several years ago, this theory was challenged by the suggested ability of bone marrow-derived progenitor cells (BMPCs) to differentiate into vascular cells during arterial remodeling.3 However, until today, the differentiation capacity of BMPCs in vivo remains highly controversial.4 In different mouse models of atherosclerotic plaque formation, it has been shown very recently that SMCs and endothelial cells (ECs) in atherosclerotic plaques are exclusively derived from the local vessel wall and not from the circulating blood.5,6
See accompanying article on page 1877
Although the homing of BM-derived cells on mechanically injured vessels has been clearly demonstrated in various animal models, the fraction of these cells expressing α-smooth muscle actin (α-SMA) showed wide diversity, ranging from negligible to substantial numbers of all vascular SMCs.3,7 Because these previous studies of mechanical injury were only followed up to 30 days after dilatation of the artery, the observations particularly focused on the time period of the inflammatory response to the injury, whereas later vascular remodeling processes up to a complete restructuring of the vessel have been left unattended. This is of critical importance, because it is well established that macrophages can also stain positive for α-SMA within an inflammatory environment.8 Moreover, in these studies, BMPCs have not been demonstrated to express highly specific markers for vascular SMCs.9 Thus, there are no data available that clearly show a differentiation of BMPCs into definite SMCs, especially beyond the inflammatory reaction of the vessel wall. Hence, the suggested long-term contribution of BM-derived cells to the neointimal cellular mass in the form of highly differentiated SMCs remains elusive.
In this study, we demonstrate that differentiation of BMPCs into SMCs or ECs is a very rare event. Instead, most of the BM-derived cells accumulating in the neointimal lesion were identified as monocytes/macrophages that disappeared at later time points. In contrast, we found that a highly proliferative stem cell antigen (Sca)-1+ and CD34+ but enhanced green fluorescence protein (eGFP) negative fraction of cells is present in the adventitial layer, suggesting that perivascular progenitor cells may very likely represent an additional source of neointimal SMCs. Thus, even though there is no doubt that circulating cells are of major impact for the healing/remodeling process of injured vessels, neointimal SMCs seem to originate from local cells of the vessel wall, not from circulating (progenitor) cells.
For an expanded Materials and Methods section, please see the supplemental data, available online at http://atvb.ahajournals.org. Key techniques involved the lethal irradiation of C57BL/6 mice followed by BM transplantation of eGFP-transgenic BM and a wire-induced injury model of the femoral artery as previously described.3 Before harvesting the dilated arteries at 3 days, 1, 2, 3, 4, 6, and 16 weeks after injury (n=8 mice per time point), the mice were perfused with 4% paraformaldehyde. Microscopy was performed with a DMRB fluorescent microscope (Leica, Wetzlar, Germany) equipped with a PIFOC piezo-element driven Z-drive (Physik Instrumente, Jena, Germany). Deconvolution of 3D widefield epifluorescence z-stacks was done under the usage of adaptive blind deconvolution algorithm with AutoQuantX (Media Cybernetics, Inc., Bethesda, Md). Confocal imaging was performed using an Eclipse TE2000-E confocal LASER scanning microscope (Nikon, Tokyo, Japan). Postprocessing and image analyses were done with MetaMorph™ (Molecular Devices, Downington, PA) and ImageJ (National Institutes of Health, Bethesda, Md).
Fluorescence-Activated Cell Sorter Analysis of BM Chimeras
The percentage of hematopoietic chimerism after irradiation and BM transplantation was assessed by flow cytometry of peripheral blood mononuclear cells at 12 weeks after transplantation. The analysis revealed that 82 to 94% of circulating mononuclear cells were eGFP+ in the recipient mice after irradiation with 9.5 Gray (88.43±4.21%, n=6) (supplemental Figure I). As compared with nonpretreated littermates, the peripheral blood cell count of the irradiated and BM-transplanted mice was within the physiological range. As a further control, immunohistochemical staining revealed a very high percentage of eGFP+ cells in the spleen of the killed mice. These cells stained positive for the pan-leukocyte marker CD45 (data not shown).
