Cell Repo s
Resou ce
Sel -Renewing Human Bone Ma ow Mesensphe es
P omo e Hema opoie ic S em Cell Expansion
Joan Ise n,
1
Bea iz Ma ı
´n-An onio,
2
Roshanak Ghazan a i,
3
Ana M. Ma ı
´n,
1
Juan A. Lo
´pez,
1
Raquel del To o,
1
Abel Sa
´nchez-Aguile a,
1
Lo ena A anz,
1
Daniel Ma ı
´n-Pe
´ ez,
1
Ma ı
´a Sua
´ ez-Lledo
´,
2
Ped o Ma ı
´n,
2
Melissa Van Pel,
4
Willem E. Fibbe,
4
Jesu
´sVa
´zquez,
1
S e an Scheding,
3,5
A
´l a o U bano-Ispizu
´a,
2
and Simo
´nMe
´ndez-Fe e
1,6,
*
1
Cen o Nacional de In es igaciones Ca dio ascula es Ca los III, Mad id 28029, Spain
2
Hema ology Depa men , Hospi al Clinic, Uni e si y o Ba celona, IDIBAPS and Ins i u e o Resea ch Josep Ca e as, Ba celona
08036, Spain
3
Lund S em Cell Cen e , Uni e si y o Lund, 22184 Lund, Sweden
4
Depa men o Immunohema ology and Blood T ans usion, Leiden Uni e si y Medical Cen e , Leiden 2333 ZA, The Ne he lands
5
Depa men o Hema ology, Ska
˚ne Uni e si y Hospi al Lund, 22184 Lund, Sweden
6
Depa men o Medicine, Icahn School o Medicine a Moun Sinai, New Yo k, NY 10029, USA
*Co espondence: [email protected]
h p://dx.doi.o g/10.1016/j.cel ep.2013.03.041
SUMMARY
S a egies o expanding hema opoie ic s em cells
(HSCs) include cocul u e wi h cells ha ecapi ula e
hei na u al mic oen i onmen , such as bone
ma ow s omal s em/p ogeni o cells (BMSCs).
Plas ic-adhe en BMSCs may be insu icien o p e-
se e p imi i e HSCs. He e, we desc ibe a me hod
o isola ing and cul u ing human BMSCs as nonad-
he en mesenchymal sphe es. Human mesen-
sphe es we e de i ed om CD45
CD31
CD71
CD146
+
CD105
+
nes in
+
cells bu could also be sim-
ply g own om e al and adul BM CD45
-en iched
cells. Human mesensphe es obus ly di e en ia ed
in o mesenchymal lineages. In cul u e condi ions
whe e hey displayed a ela i ely undi e en ia ed
pheno ype, wi h dec eased adhe ence o plas ic
and inc eased sel - enewal, hey p omo ed
enhanced expansion o co d blood CD34
+
cells
h ough sec e ed soluble ac o s. Expanded HSCs
we e se ially ansplan able in immunode icien
mice and signi ican ly inc eased long- e m human
hema opoie ic eng a men . These esul s pa e he
way o cul u e echniques ha p ese e he sel -
enewal o human BMSCs and hei abili y o suppo
unc ional HSCs.
INTRODUCTION
Hema opoie ic s em cell (HSC) ansplan a ion is ou inely pe -
o med o li esa ing p ocedu es in pa ien s wi h hema ologic
malignancies o inhe i ed me abolic/immune diso de s. Di e en
HSC sou ces, such as bone ma ow (BM), pe iphe al blood, and
co d blood, a e used o allogeneic ansplan a ion. The use o
co d blood has ecen ly been e i ed because o he sca ci y
o sui able BM dono s and because co d blood is easily and non-
in asi ely ha es ed, c yop ese ed uni s a e immedia ely a ail-
able, and pa ien s ha e a educed isk o disease ansmission
and inc eased immune ole ance. Howe e , he limi ed numbe
o HSCs p esen in co d blood has es ic ed i s use o low-
body-weigh ecipien s, whose su i al co ela es wi h HSC
dose (B oxmeye , 2011). Ex i o expansion o co d blood
HSCs would inc ease he numbe o pa ien s who would bene i
om his he apy and educe hei mo ali y isk. Howe e , HSC
expansion in cul u e emains challenging mainly due o ou
limi ed knowledge on he ac o s ha d i e HSC sel - enewal.
