Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/276365630

C-Kit+progenitorsgeneratevascularcellsfortissue-engineeredgraftsthroughmodulationoftheWnt/Klf4pathway

ARTICLEinBIOMATERIALS·AUGUST2015

ImpactFactor:8.56·DOI:10.1016/j.biomaterials.2015.04.055

READS

38

11AUTHORS,INCLUDING:

Po-WahSo

King'sCollegeLondon

60PUBLICATIONS1,173CITATIONS

SEEPROFILE

AndrianaMargariti

King'sCollegeLondon

49PUBLICATIONS972CITATIONS

SEEPROFILE

MeiMeiWong

King'sCollegeLondon

12PUBLICATIONS165CITATIONS

SEEPROFILE

YanhuaHu

King'sCollegeLondon

102PUBLICATIONS3,519CITATIONS

SEEPROFILE

Allin-textreferencesunderlinedinbluearelinkedtopublicationsonResearchGate,

lettingyouaccessandreadthemimmediately.

Availablefrom:Po-WahSo

Retrievedon:05February2016

lable at ScienceDirect

Biomaterials 60 (2015) 53e61

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

c-Kitþ progenitors generate vascular cells for tissue-engineered graftsthrough modulation of the Wnt/Klf4 pathway

Paola Campagnolo a, *, Tsung-Neng Tsai b, Xuechong Hong b, John Paul Kirton b,Po-Wah So c, Andriana Margariti b, Elisabetta Di Bernardini b, Mei Mei Wong b,Yanhua Hu b, Molly M. Stevens a, Qingbo Xu b, **

a Department of Materials, Department of Bioengineering and Institute of Biomedical Engineering, Imperial College London, United Kingdomb Cardiovascular Division, British Heart Foundation Centre, King's College London, London, United Kingdomc Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, United Kingdom

a r t i c l e i n f o

Article history:Received 5 January 2015Received in revised form22 April 2015Accepted 30 April 2015Available online

Keywords:Vascular graftStem cellsEndothelializationCell signalling

* Corresponding author. Department of Materials, RoCollege, Prince Consort Road, London, United Kingdom** Corresponding author. Department of CardiologyColdharbour Lane, London, United Kingdom.

E-mail addresses: p.campagnolo@imperial.ac.uk (kcl.ac.uk (Q. Xu).

http://dx.doi.org/10.1016/j.biomaterials.2015.04.0550142-9612/© 2015 Published by Elsevier Ltd.

a b s t r a c t

The development of decellularised scaffolds for small diameter vascular grafts is hampered by theirlimited patency, due to the lack of luminal cell coverage by endothelial cells (EC) and to the low tone ofthe vessel due to absence of a contractile smooth muscle cells (SMC). In this study, we identify a pop-ulation of vascular progenitor c-Kitþ/Sca-1- cells available in large numbers and derived from immuno-privileged embryonic stem cells (ESCs). We also define an efficient and controlled differentiation protocolyielding fully to differentiated ECs and SMCs in sufficient numbers to allow the repopulation of a tissueengineered vascular graft. When seeded ex vivo on a decellularised vessel, c-Kitþ/Sca-1-derived cellsrecapitulated the native vessel structure and upon in vivo implantation in the mouse, markedly reducedneointima formation and mortality, restoring functional vascularisation. We showed that Krüppel-liketranscription factor 4 (Klf4) regulates the choice of differentiation pathway of these cells through b-catenin activation and was itself regulated by the canonical Wnt pathway activator lithium chloride. Ourdata show that ESC-derived c-Kitþ/Sca-1-cells can be differentiated through a Klf4/b-catenin dependentpathway and are a suitable source of vascular progenitors for the creation of superior tissue-engineeredvessels from decellularised scaffolds.

© 2015 Published by Elsevier Ltd.

1. Introduction

Therapies for coronary and peripheral vascular diseases ofteninvolve vascular grafts to restorebloodperfusion. In clinical practice,the transplantation of an autologous vessel is the primary choice;however, almost 40% of patients lack autologous vessels of anadequate quality. To overcome this limitation, artificial alternativeshave been devised, including synthetic expanded polytetrafluoro-ethylene (ePTFE) and Dacron grafts and decellularisation of naturalvessels [1]. In particular, the development of decellularised vesselscaffolds allows for the creation of an ‘off-the-shelf’ materialcombining the characteristics of the natural vessels and as a

yal School of Mines, Imperial.

, King's College London, 125

P. Campagnolo), qingbo.xu@

substrate for cell adhesion and growth with low immunogenicity.Despite their promise, the application of decellularised vessel scaf-folds in the clinic has been hampered by inadequate long-termpatency due to thrombosis and neointimal hyperplasia1. We previ-ously showed in an animal model that the patency of decellularisedvessels is limited due to the lack of adequate endothelialisation [2].The outcome of the graft was greatly improved by the local appli-cation of vascular endothelial growth factor (VEGF), which inducedthe local recruitment of endothelial progenitor cells to the scaffold[2]. For the same reason, the seeding of the luminal side of thescaffold with endothelial cells has been demonstrated to improvegraft success rate by ensuring a non-thrombogenic surface of con-tact with the blood. Similarly, the seeding of smooth muscle cells(SMCs) in the media of the decellularised vessel was shown toimprove the mechanical stability of the graft and its compliance[3,4]. However, the use ofmature endothelial cells (ECs) and SMCs ishampered by their limited availability from an autologous sourceand low proliferation rate. For this reason, the exploitation of theimmuno-privileged and plastic embryonic stem cells (ESCs) is

P. Campagnolo et al. / Biomaterials 60 (2015) 53e6154

highly advantageous. Stem cells are highly proliferative and can giverise to a high number of differentiated cells including ECs and SMCs,provided themechanisms regulating the differentiation process arewell understood and controlled.

