A sonic hedgehog signaling domain in thearterial adventitia supports resident Sca1�

smooth muscle progenitor cellsJenna N. Passman*†, Xiu Rong Dong*‡, San-Pin Wu*§, Colin T. Maguire*‡, Kelly A. Hogan¶, Victoria L. Bautch*†¶,and Mark W. Majesky*†‡�**

*Carolina Cardiovascular Biology Center, Departments of ‡Medicine, �Genetics, and ¶Biology, and †Curriculum in Genetics and Molecular Biology,University of North Carolina, Chapel Hill, NC 27599

Edited by Eric N. Olson, University of Texas Southwestern Medical Center, Dallas, TX, and approved April 22, 2008 (received for review December 6, 2007)

We characterize a sonic hedgehog (Shh) signaling domain re-stricted to the adventitial layer of artery wall that supports resi-dent Sca1-positive vascular progenitor cells (AdvSca1). Usingpatched-1 (Ptc1lacZ) and patched-2 (Ptc2lacZ) reporter mice, adven-titial Shh signaling activity was first detected at embryonic day (E)15.5, reached the highest levels between postnatal day 1 (P1) andP10, was diminished in adult vessels, and colocalized with acircumferential ring of Shh protein deposited between the mediaand adventitia. In Shh�/� mice, AdvSca1 cells normally found at theaortic root were either absent or greatly diminished in number.Using a Wnt1-cre lineage marker that identifies cells of neural crestorigin, we found that neither the adventitia nor AdvSca1 cells werelabeled in arteries composed of neural crest-derived smooth mus-cle cells (SMCs). Although AdvSca1 cells do not express SMC markerproteins in vivo, they do express transcription factors thought to berequired for SMC differentiation, including serum response factor(SRF) and myocardin family members, and readily differentiate toSMC-like cells in vitro. However, AdvSca1 cells also express potentrepressors of SRF-dependent transcription, including Klf4, Msx1,and FoxO4, which may be critical for maintenance of the SMCprogenitor phenotype of AdvSca1 cells in vivo. We conclude that arestricted domain of Shh signaling is localized to the arterialadventitia and may play important roles in maintenance of resi-dent vascular SMC progenitor cells in the artery wall.

artery � differentiation � Klf4 � self-renewal � serum response factor

An adventitia surrounds most blood vessels where it functionsas a dynamic compartment for cell trafficking into and out

of the artery wall. The adventitia is a site for formation andregression of microvessels that penetrate and nourish the medialand intimal layers of vessel wall (1–3). The adventitia alsocontains perivascular nerves and lymphatic vessels and is animportant domain for NAD(P)H oxidase activity in the vesselwall (4). Adventitial cells participate in vascular growth andrepair and are important determinants of lumen size via controlof inward or outward wall remodeling processes (2, 5). Recentstudies report unexpected roles for the adventitia insofar as itsupports resident progenitor cells that can adopt both vascularand nonvascular fates (6–8). Despite these important functionalproperties, the signals and environmental cues that confer suchunique and dynamic properties to the adventitial layer remainlargely unknown.

Sonic hedgehog (Shh) is an essential morphogen and growthfactor for many developing tissues (9). In the vascular system,Shh signaling is important for specification of arterial-venousidentity of endothelial cells (10), remodeling of yolk sac bloodvessels (11), and recruitment of mural cells (12). In adult bloodvessels, Shh is angiogenic in ischemic hindlimb and cornealmicropocket assays (13, 14), promotes perineural neovascular-ization in diabetic animals (15), and protects the myocardiumagainst chronic ischemia (16). Shh stimulates production ofangiogenic factors, including VEGF-A and angiopoietin-1 by

interstitial fibroblasts (13, 14), and promotes endothelial cellchemotaxis and tube formation (17, 18). Therefore, Shh playsimportant roles in cell–cell communication during vasculardevelopment and adult wound repair.

Hedgehog (Hh) proteins signal by binding to patched-1 (Ptc1)or patched-2 (Ptc2) receptors and cdo/boc accessory proteins atthe cell surface, resulting in derepression of smoothened (smo)(19). Ptc proteins are 12-transmembrane domain receptors thatare structurally related to the resistance-nodulation-division(RND) family of bacterial membrane permeases (20) and theNiemann-Pick C1-like1 cholesterol transporter (21). Ptc proteinscontain a conserved sterol-sensing domain, and recent studiessuggest that Ptc1 represses smo by transporting activating sterolsout of the cell (22). Hh binding to Ptc1 inhibits transporteractivity thereby allowing activating sterols to accumulate (19,22). Activated smo trafficks into the primary cilium where Hhsignal mediators are concentrated (23, 24) and triggers phos-phorylation of gli factors that then move to the nucleus andstimulate gene transcription (25). Among Hh target genes arePtc1 and Ptc2; therefore, Ptc1lacZ and Ptc2lacZ activities are usedas sensitive reporters of Hh signaling in transgenic mice (26, 27).

