Ye, Xi and Abou-Rayyah, Yassir and Bischoff, Joyce and Ritchie, Alison and Sebire, Neil J. and

14 Pages·2016·2.08 MB·English

Ye, Xi and Abou-Rayyah, Yassir and Bischoff, Joyce and
Ritchie, Alison and Sebire, Neil J. and Watts, Patrick and
Churchill, Amanda J. and Bates, David O. (2016) Altered
ratios of pro- and anti-angiogenic VEGF-A variants and
pericyte expression of DLL4 disrupt the vascular
maturation in infantile haemangioma. Journal of
Pathology, 239 (2). pp. 139-151. ISSN 0022-3417
Access from the University of Nottingham repository:
http://eprints.nottingham.ac.uk/32863/15/Ye_et_al-2016-The_Journal_of_Pathology.pdf
Copyright and reuse:
The Nottingham ePrints service makes this work by researchers of the University of
Nottingham available open access under the following conditions.
This article is made available under the Creative Commons Attribution licence and may be
reused according to the conditions of the licence. For more details see:
http://creativecommons.org/licenses/by/2.5/
A note on versions:
The version presented here may differ from the published version or from the version of
record. If you wish to cite this item you are advised to consult the publisher’s version. Please
see the repository url above for details on accessing the published version and note that
access may require a subscription.
For more information, please contact [email protected]
JournalofPathology
ORIGINAL PAPER
JPathol2016;239:139–151
PublishedonlineinWileyOnlineLibrary
(wileyonlinelibrary.com)DOI:10.1002/path.4715
Altered ratios of pro- and anti-angiogenic VEGF-A variants
and pericyte expression of DLL4 disrupt vascular maturation
in infantile haemangioma
XiYe,1,2 YassirAbou-Rayyah,3 JoyceBischoff,4 AlisonRitchie,2 NeilJSebire,5 PatrickWatts,6 AmandaJChurchill1,*#
and David O Bates2,*#
1 OphthalmologyUnit,SchoolofClinicalSciences,UniversityofBristol,UK
2 CancerBiologyUnit,DivisionofCancerandStemCells,SchoolofMedicine,UniversityofNottingham,UK
3 MoorieldsEyeHospital,London,UK
4 VascularBiologyProgram,BostonChildren’sHospital,HarvardMedicalSchool,MA,USA
5 Histopathology,GreatOrmondStreetHospital,London,UK
6 UniversityHospitalofWales,Cardiff,UK
*Correspondence to: D Bates, Cancer Biology Unit, D Floor, West Block, Queen’s Medical Centre, Nottingham NG7 2UH, UK. E-mail:
[email protected]
OrAChurchillOphthalmologyUnitBristolEyeHospitalLowerMaudlinStreetBristolBS12LX,UK.E-mail:[email protected]
#Jointseniorauthors.
Abstract
Infantilehaemangioma(IH),themostcommonneoplasmininfants,isaslowlyresolvingvasculartumour.Vascular
endothelialgrowthfactorA(VEGF-A),whichconsistsofboththepro-andanti-angiogenicvariants,contributes
tothepathogenesisofIH.However,therolesofdifferentVEGF-AvariantsinIHprogressionanditsspontaneous
involution is unknown. Using patient-derived cells and surgical specimens, we showed that the relative level
of VEGF-A b was increased in the involuting phase of IH and the relative change in VEGF-A isoforms may
165
be dependent on endothelial differentiation of IH stem cells. VEGFR signalling regulated IH cell functions and
VEGF-A b inhibited cell proliferation and the angiogenic potential of IH endothelial cellsinvitroandinvivo.
165
The inhibition of angiogenesis by VEGF-A b was associated with the extent of VEGF receptor 2 (VEGFR2)
165
activationanddegradationandDelta-likeligand4(DLL4)expression.TheseresultsindicatethatVEGF-Avariants
can be regulated by cell differentiation and are involved in IH progression. We also demonstrated that DLL4
expressionwasnotexclusivetotheendotheliuminIHbutwasalsopresentinpericytes,wheretheexpressionof
VEGFR2isabsent,suggestingthatpericyte-derivedDLL4maypreventsproutingduringinvolution,independently
of VEGFR2. Angiogenesis in IH therefore appears to be controlled by DLL4 within the endothelium in a VEGF-A
isoform-dependentmanner,andinperivascularcellsinaVEGF-independentmanner.ThecontributionofVEGF-A
isoformstodiseaseprogressionalsoindicatesthatIHmaybeassociatedwithalteredsplicing.