In this study, we focused on the isolation of an abundant andclinically relevant c-Kit positive/Sca1-negative (c-Kitþ/Sca-1-)vascular progenitor population from ESCs and their differentiationto both ECs and SMCs.We then applied these cells to repopulate thedecellularised scaffold for vascular graft. The c-Kit antigen haspreviously been used to define progenitor cells in the human andmouse heart where they participate to repair after myocardialinfarction [5] and form both ECs and SMCs [6,7]. Regulating thechoice of differentiation pathway in progenitor cells is funda-mental: (a) to obtain a highly differentiated and pure population fortissue engineering; (b) for the maintenance of the physiologicalstate once the vessel is grafted and (c) to manipulate the progres-sion of the in vivo re-endothelialization of the vascular graft. Forvascular progenitor cells, the mechanisms have been only partiallyelucidated. In this paper, we focused on Krüppel-like factor 4 (Klf4),which is known to play an atheroprotective function in the vessel[8e12] and participate to anti-inflammatory [13e16] and shearstress response [17,18]. We also investigated the interaction of Klf4with Wnt/b-catenin signalling, which plays an important role invasculature development and endothelial function/remodeling[19e21] through the shuffling of b-catenin and activation of T cellfactor (TCF) target genes [13,14,22]. Understanding the differenti-ation mechanisms underpinning c-Kitþ/Sca-1- cell fate is crucialfor their use as a cell source for vascular grafts.

2. Material and methods

For expanded Materials and methods, refer to the SupplementalMethods provided.

2.1. Cell culture, isolation and differentiation

Mouse ESCs (ES-D3 cell line, American Type Culture Collection[ATCC]) and isolated c-Kitþ/Sca1-cells were cultured as previouslyreported [23]. Differentiation was induced by plating c-Kitþ/Sca1-cells on collagen IV-coated flasks in presence of differentiation me-dium (DM; a-MEM containing 10% FBS, 0.2 mM 2-mercaptoethanol,100 U/ml penicillin, and 100 mg/ml streptomycin; Invitrogen) con-tainingeither VEGF (50 ng/ml, Peprotech) or platelet-derived growthfactor (PDGF, 20 ng/ml, Sigma) for up to 21 days. The endothelialdifferentiation protocol was improved with the application of shearstress between Day 3 and 5 using an orbital shaker.

2.2. Decellularised vessel preparation and seeding

The preparation of the decellularised vessels was performed aspreviously described [2] treating the descending aorta with 0.075%SDS for 2 h. c-Kitþ ECs were seeded using a bioreactor where thegrafts were fixed between two needles and the bioreactor was setup in a standard incubator at 37 �C. Scaffolds were preconditionedfor 2 h with the indicated coating. 2 � 106 c-Kitþ EC were theninjected and allowed to seed for 12 h at a continuous rotationalmovement before initial flow was set up. For double seeded scaf-folds, a second seeding step was performed with 1 � 106 c-Kitþderived SMCs. After seeding, medium flow rate was adjustedstepwise to reach a shear stress of 30 dyn/cm2. Vessels were har-vested on Day 5 and then used for further analysis ex vivo orimmediately grafted to animals. Decellularised vessels were used asa control.

Picrosirius collagen and Miller's elastin staining was performedto assess the collagen content in the tissue-engineered vessel.

2.3. Decellularised and tissue-engineered vessel grafting to the rightcarotid artery

Mice were anesthetised and the middle portion of the carotidartery was ligated and dissected between the two ties. The graftwas interposed as previously described using a two-cuffs basedsystem [2].

A total of 56, 28 and 13 mice were transplanted respectivelywith non-seeded, c-Kitþ EC seeded, and c-Kitþ ECsþ SMCs seededgrafts, respectively. Survival was monitored for up to 56 days andtissues for patency monitoring were collected at 2, 4 and 8 weeks.

Mice underwent magnetic resonance imaging (MRI) on a 7TAgilent MRI scanner tuned to 400 MHz 1H frequency.

2.4. Lesion measurement

For en face staining, the tissue-engineered vessels were cutlongitudinally and mounted with the lumen facing up before in-cubationwith Alexa Fluor 488 phalloidin and/or 40-6-Diamidino-2-phenylindole (DAPI). Lesion measurement was performed bystaining cross-sections obtained at the centre of the graft withhematoxylin and eosin (H&E stain).