In i o, his hallma k p ope y o s em cells is main ained and
egula ed by a speci ic mic oen i onmen e e ed o as ‘‘niche’’
(Scho ield, 1978). Cocul u e o HSCs wi h s omal cells ha
ec ea e hei na u al niche is one o he cu en s a egies aimed
o expand co d blood HSCs. The cells ha o m he adul
mammalian HSC niche and hei speci ic unc ions in HSC
main enance a e a subjec o in ense in es iga ion (Me cie
e al., 2012). Candida e HSC niche cells include BM s omal
s em/p ogeni o cells (BMSCs; Me
´ndez-Fe e e al., 2010;
Oma su e al., 2010;Sacche i e al., 2007), which ha e been p o-
posed as a sou ce o eede cells o HSC cul u e. BMSCs a e
s ill oday e ospec i ely isola ed om p ima y human BM sam-
ples based on hei high adhe ence o plas ic and cul u ed om
ib oblas ic colony- o ming uni s (CFU-F; F iedens ein e al.,
1970). BMSCs cul u ed unde s anda d condi ions can suppo
he ex i o expansion o hema opoie ic p ogeni o s (B eems
e al., 1997;Ha ey and Dzie zak, 2004;Sha ma e al., 2011;Ve -
aillie, 1992) bu may be insu icien o p ese e HSCs. To gua -
an ee HSC eng a men in cu en clinical ials, a nonexpanded
co d blood uni (o pa o i ) is being co ansplan ed wi h an
expanded one. The esul s epo ed o da e indica e ha he
expanded co d blood uni ini ially con ibu es o hema opoiesis
bu only he nonexpanded co d blood uni is eng a ed in he
long e m (de Lima e al., 2012), sugges ing ha co d blood
CD34
+
cells ha ha e expanded on BMSCs g own unde s an-
da d condi ions migh no con ain unc ional long- e m HSCs.
Al e na i ely, he ela i ely lowe quan i y o T lymphocy es p e-
sen in he expanded uni migh also comp omise i s eng a men
(U bano-Ispizua e al., 2001).
1714 Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s
Mu ine BMSCs can be isola ed based on he exp ession o he
in e media e ilamen p o ein nes in. Mu ine nes in
+
BMSCs can
o m clonal mesenchymal sphe es (mesensphe es) ha a e able
o sel - enew and spon aneously di e en ia e in o mesenchymal
lineages bo h in i o and in i o. In he mouse, nes in
+
BMSCs
a e equi ed o main ain HSCs in he BM and a single mu ine
mesensphe e is able o ans e hema opoie ic ac i i y o an
ec opic bone ossicle (Me
´ndez-Fe e e al., 2010). Subendo he-
lial nes in
+
cells ha e been epo ed in human BM (Fe a o e al.,
2011). In his s udy, we aimed o cha ac e ize human BMSCs
ha a e able o o m mesensphe es, and assess hei abili y o
suppo HSCs. Ou esul s show ha nes in
+
BMSCs wi h p op-
e ies simila o hose p e iously cha ac e ized in he mouse a e
p esen in human e al and adul BM. When cul u ed and
expanded in condi ions ha p ese e hei imma u e pheno ype
(i.e., as loa ing mul ipo en sphe es), hey can p omo e ex i o
human umbilical co d blood HSC expansion, which may po en-
ially be o he apeu ic use.
RESULTS
Human BM CD45
CD31
CD71
CD105
+
CD146
+
Cells
Can Fo m Mesensphe es
We p e iously epo ed ha in he BM o Nes in-G p ansgenic
mice (Mignone e al., 2004), CD45
GFP
+
cells con ained all
he CFU-F and had he capaci y o o m mul ipo en and sel - e-
newing mesensphe es ha spon aneously ga e ise o mesen-
chymal lineages (Me
´ndez-Fe e e al., 2010). Because nes in
is an in e media e ilamen p o ein wi h in acellula localiza ion,
we sough o iden i y candida e cell su ace ma ke s ha would
allow isola ion o human BM nes in
+
cells. The ans o ming
g ow h ac o b(TGF-b) ecep o III endoglin/CD105 is a glyco-
p o ein ha is exp essed on he cell su ace o os eop ogeni o
cells (Aslan e al., 2006). We i s examined endoglin exp ession
in he BM o Nes in-G p ansgenic mice and ound ha wi hin
CD45
CD31
Te 119
GFP
+
cells, he CD105
+
popula ion dis-
played a highe sphe e- o ming e iciency as compa ed wi h
CD105
cells (Figu es S1A–S1C). Immunohis ochemical ana-
lyses o human BM biopsies also showed CD105 exp ession in
a subse o nes in
+
cells (Figu es 1A–1C, S1D, and S1E). The e-
o e, we so ed human BM CD45
cells acco ding o CD105
exp ession and ound ha CD45
CD105
+
, bu no CD45
CD105
cells, gene a ed sphe es (Figu es 1D and 1E) ha
we e mesensphe es, based on hei high con en o cells ha
we e able o gene a e clonal alkaline phospha ase
+
ib oblas ic
colonies (Figu e 1G) and obus mul ilineage di e en ia ion in o
os eoblas s (Figu es 1H and 1I), adipocy es (Figu es 1J and
1K), and chond ocy es (Figu es 1L–1O).