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtdonbehalfofPathologicalSocietyofGreatBritain
andIreland.
Keywords:haemangioma;vascularendothelialgrowthfactor;angiogenesis;endothelialcell
Received6 July2015;Revised27 February2016;Accepted3 March2016
Conlictsofinterest.DOBisaninventoronthepatentdescribingtheuseandinvestigationsofVEGF-A b.Theotherauthorshavenoconlictsof
165
interest.
Introduction vasculogenesis are VEGF-A-dependent [5] and
HemECs from some patients demonstrate a high
Infantile haemangioma (IH) is the most common neo- level of VEGF receptor 2 (VEGFR2) activation com-
plasm of infancy [1,2]. The natural history of IH is a paredwithnormalendothelialcells[6].Thus,VEGF-A
rapid proliferating phase within the irst 12 months of appearstoplaypivotalrolesinthepathogenesisofIH.
life, followed by spontaneous involution over 7–10 The VEGFA gene contains eight exons. Pre-mRNA
years [2]. Haemangioma stem cells (HemSCs) are alternative splicing controls whether the mRNA con-
believed to contribute to the pathogenesis of IH [3] tains both exon 8a and 8b or exon 8b alone. The
and have been hypothesized to be the parent cells inal translated protein product is dependent on this
of haemangioma endothelial cells (HemECs), which alternative splicing [7,8]. Proximal splicing of exon 8
expand clonally in IH [4]. HemSCs proliferation and leads to the translation of the irst six amino acids
© 2016 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.
ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninany
medium,providedtheoriginalworkisproperlycited.
140 XYeetal
in exon 8a, forming the pro-angiogenic VEGF-A a 100μg/ml streptomycin, 250ng/ml amphotericin B;
xxx
family. Distal splicing removes exon 8a from the tran- Gibco).Thissupplementedgrowthmediumiscomplete
script and results in the translation of the irst six EGM-2. The HemSCs used were from donors 125,
amino acids in exon 8b, forming the anti-angiogenic 129 and 133. HemECs were from donors 153, 158
VEGF-A b. Recently, an additional anti-angiogenic and 159. IH pericytes were cultured on plates coated
xxx
VEGF-Avariant,VEGF-Ax,generatedbyprogrammed with 0.5% gelatin and 0.3μg/cm2 ibronectin in Dul-
translational readthrough (PTR), hasbeen reported [9]. becco’smodiiedEagle’smedium(DMEM)GlutaMax,
TheVEGF-A aandVEGF-A bisoformsinthepro- low-glucose(Gibco),supplementedwith10%FBSand
165 165
andanti-angiogenicfamilieshavesimilarbindingafini- 2mMl-glutamine. Chinese hamster ovary (CHO) cells
ties to VEGFR1 and VEGFR2 [10], but VEGF-A b and normal human dermal ibroblasts (NHDFs) were
165
does not bind to neuropilin-1 (Nrp-1), the co-receptor maintained in RPMI and DMEM media respectively,
for both VEGF receptors [11]. VEGF-Ax appears to supplemented with 10% FBS and 2mMl-glutamine.
∘
havesimilarpropertiestoVEGF-A b,inthatitantag- The cells were maintained with 5% CO at 37 C in a
165 2
onizes VEGF-A a-mediated VEGFR2 activation and humidiiedchamber.
165
alsodoesnotbindtoNrp-1[9].
VEGF-A b, the most investigated isoform of the
165 Proliferatingcellcounts
anti-angiogenic variants, can inhibit tumour growth
Cells were plated at 1×104 cells/cm2 on coverslips in
and vascularization in xenograft models of colorectal,
24-wellplates.Thecellswereallowedtoattachfor24h
prostate, melanoma and other cancers [12]. The roles
andtreatedfor36hincompleteEGM-2.Thecellswere
of the pro- and anti-angiogenic VEGF-A variants have
counted, differentiated or subjected to WST1 assay or
not been studied in IH. Due to the close association
ixed with 4% paraformaldehyde, permeabilized with
of VEGF-A with IH, we set out to determine whether
PBS-Triton X (0.1%) and stained with Ki67 antibody
VEGF-A isoforms contribute to the natural progres-
(Abcamab16667)andDAPI.Thetotalnumberofcells
sion of IH and some of the associated mechanisms
andthenumberofcellspositiveforKi67inthenucleus
underlying involution, and to test the hypothesis that
werecountediniveieldsofview.
anti-angiogenicVEGF-AvariantscouldcontributetoIH
involution.
