Al ex 32(2), 2015 125
Recei ed No embe 27, 2014;
Accep ed Feb ua y 25, 2015;
epub Ma ch 2, 2015;
h p://dx.doi.o g/10.14573/al ex.1411271
Resea ch A icle
Human Vascula Model wi h De ined
S imula ion Medium – A Cha ac e iza ion
S udy
Ou i Hu ala *1, Hanna Vuo enpää *1, Ta ja Toimela 1
, Jukka Uo ila 2
, Hannu Kuokkanen 3
,
Timo Ylikomi 1,4, Je a-Riina Sa kanen 1,4 and Tuula Heinonen 1
1FICAM, Finnish Cen e o Al e na i e Me hods, School o Medicine, Uni e si y o Tampe e, Tampe e, Finland; 2Depa men
o Obs e ics and Gynecology, Tampe e Uni e si y Hospi al, Tampe e, Finland; 3Depa men o Plas ic Su ge y, Tampe e Uni e si y
Hospi al, Tampe e, Finland; 4Depa men o Cell Biology, School o Medicine, Uni e si y o Tampe e, Tampe e, Finland
Summa y
The o ma ion o blood essels is a i al p ocess in emb yonic de elopmen and in no mal physiology. Cu en ascula
modelling is mainly based on animal biology leading o species- o-species a ia ion when ex apola ing he esul s
o humans. Al hough he e a e a ew human cell based ascula models a ailable, hese assays a e insu icien ly
cha ac e ized in e ms o cul u e condi ions and de elopmen al s age o ascula s uc u es. The e o e, well cha ac e ized
ascula models wi h human ele ance a e needed o basic esea ch, emb yo oxici y es ing, de elopmen o he apeu ic
s a egies and o issue enginee ing.
We ha e p e iously shown ha he in i o ascula model based on co-cul u e o human adipose s omal cells (hASC)
and human umbilical ein endo helial cells (HUVEC) is able o induce an ex ensi e ascula -like ne wo k wi h high
ep oducibili y. In his wo k we de eloped a de ined se um- ee ascula s imula ion medium (VSM) and pe o med
u he cha ac e iza ion in e ms o cell iden i y, ma u a ion and s uc u e o ob ain a ho oughly cha ac e ized in i o
ascula model o eplace o educe co esponding animal expe imen s.
The esul s showed ha he no el ascula s imula ion medium induced an in ac and e enly dis ibu ed ascula -like
ne wo k wi h mo phology o ma u e essels. Elec on mic oscopic analysis assu ed he h ee-dimensional mic os uc u e
o he ne wo k con aining lumen. Addi ionally, ele a ed exp ession le els o he main human angiogenesis- ela ed genes
we e de ec ed.
In conclusion, wi h he newly de ined medium he ascula model can be u ilized as a cha ac e ized es sys em o
chemical es ing as well as in c ea ing ascula ized issue models.
Keywo ds: se um- ee media, angiogenesis, mesenchymal s omal cells, cocul u e echniques
his is an Open Access a icle dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion 4.0 In e na ional license (h p://c ea i ecommons.o g/
licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion and ep oduc ion
in any medium, p o ided he o iginal wo k is app op ia ely ci ed.
1 In oduc ion#
The o ma ion o he blood essel ne wo k is a i al p ocess in
g ow h and o gan de elopmen (Ca melie and Jain, 2011; Ca -
melie , 2005). In he emb yo, endo helial p ecu so cells o m
new essels ha di e en ia e in o a p imi i e ascula ne wo k
( asculogenesis) (Ca melie and Jain, 2011). Subsequen es-
sel sp ou ing (angiogenesis) c ea es a ne wo k o a e ies and
eins as well as capilla ies ha acili a e he exchange o gases
and me aboli es (Ca melie and Jain, 2011; Adams and Ali alo,
2007). To each his le el o complex o ganiza ion, he imma-
u e ascula ne wo k mus ma u e a he le el o he essel wall
*Au ho s con ibu ed equally o his wo k.
#Abb e ia ions
AA, asco bic acid; APC, allophycocyanin; BSA, bo ine se um albumin; CD144, ascula endo helial cadhe in; eGF, epide mal g ow h ac o ; eGM-2, endo helial
cell g ow h medium-2; FBS, e al bo ine se um; FI C, luo escein iso hiocyana e; FGF-2, basic ib oblas g ow h ac o ; hASC, human adipose s omal cells;
He, hepa in sodium sal ; HS, human se um; HUVeC, human umbilical ein endo helial cells; HY, hyd oco isone (co isol); IGF-I, insulin-like g ow h ac o I;
PBS, phospha e bu e ed saline; RT, oom empe a u e; PDGFRβ, pla ele de i ed g ow h ac o be a; PE, phycoe y h in; PE-CY7, phycoe y h in-cyanine; SFM,
basal se um- ee medium; αSMA, alpha smoo h muscle ac in; TRITC, e ame hyl hodamine iso hiocyana e; VEGF, ascula endo helial g ow h ac o ; VSM,
ascula s imula ion medium; W , on Willeb and ac o
Hu ala e a l .
Al ex 32(2), 2015
126
mo phology and as a whole ne wo k. Mo phological ma u a ion
in ol es ec ui men o mu al cells, deposi ion o ex acellula
ma ix and o gan-speci ic specializa ion o cells, such as in e -
endo helial junc ions and su ace ecep o s. Ma u a ion o he
ne wo k in ol es b anching and expanding he ne wo k o mee
local, issue-speci ic demands (Jain, 2003).
Since he o ma ion and main enance o he ascula ne wo k
is a complex p ocess, p oblems ela ed o i s egula ion a e com-
mon (Ucuzian and G eisle , 2007). Inadequa e essel main e-
nance o g ow h causes ischemia in myoca dial in a c ion and
neu odegene a i e o obesi y-associa ed diso de s, whe eas
excessi e ascula g ow h o abno mal emodeling p omo es
cance , in lamma o y diso de s and eye diseases (Po en e e
al., 2011). Mo eo e , gene ic s udies ha e shown ha pe u b-
ing emb yonic ascula de elopmen can ha e ad e se conse-
quences om benign ascula mal o ma ion o emb yole hali y
and congeni al de ec s (Knudsen and Kleins eue , 2011).