2.5. Statistical analysis

Data were analysed with GraphPad Prism 5.02 program usingANOVA and two-tailed student's t-test for two-groups or pair-wisecomparisons. The results were expressed as the mean ± standarderror of the mean (SEM).

3. Results

3.1. Isolated c-Kitþ/Sca-1- cells represent a novel population ofESC-derived vascular progenitor cells

ESCs cultured on collagen IV became adherent and elongated(Fig. 1ai) and contained high numbers of c-Kitþ/Sca-1- cells (about50% of the total, Fig. 1aii). c-Kitþ/Sca-1- cells were isolated by flowcytometry (FACS) and stimulated with VEGF in combination withshear stress or with PDGF-BB (Fig.1aiiieiv). In the presence of VEGF,the cells assumed cobblestone-like features (Fig. 1aiii) andexpressed EC markers CD31, Flk-1 and VE-cadherin at both mRNA(Fig. 1bi) and protein level (Fig. 1bii). Flow cytometry analysisshowed that under these culture conditions, the populationexpressing endothelial markers accounted to about the 80% of thecells (Fig. 1biii). Functional differentiation was demonstrated byMatrigel assay and Ac-LDL uptake (Fig. 1biv and SupplementalFig. 1). When cultured with PDGF-BB, the cells grew in size,assuming a polygonal shape (Fig. 1aiv), and started overexpressingSMC specific genes (Fig. 1ci) and proteins (Fig. 1cii-iii). Imageanalysis indicated a differentiation efficiency nearing 100%(Supplemental Fig. 2).

3.2. c-Kitþ derived vascular cells repopulated the decellularisedgraft and improved in vivo outcome

When suitable patient-derived vessels are unavailable for vesselgraft procedures, decellularised vessels offer an attractive alterna-tive. Our lab has developed an efficient mouse model mimickingthe clinical situation whereby the aorta is harvested and decellu-larised ex vivo before being grafted in themouse carotid artery, thussimulating the patient situation during vessel surgery [2]. In thisstudy, we aimed to exploit the potential of c-Kitþ derived cells forpre-clinical application by seeding them on the decellularisedscaffold and implanting it in the mouse model of vessel graft. We

Fig. 1. Isolation and differentiation of c-Kitþ/Sca-1- cells. a: Morphology of differentiated ESC (ai) and flow cytometry strategy for the isolation of c-Kitþ/Sca-1- subpopulation (redbox, aii). Addition of VEGF and application of shear induced endothelial phenotype (EC, aiii) while PDGF induced smooth muscle cells phenotype (SMC, aiv). b: VEGF treated cellsoverexpressed EC markers as shown by real time-PCR (bi), immunocytochemistry (bii) and flow cytometry (biii). Functional differentiation was assessed by in vitro tube formationassay (biv). c: PDGF treatment induced expression of SMC markers, as shown at mRNA level (ci) and by immunocytochemistry (cii) and western blot analysis (ciii). Data areexpressed as fold-change over Day 0 (dotted lines). Scale bars: 200 mm (a), 25 mm (bii), 100 mm (biv), 50 mm (cii) *P < 0.05; **P < 0.01; ***P < 0.001 vs. Day 0. (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version of this article.)

P. Campagnolo et al. / Biomaterials 60 (2015) 53e61 55

tested a set of conditions to determine the most efficient seedingprotocol leading to a homogenous luminal attachment of ECs on thegraft. Application of shear force was initiated at a rate of 3 dyn/cm2

and then gradually increased to reach the physiological shear stressof 30 dyn/cm2. The protocol proved crucial for the cell attachmentto the lumen; the best coverage was obtained when the flow was

P. Campagnolo et al. / Biomaterials 60 (2015) 53e6156

increased hourly over the first 24 h (Fig. 2a and b). Next, thedecellularised vessel was coated with gelatin or fibronectin andresults showed that cell attachment was vastly improved aftercoating, upon the application of the shear (Fig. 2c). Under theseconditions, c-Kitþ ECs efficiently attached to the luminal surface ofthe vessel, producing a continuous monolayer of CD31þ ECs(Fig. 2d) that responded to the application of physiological shearstress by aligning to the direction of the flow and becoming elon-gated (Fig. 2e). Collagen and elastin staining showed that cellseeding improved the extracellular matrix content (Fig. 2f).

When decellularised vessels seeded with c-Kitþ ECs wereimplanted in a model of vascular graft, we observed increased re-sidual lumen area (Fig. 2g) and reduced neointimal formation(Fig. 2h). Overall, 100% of the seeded grafts were patent at 4 weeks,compared with 60% of the control grafts (Supplemental Table 1). Asreported in Supplemental Table 2, the major complication observedin mice grafted with decellularised vessel was thrombus formation,which was completely abolished by graft seeding. To furtherimprove the design of the graft and to ameliorate the performancein terms of graft rupture (accounting for 28.6% of the graft failuresin c-Kit-EC seeded vessels, Supplemental Table 2), we seeded thedecellularised vessels with a combination of c-Kitþ derived SMCs.Seeding of the graft with ECs only or ECsþ SMCs significantlyincreased the survival of the mice (Fig. 2i) by reducing graft oc-clusion (Fig. 2j). Survival rate was increased to over 80% in thedouble seeded vessels (Fig. 2i). MRI confirmed the superior per-formance of ECsþ SMCs seeded vessels (Fig. 2k).