The origins and functions of the adventitia are poorly under-stood. Hu et al. (6) reported that stem cell antigen-1 (Sca1)-positive cells with a potential to differentiate into smooth musclecells (SMCs) were found in the aortic adventitia of adultApoE�/� mice (AdvSca1 cells). When transferred to the adven-titial side of experimental vein grafts, AdvSca1-derived SMCsmade up 30% of cells in the graft neointima after 4 weeks (6).The adventitia responds to balloon catheter injury with rapidincreases in SM�Actin and SM22� expression, particularly at themedia–adventitia interface (5). Endothelial progenitor cells arereported to cluster between the media and adventitia in humanarteries (7, 8). These cells formed capillary sprouts in ex vivo ringculture assays and were recruited by tumor cells for capillaryvessel formation in vivo. Given the role of Shh signaling in themaintenance of resident stem and progenitor cells in skin (28),nervous system (29), lymphoid tissue (30), and hematopoieticcells (31), we sought to test for similar roles for Hh signaling inthe adventitial tissue surrounding blood vessels.

Author contributions: J.N.P. and M.W.M. designed research; J.N.P., X.R.D., S.-P.W., C.T.M.,and K.A.H. performed research; V.L.B. contributed new reagents/analytic tools; J.N.P.,X.R.D., and M.W.M. analyzed data; and J.N.P. and M.W.M. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

§Present address: Department of Molecular and Cellular Biology, Baylor College of Medi-cine, BCM-130, One Baylor Plaza, Houston, TX 77030.

**To whom correspondence should be addressed. E-mail: mmajesky@med.unc.edu.

This article contains supporting information online at www.pnas.org/cgi/content/full/0711382105/DCSupplemental.

© 2008 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0711382105 PNAS � July 8, 2008 � vol. 105 � no. 27 � 9349–9354

MED

ICA

LSC

IEN

CES

ResultsHh Signaling in Postnatal Blood Vessels. We examined �-gal activityin blood vessels of PtclacZ transgenic mice at postnatal day 2 (P2).In whole-mount hearts, Ptc1lacZ and Ptc2lacZ activities weredetected in coronary arteries, but not in myocardium or epicar-dium (Fig. 1A). The ascending aorta and pulmonary trunk werestrongly positive at this time (Fig. 1 A). Both Ptc1lacZ and Ptc2lacZ

activities declined with age and were detected at low levels incoronary arteries and transverse aorta from adult mice (Fig. 1B).In aortic cross-sections at P2, strong Ptc1lacZ activity was re-stricted to the adventitia, with only an occasional positive cellfound in the media or intima (Fig. 1C). Adventitial Ptc1lacZ

activity was also found in intercostal (Fig. 1D) and femoral (Fig.1F) arteries. In the mesenteric arcade, Ptc1lacZ staining ofarteries was more intense than that of veins (Fig. 1E), whereaslymphatic vessels were negative [supporting information (SI)Fig. S1]. Ptc2lacZ activity exhibited very similar patterns (Fig. 1D–F Right and Fig. S2). Thus, active Hh signaling was found inlarge and medium-sized arteries and veins in the perinatal periodand was localized to the adventitia of these vessels.

Shh in the Adventitia. Shh protein was localized to the interfacebetween the media and adventitia in large and medium-sizedarteries 2 days after birth (Fig. 2A). This pattern of Shhdistribution colocalized with both Ptc1lacZ (Figs. 1C and 2B Inset)and Ptc2lacZ activities. We could not detect Indian hedgehog(Ihh) in postnatal arteries and found only low levels of deserthedgehog (Dhh) in the endothelium (data not shown). Theseresults indicate that Ptc1lacZ and Ptc2lacZ activities in the arterywall correlate with localized deposition of Shh protein in acircumferential ring in the extracellular matrix (ECM)-rich spacebetween the media and adventitia (Fig. 2 A).

Sca1-Positive Cells in the Adventitia. A previous report (6) de-scribed Sca1� cells in the adventitia surrounding the aortic rootin adult ApoE�/� mice. One possible role of Shh may thereforebe to maintain Sca1� progenitor cells within this domain ofartery wall. As shown in Fig. 2C, Sca1� cells are restricted to theadventitia (AdvSca1 cells) and exhibit little or no overlap withSM�Actin-positive cells in the media. A similar distribution was

found in mesenteric and femoral arteries (Fig. 2 E and F). Closeproximity was found between Shh protein and AdvSca1 cells(Fig. 2D). Extensive overlap was observed between �-gal-positive cells and AdvSca1 cells in the aortic adventitia of Ptc1lacZ

mice (Fig. 2D Inset). These findings suggest that a zone of Shhsignaling in the adventitia identifies a unique domain of arterywall within which resident Sca1� vascular progenitor cells arefound in vivo.