Angiogenesis models a e impo an ools o s udying he
mechanisms o angiogenesis and he he apeu ic s a egies o
modula e neo ascula iza ion (Ucuzian and G eisle , 2007). Due
o an inc easing amoun o compounds a ec ing he ascula
sys em, accu a e asculogenesis and angiogenesis models a e
needed o chemical sa e y es ing and o d ug de elopmen
(Sa kanen e al., 2011; Bishop e al., 1999). Cu en ly, p eclini-
cal animal models a e dominan ly used o angiogenesis es ing
al hough hey a e no conside ed op imal in e icacy o ele ance
o humans. The mos commonly used in i o angiogenesis as-
says include he chick cho ioallan oic memb ane (CAM) assay,
Ma igel plug assay, zeb a ish emb yo sys em, co neal mic opo-
cke assay, a /mouse hind limb ischemia model and a ao ic ing
assay (No by, 2006; Aue bach e al., 2003). Despi e he ad an-
age o p o iding mo e in o ma ion on complex cellula in e ac-
ions compa ed o in i o models, animal models a e bu dened
by se e al disad an ages, such as a iabili y, animal-speci ici y
and e hical conce ns (No by, 2006). Human cell based models
ha e he po en ial o be aluable ools in p edic ing e ec s in
man. Howe e , he human ele ance o hese in i o models
needs o be con i med in e ms o cell iden i y, physiological a -
chi ec u e and unc ionali y (Bale e al., 2014; Ha ung, 2011).
In addi ion, he de elopmen al s age o he model sys em and a
de ined medium composi ion a e c i ical, especially when oxi-
cological applica ions a e conside ed. Se um- ee medium wi h
xeno- ee and de ined supplemen s is conside ed essen ial o in
i o models o dec ease he a ia ion be ween expe imen s due
o unknown componen s in he medium, and u he , unknown
binding p ope ies o hese componen s (Shen e al., 2013; B un-
ne e al., 2010; an de Valk e al., 2010). Cu en ly, cul u e
media a e s ill commonly supplemen ed wi h se um, al hough
i has a highly uncha ac e ized composi ion, including a ious
cy okines and g ow h ac o s, as well as a lo - o-lo a iabili y
(Lind oos e al., 2011; B unne e al., 2010).
We and o he s ha e shown ha adipose s omal cells and um-
bilical ein endo helial cells a e capable o sel -assembling in o a
dense, h ee-dimensional ascula -like ne wo k (Sa kanen e al.,
2012; Me eld-Clauss e al., 2010; Ve seijden e al., 2010). While
adipose s omal cells sec e e ac o s ha induce endo helial cell
(EC) sp ou ing and lumen o ma ion (Rubina e al., 2009; T ak-
ue e al., 2008; Rehman e al., 2004; Kil oy e al., 2007; Bishop
e al., 1999), he suppo ing s omal cells also enhance ascula
basemen memb ane and lumen o ma ion (Me eld-Clauss e
al., 2010; Newman e al., 2013; S a man e al., 2009).
The aim o his s udy was o de elop a de ined medium and
u he cha ac e ize he in i o ascula model de eloped by us,
which is composed o human adipose s omal cells (hASC) and
human umbilical ein endo helial cells (HUVEC) (Sa kanen e
al., 2012). Ou esul s showed ha he new ascula s imula-
ion media (VSM) de eloped in his s udy p oduces an ex en-
si e ascula -like ne wo k wi h ma u e p ope ies and p o ides
a alid al e na i e o comme cial EGM-2 medium when com-
pa ing ascula -like ne wo k o ma ion capaci y. This ascula
model has he po en ial o be used in he sa e y and e icacy
assessmen o angiogenic compounds. In addi ion, he ascula -
like ne wo k combined wi h a ge cells, such as ca diomyocy es
(Vuo enpää e al., 2014), can be used as a issue enginee ing
pla o m o c ea e ascula ized issue models.
2 Ma e ials and me hods
This s udy con o ms o he p inciples ou lined in he Decla a ion
o Helsinki. The human adipose issue samples we e ob ained
om su gical ope a ions and human umbilical co ds we e e-
cei ed om caesa ean sec ions wi h indi idual w i en in o med
consen a Tampe e Uni e si y Hospi al, Tampe e, Finland. The
use o hASC and HUVEC we e app o ed by he E hics Com-
mi ee o he Pi kanmaa Hospi al Dis ic , Tampe e, Finland
wi h pe mi numbe s R03058 and R08028, espec i ely.
Isola ion and cul u e o human adipose s omal cells
hASC we e isola ed om human adipose issue by using a me-
chanical and enzyma ic p ocedu e desc ibed p e iously (Sa -
kanen e al., 2012). B ie ly, human adipose issue specimens
we e mechanically cu in o small pieces and enzyma ically di-
ges ed wi h 0.15% collagenase I (In i ogen, Paisley, Sco land,
UK) in Dulbecco’s Modi ied Eagle’s Medium Nu ien Mix u e
F-12 (DMEM/F12, Gibco, In i ogen, Ca lsbad, CA, USA).
hASC we e cul u ed in DMEM/F12 supplemen ed wi h 10%
human se um (HS, Lonza G oup L d, Basel, Swi ze land) and
1% L-Glu amine (Gibco). The cells we e es ed o mycoplasma
con amina ion (MycoAle ® Mycoplasma De ec ion Ki , Lonza
G oup L d) be o e expe imen al use.
Isola ion and cul u e o human umbilical ein endo helial cells
HUVEC we e isola ed om human umbilical co d eins using
0.05% collagenase I as desc ibed p e iously (Sa kanen e al.,
2011). The cells we e cul u ed in EGM™-2 Endo helial Cell
G ow h Medium-2 (EGM-2, Lonza G oup L d). Be o e use he
cells we e es ed o mycoplasma con amina ion (MycoAle ®
Mycoplasma De ec ion Ki , Lonza).
Es ablishmen o he co-cul u e o ming he ascula -like
ne wo k
Co-cul u e o hASC and HUVEC was es ablished as desc ibed
p e iously (Sa kanen e al., 2012). B ie ly, hASC we e seeded in
Hu ala e a l .