Totalcellcount
Cells were plated at 1×104 cells/cm2 on coverslips in
Materials and methods 24-well plates. Cells were allowed to attach overnight
under normal growth conditions before administration
of treatments. One well was removed and its cells
Humantissuesandimmunohistochemistry
ixed in 4% paraformaldehyde and stained with DAPI
Formalin-ixedandparafin-embeddedhumanIHspec-
nuclear stain every 24h, and cell number was counted
imens were biopsied or excised as part of clinical
microscopicallyiniverandomieldsofview/coverslip.
care (and conirmed diagnostically with histological
Cell numbers were normalized to the number of cells
morphology and characteristic GLUT-1 expression).
attachedafter24h.
Humantissueswerecollectedwithethicalapproval(No.
07/H0102/45).ThestageofIHwasdividedintoprolifer-
WST-1proliferationassay
atingorinvolutingphases.IHexcisionsfromchildrenup
toandincludingage1yearwereclassiiedasproliferat- Cellswereplatedat3×103cells/wellofasix-wellplate.
ing,whilstexcisionsfromchildren3yearsorolderwere The cells were left to attach in complete EGM-2 for
classiiedasinvoluting.Ifpatientsampleswerederived 5h.Themediumwasreplacedwithtreatmentcontaining
fromchildrenaged1–3years,thehistologyofthespec- EGM-2andgrownfor36h.Mediaandtreatmentswere
imenswasexaminedandthestageofIHwasconirmed removed and replaced with EBM-2 supplemented with
bypathologists. 1%FBSand10%WST-1reagent(RocheAppliedBio-
science)andincubatedundernormalcultureconditions
for2h.
Cellculture
IHstemcells(HemSCs),IHendothelialcells(HemECs)
Endothelialdifferentiationprotocol
and IH pericytes were isolated as described [3,4,13]
underthehumansubjectprotocolapprovedbytheCom- HemSCs were plated on ibronectin-coated plates at
mittee on Clinical Investigation at Boston Children’s 2×104 cells/cm2 in complete EGM-2 for 12h. Hem-
Hospital. Informed consent was obtained according to SCs were then cultured in endothelial differentiating
theDeclarationofHelsinki.HemSCsandHemECswere medium [14] consisting of serum-free EBM-2, 1×
grown on ibronectin (0.3μg/cm2; Millipore)-coated insulin–transferrin–selenium (Gibco), linoleic acid
plates in endothelial basal medium 2 (EBM-2; Lonza), at 1:5000 dilution, 60μm ascorbic acid-2-phosphate,
with 20% foetal bovine serum (FBS), EBM-2 Single- 1μm dexamethasone (Sigma-Aldrich) and VEGF-B at
Quotwithouthydrocortisone(Lonza),2mMl-glutamine 10ng/ml (PeproTech) for 14 days. Fresh endothelial
and1× antibiotic–antimycotic (100 U/ml penicillin, differentiatingmediumwasaddedevery2days.
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
VEGF-AvariantsandDLL4ininfantilehaemangioma 141
Adenovirus products were run on an ethidium bromide agarose gel
andimagedusingaGeldocEZsystem(Bio-Rad).
Chinese hamster ovary (CHO) cells were infected
Quantitative PCR was carried out using lightcycler
with recombinant, replication-deicient adenoviruses
480 SYBR Green master (Roche Applied Science)
expressing GFP or sDLL4 under a CMV promoter
and 0.25μm gene speciic primers, using a Lightcycler
(VectorBiolabs)[15]at100MOI.
480 real-time PCR thermal cycler (Roche Applied
Science). All reactions were carried out using the
Proteinexpression
protocol speciied for PCR for 45 cycles; CT val-
Immunoluorescence ues<35 were deemed acceptable. Human DLL4
primers were purchased from Qiagen (QT00081004);
Parafin-embedded sections were dewaxed and rehy- other primer sequences are listed in Table S1 (see
drated. Antigen retrieval was carried out in citrate supplementary material). Relative mRNA levels were
buffer(Sigma).Primaryantibodies(PanVEGF-A,Santa calculated, normalized to B -microglobulin (B M)
2 2
CruzSC-152;VEGF-Axxxb,R&DSystemsMAB1045; mRNA.