Time-Course Analysis on NI Formation
Following wire-induced dilatation of the femoral artery, the size of the NI increased over a time period of 4 weeks (NI/media ratio of 2.13+0.26), whereas the ratio of eGFP+ cells/all neontimal cells constantly decreased toward later time points (Figure 1). As previously described, the wire-induced dilatation of the femoral artery caused a complete disruption of the endothelial layer and a substantial loss of medial SMCs.10 Platelets were the first cells adhering to the injured luminal surface, followed by the recruitment of leukocytes that peaked at 2 weeks after injury (supplemental Figure II). At this time point, the eGFP+ cells accounted for 68.95±2.74% of all neointimal cells. Subsequently, medial cells showed high proliferative indices, as determined by proliferating cell nuclear antigen staining (Figure 2). A broad accumulation of α-SMA-expressing cells in the neointimal lesion was observed at 3 to 4 weeks after dilatation. Interestingly, the neointimal lesion development in nonirradiated littermates occurred earlier with a peak in leukocyte recruitment at 1 week and a SMC accumulation already at 2 weeks after dilatation (supplemental Figure III). The neointimal size was dose-dependently attenuated in irradiated and BM-transplanted mice (supplemental Figure III), but this effect was independent on the use of wild-type BM or eGFP-transgenic BM (data not shown). Importantly, the fractions of eGFP+-mononuclear cells in mice irradiated with 10.5 Gray (89.74±5.09%) revealed no significant differences compared with mice irradiated with 9.5 Gray (88.43±4.21%), and we detected abundant staining of BM-derived cells in the adventitia/perivascular tissue.
Neointimal SMCs exhibited a dedifferentiated phenotype with relatively low levels of α-SMA at 4 weeks after injury. At later time points, a more differentiated population of neointimal SMCs was restored, as shown by the strong expression of α-SMA, smooth muscle-myosin heavy chain, and calponin. In contrast, the number of eGFP+ cells in the neointimal lesion further declined and accounted for only 2.16±0.78% of all neointimal cells at 16 weeks after injury (Figure 2).
Differentiation of BMPCs Into Vascular SMCs
Following wire-induced injury, the complete lesion range of each artery was carefully analyzed for the coexpression of α-SMA, SM-MHC, or calponin in eGFP+ cells, in order to identify SMCs derived from BMPCs (n=6 cross-sections throughout the dilated area of each artery; n=8 mice per time point). Deconvolution analysis of high-resolution z-axis image stacks revealed that 0.9±2.1% of all eGFP+ cells coexpressed α-SMA at 4 weeks after injury (Figure 3A and 3B and supplemental Figure IV). However, 3D reconstructions of z-stacks obtained by confocal microscopy revealed that the expression of α-SMA was only very faint in BM-derived cells (see supplemental video). Therefore, the expression of α-SMA in BM-derived cells was found to be a very rare event during NI formation despite the fact that high numbers of eGFP+ cells were present in the neointimal lesion at 4 weeks after injury. Importantly, we could not detect BM-derived cells expressing markers for highly differentiated SMCs, like SM-MHC or calponin, at any time point after dilatation. At later time points, the remaining eGFP+ cells in the vascular wall were predominantly located in the medial or adventitial layer and did not coexpress α-SMA, SM-MHC, or calponin at all (Figure 2F and 2G, supplemental Figure VI).
Differentiation of BMPCs Into ECs
After disruption of the endothelial layer and complete denudation of the injured vessel section, we further analyzed a possible differentiation of BMPCs into ECs. The reendothelialization mainly started at the edges of the dilated/denuded region and was largely completed at 3 weeks after dilatation (data not shown). The costaining for eGFP and CD31 in all analyzed cross-sections from all dilated arteries at any time point detected only 5 BM-derived (eGFP+) cells coexpressing CD31 (Figure 3C and 3D). However, because CD31 has also been shown to be expressed in cells of myeloid origin, these double positive cells may not represent genuine ECs but endothelial-like myeloid cells.11 To further confirm the presence of differentiated ECs of BM origin, immunohistochemical analysis for the expression of von Willebrand factor was performed. However, we were not able to detect any eGFP+ cells coexpressing von Willebrand factor (data not shown).