Together, these data indicate that the application of progenitor-derived vascular cells for repopulation of vascular grafts is feasibleand that seeding dramatically reduces lumen occlusion and ulti-mately mortality.

3.3. Klf4 induces EC differentiation and inhibits SMC differentiation

Data presented so far demonstrated that c-Kit-derived vascularprogenitor cells can give rise to both ECs and SMCs and that theirapplication to tissue-engineered vascular grafts can significantlyimprove in vivo performance and survival. Next we studied theunderlying molecular mechanisms determining the choice of dif-ferentiation pathway of vascular progenitor cells. The under-standing of the regulation of progenitor cell differentiation isfundamental for the development of strategies aimed at thereduction of graft complications due to the over-proliferation ofSMC and the lack of EC coverage.

We established that Krüppel-like transcription factor 4 (Klf4)was regulated upon differentiation. In particular, Klf4 expressionwas strongly induced during EC differentiation and downregulatedduring SMC differentiation (Fig. 3a). Upon adenoviral over-expression of Klf4 (Supplemental Fig. 3), we observed increased ECcommitment of the c-Kitþ derived vascular progenitors as shownby increased tube-formation capacity (Fig. 3b). Klf4 overexpressionled to increased expression of EC markers CD31, VE-cadherin andeNOS at both RNA and protein level, and a concomitant reduction inSMC markers (Fig. 3c and d). Additionally, lentiviral knockdown ofKlf4 (Supplemental Fig. 3) resulted in increased expression of SMCand the inhibition of EC markers at both mRNA and protein levels(Fig. 3e and f). These results highlight the central role of Klf4 in thedetermination of the differentiation fate in c-Kit-derived vascularprogenitor cells.

3.4. Klf4 and canonical Wnt pathway interact to determine c-Kitþcell differentiation fate

The canonical Wnt pathway has been implicated in the fatedecision of stem cell-derived vascular cells [17,18,21,24]. Klf4 has

also been shown to interact with b-catenin in the nucleus in thecontext of tumour growth [25,26]. In our study, we found that Klf4overexpression led to a nuclear accumulation of b-catenin as shownby confocal imaging and western blot analysis (Fig. 4a and b); thiseffect was reversed when Klf4 expression was inhibited (Fig. 4c).The reported downstream effectors of the Wnt pathway Axin2, ID2and Cyclin D1 were altered by Klf4 overexpression and knockdown(Supplemental Fig. 4).

Treatment of cells with lithium chloride (LiCl), a Wnt pathwayand b-catenin activator, resulted in an increase of Klf4 expressionand EC markers and a reduction of SMC markers expression(Fig. 4c), mimicking the effects of the exogenous upregulation ofKlf4 described above. Conversely, when b-catenin expression wasknocked down, EC differentiation was impaired: EC markerexpression was reduced alongside increased levels of SMC markers(Fig. 4d and e).

To confirm the importance of Klf4 in the activation of the Wntpathway, we established that the activity of the b-catenin-depen-dent TCF promoter was ablated by knockdown of Klf4, as shown byTOP-flash luciferase assay (Supplemental Fig. 5). Interestingly, LiCl-dependent EC marker upregulation was reduced in Klf4-deficientcells (Fig. 4f). Conversely, b-catenin knockdown blunted the Klf4-induced upregulation of CD31 expression and downregulation ofSMA (Fig. 4g). Our results reveal an interesting relationship be-tween Klf4 expression and b-catenin activation, whereby LiCl in-duction of EC differentiation is mediated by Klf4 and Klf4 requiresb-catenin to upregulate EC markers. This observation is in line witha previously reported mechanism showing that canonical Wntactivation directly increases Klf4 expression [16], indicating thatKlf4 might partially mediate b-catenin activation. However, at thisstage further studies are required to elucidate in the exactmechanism.

4. Discussion

Acute coronary obstruction requires arterial bypass surgerywhere an autologous vessel is grafted to the narrowed artery[15,16,27]. When autologous grafts are unavailable, the decellular-isation of native vessels can potentially provide alternative materialendowed with comparable characteristics [28e32]. However, thedelayed endothelialisation dramatically increases the potential forneointimal formation [2,8e12], which creates the need for a suit-able source of transplantable ECs. Stem cell-derived vascular cellsallow the generation of a greater number of ECs compared tomature sources that could be exploited in the clinical practiceowing to their privileged immunogenic profile [19e21,33]. In thisstudy, we provide evidence that ESC-derived c-Kitþ/Sca-1- cells area suitable source of ECs and SMCs and can therefore be consideredbona fide vascular progenitor cells. In particular, we demonstratedthe clinical relevance of ESC-derived cells in tissue engineering byseeding them on a decellularised scaffold using a bioreactor andapplying it to a model of vascular graft. The optimised graft prep-aration and seeding allowed the creation of an improved cellu-larised graft displaying a continuous luminal layer of endothelialcells and a robust SMC layer. This improved survival and patencyand reduced graft failure by avoiding thrombus, neointima devel-opment, and graft rupture [32,34,35]. Previously, our group hasshown that decellularised vessels seeded with endothelial andsmooth cells derived from adult reprogrammed cells displayedincreased patency and survival, compared to their non-seededcounterparts [4]. In the present study, we explored the applica-tion of embryonic derived c-Kitþ cells as a cell source of endothelialand smooth muscle cells. This approach has the potential advan-tage to provide a relatively easier protocol to obtain larger numberof vascular cells with high purity. Additionally, the implantation of