VEGF-A a [16]; CD31, Abcam ab28364; DLL4,
XXX
Abcam ab183532; human CD31, Cell Signaling 3528;
Invitroco-cultureangiogenesisassay
and NG2, Abcam ab83178) were incubated overnight
∘
at 4 C. Appropriate AlexaFluor secondary antibodies NHDFs were grown on 19mm glass coverslips until
(MolecularProbes)wereincubatedfor1hatroomtem- 70–80% conluent. HemECs (3×104) were plated on
perature. Five ields of view were recorded from each theNHDFmonolayerandallowedtosettlefor6h.The
patient’ssection.Allanalyseswerecarriedoutmasked, mediumwasreplacedwithtreatmentcontainingEBM-2
ie patient groups were revealed only after the analy- supplemented with 4% FBS, and the treatment was
ses have been completed. Immunoluorescence images refreshedevery2daysfor10days.Thecellswereixed
wereacquiredusingaLeicaDMRBluorescencemicro- with70%ethanolandstainedforVE-Cadherin(Abcam
scope itted with an Olympus DP72 camera, using ab33168) and the nuclei stained with DAPI. Images
OlympusCellSenssoftware.Forco-localization,images wereacquiredfromiveieldsofviewat×10objective
weretakenusingaLeicaSP8confocalmicroscopeand magniication. Area of VE-Cadherin staining, tubule
LEICALASsoftware. lengthandnumberofbranchpointsweremeasuredand
normalizedtotheareaoftheield.
Immunoblotting
Murinehaemangiomamodel
Cells were plated in six-well plates and grown until
70–75% conluent. The cells were serum-starved with Themurinemodelwasproducedasdescribed[5],using
EBM-2 medium for 12–16h prior to VEGF-A treat- 2×106 HemSCs and 2×106 HemECs resuspended in
ments,thenlysedinlysisbuffer(200mMNaCl,75mM 200μl Matrigel (Corning) and injected subcutaneously
Trizmabase,1mMEDTA,1.5%TritonX,0.75%Np-40, into the left lank of CD1 nude mice (see supplemen-
15mM NaF and 1.5mM Na VO ; Sigma-Aldrich) on tary material, Supplementary materials and methods,
3 4
icefor5minandsubjectedtostandardimmunoblotting for details of mouse housing conditions). The mice
(antibodydetailsandblottingconditionsaredescribedin were also weighed weekly and checked daily by an
Supplementary materials and methods, see supplemen- experiencedtechnician.Implantsweremeasuredweekly
tarymaterial). using Vernier callipers and the volume (V) was cal-
culated using the formula: V=ab2π/6, where a is the
longest diameter and b is the longest diameter perpen-
VEGF-AELISA
diculartoa.Attermination,theexplantswereremoved
VEGF-AandVEGF-A bproteinlevelswereassessed and ixed with neutral-buffered formalin and embed-
165
usinghumanVEGF-AandVEGF-A bDuoSet(R&D dedinparafinorcryopreservedinOCT.Three5–7μm
165
Systems),followingthemanufacturer’sinstructions. non-consecutive sections positioned 20–30μm apart
were stained with haematoxylin and eosin (H&E), or
human-speciicCD31formicrovesseldensitycounting,
Polymerasechainreaction(PCR)
or with oil red O for fat deposition scoring. The analy-
RNA was extracted with TRI-reagent (Invitro- seswereconductedwithoutknowledgeofthetreatment
gen); all materials, except primers, were purchased groups.
from Promega; 1μg RNA was DNase-treated and
reverse-transcribedwithM-MLVreversetranscriptase.
Statisticalanalyses
PCR was carried out using a C1000 thermal cycler
(Bio-Rad).Reactionswerecarriedoutusingthefollow- Data are presented as mean±standard error of the
∘
ingsteps:95 Cfor10min(hotstartinitiation),followed mean (SE). Statistical analyses were carried out using
∘
by repeating 95 C for 30s, annealing temperature for two-tailedStudent’st-testorANOVAwithTukey’spost
∘
30sand72 Cfor30sforupto35cycles.Ainalexten- hoc test where appropriate, unless stated otherwise;
sionof72∘Cfor10minwasthencarriedout.Ampliied p<0.05wasconsideredsigniicant.