Further Characterization of eGFP+-Cells During NI Formation
The absolute number of eGFP+ cells in the NI, as well as the eGFP+ cells/all neontimal cells ratio, continuously declined at later time points after injury due to both a loss of eGFP+ cells and increasing numbers of locally derived, proliferating SMCs in the NI (Figure 4). To further characterize the accumulating eGFP+ cells, we performed immunohistochemical staining with different leukocyte markers. During the first weeks after injury, nearly all accumulating cells were positive for the pan-leukocyte marker CD45 (data not shown). The fraction of monocytes/macrophages (MoMa-2-expressing cells) of all eGFP+ cells in the NI increased from 0.31±0.04 at 3 days to 0.88±0.11 at 16 weeks after dilatation, so that the neointimal eGFP+ cells at late time points were predominantly identified as macrophages (Figure 5B). Importantly, the absolute number of macrophages in the neointimal lesion declined from 97.88±9.6 macrophages at 2 weeks to 5.13±1.89 macrophages at 16 weeks after injury (Figure 5C). This decline correlated with the decrease of all neointimal eGFP+ cells, respectively (Figure 4). Indeed, the inflammatory cells were not primarily targeted to the NI per se, because the majority of eGFP+ cells was detected in the perivascular tissue. Interestingly, large portions of the medial layer lacking SMCs were also impregnated with eGFP+ cells that did not coexpress α-SMA (Figure 2 and supplemental Figure II). Moreover, there often seemed to be completely distinct regions of either BM-derived macrophages or locally derived vascular SMCs as shown by staining for CD68 (monocytes/macrophages) and α-SMA in 2 adjacent slides of the same vessel (supplemental Figure VI). Thus, BM-derived cells can be considered to contribute to vascular remodeling by paracrine actions rather than being a relevant source of subsequently differentiated vascular cells.
Possible Role for Perivascular Stem Cells During NI Formation
Because neointimal cells seem to be derived primarily from local cells of the vessel wall, we further analyzed the role of non-BM-derived perivascular progenitor cells during NI formation. In nondilated arteries, 42.84% of adventitial cells were identified as Sca-1+ cells, whereas hardly any eGPF+ cells could be detected within the vessel wall. The fraction of resident Sca-1+ cells did virtually not change at 1 and 2 weeks after injury, but we observed eGPF+ cells in the adventitia that also expressed Sca-1 at these time points. At 3 and 4 weeks after dilation, we detected the highest rates of proliferating cells within the adventitia and also found increased absolute numbers of adventitial Sca-1+ cells. At 4 weeks after injury, the absolute number of Sca-1+ cells increased from 47.86±7.18 cells/slide in noninjured arteries to 158.5±16.656 cells/slide, and the fraction of proliferating adventitial cells rose to 14.94% of all adventitial cells (Figure 6 and supplemental Figure VII). Importantly, we observed singular Sca-1+ cells and CD34+ cells migrating into the NI at 3 and 4 weeks after dilation (Figure 6E). Otherwise, the expression of Sca-1 was restricted to cells within the adventitia.
Impact of Fixation and Image Acquisition Techniques for the Detection of Differentiated BMPCs
Because the endogenous eGFP signal was already very strong in the transplanted cells, an anti-eGFP antibody only showed a moderate enhancement of the signal (data not shown). However, such antibodies have been used to generate previously published data.3 Immediate perfusion and subsequent fixation with 4% paraformaldehyde were indispensable for maintaining the cell-specific eGFP signal. In contrast to direct fixation, we observed a diffusion of the tracer molecule throughout artery cross-sections. This unspecific staining was further enhanced by the use of an anti-eGFP antibody (supplemental Figure VIIIA). Moreover, the use of inadequate filter blocks with overlapping emission wavelengths revealed false-positive results. In areas with strong red fluorescence (staining for α-SMA), signals of this staining were also detected in the green channel, where eGFP signals were assessed (supplemental Figure VIIIB). These results indicate that proper fixation, staining, and acquisition techniques are mandatory for the accurate assessment of the contribution of BM-derived cells in this model and might explain the confusing results in some previous reports.