Al ex 32(2), 2015 127
EGM-2 (Lonza) a 20,000 cells/cm2. A e 1-3 h HUVEC we e
seeded on op o hASC a 4,000 cells/cm2. The hASC we e used
a passage 2 and HUVEC a passage 4 in he co-cul u e (passage
numbe inc eased a seeding). The day a e pla ing, he s imula-
ion media we e applied o he co-cul u e (see Tab. 1): 1) EGM-2
con aining epide mal g ow h ac o , ascula endo helial g ow h
ac o A (VEGF), ib oblas g ow h ac o 2 (FGF-2), insulin-
like g ow h ac o I (IGF-I), asco bic acid (AA), hepa in, hyd o-
co isone (HY), an ibio ic mix: 30 µg/ml gen amicin and 15 ng/
ml ampho e icin and 2% e al bo ine se um (FBS); 2) EGM-2/
HS, whe e 2% FBS was eplaced wi h 2% HS; 3) se um ee ba-
sal medium (SFM); 4) Vascula s imula ion medium (VSM).
In his i s media compa ison s udy, he AA, HY and hepa in
in VSM we e aken om he EGM-2 ki and used acco ding
o he manu ac u e ’s p o ocol (concen a ions o hese a e no
publicly a ailable). A e ha , concen a ions o AA, hepa in
sodium sal om bo ine in es inal mucosa (HE) and HY pu -
chased om Sigma we e op imized o VSM. In he concen a-
ion op imiza ion s udy he es ed concen a ions o AA we e 0,
50, 100, 200, 500, 1000 and 2000 µg/ml; o HY 0, 20, 200, 1000,
and 2000 ng/ml and o HE 0, 50, 500, 10,000 and 50,000 ng/ml.
The concen a ions o 10 ng/ml VEGF and 1 ng/ml FGF-2, used
o VSM, had been op imized in ou p e ious s udy (Sa kanen
e al., 2012). The co-cul u es we e g own o 6 days p io o
immunocy ochemis y o quan i a i e eal- ime PCR (qPCR)
p ocessing. S imula ion medium was changed once du ing he
6 day cul u e.
Quan i a i e eal- ime PCR
Genes ac i a ed in co-cul u e o hASC and HUVEC in EGM-
2 medium e sus co-cul u e in VSM we e analyzed by qPCR.
To al RNA was ex ac ed a day 6 using he RNAeasy miniki
(Qiagen) ollowing he manu ac u e ’s p o ocol. A s ep o elimi-
na e genomic DNA con amina ion was included in he isola ion
and pe o med wi h RNase- ee DNase se (Qiagen). Re e se
ansc ip ion o o al RNA o cDNA was pe o med using RT2
Fi s S and Ki (Qiagen) ollowing he manu ac u e ’s ins uc-
ions.
Human Angiogenesis RT2 P o ile ™ PCR A ay (Qiagen, Va-
lencia, Cali o nia, USA) was used o p o ile he exp ession o
84 key angiogenesis- ela ed genes ( he comp ehensi e lis o
genes included in he a ay can be ound in h p://www.sabio-
sciences.com). The a ay was pe o med acco ding o he man-
u ac u e ’s p o ocol using BioRad CFX96 Real Time Sys em
(BioRad Labo a o ies, USA). The a ay con ained i e house-
keeping genes and con ols, including genomic DNA con ol,
Re e se T ansc ip ion Con ol and Posi i e PCR con ols. The
co-cul u e o hASC and HUVEC g own in EGM-2 medium was
used as a con ol. Th ee independen expe imen s wi h he a ay
we e pe o med using he same cells.
Immunocy ochemis y
To analyze he ascula -like ne wo k o ma ion and di e en
cell ypes p esen in hASC and HUVEC co-cul u e, immunocy-
ochemical s aining was pe o med. In p ima y an ibody s ain-
ing, endo helial cell speci ic an ibody o abbi an i-human on
Willeb and ac o IgG (an i-VWF, 1:100, F3520, Sigma) wi h
common pe icy ic ma ke α-human smoo h muscle ac in (mon-
oclonal an i-SMA clone 1A4, 1:200, M0851, DAKO), ascula
smoo h muscle cell ma ke smoo h muscle myosin hea y chain
(an i-SMMHC, clone hSM-V, 1:800, M7786, Sigma), con ac-
ile smoo h muscle cell ma ke calponin (an i-calponin, clone
hCP, 1:800, C2687, Sigma), pe icy ic and smoo h muscle cell
p ogeni o ma ke pla ele de i ed g ow h ac o ecep o -β
(an i-PDGFRβ, clone PDGFR-B2, 1:800, P7679, Sigma), as-
cula endo helial cadhe in (CD144,Clone 55-7H1, 1:50, 555661,
BD Pha mingen), monoclonal occludin (clone 1G7, 1:300,
WH0004950M1, Sigma) o basemen memb ane ma ke colla-
gen IV (an i-COLIV, clone COL-94, 1:500, C1926, Sigma), all
an i-human and all p oduced in mouse, excep W , we e used.
Co-cul u e was ixed wi h 70% e hanol a day 6. A e ixa-
ion, he cells we e pe meabilized wi h 0.5% T i on-X100 (MP
Biochemicals, Ohio, USA) and non-speci ic binding si es we e
blocked wi h 10% bo ine se um albumin (BSA, Roche Diagnos-
ics Co po a ion, Indianapolis, USA). P ima y an ibody in 1%
BSA was applied o he cells. Seconda y an ibodies used we e
polyclonal goa an i- abbi IgG e ame hyl hodamine (TRITC,
1:50, T6778 Sigma), an i- abbi IgG A568 (1: 400, A11011, In-
i ogen) and an i-mouse IgG luo escein iso hiocyana e (FITC,
1:100, F4143, Sigma). A e immunocy ochemical s aining he
Tab. 1: S imula ion media es ed in ascula model
Ac onym Basal medium Se um G ow h ac o s Supplemen a ion
eGM-2 eBM-2 2% FBS VeGF, FGF-2, IGF-I, hyd oco isone, asco bic acid, hepa in,
epide mal g ow h ac o GA-1000
eGM-2/HS eBM-2 2% HS VeGF, FGF-2, IGF-I, hyd oco isone, asco bic acid, hepa in,
epide mal g ow h ac o GA-1000
basal SFM DMeM/F12 – I S, BSA, NaP, l-glu amine, 3,
VSM DMeM/F12 – VeGF, FGF-2 I S, BSA, NaP, l-glu amine, 3,
hyd oco isone, asco bic acid,
hepa in sodium sal
0.1% I S (insulin- ans e in-sodium seleni e media supplemen ), 1.28 mM l-glu amine, 1% BSA (bo ine se um albumin),
2.8 mM NaP (sodium py u a e), 100 IU/ml P/0.1 mg/ml S, 0.1 nM 3 (3,3’,5- iiodo-l- hy onine sodium sal ), HS (human se um),
FBS ( e al bo ine se um), VEGF ( ascula endo helial g ow h ac o A), FGF-2 ( ib oblas g ow h ac o 2), IGF-I (insulin-like g ow h
ac o I), GA-1000 (30 µg/ml gen amicin and 15 ng/ml ampho e icin), VSM ( ascula s imula ion medium), SFM (se um ee medium)
Hu ala e a l .