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
142 XYeetal
Results HemSCs by treatment for 2 weeks with a different
member of the VEGF family, VEGF-B, a protocol
previously shown to induce an endothelial phenotype
Increaseintheproportionofanti-angiogenic
viaVEGFR1activation[14].Weconirmedendothelial
VEGF-AisoformscoincideswithIHinvolution
differentiation by morphological changes, reduction of
Immunoluorescence staining of human IH sections the mesenchymal marker CD90 and also acquisition of
for total VEGF-A showed no differences in VEGF-A theendothelialmarkerVE-cadherin(seesupplementary
between the proliferating and the involuting phase of material, Figure S5). The differentiated cells were
IH(Figure1A).Incontrast,stainingwithVEGF-A a- then returned to complete EGM-2 medium for 72h
xxx
and VEGF-A b-speciic antibodies (see supplemen- and we found that the differentiated cells produced
xxx
tary material, Figure S1) [8,16] showed a signiicant a lower level of VEGF-A using ELISA (Figure 2E;
reductioninthelevelofthepro-angiogenicVEGF-A a p<0.01) and western blotting (Figure 2G). Further-
xxx
isoformduringinvolution(Figure1B)andanincreasein more,wealsofoundthatthelevelofVEGF-A bwas
165
the level of the anti-angiogenic VEGF-A b isoforms signiicantly increased compared with undifferentiated
xxx
(Figure 1C). These data suggest that there is an alter- cells, using ELISA (Figure 2F; p<0.05) and western
ation in the ratio of VEGF-A isoforms, with a relative blotting (Figure 2G), suggesting that the differenti-
increase in the proportion of anti-angiogenic variants ation process may be at least partially responsible
duringinvolution. for the differential VEGF-A isoform expression. An
additional band at a slightly higher molecular weight
VEGFRfacilitatesIHcellproliferationandchanges than that of VEGF-A a or VEGF-A b was consis-
165 165
inVEGF-AisoformsinIHareassociated tently increased following differentiation (Figure 2G).
This band was detected using both total VEGF-A
withdifferentiationofHemSCsintoanendothelial
and VEGF-A b antibodies and may have been the
lineage 165
anti-angiogenic VEGF-Ax protein recently described
We questioned whether VEGFR signalling is involved [9].
inIHcellfunctions.WeusedacommonlyusedVEGFR VEGF-B acts by signalling through VEGFR1. To
inhibitor, PTK787 (see supplementary material, Figure determine whether acute VEGFR1 activation could
S2), and identiied that proliferation of both HemSCs induce VEGF-A b expression directly, and hence
165
andHemECswassigniicantly inhibitedbyPTK787at up-regulate VEGF-A b prior to differentiation, we
165
concentrations speciic to VEGFRs (Figure 1D), sup- stimulatedHemSCswithVEGF-Bfor36h(insuficient
porting the notion that active VEGF-A signalling is to result in differentiation). No signiicant changes
importantforIH. in VEGF-A, VEGF-A b, SRSF2 expression (see
165
CD133 was used to isolate multipotent HemSCs supplementary material, Figure S6A, B) or SRSF6
from proliferating IH [3,5], which are lost during IH phosphorylation were observed (see supplementary
involution (see supplementary material, Figure S3A), material, Figure S6C, D). SRSF2 and −6 are other
consistent with previous observations [17]. HemSCs splice factors involved in VEGF-A splicing [18,19].
express higher levels of VEGF-A compared with Only the phosphorylation of SRSF1, which has been
NHDFs, HUVECs, HemEPCs [5] and, in this case, associatedwithproximalsplicingofVEGF-Aincancer
also the HemECs (Figure 2A, D). Using primers span- cells and podocytes, was increased by VEGF-B over
ning exons 7 and 8b, which distinguish the pro- and this time period (see supplementary material, Figure
anti-angiogenic VEGF-A variants (Figure 2B), we S6C,D).
found that the relative VEGF-A b:VEGF-A a ratio
165 165
was higher in HemECs when compared with HemSCs,
VEGF-A binhibitsVEGF-A amediatedHemEC
at both the cDNA and protein levels (Figure 2C, D). 165 165
angiogenesisandIHcellproliferation
VEGF-A b was detected in HemSC protein lysates
165
by ELISA, but not in the mRNA by RT–PCR. This WefoundVEGF-A bwasmorelocalizedtothestruc-
xxx
may be explained by the competitive ampliication tured endothelia with deined lumen structure, a typ-
between VEGF-A a and VEGF-A b. Using recom- ical feature of late proliferating and involuting phase
165 165
binant VEGF-A cDNAs, we found that when the (Figure 3Aiv–ix) compared with the disorganized vas-
VEGF-A a variant concentration was>three-fold culature present in the early stages of the proliferating
165
higher than VEGF-A b, VEGF-A b detection was phase (Figure 3Ai–iii); thus, the relative VEGF-A iso-
165 165
completely lost by PCR (see supplementary material, formratiosanddistributionmayhaveadirectfunctional
Figure S4). This may also explain the inconsistencies roleinregulatingIHangiogenesis/vasculogenesis.