Fig. 2. Ex vivo seeding of decellularised vessel scaffolds and in vivo grafting. Two protocols were followed to increase the shear from 3 to 30 dyn/cm2 in the tissue engineered graftsafter luminal seeding of ECs: ‘daily step-wise’ (protocol A) or ‘hourly step-wise’ (protocol B) (a); cell retention was quantified in (b). Effect of PBS or gelatin or fibronectin coating oncell retention in static (empty bars) and shear (grey bars) was quantified in (c). c-Kitþ ECs formed a homogeneous CD31þ luminal monolayer on the decellularised vessel (d) andelongate in response to shear (e). Picosirius staining for collagen and Miler's elastin staining of decellularised vessels showed improved matrix deposition after cell seeding (f).In vivo implantation of vessels seeded with c-Kitþ ECs showed improved lumen patency (g) and reduced neointima formation (h). Both the seeding with ECs and with thecombination of ECs and SMCs (ECsþ SMCs) increased the survival of grafted animals (i) and reduced vessel obstruction (j). MRI detection of the blood flow in the normal carotid (k,i), an occluded carotid grafted with a decellularised vessel (k, ii) and a patent carotid grafted with ECsþ SMCs seeded scaffolds at Day 2 (k, iii), and 2 weeks (k, iv). Red arrowsindicate the position of the grafts. Scale bars: 50 mm (d, f, l) and 100 mm (e). *P < 0.05; **P < 0.01; ***P < 0.001. (For interpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)

P. Campagnolo et al. / Biomaterials 60 (2015) 53e61 57

Fig. 3. Effect of Klf4 expression in differentiating c-Kitþ cells. VEGF treatment induced and PDGF decreased Klf4 expression (a, vs. Day 0, dotted line). Overexpression of Klf4(AdKlf4) increased EC differentiation and reduced SMC markers as shown by tube-formation (b), PCR (c) and western blot (d), as compared to control virus (AdCTL). Klf4 knockdown(ShKlf4) reduced endothelial marker expression and increased levels of SMC markers at RNA (e) and protein level (f) *P < 0.05, **P < 0.001, ***P < 0.0001 vs. control.

P. Campagnolo et al. / Biomaterials 60 (2015) 53e6158

the vascular grafts prepared using c-Kitþ progenitors-derived ECand SMC determined better survival rate (over 80%) compared tothe adult cell-derived grafts (60%). Furthermore, the use of anabundant population of mouse progenitor cells characterised bythe expression of the c-Kit antigen makes our findings easilytranslatable to their human counterpart. The c-Kit antigen isexpressed in several progenitor populations in adult human tissues

including the heart where they act as a reservoir of cardiomyocytes,ECs and SMCs [5,36].

A thorough understanding of the molecular mechanismsimplicated in the differentiation of stem cells into the vascular celllines used for the creation of the grafts will help the development ofstrategies to improve the long-term performance of vascular grafts.In this study, we present a novel mechanism involving the

Fig. 4. Klf4 and canonical Wnt pathway are interconnected in the determination of c-Kitþ cells differentiation. Confocal imaging showing the membrane localisation of b-catenin incontrol cells (AdCTL) and nuclear accumulation in Klf4 overexpressing cells (AdKlf4, a). Western blot confirmed b-catenin activation by AdKlf4 (b) and inhibition by ShKlf4 (c). LiCl(b-catenin activator) induced Klf4 and EC marker expression and reduced the SMC marker expression in a concentration-dependent manner (c: vs. NaCl, dotted line). Knockdown ofb-catenin (Shbcat) impaired VEGF-dependent EC differentiation and increased SMC markers (d, vs. control virus: dotted line and e). LiCl-induced increase of EC markers wasinhibited by Klf4 knockdown (f, ShKlf4). Klf4-induced EC marker upregulation and SMC marker downregulation is ablated by b-catenin knockdown (g). *P < 0.05; **P < 0.01. Scalebars: 10 mm (a). *P < 0.05; **P < 0.01 and ***P < 0.001.