Al ex 32(2), 2015
128
ascula -like ne wo k was analyzed and pho og aphed wi h
Nikon Eclipse TS100 in e ed luo escence mic oscope (Nikon,
Tokyo, Japan) and Nikon digi al sigh DS-U2 came a (Nikon).
Images we e u he p ocessed wi h NIS Elemen s (Nikon, To-
kyo, Japan) and Adobe Pho oshop CS3-so wa e (Adobe Sys-
ems Inco po a ed, San Jose, CA, Uni ed S a es).
Flow cy ome ic su ace ma ke exp ession analysis o
HUVEC and hASC
hASC we e cul u ed in hASC medium (passage 1) o 6-7 days
and HUVEC we e cul u ed in EGM-2 medium (passage 3) o
3 days p io o su ace ma ke exp ession analysis using a BD
FACSCan o II low cy ome e (BD Biosciences, E embodegem,
Belgium). Fo he low cy ome y analysis cells we e di ided
in o 5 ml polys y ene ound bo om FACS ubes (BD, New Je -
sey, USA) a 250,000 cells pe ube. The cells we e washed once
wi h wa m s aining bu e (1% BSA in PBS) and cen i uged
a 131 x g o 5 min, a e which hey we e s ained ei he o
su ace ma ke s o o in acellula ma ke s.
Fixa ion and pe meabiliza ion we e only pe o med o s ain-
ing o in acellula ma ke s. The ixa ion was conduc ed by
incuba ing he samples o 30 min in 2% pa a o maldehyde
in PBS a oom empe a u e (RT). The cells we e hen cen i-
uged a 500 x g o 5 min. Pe meabiliza ion o he cells was
pe o med by 10 min incuba ion in 0.1% T i on-X100 in PBS a
RT a e which he cells we e cen i uged a 500 x g o 5 min
and washed once o wice wi h s aining bu e be o e addi ion
o an ibodies.
The labelled mouse an i-human an ibodies used we e in a-
cellula W -A2-allophycocyanin IgG2b (APC, #IC27641A)
and eNOS- phycoe y h in IgG1 (PE, #560103), and su ace
ma ke s CD144-FITC IgG1 (#560411), CD73-Phycoe y h in-
Cyanine IgG1 (PE-CY7, #561258), CD309-PE IgG1 (#560872),
CD68-FITC IgG2b (#562117), NG2-PE IgG1 (#FAB2585P),
CD90-FITC IgG1 (#561969), CD105-V450 IgG1 (#561447),
CD34-APC IgG1 (#561209), CD140b-PE IgG2a (#558821),
CD31-V450 IgG1 (#561653), CD45-PE IgG1 (#560975),
CD14-FITC IgG1 (#561712). Iso ype con ols mouse IgG2b-
APC (#IC0041A), Mouse IgG1-PE-CY7 (#557872), mouse
IgG1-PE (#559320), mouse IgG1-FITC (#555748), mouse
IgG2b-FITC (#556655), mouse IgG1-V450-(#642268), mouse
IgG1-APC (#550854), mouse IgG2a-PE (#551438). All an i-
bodies and hei co esponding iso ype con ols used in he low
cy ome y analysis we e pu chased om BD excep NG2, W
and IgG2b-APC, which we e pu chased om R&D Sys ems.
Labelled an ibodies we e added in o cell suspension in cold
s aining bu e and incuba ed on ice o 30 min in he da k. A e
incuba ion he cells we e washed once wi h s aining bu e and
wice wi h PBS. Su ace ma ke s ained cells we e cen i uged
a 200 x g o 5 min and in acellula ma ke s ained cells we e
cen i uged o 5 min a 500 x g.
Flow cy ome y analysis was pe o med wi h cells suspended
in ice cold PBS and 5,000 e en s we e analyzed pe sample.
Compensa ion was done wi h compensa ion pa icles, i.e., BD™
CompBeads (BD) acco ding o he manu ac u e ’s ins uc ions.
The esul s we e analyzed wi h BD FACSDi a™ So wa e
(BD). The posi i e exp ession was ob ained by ga ing 98% o
he e en s iso ype con ol esul s and hen in e ing he ga e o
ob ain a pe cen age o posi i ely s ained cells in he samples.
The esul s we e calcula ed as pe cen ages wi h SD.
Elec on mic oscopy
hASC and HUVEC co-cul u e was pe o med in 24-well UpCell
pla es (The moFishe ) o ansmission elec on mic oscopy
(TEM) and on glass co e slips coa ed wi h 0.1% gela in o
scanning elec on mic oscopy (SEM). Co-cul u e was main-
ained o 6 days and washed wice wi h PBS p io o ixa ion.
SEM specimens we e ixed in 2% glu a aldehyde in 0.1 M
phospha e bu e a e which he SEM samples we e dehyd a -
ed in alcohol and d ied in Bal ec c i ical poin d ye (Bal ec,
CPD030, Balze s, Liech ens ein). A laye o pla inum was spu -
e ed on o he specimens wi h a spu e coa e (Aga Scien i ic,
S ans ed, England). Specimens we e examined in a Zeiss Ul a
Plus scanning elec on mic oscope, (Ca l Zeiss MT – Nano ech-
nology Sys em Di ision, Ca l Zeiss NTS GmbH, Obe kochen,
Ge many) using 5 kV as an accele a ing ol age.
TEM specimens we e de ached om he empe a u e sensi-
i e 24-well UpCell pla e and ans e ed o ixa i e wi h 1%
glu a aldehyde, 4% o maldehyde mix u e in 0.1 M phospha e
bu e o 10 min. The cell shee was imme sed in 2% aga ose in
dis illed wa e and pos ixed in 1% osmium e oxide, dehyd a -
ed in ace one and embedded in Epon LX 112 (Ladd Resea ch
Indus ies, Ve mon , USA). Thin sec ions we e cu wi h Leica
Ul acu UCT ul amic o ome, s ained in u anyl ace a e and lead
ci a e and examined in a Philips CM100 ansmission elec on
mic oscope. Images we e cap u ed by a Mo ada CCD came a
(Olympus So Imaging Solu ions GMBH, Muns e Ge many).