Analysis of Isolated AdvSca1 Cells. AdvSca1 cells were isolated byimmunoselection and examined for progenitor cell markers.AdvSca1 cells exhibited a CD34�/c-kit�/CD140b� marker pro-file (Fig. 3A) and were negative for CD45 and CD68 (data notshown), similar to the profile reported by Hu et al. (6). Freshlyisolated AdvSca1 cells consistently expressed markers of Shhsignaling including ptc1, ptc2, smo, gli1, gli2, and gli3 (Fig. 3A).The distribution of Shh in the extracellular space is governed bymultiple binding and transport proteins including hedgehoginteracting protein-1 (Hhip1), cdo, and boc (reviewed in ref. 19),all of which are expressed by AdvSca1 cells (Fig. 3A). In addition,AdvSca1 cells contain mRNA for Shh, with few, if any, tran-scripts detectable for Dhh or Ihh (Fig. 3A). These results suggestthat AdvSca1 cells both produce and respond to Shh, and that therestricted distribution of Shh in the artery wall is, in part, locallycontrolled by factors produced by AdvSca1 cells themselves.

Hu et al. (6) reported that AdvSca1 cells differentiate intoSMCs when incubated with PDGF-BB (6). We found that whencultured in DMEM containing 10% serum many AdvSca1 cells(�30–50%) lost expression of Sca1, gained expression ofSM�Actin, SM22�, calponin, and SM-MHC, and adopted anelongated, mesenchymal cell shape (Fig. 3 D and E). To rule outthe possibility that up-regulation of SMC markers resulted fromexpansion of a pool of medial SMCs carried over in the AdvSca1cell isolation, we repeated our experiments in the presence ofinhibitors of cell proliferation (aphidicolin or hydroxyurea).After verifying growth arrest, we observed up-regulation of SMCmarker proteins over a similar time course and to a similar extentas seen in the absence of cell cycle inhibitors (Fig. S3). Weconclude that AdvSca1 cells can directly differentiate into

Fig. 1. Hh signaling in artery wall. (A, B, and D–F ) Tissues were isolated from PtclacZ mice at P2 and stained for �-gal activity. All major arteries examined,including coronary arteries, aorta, pulmonary trunk (A), intercostal (D), mesenteric (E), and femoral arteries (F ) were positive for Ptc1lacZ or Ptc2lacZ activity asshown. (B) Adult hearts displayed low levels of �-gal activity in coronary arteries, aorta, and pulmonary artery. (C) A cross-section of the aorta at P2 revealsPtc1lacZ-positive cells are restricted to the adventitia. Ica, intercostal artery; A, artery; V, vein; N, nerve. (Magnification: A and D–F, �7; B, �3; C, � 40.)

9350 � www.pnas.org�cgi�doi�10.1073�pnas.0711382105 Passman et al.

SMC-like cells, and that this differentiation does not require cellproliferation.

AdvSca1 cells do not express SMC marker proteins in vivo(Figs. 2 and 4). However, they contain mRNAs for transcriptionfactors involved in SMC differentiation, including serum re-sponse factor (SRF), myocardin family members, and cysteine-rich LIM proteins Csrp1 and Csrp2 (Fig. 3A). In addition, asubset of freshly isolated AdvSca1 cell nuclei were immunopo-sitive for SRF protein (Fig. 3B). However, AdvSca1 cells alsoexpress potent silencers of SRF-dependent transcription, includ-ing Msx1, Klf4, and FoxO4 (Fig. 3D and data not shown). Eachof these factors can inhibit SRF and myocardin-dependent SMCgene expression (32, 33), and they are down-regulated as Ad-vSca1 cells acquire SMC markers and increase SRF coactivatorexpression (Fig. 3D). Therefore, a subset of AdvSca1 cells maybe maintained as SMC progenitors by expression of silencers ofSRF-dependent transcription. When AdvSca1 cells were incu-bated with VEGF-A (10 ng/ml, 9 days), PECAM1-positiveendothelial cell clusters appeared in elongated cords that oftencolocalized with SM�Actin-positive cells (Fig. S4). When incu-bated with BMP2 (50 ng/ml, 20 days), colonies formed thatstained with alizarin red or von Kossa’s stain consistent withdifferentiation of a subset of AdvSca1 cells to osteogenic cells(Fig. S4). Under conditions described above, 30–50% of Ad-vSca1 cells differentiated to SMC-like cells, �10% producedendothelial cells, and 1–5% formed osteogenic colonies in vitro.

Developmental Origins of AdvSca1 Cells. Bone marrow reconstitu-tion experiments previously showed that AdvSca1 cells are not

marrow-derived (6). To gain insight into the origins of AdvSca1cells in vivo we examined mice from embryonic day 12.5 (E12.5)to 12 weeks after birth. AdvSca1 cells first appeared in theperivascular space between the ascending aorta and pulmonarytrunk between E15.5 and E18.5 (Fig. 4C, arrows), well after thetunica media had acquired SMCs and organized a full comple-ment of elastic lamellae. AdvSca1 cells were present in thisposition at all subsequent time points examined, and theyincreased in number with postnatal growth of the artery wall(Fig. 4 D–F). We then used Wnt1-cre transgenic mice to labelcells of neural crest origin (34). Wnt1-cre-activated �-gal reporterlabeled neural crest-derived SMCs in proximal aorta, pulmonarytrunk, common carotid arteries, and ductus arteriosus, but didnot label AdvSca1 cells in these vessels (Fig. 4 G and H).