P. Campagnolo et al. / Biomaterials 60 (2015) 53e61 59

P. Campagnolo et al. / Biomaterials 60 (2015) 53e6160

transcription factor Klf4 and the canonical Wnt pathway that couldbe exploited in future to improve the levels of endothelialisation ofvascular grafts and control SMC overgrowth leading to neointimalgrowth. In particular, we described the central role of the tran-scription factor Klf4 in driving differentiation preferentially to-wards the EC lineage and therefore reducing SMC differentiation.This was in line with previous findings showing that Klf4 is upre-gulated by shear stress [18], a typical EC differentiation stimulus,and that Klf4 regulates eNOS [10,14,31,37] and VE-cadherinexpression [16] and inhibits SMC maturation [38]. Furthermore, arecent publication showed that Klf4 played a fundamental role inthe maintenance of the vascular phenotype of endothelial pro-genitor cells [39]. Interestingly, the Klf4-induced canonical Wntpathway activation and effect on differentiation was mimicked byLiCl, a Wnt activator. These results are in line with previous studiesdescribing the importance of Wnt in the development of the em-bryonic vasculature [20,21]. These in vitro studies create the foun-dation for future investigations on the application of Klf4 and b-catenin overexpression/stimulation in the context of vascular graftin vivo in order to ameliorate the graft endothelialization, as pre-viously demonstrated for VEGF administration [2].

The data presented described for the first time the vasculogenicpotential of ESC-derived c-Kitþ cells and their suitability as a sourceof vascular cells for the creation of alternative grafts. We provided anovel insight into the mechanism regulating their differentiation,thus providing the translational potential for this stem cell therapy.

Conflicts of interest

None.

Acknowledgement

This work was supported by British Heart Foundation grant RG/14/6/31144 and the Oak Foundation. We would like to thank MrZhongyi Zhang for the exceptional technical work.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.biomaterials.2015.04.055.

References

[1] D.G. Seifu, A. Purnama, K. Mequanint, D. Mantovani, Small-diameter vasculartissue engineering, Nat. Rev. Cardiol. N. D. 10 (2013) 410e421.

[2] T.-N. Tsai, J.P. Kirton, P. Campagnolo, L. Zhang, Q. Xiao, Z. Zhang, W. Wang,Y. Hu, Q. Xu, Contribution of stem cells to neointimal formation of decellu-larized vessel grafts in a novel mouse model, Am. J. Pathol. 181 (2012)362e373.

[3] Z. Gong, L.E. Niklason, Small-diameter human vessel wall engineered frombone marrow-derived mesenchymal stem cells (hMSCs), FASEB J. 22 (2008)1635e1648.

[4] E. Karamariti, A. Margariti, B. Winkler, X. Wang, X. Hong, D. Baban,J. Ragoussis, Y. Huang, J.-D.J. Han, M.M. Wong, C.M. Sag, A.M. Shah, Y. Hu,Q. Xu, Smooth muscle cells differentiated from reprogrammed embryoniclung fibroblasts through DKK3 signaling are potent for tissue engineering ofvascular grafts, Circ. Res. 112 (2013) 1433e1443.

[5] A.P. Beltrami, L. Barlucchi, D. Torella, M. Baker, F. Limana, S. Chimenti,H. Kasahara, M. Rota, E. Musso, K. Urbanek, A. Leri, J. Kajstura, B. Nadal-Ginard,P. Anversa, Adult cardiac stem cells are multipotent and support myocardialregeneration, Cell 114 (2003) 763e776.

[6] F. Limana, A. Zacheo, D. Mocini, A. Mangoni, G. Borsellino, A. Diamantini, R. DeMori, L. Battistini, E. Vigna, M. Santini, V. Loiaconi, G. Pompilio, A. Germani,M.C. Capogrossi, Identification of myocardial and vascular precursor cells inhuman and mouse epicardium, Circ. Res. 101 (2007) 1255e1265.

[7] E. Gambini, G. Pompilio, A. Biondi, F. Alamanni, M.C. Capogrossi, M. Agrifoglio,M. Pesce, C-kitþ cardiac progenitors exhibit mesenchymal markers andpreferential cardiovascular commitment, Cardiovasc. Res. 89 (2011) 362e373.

[8] K. Takahashi, K. Tanabe, M. Ohnuki, M. Narita, T. Ichisaka, K. Tomoda,S. Yamanaka, Induction of pluripotent stem cells from adult human fibroblasts

by defined factors, Cell 131 (2007) 861e872.[9] T. Yoshida, K.H. Kaestner, G.K. Owens, Conditional deletion of kruppel-like

factor 4 delays downregulation of smooth muscle cell differentiationmarkers but accelerates neointimal formation following vascular injury, Circ.Res. 102 (2008) 1548e1557.

[10] P. Kunes, Z. Holubcov�a, J. Krejsek, Occurrence and significance of the nucleartranscription factor Krüppel-like factor 4 (KLF4) in the vessel wall, ActaMedica (Hradec Kralove) 52 (2009) 135e139.

[11] Y. Liu, S. Sinha, G. Owens, A transforming growth factor-b control elementrequired for SM a-actin expression in vivo also partially mediates GKLF-dependent transcriptional repression, J. Biol. Chem. 278 (2003) 7.

[12] Y. Liu, S. Sinha, O.G. McDonald, Y. Shang, M.H. Hoofnagle, G.K. Owens, Krup-pel-like factor 4 abrogates myocardin-induced activation of smooth musclegene expression, J. Biol. Chem. 280 (2005) 9.