Quan i a i e analysis o ascula -like ne wo k o ma ion
Vascula -like ne wo ks, i.e., W -posi i e ubule s uc u es
o med in di e en s imula ion media, we e imaged using Cell-
IQ (Chipman ech., Tampe e, Finland) wi h 10x objec i e and
5x5 g id. The quan i a ion o he a ea o he ascula -like ne -
wo k was pe o med using ImageJ so wa e (Na ional ins i u es
o heal h, NIH, Ma yland, USA) o he image analysis. Im-
ages we e i s con e ed o 8-bi g ay scale, hen backg ound
was sub ac ed and inally he bina y h eshold unc ion was
adjus ed o ob ain he bes con as o he ascula -like ne wo k
agains he backg ound. Wi h hese se ings, he o al a ea o
ascula -like ne wo k was calcula ed as he o al numbe o pix-
els in images wi h se h eshold.
S a is ical analysis
S a is ical analyses we e pe o med and g aphs p ocessed wi h
G aphPad P ism 5.0 (G aphPad So wa e, Inc., San Diego, CA,
USA). The esul s conce ning ascula -like ne wo k o ma-
ion we e subjec ed o one-way ANOVA ollowed by Dunne ’s
pos - es when applicable. The esul s we e epo ed as o al
a ea ± SD and di e ences we e conside ed signi ican when
p < 0.05*, p < 0.01** and p < 0.001***.
PCR esul s we e analyzed wi h he PCR A ay Da a Analy-
sis Web Po al (h p://www.SABiosciences.com/pc a ayda aa-
nalysis.php). The ollowing o mula was used o calcula e he
ela i e amoun o he ansc ip s o he co-cul u e o hASC and
Hu ala e a l .
Al ex 32(2), 2015 129
CD144 and CD31 and mac ophage/monocy e ma ke s CD68
and CD14 we e e y low (< 1.4%).
HUVEC we e shown o exp ess endo helial ma ke CD31
and a specialized ascula endo helial ma ke CD144 a a high
le el (> 70%). Angiogenic endo helial ma ke CD105 and cell
su ace enzyme CD73 we e ound in he HUVEC popula ion
(57-68%). Ma ke s o mac ophages/monocy es (CD68) and
mu al cells (NG2) we e low (≤ 2%).
3.2 VSM induced op imal ascula -like
ne wo k o ma ion
O he di e en s imula ion media (Tab. 1), VSM induced an
op imal ascula -like ne wo k o ma ion (Fig. 1). The imaging
analysis showed mo phological di e ences in he ascula -like
ne wo k in VSM compa ed o EGM-2 medium. VSM p oduced
a uni o mly dis ibu ed ascula -like ne wo k wi h connec ed
b anches and ewe cell agg ega es. Low se um (2%) and se-
HUVEC in he VSM compa ed o ansc ip s o he co-cul u e
in EGM-2 medium: ΔΔCT = ΔCT (VSM) − ΔCT (EGM-2). A
wo- old change compa ed o housekeeping gene GAPDH was
conside ed signi ican .
3 Resul s
3.1 Pheno ypic cha ac e iza ion
Pheno ypic cha ac e iza ion o he cells used in he ascula
model was pe o med by analyzing he su ace and in acellula
ma ke s o hASC and HUVEC sepa a ely (Tab. 2). The mesen-
chymal s em cell ma ke s CD73, CD90 and CD105 we e highly
exp essed (> 64%) in hASC. Also, pe icy e ma ke PDGFR-β
(CD140b) was s ongly exp essed (> 52%) in he hASC popu-
la ion. Hema opoie ic ma ke CD34 was mode a ely exp essed
(< 40%), whe eas he exp ession le els o endo helial ma ke s
Tab. 2: Pheno ypic cha ac e iza ion o human adipose s omal cells (hASC p1) and human umbilical ein
endo helial cells (HUVEC p3)
The exp ession o su ace and in acellula ma ke s was analysed by low cy ome y (5000 e en s pe sample).
Resul s a e shown as pe cen age o posi i e cells.
Cells P o ein An igen Posi i e cells % SD n
hASC W on Willeb and ac o 7.5 4.3 11
CD144 ascula endo helial cadhe in 1.4 2.9 9
eNOS endo helial ni ic oxide syn hase 0.7 1.1 8
CD140b PDGF Recep o b 52.8 16.0 6
CD45 leukocy e common an igen 16.4 14.9 6
CD14 exp essed on monocy es/mac ophages 0.7 0.8 7
CD68 exp essed on mac ophages and monocy es 1.1 1.4 8
CD309 VeGF ecep o 2 1.2 3.5 6
CD31 pla ele endo helial cell adhesion molecule 0.9 3.1 6
NG2 chond oi in p o eoglycan 31.8 29.1 8
CD90 hy-1 82.9 15.5 10
CD105 endoglin 64.1 27.1 7
CD73 ec o-5'-nucleo idase 88.8 6.7 12
CD34 sialomucin-like adhesion molecule 40.9 34.3 10
HUVEC W on Willeb and Fac o 45.6 13.2 13
CD144 ascula endo helial cadhe in 70.6 14.7 17
CD309 VeGF ecep o 2 31.4 17.0 10
CD73 ec o-5'-nucleo idase 68.7 17.1 14
CD68 exp essed on mac ophages and monocy es 1.9 10.1 11
eNOS endo helial ni ic oxide syn hase 48.4 18.4 13
CD105 endoglin 57.4 26.6 9
CD34 sialomucin-like adhesion molecule 43.1 18.6 9
CD31 pla ele endo helial cell adhesion molecule 81.3 13.7 8
CD140b PDGF Recep o b 7.2 8.7 8
NG2 chond oi in p o eoglycan 2.0 1.1 7

Hu ala e a l .
Al ex 32(2), 2015
130
HUVEC in EGM-2 medium (con ol) and in VSM. Mode a e
di e ences in he exp ession o angiogenesis- ela ed genes we e
de ec ed be ween EGM-2 and VSM. In VSM, nine genes we e
up- egula ed (Angp 1, F3, FIGF, IGF-I, LEP, MDK, MMP2,
MMP9, PGF) and nine (CCL11, CXCL9, FN1, IL6, IL8, SER-
PINE1, TGFB2, THBS2, TIMP1) we e down- egula ed in com-
pa ison o EGM-2 medium. Fold changes and s a is ical sig-
ni icances o he up- and down- egula ed angiogenesis- ela ed
genes a e shown in Table 3.