in VEGF-A b detection using PCR overall, as its We used VEGF-A a as a positive control and
165 165
ampliication is heavily inluenced by the level of induced angiogenesis of HemECs in vitro, using
VEGF-A a. a co-culture angiogenesis assay. VEGF-A b did
165 165
FollowingobservationsthatHemECsproducehigher not induce HemECs angiogenesis and inhibited
levels of VEGF-A b compared with HemSCs, we VEGF-A a-mediated tube formation, extension and
165 165
hypothesized that the changes in VEGF-A isoforms branching (Figure 3B–E). A previous study showed
may be differentiation-dependent. We differentiated thatVEGF-AknockdowninhibitsHemSCsproliferation
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
VEGF-AvariantsandDLL4ininfantilehaemangioma 143
Figure1.ChangesinVEGF-AisoformlevelscoincidewithIHinvolution.(A)Tissuefromproliferatingphase(n=8patients)andinvoluting
phase(n=11patients)ofIHwerestainedwithatotalVEGF-AantibodyandshowednodifferencesintotalVEGF-Abetweentheproliferating
andinvolutingphasesofIH.(B)TheareastainedbyaVEGF-A aantibodyandnormalizedtocellnumbershowedasigniicantreductionof
xxx
VEGF-A aintheinvoluting(n=10)phaseofIHcomparedwiththeproliferating(n=5)phase(p<0.05,two-tailedt-test).(C)VEGF-A b
xxx xxx
wasincreasedintheinvolutingphasecomparedwiththeproliferatingphaseofhaemangioma(p<0.05,two-tailedt-test).(D)HemSCsand
HemECsweretreatedwithincreasingconcentrationsofPTK787,aVEGFRinhibitor,whichsigniicantlyinhibitedproliferationofHemSCsand
HemECsinaconcentration-dependentmanner(p<0.001,one-wayANOVA,Bonferroniposthoctest;n=5).Scalebar=50μm;*p<0.05,
***p<0.001;NS,notsigniicant
and the vascularization of IH implants in vivo [5]. We murinemodelhasbeendescribedtoacquireadipocytes
also obtained a similar result by treating the HemSCs over time [3], a characteristic feature of human IH
and cell implants with bevacizumab (see supplemen- pathology. The implants treated with recombinant
tary material, Figure S7A), suggesting the importance human VEGF-A b (rhVEGF-A b) also showed an
165 165
of VEGF-A signalling in the growth and angiogene- increase in fat deposition by oil red O staining (see
sis of this IH model. The effect of VEGF-A b on supplementarymaterial,FigureS7B).
165
angiogenesis was also tested using this in vivo model, We identiied that VEGFR inhibition using PTK787
and rhVEGF-A b reduced the microvessel density [20] reduced proliferation of both HemECs and Hem-
165
in the cell–Matrigel implants (Figure 4A, B). This SCs (Figure 1D). Therefore, we investigated whether
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
144 XYeetal
Figure2.HemECsandendothelialdifferentiationofHemSCsareassociatedwithVEGF-A bup-regulation.(A)DeterminationofmRNA
165
levels of VEGF-A in HemECs and HemSCs by RT–qPCR (n=3 patients). (B) Diagram of RT–PCR primers spanning exons 7 and 8b, thus
amplifying both VEGF-A a and VEGF-A b. (C) RT–PCR for HemECs and HemSCs and plasmids containing recombinant cDNA for
165 165
VEGF-A a and VEGF-A b, or a mixture of equal amounts of both (n=3 patients): MWM, molecular weight marker; −RT, no reverse
165 165
transcriptase.(D)AssessmentoftotalVEGF-AandVEGF-A bproteininlysatefromHemECsandHemSCs,usingsandwichELISA(n=3
165
patients). (E) Measurement of total VEGF-A in differentiated and undifferentiated HemSCs by ELISA. (F) Measurement of VEGF-A b
165
in differentiated and undifferentiated HemSCs by ELISA. (G) Immunoblots of protein extracted from undifferentiated or differentiated
HemSCs,showingreducedintensityatbandsizecorrespondingtobothVEGF-A aandVEGF-A b,usingatotalVEGF-Aantibody(arrow)
165 165
(p<0.05;n=3),andincreasedintensityofbandcorrespondingtoVEGF-A bmolecularweight,usingtheVEGF-A bantibody(p<0.05;
165 165
n=3)(arrow):anadditionalband(#)atahighermolecularweightthanVEGF-A aorVEGF-A bwasconsistentlydetectedfollowing
165 165
differentiationofHemSCs;allanalysedusingtwo-tailedt-test
endogenousVEGF-A bhasaneffectonIHcellprolif- act in an autocrine manner to limit the proliferation of
165
eration. Inhibition of VEGF-A b was achieved using HemECs, which produce a relatively high proportion
165
a previously characterized antibody [8]. VEGF-A b of anti-angiogenic VEGF-A (Figure 2C, D). HemSCs