[13] N. Ohnesorge, D. Viemann, N. Schmidt, T. Czymai, D.S.E. Spiering,M. Schmolke, S. Ludwig, J. Roth, M. Goebeler, M. Schmidt, Erk5 activationelicits a vasoprotective endothelial phenotype via induction of Kr√�ppel-likefactor 4 (KLF4), J. Biol. Chem. 285 (2010) 26199e26210.

[14] G. Villarreal Jr., Y. Zhang, H.B. Larman, J. Gracia-Sancho, A. Koo, G. Garcia-Cardena, Defining the regulation of KLF4 expression and its downstreamtranscriptional targets in vascular endothelial cells, Biochem. Biophys. Res.Commun. 391 (2010) 984e989.

[15] G. Orlando, K.J. Wood, P. De Coppi, P.M. Baptista, K.W. Binder, K.N. Bitar,C. Breuer, L. Burnett, G. Christ, A. Farney, M. Figliuzzi, J.H. Holmes, K. Koch,P. Macchiarini, S.-H. Mirmalek Sani, E. Opara, A. Remuzzi, J. Rogers, J.M. Saul,D. Seliktar, K. Shapira-Schweitzer, T. Smith, D. Solomon, M. Van Dyke, J.J. Yoo,Y. Zhang, A. Atala, R.J. Stratta, S. Soker, Regenerative medicine as applied togeneral surgery, Ann. Surg. 255 (2012) 867e880.

[16] C.E. Cowan, E.E. Kohler, T.A. Dugan, M.K. Mirza, A.B. Malik, K.K. Wary, Kruppel-like factor-4 transcriptionally regulates VE-cadherin expression and endo-thelial barrier function, Circ. Res. 107 (2010) 959e966.

[17] S.M. McCormick, S.G. Eskin, L.V. McIntire, C.L. Teng, C.-M. Lu, C.G. Russell,K.K. Chittur, DNA microarray reveals changes in gene expression of shearstressed human umbilical vein endothelial cells, Proc. Natl. Acad. Sci. 98(2001) 8955e8960.

[18] P.R. Clark, T.J. Jensen, M.S. Kluger, M. Morelock, A. Hanidu, Z. Qi, R.J. Tatake,J.S. Pober, MEK5 is activated by shear stress, activates ERK5 and induces KLF4to modulate TNF responses in human dermal microvascular endothelial cells,Microcirculation 18 (2011) 102e117.

[19] S.-F. Yet, M.M. McA'Nulty, S.C. Folta, H.-W. Yen, M. Yoshizumi, C.-M. Hsieh,M.D. Layne, M.T. Chin, H. Wang, M.A. Perrella, M.K. Jain, M.-E. Lee, Human EZF,a Kruppel-like zinc finger protein, is expressed in vascular endothelial cellsand contains transcriptional activation and repression domains, J. Biol. Chem.273 (1998) 1026e1031.

[20] D.-H. Yang, J.-Y. Yoon, S.-H. Lee, V. Bryja, E.R. Andersson, E. Arenas, Y.-G. Kwon, K.-Y. Choi, Wnt5a is required for endothelial differentiation of em-bryonic stem cells and vascularization via pathways involving both Wnt/beta-catenin and protein kinase calpha, Circulation Res. 104 (2009) 372e379.

[21] M. Corada, D. Nyqvist, F. Orsenigo, A. Caprini, C. Giampietro, M.M. Taketo,M.L. Iruela-Arispe, R.H. Adams, E. Dejana, The Wnt/beta-catenin pathwaymodulates vascular remodeling and specification by upregulating Dll4/notchsignaling, Dev. Cell. 18 (2010) 938e949.

[22] C. Skurk, H. Maatz, E. Rocnik, A. Bialik, T. Force, K. Walsh, Glycogen-synthasekinase3/beta-catenin axis promotes angiogenesis through activation ofvascular endothelial growth factor signaling in endothelial cells, CirculationRes. 96 (2005) 308e318.

[23] L. Zeng, Q. Xiao, A. Margariti, Z. Zhang, A. Zampetaki, S. Patel, M.C. Capogrossi,Y. Hu, Q. Xu, HDAC3 is crucial in shear- and VEGF-induced stem cell differ-entiation toward endothelial cells, J. Cell Biol. 174 (2006) 1059e1069.

[24] A. Margariti, A. Zampetaki, Q. Xiao, B. Zhou, E. Karamariti, D. Martin, X. Yin,M. Mayr, H. Li, Z. Zhang, E. De Falco, Y. Hu, G. Cockerill, Q. Xu, L. Zeng, Histonedeacetylase 7 controls endothelial cell growth through modulation of beta-catenin, Circulation Res. 106 (2010) 1202e1211.

[25] W. Zhang, X. Chen, Y. Kato, P.M. Evans, S. Yuan, J. Yang, P.G. Rychahou,V.W. Yang, X. He, B.M. Evers, C. Liu, Novel cross talk of kruppel-like factor 4and b-catenin regulates normal intestinal homeostasis and tumor repression,Mol. Cell. Biol. 26 (2006) 2055e2064.

[26] P.M. Evans, X. Chen, W. Zhang, C. Liu, KLF4 interacts with beta-catenin/TCF4and blocks p300/CBP recruitment by beta-catenin, Mol. Cell. Biol. 30 (2010)372e381.