3.4 Ma u a ion s age, ex acellula ma ix
p oduc ion and 3D p ope ies o he ascula -like
ne wo k o med in VSM
Immunocy ochemical s aining o he ascula -like ne wo k
o med in VSM showed he p esence o pla ele de i ed g ow h
ac o ecep o -β (PDGFRβ) posi i e cells (Fig. 3D). Also,
smoo h muscle ac in (SMA), smoo h muscle myosin hea y
chain (SMMHC) and con ac ile smoo h muscle calponin posi-
i e cells su ounded he ubule s uc u es (Fig. 3A-C). CD144
and occludin posi i e s aining indica ed he p esence o adhe -
ence junc ions be ween endo helial cells (Fig. 3E-F).
um- ee medium enabled s able cell a achmen and induced a
dense, connec ed ascula -like ne wo k wi h in ac ubule walls.
Human se um was ound o induce a dense and mo e connec ed
ascula -like ne wo k han FBS (Fig. 1).
In he supplemen op imiza ion se e al di e en concen a-
ions o asco bic acid, hepa in sodium sal and hyd oco isone
we e es ed and compa ed o ascula -like ne wo k o med
in EGM-2 medium. An op imal ascula -like ne wo k o -
ma ion was ob ained wi h he AA concen a ion 100 µg/ml
(Fig. 2A), HE concen a ion 0-500 ng/ml (Fig. 2B) and HY
concen a ion 0.2 µg/ml (Fig. 2C). The esul s showed ha a
ascula -like ne wo k was o med in he absence o HE sup-
plemen wi h mino mo phological di e ences. Howe e , low
concen a ions o HE (50-500 ng/ml) inc eased he o al a ea
o he ascula -like ne wo k signi ican ly compa ed o EGM-
2 medium (Fig. 2B).
3.3 Exp ession o angiogenesis- ela ed genes
in VSM compa ed o EMG-2
Exp ession o 84 angiogenesis- ela ed genes was analyzed
om a ascula -like ne wo k o med in co-cul u e o hASC and
Fig. 1: Mo phology o ascula -like ne wo k in
di e en s imula ion media wi h di e en se um
concen a ions
Vascula -like ne wo ks we e s ained agains Col IV (FI C,
g een) and W ( RI C, ed). Co-localiza ion is shown in
he me ged image. Vascula -like ne wo k o med in (A)
comme cial eGM-2 medium wi h 2% FBS, (B) se um- ee
VSM, (C) eGM-2 wi h 2% HS and (D) basal SFM. Scale
ba s 100 µm in each igu e. Col IV= collagen ype IV,
VWF= on Willeb and ac o , VSM= ascula s imula ion
medium, FBS= e al bo ine se um, HS= human se um,
SFM=basal se um ee medium.
Fig. 2: Vascula -like ne wo k o ma ion (a ea in pixels) in di e en
concen a ions o (A) asco bic acid, (B) hepa in sodium sal and
(C) hyd oco isone in VSM
Vascula -like ne wo k o med in VSM supplemen ed wi h di e en
concen a ions o asco bic acid, hepa in sodium sal o hyd oco isone we e
compa ed o eGM-2 medium and di e ences p esen ed as ollows:
p < 0.05*, p < 0.01** and p < 0.001***. An op imal ascula -like ne wo k
o ma ion was ob ained wi h 100 µg/ml asco bic acid and 0.2 µg/ml
hyd oco isone. Hepa in sodium sal could be used in concen a ions
0-500ng/ml. Resul s a e depic ed as o al a ea wi h s anda d de ia ions,
n = 5 in each supplemen and concen a ion.
Hu ala e a l .
Al ex 32(2), 2015 131
The pheno ypic analysis o he building blocks o he ascula
model, i.e., he hASC and HUVEC, was pe o med using low
cy ome y. HUVEC showed a s ong exp ession o CD105 (en-
doglin), sugges ing an ac i e a he han quiescen s a e o he
cells (Be nabeu e al., 2009). In e es ingly, a ecen s udy epo -
ed a mino ole o endoglin in asculogenesis whe eas VEGF-
induced angiogenesis was se e ely impai ed in an endoglin de-
icien mouse emb yonic s em cell model (Liu e al., 2014). The
high exp ession o CD144 and CD31 suppo s he ac i e s a us
o HUVEC in he o ma ion o in e cellula junc ions. The le el
o CD73, a 5'-ec onucleo idase ha no mally supp esses p o-
in lamma o y esponses in human endo helial cells (G unewald
and Ridley, 2010) was high in ou HUVEC analysis. Pe icy e
ma ke s NG2 and CD140b (PDGFRβ) we e low, indica ing he
absence o mu al cells in he HUVEC popula ion.
This s udy con i ms he ea lie esul s shown by us (Sa kanen
e al., 2012) and o he s (Lind oos e al., 2010; T ak ue e al.,
2008) ha he s omal- ascula ac ion ex ac ed om adipose
issue is he e ogenic. hASC exp essed mesenchymal s em cell
ma ke s CD90, CD105 and CD73 a high le els, showing he
Elec on mic oscopic analysis con i med he mic os uc u e
o he ubules o med in VSM. Tubules we e shown o ha e
a lumen, basemen memb ane and junc ions (Fig. 4A-C). The
h ee dimensional s uc u e o he ascula -like ne wo k and
shape o he ubules can be seen in he close-up image Figu e
4D.
4 Discussion
P esen chemical sa e y es ing and non-clinical d ug de elop-
men ely mainly on animal biology al hough ele an sa e y
es ing da a should be o human o igin. Fu he , in i o es
sys ems should be ho oughly cha ac e ized in hei human el-
e ance and unc ionali y. In his s udy, we de eloped a no el
ascula s imula ion medium (VSM) and u he cha ac e ized
he in i o human ascula model de eloped by us (Sa kanen e
al., 2012). This de ined in i o ascula model has he po en ial
o be used o sa e y assessmen o compounds and o issue
enginee ing applica ions.