Analysis of AdvSca1 Cells in Shh�/� Embryos. To determine whetherShh was required for the formation or maintenance of AdvSca1cells, we examined Shh�/� mice. Two anatomically defined siteswere studied. The presence of AdvSca1 cells at the aortic root,where they are first detected during development (Fig. 4C) andare most abundant in adult ApoE�/� mice (6), was confirmed inWT mice (Fig. 5A). By contrast, we could not detect AdvSca1cells at the aortic root in Shh�/� embryos (Fig. 5B). A smallnumber of AdvSca1 cells were found in the transverse portion ofthoracic aorta (Fig. 5 E and F). An absence of septation of thecommon truncus arteriosus in Shh-deficient embryos has beendescribed (35). AdvSca1 cells were present in the descendingthoracic aorta of Shh�/� embryos, but always in reduced num-bers (Fig. 5 C–F and Fig. S5). The presence of residual AdvSca1cells suggests that factors other than Shh are involved in homingand/or maintenance of these cells. Because Dhh expression wasdetected in the aorta, low levels of Dhh may allow some of thesecells to escape a requirement for Shh in the artery wall.

Fig. 2. Distribution of Shh and Sca1� cells in the adventitia. (A) Shh proteinis found outside the aortic external elastic lamina (marked by green autofluo-rescence) at P2. (B) Shh protein (red) is found between the media (SM�Actin,green) and adventitia of descending aorta. (B Inset) Shh (red) is colocalizedwith �-gal in aortic sections from Ptc1lacZ mice. (C) Sca1� cells (red) resideoutside the media (SM�Actin, green) in P2 aorta. (D) Sca1� cells (green) arefound in close proximity to Shh (red) and are embedded within a �-gal-positivedomain (Inset) in aortic adventitia from Ptc1lacZ mice. (E) Mesenteric vesselsstained for SM�Actin (green) and Shh (red). (F ) Femoral arteries stained forSM�Actin (green) and Sca1 (red). Images are stacked Z-plane sections fromconfocal microscopy. Adventitial localization of Sca1� cells correlates with thedistribution of Shh protein and active Shh signaling in vivo.

Fig. 3. Analysis of AdvSca1 cells. (A) RT-PCR analysis of AdvSca1 cells fromadult arteries. (B) Freshly isolated AdvSca1 cells (red) contain SRF (green)colocalized with DAPI-stained nuclei (blue) (arrows). (C) Freshly isolatedAdvSca1 cells (red) contain Klf4 (green) colocalized with DAPI-stained nuclei(blue) (arrows). (D) RT-PCR analysis of AdvSca1 cells cultured for 0, 12, or 28days. M is total RNA from aortic media used as a positive control for SMCmarkers. This analysis was performed three times with similar results. (E)AdvSca1 cells cultured in serum-containing medium for 9 days. Many cellsdown-regulate ScaI (red) and up-regulate SM�Actin (green). (Magnifications:B and C, �40; E, �10.)

Passman et al. PNAS � July 8, 2008 � vol. 105 � no. 27 � 9351

MED

ICA

LSC

IEN

CES

DiscussionWe report the characterization of a Shh signaling domain in theadventitial layer of vessel wall. This domain is strongly positivefor both Ptc1lacZ and Ptc2lacZ activities and colocalizes with a ringof Shh protein deposited between the media and adventitia.Embedded within this domain is a population of Sca1-expressingvascular progenitor cells, termed AdvSca1 cells, that can differ-entiate to SMC-like cells ex vivo. The number of AdvSca1 cellsin the aortic root is greatly diminished in Shh�/� mice at E18.5.These findings suggest that Shh signaling may play importantroles in development of the adventitia and maintenance of aresident population of Sca1� vascular progenitor cells in theartery wall.

Hh Signaling in Postnatal Artery Wall. Ptc1lacZ and Ptc2lacZ activi-ties in the adventitia of large and medium-sized arteries weregreatest between 1 and 10 days after birth. Although the basicstructure of major arteries is established at birth, substantialgrowth and remodeling occur in the perinatal period. Forexample, the number of SMCs in rat aorta increases 2.5-fold,wall thickness doubles, and collagen and elastin contentsincrease �3-fold in the first month after birth (36). Thesechanges in wall structure are adaptations to large perinatal

increases in blood f low and blood pressure that are accom-panied by corresponding increases in smooth muscle contrac-tile protein mass (37). After 3 months of age, growth of theartery wall is diminished, cell proliferation rates are low, andwall structure has reached an adult state. We found PtclacZ

activities and Shh protein levels were maintained at low levelsin adult mouse arteries, suggesting that adventitial Shh sig-naling activity correlates with rapid postnatal growth andremodeling of the artery wall.