Inhibiting SMC proliferation is a highly effective way to prevent luminal stenosis due to neointimal lesion formation. Drug-eluting stents antagonize the stenotic process of NI formation and have been shown to reduce the rate of restenosis after percutaneous coronary intervention, but delayed reendothelialization has increased the risk of in-stent-thrombosis and defines the requirement for prolonged antiplatelet therapy.12 Thus, elucidating the exact pathophysiological mechanisms and especially the origin of vascular cells that contribute to neointimal lesion formation is essential to optimize future invasive therapeutic strategies. The results of the present study provide evidence that the definite differentiation of BMPCs cells into SMCs or ECs is only an exceedingly rare event, ie, the exception rather than the rule. Moreover, most of the BM-derived cells found in the NI were monocytes/macrophages, and there was no apparent substantial long-term contribution of these cells to the cellular mass of the NI.
Because we only found solitary α-SMA-expressing BM-derived cells in the neointimal lesion, our results are in direct contrast to previous reports claiming a BM origin for nearly half of the vascular SMCs as well as a substantial fraction of ECs.3 However, the suggested differentiation of BMPCs into vascular cells during atherosclerosis or NI formation has always been a matter of debate.4 One reason for some of the controversy is likely to be related to methodological problems and “false-positive” staining. Our results highlight the critical fact that only high-resolution microscopy as used in this study can distinguish between a genuine double positive staining and “pseudo” marker colocalization due to the superimposition of vascular cells with adjacent BM-derived cells. Moreover, the failure to immediately and adequately fix tissues results in the diffusion of the eGFP tracer molecule to adjacent cells. This phenomenon may then be easily misinterpreted as a faint eGFP signal in locally derived SMCs that in fact do not express eGFP.
Another explanation might be that infiltrating macrophages can indeed temporarily express α-SMA in an inflammatory environment.8 These cells resemble an intermediate phenotype of the macrophage at the interim state between late inflammation and scar formation and thus lack markers for highly differentiated SMCs.8 These intermediate phenotype macrophages may also possibly have confounded the data of previous studies. This hypothesis is most convincingly supported by the inability of these studies as well as our study to detect more specific markers for differentiated SMCs, such as SM-MHC or calponin in BM-derived cells in neointimal or atherosclerotic lesions.9 Especially, cross-immunoreactivity of antibodies detecting SM-MHC with nonmuscle MHC is an issue of concern. Thus, we very carefully assessed the specificity of the used SM-MHC antibody (supplemental Figure V). With the use of this antibody, however, we were not able to detect SM-MHC expression in BM-derived cells at any time point. Indeed, in a recent review, Tanaka and Sata13 speculated that the BM-derived cells expressing α-SMA that were described and quantified in their initial observations might represent “SMC-like macrophages” rather than differentiated SMCs. However, there was still a lack of conclusive data supporting this hypothesis. Although a study of an atherosclerotic mouse model has already questioned the BM-derived origin of SMCs in atherosclerosis,5 it has been claimed that the accumulation of differentiated BMPCs correlates with the severity of the injury model.9 Consequently, even though a differentiation of BM cells into SMC is missing during the chronic process of atherosclerosis development, a substantial contribution of BMPCs to neointimal lesion formation may exist following the acute denudation and dilatation by wire-induced injury, which resembles the acute severe injury during percutaneous coronary intervention. However, our data now provide the first real evidence of a missing long-term contribution and terminal differentiation of BM-derived cells into SMCs in neointimal lesions beyond the state of the inflammatory response. Moreover, also an exceedingly rare differentiation of BM-derived cells into ECs was observed. These cells may not represent genuine ECs but rather CD31-expressing “endothelial-like” myeloid cells, because cells of myeloid origin have also been shown to express CD31.11 We were not able to detect further EC-specific markers like von Willebrand factor in BM-derived cells.