Tab. 3: Fold changes and s a is ical signi icance o exp ession o angiogenesis- ela ed genes in ascula -like ne wo k o med
in VSM compa ed o EGM-2 medium
Genes Ta ge Fold change 2^(- ΔΔCT) S a is ical signi icance
Up- egula ed genes
ANGPT1 angiopoie in 1 2.964 *
F3 coagula ion ac o III 3.255 *
FIGF ascula endo helial g ow h ac o D 3.043 ns
IGF-I insulin-like g ow h ac o I 16.555 ns
LEP lep in 2.932 ns
MDK neu i e g ow h-p omo ing ac o 2 2.06 ns
MMP2 ma ix me allopep idase 2 2.011 *
MMP9 ma ix me allopep idase 9 3.919 *
PGF placen al g ow h ac o 2.319 *
ANGPT2 angiopoie in 2 1.769 **
Down- egula ed genes
CC L11 chemokine ligand 11 2.613 ns
CXCL9 chemokine ligand 9 2.197 ns
FN1 ib onec in 1 2.295 *
IL6 in e leukin 6 2.331 ns
IL8 in e leukin 8 2.546 ns
SERPINE1 se pin pep idase inhibi o 3.465 **
TGFB2 ans o ming g ow h ac o be a 2 2.080 ns
THBS2 h ombospondin 2 2.085 **
TIMP1 IMP me allopep idase inhibi o 1 5.01 **
VEFG-A ascula endo helial g ow h ac o A 1.804 ns
TGFβ1 ans o ming g ow h ac o be a 1 1.807 *
FGF-2 ib oblas g ow h ac o 2 1.917 ns
Hu ala e a l .
Al ex 32(2), 2015
132
he e is a mode a e exp ession o CD34; hese a e ma ke s o
mesenchymal s em cells and endo helial p ogeni o s. A hASC
popula ion o CD90+/CD34+ cells was shown o be capable
o di e en ia ing in o endo helial cells and o m capilla y-like
s uc u es (De F ancesco e al., 2009).
The exp ession o ma u e endo helial cell ma ke s CD31 and
CD144 in he hASC popula ion was e y low. In e es ingly,
he e is a epo sugges ing ha adipose s omal cells may en-
hance endo helial di e en ia ion o p ogeni o cells (Rubina e
al., 2009). Howe e , ano he essen ial ma ke o endo helial
p ogeni o cells, CD309 (VEGFR2) (Yode , 2012), showed e y
low exp ession in he six hASC lines ha we e analyzed in his
s udy. The e o e, he p esence o an endo helial p ogeni o cell
popula ion in he s omal- ascula ac ion needs o be s udied
u he .
Se um- ee VSM was shown o p oduce a ep oducible, ex-
ensi e ma u e ascula -like ne wo k and p o ide a sui able al-
e na i e o he comme cial EGM-2 medium (p e iously used
in he model, con aining 2% FBS). Se um had a s ong co e-
la ion wi h ascula -like ne wo k o ma ion. A highe se um
(10%) concen a ions he ne wo k was sho e wi h b oken u-
bule walls, and andom de achmen o he cell laye occu ed
(da a no shown). In low- o se um ee (0-2%) en i onmen ,
p esence o a cell popula ion wi h mul ilineage di e en ia ion
po en ial. Pe icy e ma ke s CD140b (PDGFRβ) and NG2 we e
exp essed a high o mode a e le el, suppo ing he inding ha
pe icy es om he hASC popula ion a e lining he ascula -like
ne wo k in ou model.
We also analyzed whe he he popula ion o endo helial p o-
geni o cells capable o asculogenesis could be ound in hASC
as sugges ed by some ea lie epo s (Sa kanen e al., 2012;
Mi an ille e al., 2004; Heyda khan-Hag all e al., 2008). The
esul s showed ha , in addi ion o he high exp ession o CD90,
Fig. 3: Vascula -like ne wo k o med in se um- ee VSM
shows ascula ma u a ion ma ke s a day 6
α- W -s aining is shown in ed in all igu es. (A) Calponin-,
(B) SMA- and (C) SMM-posi i e cells su ounding he ascula -like
ne wo k. (D) α-PDGFRβ s ained pe icy es a e loca ed along
he ubules. (e) CD144- and (F) occludin-posi i e junc ions can be
ound in he ubule s uc u es. Scale ba 100 μm in each image.
VSM= ascula s imula ion medium, W = on Willeb and ac o ,
SMA = smoo h muscle ac in, SMM = smoo h muscle myosin
hea y chain, PDGFRβ = pla ele de i ed g ow h ac o ecep o β,
CD144 = ascula endo helial cadhe in.
Fig. 4: T ansmission (A, B, C) and scanning (D)
elec on mic oscopy images o he ascula -like
ne wo k cul u ed in VSM
(A) Ma u e ubule wi h lumen = L and ex acellula ma ix
= ECM. The lumen is illed wi h deb is om he apop o ic
cells. Scale ba 2 µm. (B) Close-up image wi h basemen
memb ane = BM, adhe ence junc ion = AJ and lumen o
he ascula -like ubules. Scale ba 2 µm. (C) ubule in
ea lie s age o ascula de elopmen showing endo helial
cell wi h la ge nucleus (whi e a ow), lumen, adhe ence
junc ions and eCM. Scale ba 2 µm. D) Scanning elec on
mic oscopy image shows he 3D s uc u e and he ubule
shape o he ascula -like ne wo k. Scale ba 2 µm.
VSM = ascula s imula ion medium.
Hu ala e a l .
Al ex 32(2), 2015 133
cula -like ne wo k. Howe e , i p oduced cell agg ega es a
high concen a ions whe eas a lack o hyd oco isone induced
a spa se ascula -like ne wo k. Ou esul s sugges ha hyd o-
co isone ac s as a mi ogen in he ascula model. O he sup-
plemen s o he VSM included ITS (insulin, ans e in, selenic
acid medium supplemen ), BSA, sodium py u a e, L-glu amine
and T3 (3,3',5- iiodo-L- hy onine). The concen a ions o hese
supplemen s we e epo ed ea lie by us (Vuo enpää e al., 2014)
and we e used in his s udy wi h mino modi ica ions.
Lumen o ma ion is c i ical o he ans o ma ion o co ds in-
o a pe usable ascula sys em (Cha pen ie and Conlon, 2014).