Distribution of Shh in the Adventitia. Ptc1lacZ and Ptc2lacZ activitieswere colocalized with Shh protein deposited between the mediaand adventitia. Movement of Hh proteins in the extracellularspace is limited by covalent lipid modifications at both amino-and carboxyl-terminal domains (38, 39). Lipid-modified Hhproteins are secreted as multimers that interact with heparansulfate proteoglycans (HSPGs), which further restricts theirmovement (40, 41). Our data suggest that Shh protein is con-centrated between the media and adventitia because of localsynthesis, secretion, and retention of Shh multimers by ECMcomponents, such as the HSPG perlecan (42), in the artery wall.PtclacZ-positive cells are located throughout the adventitia,whether directly adjacent to Shh protein at the media-adventitiainterface or several cell layers removed. We do not yet knowwhether this pattern of PtclacZ activity in vivo is the result ofparacrine secretion of Shh by AdvSca1 cells themselves, ECMretention of Shh produced from another source, or some com-bination of both. However, freshly isolated AdvSca1 containreadily detectable amounts of Shh mRNA, suggesting that localsynthesis and secretion of Shh is a normal function of the arterialadventitia.

Fig. 4. AdvSca1 cells in developing aorta. (A–F ) Tissues were obtained at thetime points indicated. Cross-sections through the aortic root were stained forSca1 (red), SM�Actin (green), and DAPI (blue). AdvSca1 cells first appearbetween E15.5 and E18.5 in the adventitial space between aorta and pulmo-nary trunk (white arrows in C) and persist in this location into adulthood. (G)Cross-section from Wnt1-cre/R26R (5 weeks old) aortic root stained for �-galactivity and nuclear fast red. SMCs (blue) but not adventitial cells originatefrom neural crest. Dashed box is shown at higher magnification in H. (H)Cross-section through aortic root shows no overlap between �-gal-positiveand AdvSca1� cells. Ao, aorta; PT, pulmonary trunk. (Magnification scale barin E applies to A–F; G, �10.)

Fig. 5. AdvSca1 cells in Shh�/� arteries. (A–D) Aortic tissues from WT (A andC) and Shh�/� (B and D) embryos at E18.5. Cross-sections through aortic root(A and B) and descending aorta (C and D) were immunostained for Sca1 (red),SM�Actin (green), and DAPI (blue). (E) Solid lines in indicate relative positionsof sections in A–D. (F ) Dotted line boxes in A–D correspond to aortic segmentsshown here. AdvSca1 cells are absent in aortic root, greatly reduced in as-cending and transverse aorta (termed ascending), and diminished in descend-ing aorta of Shh�/� embryos.

9352 � www.pnas.org�cgi�doi�10.1073�pnas.0711382105 Passman et al.

Analysis of Isolated AdvSca1 Cells. Embedded within the PtclacZ-positive adventitial domain, we found a population of Sca1-expressing cells that were negative for marker proteins ofdifferentiated SMCs. However, when removed from the ad-ventitia and placed in serum-containing medium, a significantfraction (�30–50%) of these cells lost Sca1 from the cellsurface, acquired a well spread mesenchymal cell shape, andexpressed the SMC marker proteins SM�Actin, SM22�, cal-ponin, and SM-MHC by 12–28 days. When AdvSca1 cells wereexposed to BrdU to label replicating cells, a significant numberof Sca1–BrdU double positive cells were found after 7 days,indicating that some AdvSca1 self-renew while, at the sametime, others differentiate to SMC-like cells under the sameconditions (Fig. S3). These results suggest that AdvSca1 cellsare a heterogeneous mixture of cells that can either self-renewor respond to extracellular signals and differentiate to SMC-like cells in vitro.

Maintenance of a Progenitor Phenotype by Transcriptional Corepres-sors. AdvSca1 cells do not express SMC marker proteins in vivo.It was intriguing, therefore, to find that AdvSca1 cells con-tained transcripts for SRF and myocardin family members andwere immunopositive for SRF when examined immediatelyafter isolation from the artery wall. These transcription factorsare strong activators of SMC gene expression (37, 43–45) andare necessary for vascular SMC differentiation in vivo. YetSRF also interacts with potent transcriptional corepressors,and it is the competition between coactivators and corepres-sors that determines whether SRF-dependent target genes aretranscribed (44–48). Indeed, we found that the transcriptionalcorepressors Msx1, Klf4, and FoxO4 were also expressed byAdvSca1 cells. Msx1 forms a ternary complex with SRF andmyocardin and inhibits binding of SRF-myocardin to CArGbox motifs in SMC target genes (32). Klf4 is a zinc finger-containing protein that binds to a GC-rich element (TCE) andinhibits SRF-dependent transcription of multiple SMC markergenes (47). The forkhead transcription factor FoxO4 interactswith myocardin and inhibits its coactivator function for SRF-dependent SMC gene expression (49). These results are rem-iniscent of a report by Matsuura et al. (50), showing that Sca1�

cells from adult mouse hearts express Nkx2–5 and GATA4,two transcription factors important for cardiac myocyte dif-ferentiation, yet they do not express markers of differentiatedmyocardial cells. Heart-derived Sca1� cells differentiate tobeating cardiomyocytes when exposed to oxytocin in vitro, thusconfirming their cardiogenic potential (50). These findings inheart and artery wall suggest a model in which Sca1� progen-itor cells in postnatal tissues are specified for certain cell fatesby expression of transcription factors that are required forthose fates. The maintenance of a SMC progenitor phenotypemay therefore critically depend on expression of transcrip-tional corepressors that block SRF-dependent transcriptionand recruit chromatin remodeling complexes to silence geneexpression.