165
neutralizationincreasedthegrowthofHemECs,butnot produce a lower proportion of VEGF-A b compared
165
HemSCs, over time compared with non-speciic IgG with HemECs (Figure 2C, D), and supplementing
(Figure 4C, D). This suggests that VEGF-A b may the growth media with rhVEGF-A b, signiicantly
165 165
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
VEGF-AvariantsandDLL4ininfantilehaemangioma 145
Figure3.VEGF-A b regulates sprouting angiogenesis in IH. (A) VEGF-A b, CD31 and nuclear (DAPI) staining of human IH sections:
165 xxx-
histologicalfeaturesofproliferatingIHmovefromearly,barelydiscernableandmostlylumen-lessmicrovessels(i–iii)tomorestructured
vesselslateintheproliferatingphase/earlyinvolutingphase(iv–vi)andinallymicrovesselswithdeinedenlargedluminaintheinvoluting
phase(vii–ix);VEGF-A bstainingbecomesmorelocalizedtothemicrovesselsastheendotheliamatureandorganizeintodeinedluminal
xxx
structures.(B)HemECsweretreatedwithVEGF-A a,VEGF-A borbothonamonolayerofNHDFs,usinganinvitroangiogenesisassay:
165 165
HemECswerestainedwithVE-cadherin.(C–E)Quantiicationof(B):2.5nMVEGF-A ainducedtubearea(C),length(D)andbranching(E),
165
comparedwiththeuntreated,2.5nMVEGF-A borHemECstreatedwith2.5nMVEGF-A aandVEGF-A b(n=3;p<0.001,one-way
165 165 165
ANOVA).Scalebars=100μm
inhibited their proliferation (measured using Ki67 as Pro-andanti-angiogenicVEGF-Aisoforms
a proliferation marker) (Figure 4E, F). The inhibitory differentiallyregulateVEGFR2activation,
effect of both VEGF-A b and bevacizumab on Hem- degradationanddownstreamsignalling
165
SCs,butnotHemECs,wasfurtherconirmedusingthe
VEGF-A mediates the proliferation and angiogene-
WST-1proliferationassay(seesupplementarymaterial,
sis of endothelial cells via VEGFR2[21]. Whilst the
Figure S8A, B). This suggests that VEGF-A b from
165 signalling of VEGFR2 stimulated by VEGF-A a
165
othersources,suchasthoseproducedbyHemECs,may and VEGF-A b has been investigated in human
165
inhibit HemSCs proliferation in a paracrine manner, microvascular endothelial cells (HMVECs), it has not
displaying an intricate system of proliferation and been conducted in HemECs, an endothelial cell with
angiogeniccontrolbetweenthetwodifferentcelltypes. progenitor-likegeneexpressionandproperties[22],nor
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
146 XYeetal
Figure4.VEGF-A binhibitsangiogenesisinamurinemodelofIHandinhibitsIHcellproliferationinvitro.(A)HemEC–HemSC–Matrigel
165
implants treated with 15μgs.c. rhVEGF-A b (n=5) or saline (n=8) twice weekly for 2 weeks. The implants were sectioned, stained
165
with H&E and the blood-illed lumina were counted (arrow). Implants treated with rhVEGF-A b showed lower microvessel density
165
thansaline-treatedones(n=8;p<0.01).(B)ImplantstreatedwithrhVEGF-A b(n=6)formedfewerhumanCD31+ microvesselsthan
165
saline-treatedones(n=8;p<0.001,two-tailedStudent’st-test).(C,D)HemSCs(C)andHemECs(D)weretreatedwith1μg/mlVEGF-A b
165
neutralizingantibodyfor24–96hincompleteEGM-2:cellnumbersat48,72and96hwerenormalizedtothenumberofcellsat24h
(n=3;p<0.05,two-wayANOVA,BonferroniposthoctestcomparedwithIgG).(E)HemECsweretreatedincompleteEGM-2withspeciied
concentrationsofrhVEGF-A bfor48h;cellswerestainedforDAPIandKi67.(F)Quantiicationof(E):thepercentageofproliferating
165
HemECswasunaffectedby5or10nMrhVEGF-A b;thepercentageofproliferatingHemSCswasreducedwhencellsweretreatedwith
165
10nMVEGF-A bcomparedwithuntreatedones.Scalebars=50μm;*p<0.05,**p<0.01,***p<0.01;NS,notsigniicant
165
inHemSCs.WefoundthatbothHemSCsandHemECs VEGF-A b induced a reduction of VEGFR2 over
165
respond, in a similar manner to the VEGF-A isoforms, time. However, the reduction caused by VEGF-A b
165
as normal endothelial cells, wherein VEGF-A a stimulation was more prominent than that induced
165
activates VEGFR2 as well as downstream signalling, byVEGF-A awithprolongedstimulations(Figure5C,
165
whichisimportantforIHcellproliferation,andthatthis D), suggesting that VEGF-A b may preferentially
165
couldbeblockedwithVEGF-A b(Figure5A,B;see direct VEGFR2 for degradation, as opposed to recy-
165
also supplementary material, Figure S9). Others have cling.