[27] K.A. Eagle, R.A. Guyton, R. Davidoff, F.H. Edwards, G.A. Ewy, T.J. Gardner,J.C. Hart, H.C. Herrmann, L.D. Hillis, A.M. Hutter, B.W. Lytle, R.A. Marlow,W.C. Nugent, T.A. Orszulak, E.M. Antman, S.C. Smith, J.S. Alpert, J.L. Anderson,D.P. Faxon, V. Fuster, R.J. Gibbons, G. Gregoratos, J.L. Halperin, L.F. Hiratzka,S.A. Hunt, A.K. Jacobs, J.P. Ornato, ACC/AHA 2004 guideline update for coro-nary artery bypass graft surgery: summary article, Circulation 110 (2004)1168e1176.

[28] Q. Xiao, L. Zeng, Z. Zhang, A. Margariti, Z.A. Ali, K.M. Channon, Q. Xu, Y. Hu,Sca-1þ progenitors derived from embryonic stem cells differentiate intoendothelial cells capable of vascular repair after arterial injury, Arterioscler.Thromb. Vasc. Biol. 26 (2006) 2244e2251.

[29] Q. Xiao, L. Zeng, Z. Zhang, Y. Hu, Q. Xu, Stem cell-derived Sca-1þ progenitorsdifferentiate into smooth muscle cells, which is mediated by collagen IV-integrin {alpha}1/beta1/{alpha}v and PDGF receptor pathways, Am. J. Phys-iol. Cell. Physiol. 292 (2007) C342eC352.

P. Campagnolo et al. / Biomaterials 60 (2015) 53e61 61

[30] N.D. Martin, P.J. Schaner, T.N. Tulenko, I.M. Shapiro, C.A. DiMatteo,T.K. Williams, E.S. Hager, P.J. DiMuzio, In vivo behavior of decellularized veinallograft, J. Surg. Res. 129 (2005) 17e23.

[31] B. Shen, R.S. Smith, Y.-T. Hsu, L. Chao, J. Chao, Kruppel-like factor 4 is a novelmediator of kallistatin in inhibiting endothelial inflammation via increasedendothelial nitric-oxide synthase expression, J. Biol. Chem. 284 (2009) 7.

[32] M. Pasic, W. Müller-Glauser, L.K. Segesser von, M. Lachat, T. Mihaljevic,M.I. Turina, Superior late patency of small-diameter Dacron grafts seeded withomental microvascular cells: an experimental study, Ann. Thorac. Surg. 58(1994), 677e83ediscussion683e4.

[33] M. Drukker, H. Katchman, G. Katz, S. Even-Tov Friedman, E. Shezen,E. Hornstein, O. Mandelboim, Y. Reisner, N. Benvenisty, Human embryonicstem cells and their differentiated Derivatives are less susceptible to immunerejection than adult cells, Stem Cells 24 (2006) 221e229.

[34] V. Bhattacharya, P.A. McSweeney, Q. Shi, B. Bruno, A. Ishida, R. Nash, R.F. Storb,L.R. Sauvage, W.P. Hammond, M.H.-D. Wu, Enhanced endothelialization andmicrovessel formation in polyester grafts seeded with CD34þ bone marrowcells: presented to the Western Vascular Society, Whistler, BC, Canada, Blood2000 (95) (September 11, 1998) 581e585.

[35] R.M. Rao, L. Yang, G. Garcia-Cardena, F.W. Luscinskas, Endothelial-dependentmechanisms of leukocyte recruitment to the vascular wall, Circulation Res.101 (2007) 234e247.

[36] G. Suzuki, V. Iyer, T.-C. Lee, J.M. Canty, Autologous mesenchymal stem cellsmobilize cKitþ and CD133þ bone marrow progenitor cells and improveregional function in hibernating myocardium/novelty and significance, Cir-culation Res. 109 (2011) 1044e1054.

[37] G. Zhou, A. Hamik, L. Nayak, H. Tian, H. Shi, Y. Lu, N. Sharma, X. Liao, A. Hale,L. Boerboom, R.E. Feaver, H. Gao, A. Desai, A. Schmaier, S.L. Gerson, Y. Wang,G.B. Atkins, B.R. Blackman, D.I. Simon, M.K. Jain, Endothelial Kruppel-likefactor 4 protects against atherothrombosis in mice, J. Clin. Invest. 122(2012) 4727e4731.

[38] S. Yamaguchi, K. Yamahara, K. Homma, S. Suzuki, S. Fujii, R. Morizane,T. Monkawa, Y. Matsuzaki, K. Kangawa, H. Itoh, The role of microRNA-145 inhuman embryonic stem cell differentiation into vascular cells, Atherosclerosis219 (2011) 468e474.

[39] X. Li, Y. Song, D. Wang, C. Fu, Z. Zhu, Y. Han, C. Li, N. Wang, Y. Zhu, LIFmaintains progenitor phenotype of endothelial progenitor cells via Kruppel-like factor 4, Microvasc. Res. 84 (2012) 270e277.