Origins of AdvSca1 Cells. The origins of AdvSca1 cells are unclear.Hu et al. (6) found no evidence that AdvSca1 cells arise frombone marrow. We investigated this question for aortic AdvSca1cells by using Wnt1-cre transgenic mice in which �-gal ispermanently expressed in neural crest-derived cells (34). Ad-vSca1 cells directly adjacent to Wnt1-cre-positive aortic SMCsdid not express the neural crest marker. Moreover, our devel-opmental time course studies showed that AdvSca1 cells didnot appear in aortic arch arteries until between E15.5 andE18.5, much later than the migration of neural crest-derivedSMC progenitors into the aortic arch complex �E9.5. Twoother lineages contribute SMCs to the aorta and other greatvessels, i.e., paraxial somitic mesoderm and secondary heart

field (51). Lineage-specific cre mice are available for both ofthese progenitors to test whether AdvSca1 cells arise fromeither of these sources.

Summary. We characterize a novel domain of Shh signaling inthe arterial adventitia and show that a population of Sca1�

vascular progenitor cells resides within this domain. Becausethe adventitia lies between the vessel wall and surroundingtissues, progenitor cells within the adventitia could, in prin-ciple, contribute to remodeling, repair, or disease processes ineither the vessel wall or in neighboring nonvascular tissues.Because essentially all tissues of the body contain bloodvessels, the maintenance and fate of adventitial progenitorcells described here and by Hu et al. (6) may be relevant to awide variety of normal and pathogenic processes. Future workwill examine the role of Hh signaling in responses of theadventitia to tissue injuries involving both vessels and sur-rounding nonvascular cell types.

Materials and MethodsAnimals Used. All protocols were approved by the Institutional Animal Careand Use Committee at the University of North Carolina. Mice used includePtc1lacZ (Ptctm1Mps) (26), Ptc2lacZ (Ptc2tm1Dgen; Jackson Laboratories; 005827),Wnt1cre (Jackson Laboratories; 003829), Shhtm1Amc (Jackson Laboratories;003318), Rosa26 (52), CD1 (Charles River), 129/SvEvJ and C57BL/6J (JacksonLaboratories). Noon on the day of vaginal plug was designated E0.5.

Whole-Mount �-Gal Activity. Embryos or tissues were fixed in fresh 0.2%glutaraldehyde in 0.1 M sodium phosphate, 5 mM EGTA, 2 mM MgCl2, pH 7.3,for 1 h on ice, washed three times with rinse buffer (10 mM sodium phosphate,2 mM MgCl2, and 0.1% Triton X-100, pH 7.4) for 30 min each at roomtemperature, and stained as described in SI Text.

Immunofluorescence Staining. Tissues were fixed in 4% paraformaldehyde(PFA), rinsed in PBS, saturated with 20% sucrose for cryoprotection, embed-ded in agar, and frozen in OCT. Cryosections (12 �m) were fixed and stainedas described in SI Text. Cultured cells were fixed in 4% PFA for 15 min at roomtemperature, then stained as described in SI Text. Primary antibodies andimaging methods are described in SI Text.

AdvSca1 Cell Isolation and Culture. The ascending aorta, aortic arch, and �1cm of the descending aorta were dissected from adult mice and placed inPBS on ice. Endothelium was removed with a cotton swab. Adventitia wasdissected and rinsed in HBSS, digested with 14 mg/ml collagenase type 2(Worthington) and 0.75 mg/ml elastase (Roche) in HBSS for 2 h at 37°C withgentle rocking, and filtered (70 �m). Cells in the filtrate were pelleted at300 � g then rinsed in PBS � 0.5% BSA. Aortic Sca1� cells were isolated byusing anti-Sca1 immunomagnetic MicroBeads and a MACS cell separationsystem (Miltenyi). Cells were passed over two consecutive columns toincrease purity of the isolation. A typical Sca1� fraction thus prepared is1– 4% of total cells, with 80 –95% of isolated cells positive for Sca1 antigen,and �1% of isolated cells positive for SM�Actin by immunostaining.Isolated Sca1� cells were cultured in DMEM (Sigma) plus 10% FBS (HyClone)and 1� antibiotic/antimycotic solution (Gibco) at 37°C, 5% CO2. Cells wereseeded in 48-well tissue culture plates at 8 � 103 cells per well. Growthfactors tested were VEGF-A (10 ng/ml; Clonetics) and BMP2 (50 ng/ml,PeproTech).

RT-PCR Analysis. Total cellular RNA was isolated by guanidinium isothiocya-nate denaturation and phenol/chloroform extraction as described (53). Two-step RT-PCR was carried out with the GeneAmp RNA PCR kit (Applied Biosys-tems) according to the manufacturer’s instructions. Primer sequences and PCRconditions are listed in Table S1.