demonstratedthatthelackofNrp-1signallingpromotes The Notch–Delta signalling pathway is one of the
VEGFR2 degradation [23], and that VEGF-A b, most studied pathways in sprouting angiogenesis.
165
which does not bind to Nrp-1 [11], promotes Whilst Delta-like ligand 4 (DLL4) expression can be
VEGFR2 degradation and/or prevents recycling of the induced by VEGF-A a in normal endothelial cells
165
receptor. In HemECs, both VEGF-A a and [24] and in HemECs (Figure 5E, F), VEGF-A b did
165 165
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com
VEGF-AvariantsandDLL4ininfantilehaemangioma 147
Figure5.VEGFR2signalling,degradationanddownstreamtargetsaredifferentiallyregulatedbyVEGF-Aisoforms.Cellswereserum-starved
priortotreatmentwith2.5nMVEGF-A a,VEGF-A borboth.(A,B)VEGF-A ainducedVEGFR2signallinganddownstreamactivation
165 165 165
ofERK1/2inbothHemSCsandHemECs;VEGF-A bweaklyactivatedVEGFR2andERK1/2inHemSCsandHemECs;VEGF-A binhibited
165 165
VEGF-A a-mediatedVEGFR2andERK1/2phosphorylationinbothHemSCsandHemECs.(C)HemECsweretreatedwith2.5nMVEGF-A a
165 165
andVEGF-A bandproteinextractedatthetimepointsshown.(D)Quantiicationof(C):treatmentwithVEGF-A b,butnotVEGF-A a,
165 165 165
signiicantly reduced VEGFR2 levels; n=3; p<0.05, one-way ANOVA; *p<0.05 compared with VEGF-A b; #p<0.05 compared with
165
time 0. (E) HemECs were treated as above and RNA extracted and subjected to RT–qPCR for DLL4 expression; p<0.001, ANOVA. (F)
VEGF-A a-induced DLL4 expression compared with untreated (n=3; p<0.05, one-way ANOVA); VEGF-A b, alone or together with
165 165
VEGF-A a,didnotinduceDLL4expressionattheproteinlevel;*p<0.05,**p<0.01,***p<0.01;NS,notsigniicant
165
not induce DLL4 expression to the same extent and material, Figure S10), using an adenovirus vector, and
was able to ameliorate VEGF-A a-mediated DLL4 used the conditioned medium of the infected cells to
165
up-regulation in HemECs (Figure 5E, F). We tested treat the HemECs in the in vitro angiogenesis assay.
whetheractivationofNotchsignallingusingthesoluble sDLL4 blocked the angiogenic tube formation, exten-
portion of DLL4 (sDLL4), previously described [25], sion and branching ability of HemECs under all of the
can affect the angiogenic potential of HemECs. We conditions tested, including those that were stimulated
overexpressedsDLL4inCHOcells(seesupplementary withVEGF-A a(Figure6A–D).
165
©2016TheAuthors.TheJournalofPathologypublishedbyJohnWiley&SonsLtd JPathol2016;239:139–151
onbehalfofPathologicalSocietyofGreatBritainandIreland.www.pathsoc.org www.thejournalofpathology.com