In fact, a growing body of evidence confirms the lack of developmental plasticity of adult BMPCs.14 Circulating endothelial progenitor cells were shown to mainly represent monocytic cell lineages that do not have the ability to substantially “transdifferentiate” into genuine vascular cells in vivo.15 This was confirmed by the inability of BMPCs to differentiate into ECs in a time-course experiment of tumor-induced angiogenesis.16 In addition, there was no evidence for BM-derived renewal of the endothelium in a mouse model of atherosclerosis.6 Because we now demonstrate that this is also true for vascular cells after wire-induced injury of the mouse femoral artery, the properties of BMPCs to accelerate vascular SMC or EC proliferation seem to be rather attributed to their paracrine effects.
This hypothesis is also supported by the results of recent clinical trials that assessed the effects of injecting BMPCs into coronary arteries of patients with acute myocardial infarction following percutaneous coronary intervention.17 Importantly, this treatment did not aggravate the development of restenosis, thus indirectly implying that there is not a substantial differentiation of BM-derived cells into functional vascular SMCs in humans.18 Indeed, a clinical trial on NI formation in human cardiac allograft vasculopathy recently confirmed the exclusive local origin of neointimal SMCs. In accordance with the data from our animal model, the authors did not find any long-term contribution of BMPCs to neointimal lesions.19
The implications of our results lead us back to the original hypothesis of a local origin of vascular SMCs in NI formation. Many years ago, the groups of Benditt and Schwartz provided evidence for a monoclonal origin of SMCs in atherosclerotic lesions.20 This key finding has recently been complemented by the identification of ABCG-2-expressing stem cells in the medial layer of the mouse aorta and femoral artery, suggesting a particular involvement of this developmental subpopulation of medial SMCs.21 We now demonstrate that a Sca-1+ and CD34+ but eGFP negative fraction of cells within the adventitial layer exerts high proliferative indices during NI formation and very likely represents an additional source of neointimal SMCs. In vitro experiments from Passman et al22 indicate that 30 to 50% of adventitial Sca-1+ cells can differentiate into smooth muscle-like cells and thus lose Sca-1 from their cell surface, whereas other adventitial Sca-1+ cells self-renew at the same time. Although we confirm these data in our in vivo model, we do not, in fact, know whether these cells represent genuine progenitor cells and thus have the ability to account for highly differentiated SMCs. However, the high proliferative indices of adventitial cells observed in our time-course experiment of NI formation, as well as recent data from a transplant atherosclerosis model, certainly support this possibility.23 Further research is needed to clarify the contribution of perivascular progenitor cells compared with proliferating medial SMCs or medial progenitor cells in the development of a neointimal lesion.
In conclusion, we clearly show that BM-derived cells do not account for highly differentiated vascular SMCs detected during NI formation. Instead, the very rare numbers of BM-derived cells that faintly and temporarily express α-SMA in the neointimal lesions very likely represent SMC-like macrophages in an interim state. Consequently, these cells do not contribute to a completely restructured vessel in the long term but are only transiently present in the NI at time points that are correlated with a pronounced inflammatory response. Nevertheless, targeting these cells might interfere with the inflammatory response and thus might represent an interesting therapeutic strategy to prevent the activation of local cells and thus NI formation. These findings newly define the role of BM-derived cells for the development of neointimal lesions after vascular injury and substantially add to the current understanding of the pathogenisis of NI formation and restenosis as its clinical correlate.
We thank Dr Karin Hersemeyer for supporting the confocal microscopy analysis and Stefanie Wolfram for her excellent technical assistance.
Sources of Funding
This study was supported by the Deutsche Forschungsgemeinschaft Grant SFB 547 A10 and Exzellenzcluster 147 “Cardio-Pulmonary Systems.”
Received on: February 28, 2010; final version accepted on: June 2, 2010.