ACKNOWLEDGMENTS. We thank Da-Zhi Wang, Frank Conlon, and JimFaber for helpful discussions. This work was carried out by J.N.P. in partialfulfillment of requirements for the PhD degree in the Curriculum in Ge-netics and Molecular Biology, University of North Carolina. This work wassupported by National Institutes of Health Grants HL-19242 (to M.W.M.),HL-07816 (to C.M.), GM-00851 (to J.N.P.), and HL43174 and HL71993 (toV.L.B.), National Research Service Award F32 HL68484 (to K.H.), AmericanHeart Association Grant 0715320U (to J.N.P. and M.W.M.), and the CarolinaCardiovascular Biology Center.

Passman et al. PNAS � July 8, 2008 � vol. 105 � no. 27 � 9353

MED

ICA

LSC

IEN

CES

1. Heistad D, Marcus M, Larsen G, Armstrong M (1981) Role of vasa vasorum in nourish-ment of the aortic wall. Am J Physiol 240:H781–H787.

2. Sartore S, et al. (2001) Contribution of adventitial fibroblasts to neointima formationand vascular remodeling: From innocent bystander to active participant. Circ Res89:1111–1121.

3. Moulton K, et al. (2003) Inhibition of plaque neovascularization reduces macrophageaccumulation and progression of advanced atherosclerosis. Proc Natl Acad Sci USA100:4726–4741.

4. Haurani M, Pagano P (2007) Adventitial fibroblast reactive oxygen species as autacrineand paracrine mediators of remodeling: Bellwether for vascular disease? CardiovascRes 75:679–689.

5. Ryan S, Koteliansky V, Gotwals P, Lindner V (2003) Transforming growth factor-�-dependent events in vascular remodeling following arterial injury. J Vasc Res 40:37–46.

6. Hu Y, et al. (2004) Abundant progenitor cells in the adventitia contribute to athero-sclerosis of vein grafts in ApoE-deficient mice. J Clin Invest 113:1258–1265.

7. Zengin E, et al. (2006) Vascular wall resident progenitor cells: A source for postnatalvasculogenesis. Development 133:1543–1551.

8. Pasquinelli G, et al. (2007) Thoracic aortas from multi-organ donors are suitable forobtaining resident angiogenic mesenchymal stromal cells. Stem Cells 25:1627–1634.

9. McMahon A, Ingham P, Tabin C (2003) Developmental roles and clinical significance ofhedgehog signaling. Curr Top Dev Biol 53:1–114.

10. Lawson N, Vogel A, Weinstein B (2002) Sonic hedgehog and vascular endothelialgrowth factor act upstream of the notch pathway during arterial endothelial differ-entiation. Dev Cell 3:127–136.

11. Byrd N, et al. (2002) Hedgehog is required for murine yolk sac angiogenesis. Develop-ment 129:361–372.

12. Nagase M, Natase T, Koshima I, Fujita T (2006) Critical time window of hedgehog-dependent angiogenesis in murine yolk sac. Microvasc Res 71:85–90.

13. Pola R, et al. (2001) The morphogen sonic hedgehog is an indirect angiogenic agentupregulating two families of angiogenic growth factors. Nat Med 7:706–711.

14. Pola R, et al. (2003) Postnatal recapitulation of embryonic hedgehog pathway inresponse to skeletal muscle ischemia. Circulation 108:479–485.

15. Kusano K, et al. (2004) Sonic hedgehog induces arteriogenesis in diabetic vasa nervo-rum and restores function in diabetic neuropathy. Arterioscler Thromb Vasc Biol24:2102–2107.

16. Kusano K, et al. (2006) Sonic hedgehog myocardial gene therapy: Tissue repair throughtransient reconstitution of embryonic signaling. Nat Med 11:1197–1204.

17. Kanda S, et al. (2003) Sonic hedgehog induces capillary morphogenesis by endothelialcells through phosphoinositide 3-kinase. J Biol Chem 278:8244–8249.

18. Vokes S, et al. (2004) Hedgehog signaling is essential for endothelial tube formationduring vasculogenesis. Development 131:4371–4380.

19. Wang Y, McMahon A, Allen B (2007) Shifting paradigms in hedgehog signaling. CurrOpin Cell Biol 19:159–165.

20. Taipale J, Cooper M, Maiti T, Beachy P (2002) Patched acts catalytically to suppress theactivity of smoothened. Nature 418:892–897.

21. Incardona J (2005) From sensing cellular sterols to assembling sensory structures. DevCell 8:798–799.

22. Corcoran R, Scott M (2006) Oxysterols stimulate sonic hedgehog signal transductionand proliferation of medulloblastoma cells. Proc Natl Acad Sci USA 103:8408–8413.

23. Huangfu D, et al. (2003) Hedgehog signaling in the mouse requires intraflagellartransport proteins. Nature 426:83–87.