Elec on mic oscopic analysis assu ed he o ma ion o 3D as-
cula mic os uc u e including lumen in he no el VSM. Tubules
a di e en s ages o lumen de elopmen we e de ec ed, indica -
ing he ongoing p ocess o ne wo k o ma ion a day 6. The ini-
ia ion o lumen de elopmen is igge ed by apical-basal pola -
i y o he endo helial cells in which CD144 plays a c i ical ole
in p omo ing he localiza ion o pola i y ma ke s (Cha pen ie
and Conlon, 2014). The elec on mic oscopic analysis as well as
immunocy ochemis y esul s o his s udy showed ha hASC
and HUVEC co-cul u e ac i ely p oduces ex acellula ma ix
(ECM) componen s, including ib illins, hus c ea ing na u al
3D sca old a ound hem. The ecip ocal in e ac ion be ween
ECM s oma and ascula ne wo k is impo an in di ec ing es-
sel g ow h (Hoying e al., 2014; Du e al., 2014). The hASC and
HUVEC co-cul u e gi es mechanical suppo o o he a ge
cells, e.g., ca diomyocy es, and, addi ionally, he mic oen i on-
men o med by he co-cul u e enhances a ge cell iabili y as
epo ed p e iously (Vuo enpää e al., 2014). This 3D ascula
model p o ides a mo e in i o-like es sys em wi hou an a i i-
cial sca old ha may in e e e wi h he cell-cell in e ac ions o
a ec he oxicological applica ions o he model.
In he geno ypic analysis 84 human angiogenesis- ela ed
genes we e s udied. Nine genes we e down- egula ed and nine
genes up- egula ed in VSM compa ed o EGM-2 medium. The
exp ession o angiopoie in 1, a ma ke o ma u e ubules, was
signi ican ly highe in VSM compa ed o EGM-2 medium,
whe eas angiopoie in 2, a ma ke o he ea ly s age, i.e., sp ou -
ing angiogenesis, was sligh ly up egula ed in VSM. VEGF-A and
FGF-2, indica o s o ea ly s age angiogenesis, we e mode a ely
ye no signi ican ly down- egula ed in VSM compa ed o EGM-
2. Since bo h media included VEGF and FGF-2, he exp ession
o hese g ow h ac o s was appa en ly unnecessa y o he
cells. On he con a y, placen al g ow h ac o (PGF) showed
modes up- egula ion in VSM. PGF and i s ecep o VEGFR-1
a e minimally exp essed in adul quiescen ascula u e, bu a e
ma kedly up- egula ed du ing pa hological condi ions (Ca me-
lie e al., 2001). TGFβ1, a ma ke o ubule ma u a ion, was
sligh ly down- egula ed in VSM. Vessel ma u a ion elies pa ly
on TGF-β signaling and TGF-β s imula es mu al cell p oli -
e a ion and mig a ion and p oduc ion o ECM (Po en e e al.,
2011). The gene exp ession analysis con i ms he inding ha
VSM enhances he ma u a ion o he ascula -like ne wo k.
In e es ingly, he exp ession o insulin like g ow h ac o
(IGF-I) showed he highes up- egula ion in VSM. IGF-I ea -
men in myoca dial in a c ion has been shown o inc ease ci -
cula ing angiogenic g ow h ac o s, hus p o iding p o ec ion
ne wo k o ma ion inc eased, ubules we e highly b anched and
ubule walls emained in ac . In se um- ee condi ions we de-
ec ed hinne and mo e b anched ubules as also epo ed by
Yang and Xiong (Yang and Xiong, 2012). Se um- ee medium
is ideal o use in d ug de elopmen , since se um is a complex
mix u e o componen s wi h unknown composi ion and p o ein
binding a ini ies (Shen e al., 2013). Fu he mo e, a issue con-
s uc aimed o clinical he apy should be cul u ed in human
se um o , p e e ably, in a se um- ee en i onmen o a oid he
isk o in ec ion and se e e immune eac ions in he ecipien
(Pa ikoski e al., 2013; Lind oos e al., 2011, 2010; Holm e
al., 2010).
Fe al bo ine se um (FBS), al hough being he mos widely
used g ow h supplemen , holds ambiguous, unknown e ec s o
cell cul u e and also aises e hical conce ns due o he numbe
o bo ine e uses needed o se um p oduc ion (Go ipamula
e al., 2013; B unne e al., 2010). In his s udy, mic oscopical
analysis showed ha HS was mo e induc i e o ascula -like
ne wo k o ma ion han FBS. Replacemen o FBS wi h HS has
been epo ed o suppo equal o highe p oli e a ion a es and
di e en ia ion capaci y o adipose s omal cells (Lind oos e al.,
2011; B unne e al., 2010). Fu he mo e, human mesenchymal
s em cells ha e been shown o main ain hei immunopheno ype
and mul ilineage po en ial in se um- ee medium (Pa ikoski e
al., 2013; Ma k e al., 2013).
In his s udy we de eloped a de ined ascula s imula ion
medium since comme cial media p oducing companies do no
necessa ily publish he speci ic composi ion o hei media.
This migh complica e he use o comme cial media in oxico-
logical s udies as medium componen s may in e ac wi h es
compounds. In addi ion o commonly used VEGF and FGF-2
de ined by us ea lie (Sa kanen e al., 2011), we ound ha he
ne wo k o ma ion can be u he imp o ed by addi ion o AA,
HE and HY. Asco bic acid ( i amin C) is an essen ial nu ien o
human endo helial cells necessa y o hei e ec i e mig a ion
and o he syn hesis o collagen ype IV, an impo an compo-
nen o basemen memb ane (Telang e al., 2007). In his s udy,
he o ma ion o basemen memb ane a ound he ubules was
impai ed in he absence o AA (da a no shown). Howe e , AA
inhibi ed angiogenesis a high concen a ions (1000-2000 µg/
ml) as epo ed p e iously (Miki o a e al., 2008). The e o e,
AA is essen ial o collagen IV o ma ion in basemen mem-
b ane a low concen a ions.
Al hough HE is needed o he a achmen o some g ow h ac-
o s o hei cell su ace ecep o s (Ashika i-Hada e al., 2005)
and was ound o be bene icial o he ne wo k o ma ion, i did
no induce a signi ican ad an age o he mo phology o b anch-
ing o he ubules. Since HE is an animal de i ed subs ance i
should a he be a oided whe e clinical applica ions a e con-
ce ned. The lack o need o added HE can be explained by he
sec e ion o pe lecan by HUVEC (Mu ikipudi e al., 2013; Sch-
lessinge e al., 2000). Ou esul s co ela e wi h ea lie s udies
(Kho ana e al., 2003; Jung e al., 2001) showing ha HE inhib-
i s ascula -like ne wo k o ma ion a high concen a ions.
Hyd oco isone, al hough no p o-angiogenic i sel , has a
bene icial e ec on angiogenesis (Goding, 2009). In his s udy,
hyd oco isone inc eased he numbe o b anches in he as-