Sartore S, Chiavegato A, Faggin E, Franch R, Puato M, Ausoni S, Pauletto P. Contribution of adventitial fibroblasts to neointima formation and vascular remodeling: from innocent bystander to active participant. Circ Res. 2001; 89: 1111–1121.
Hoofnagle MH, Thomas JA, Wamhoff BR, Owens GK. Origin of neointimal smooth muscle: we’ve come full circle. Arterioscler Thromb Vasc Biol. 2006; 26: 2579–2581.
Bentzon JF, Weile C, Sondergaard CS, Hindkjaer J, Kassem M, Falk E. Smooth muscle cells in atherosclerosis originate from the local vessel wall and not circulating progenitor cells in apoE knockout mice. Arterioscler Thromb Vasc Biol. 2006; 26: 2696–2702.
Hagensen MK, Shim J, Thim T, Falk E, Bentzon JF. Circulating endothelial progenitor cells do not contribute to plaque endothelium in murine atherosclerosis. Circulation. 121: 898–905.
Rodriguez-Menocal L, St-Pierre M, Wei Y, Khan S, Mateu D, Calfa M, Rahnemai-Azar AA, Striker G, Pham SM, Vazquez-Padron RI. The origin of post-injury neointimal cells in the rat balloon injury model. Cardiovasc Res. 2009; 81: 46–53.
Tanaka K, Sata M, Hirata Y, Nagai R. Diverse contribution of bone marrow cells to neointimal hyperplasia after mechanical vascular injuries. Circ Res. 2003; 93: 783–790.
Costa MA, Simon DI. Molecular basis of restenosis and drug-eluting stents. Circulation. 2005; 111: 2257–2273.
Woodfin A, Voisin MB, Nourshargh S. PECAM-1: a multi-functional molecule in inflammation and vascular biology. Arterioscler Thromb Vasc Biol. 2007; 27: 2514–2523.
Jensen LO, Maeng M, Kaltoft A, Thayssen P, Hansen HH, Bottcher M, Lassen JF, Krussel LR, Rasmussen K, Hansen KN, Pedersen L, Johnsen SP, Soerensen HT, Thuesen L. Stent thrombosis, myocardial infarction, and death after drug-eluting and bare-metal stent coronary interventions. J Am Coll Cardiol. 2007; 50: 463–470.
Wagers AJ, Sherwood RI, Christensen JL, Weissman IL. Little evidence for developmental plasticity of adult hematopoietic stem cells. Science. 2002; 297: 2256–2259.
Yoder MC, Mead LE, Prater D, Krier TR, Mroueh KN, Li F, Krasich R, Temm CJ, Prchal JT, Ingram DA. Redefining endothelial progenitor cells via clonal analysis and hematopoietic stem/progenitor cell principals. Blood. 2007; 109: 1801–1809.
Purhonen S, Palm J, Rossi D, Kaskenpaa N, Rajantie I, Yla-Herttuala S, Alitalo K, Weissman IL, Salven P. Bone marrow-derived circulating endothelial precursors do not contribute to vascular endothelium and are not needed for tumor growth. Proc Natl Acad Sci U S A. 2008; 105: 6620–6625.
Assmus B, Walter DH, Lehmann R, Honold J, Martin H, Dimmeler S, Zeiher AM, Schachinger V. Intracoronary infusion of progenitor cells is not associated with aggravated restenosis development or atherosclerotic disease progression in patients with acute myocardial infarction. Eur Heart J. 2006; 27: 2989–2995.
Sainz J, Al Haj Zen A, Caligiuri G, Demerens C, Urbain D, Lemitre M, Lafont A. Isolation of “side population” progenitor cells from healthy arteries of adult mice. Arterioscler Thromb Vasc Biol. 2006; 26: 281–286.
Passman JN, Dong XR, Wu SP, Maguire CT, Hogan KA, Bautch VL, Majesky MW. A sonic hedgehog signaling domain in the arterial adventitia supports resident Sca1+ smooth muscle progenitor cells. Proc Natl Acad Sci U S A. 2008; 105: 9349–9354.