24. Corbit K, et al. (2005) Vertebrate smoothened functions at the primary cilium. Nature437:1018–1021.

25. Rohatgi R, Milenkovic L, Scott M (2007) Patched1 regulates hedgehog signaling at theprimary cilium. Science 317:372–376.

26. Goodrich L, Milenkovic L, Higgins K, Scott M (1997) Altered neural cell fates andmedulloblastoma in mouse patched mutants. Science 277:1109–1113.

27. Motoyama J, Takabatake T, Takeshima K, Hui C (1998) Ptch2, a second mouse Patchedgene is coexpressed with sonic hedgehog. Nat Genet 18:104–106.

28. Adolphe C, et al. (2004) An in vivo comparative study of sonic, desert, and Indianhedgehog reveals that hedgehog pathway activity regulates epidermal stem cellhomeostasis. Development 131:5009–5019.

29. Ahn S, Joyner A (2005) In vivo analysis of quiescent adult neural stem cells respondingto sonic hedgehog. Nature 437:894–897.

30. Uhmann A, et al. (2007) The hedgehog receptor patched controls lymphoid lineagecommitment. Blood 110:1814–1823.

31. Baron M (2003) Embryonic origins of mammalian hematopoiesis. Exp Hematol31:1160–1169.

32. Hayashi N, Nakamura S, Nishida W, Sobue K (2006) Bone morphogenetic protein-induced Msx1 and Msx2 inhibit myocardin-dependent smooth muscle gene transcrip-tion. Mol Cell Biol 26:9456–9470.

33. Liu Y, et al. (2005) Kruppel-like factor 4 abrogates myocardin-induced activation ofsmooth muscle gene expression. J Biol Chem 280:9719–9727.

34. Jiang X, Rowitch D, Soriano P, McMahon A, Sucov H (2000) Fate of the mammaliancardiac neural crest. Development 127:1607–1616.

35. Washington Smoak I, et al. (2005) Sonic hedgehog is required for cardiac outflow tractand neural crest cell development. Dev Biol 283:357–372.

36. Olivetti G, Anversa P, Melissari M, Loud A (1980) Morphometric study of early postnataldevelopment of the thoracic aorta in the rat. Circ Res 47:417–424.

37. Owens G, Kumar M, Wamhoff B (2004) Molecular regulation of vascular smooth musclecell differentiation in development and disease. Physiol Rev 84:767–801.

38. Mann R, Beachy P (2004) Novel lipid modifications of secreted protein signals. AnnuRev Biochem 73:891–923.

39. Pepinsky R, et al. (1998) Identification of a palmitic acid-modified form of human sonichedgehog. J Biol Chem 273:14037–14045.

40. Zhu A, Scott M (2004) Incredible journey: How do developmental signals travel throughtissue? Genes Dev 18:2985–2997.

41. Eugster C, Panakova D, Mahmoud A, Eaton S (2007) Lipoprotein-heparan sulfateinteractions in the hh pathway. Dev Cell 13:57–71.

42. Carrasco H, Olivares G, Faunes F, Oliva C, Larrain J (2005) Heparan sulfate proteoglycansexert positive and negative effects in Shh activity. J Cell Biochem 96:831–838.

43. Wang D, Olson E (2004) Control of smooth muscle development by the myocardinfamily of transcriptional coactivators. Curr Opin Genet Dev 14:558–566.

44. Miano J, Long X, Fujiwara K (2007) Serum response factor: master regulator of the actincytoskeleton and contractile apparatus. Am J Physiol 292:C70–C81.

45. Chang D, et al. (2003) Cysteine-rich LIM-only proteins CRP1 and CRP2 are potentsmooth muscle differentiation cofactors. Dev Cell 4:107–118.

46. Liu N, Olson E (2006) Coactivator control of cardiovascular growth and remodeling.Curr Opin Cell Biol 18:715–722.

47. Kawai-Kowase K, Owens G (2007) Multiple repressor pathways contribute to pheno-typic switching of vascular smooth muscle cells. Am J Physiol 292:C59–C69.

48. Zhou J, Hu G, Herring BP (2005) Smooth muscle-specific genes are differentiallysensitive to inhibition by Elk-1. Mol Cell Biol 25:9874–9885.

49. Liu Z, Wang Z, Yanagisawa H, Olson E (2005) Phenotypic modulation of smooth musclecells through interaction of FoxO4 and myocardin. Dev Cell 9:261–270.

50. Matsuura K, et al. (2004) Adult cardiac sca-1-positive cells differentiate into beatingcardiomyocytes. J Biol Chem 279:11384–11391.

51. Majesky M (2007) Developmental basis of vascular smooth muscle diversity. Arterio-scler Thromb Vasc Biol 27:1248–1258.

52. Soriano P (1999) Generalized lacZ expression with the ROSA26 Cre reporter strain. NatGenet 21:70–71.

53. Landerholm T, et al. (1999) A role for serum response factor in coronary smooth muscledifferentiation from proepicardial cells. Development 126:2053–2062.

9354 � www.pnas.org�cgi�doi�10.1073�pnas.0711382105 Passman et al.