We cha ac e ized he immunopheno ype o human BM mes-
ensphe e-ini ia ing cells. In he human BM, CD45
CD31
Figu e 1. Human BM CD45
CD105
+
nes in
+
Cells Can Fo m Mesensphe es in Cul u e
Cha ac e iza ion o human BM mesensphe e-
o ming cells.
(A–C) Coexp ession o nes in and CD105/endoglin
in a subse o human BM cells. Immunohis o-
chemis y o heal hy human BM biopsy showing
exp ession o (A) NESTIN (g een), (B) CD105 ( ed),
and (C) composi e image wi h nuclei coun e -
s ained wi h DAPI (blue).
(D–F) Human BM mesensphe es de i ed om
CD45
CD105
+
cells.
(D and E) Rep esen a i e low diag ams showing
he exp ession o (D) CD45 in BM nuclea ed cells
and (E) CD105 in BM CD45
cells.
(F and G) Wi hin human BM s oma, only CD105
+
cells could o m mesensphe es (F), each o which
yielded 175 ±12 alkaline phospha ase
+
(pink)
clonal adhe en ib oblas ic colonies (G).
(H–O) Mul ilineage di e en ia ion o human BM
mesensphe es in o (H and I) os eoblas s, (J and K)
adipocy es, and (L–O) chond ocy es. qPCR o
genes associa ed wi h (H) os eoblas ic (RUNX2 and
BGLAP), (J) adipogenic (CFD and ADIPOQ) and (L)
chond ogenic (SOX9,COL2A1, and COL11A2) di -
e en ia ion is shown in human BM mesensphe es
(Mes) and hei p ogeny cul u ed o 4 weeks in
di e en ia ion media (Di ; n = 3–22). *p < 0.05; un-
pai ed wo- ailed es . E o ba s indica e SEM.
(I, K, and M–O) His ological analyses show a ully
di e en ia ed pheno ype o (I) aliza in ed
+
os eo-
blas s, (K) oil ed O
+
adipocy es, and (M) Alcian
blue
+
o (N and O) oluidine blue
+
chond ocy es.
(O) Rep esen a i e oluidine-blue-s ained pa a in
sec ion o a sphe e.
See also Figu e S1.
Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s 1715
CD146
+/low
cells ha e been epo ed o con ain all CFU-F (Sac-
che i e al., 2007;To min e al., 2011). Hence, CD45
CD31
CD71
(nonhema opoie ic/endo helial/e y h oid) cells we e
so ed acco ding o CD146 and CD105 exp ession (Figu es 2A,
2B, and S2), and compa ed in e ms o hei capaci y o o m clo-
nogenic ib oblas ic colonies and mesensphe es a clonal den-
si y. All clonogenic and sphe e- o ming capaci ies we e
es ic ed o he CD105
+
CD146
+
subse , and CD105
CD146
and CD105
+
CD146
cells did no gene a e any p ogeny. The
equency o clonogenic ib oblas ic colonies (3.4% ±2.1%)
was close o he sphe e- o ming e iciency (5.9% ±2.3%) ob-
ained om CD45
CD31
CD71
CD105
+
CD146
+
cells pla ed
a low densi y. Sphe e- o ming e iciency (12.8% ±4.7%) was
doubled when CD45
CD31
CD71
CD105
+
CD146
+
cells
we e pla ed by single-cell deposi ion (Figu es 2C–2I). The
sphe es showed a compac s uc u e wi h a hick co e con ain-
ing ib oblas ic-shaped cells embedded in a dense eosinophilic
ex acellula ma ix (Figu es 2J and 2K).
We nex add essed whe he human mesensphe es could be
ob ained om e al BM and whe he hese speci ic cul u e con-
di ions could selec o he g ow h o sphe e- o ming cells
wi hou he need o mul iple selec ion ma ke s. Fe al human
BM mononuclea cells we e immunomagne ically deple ed o
CD45
+
cells and pla ed in mesensphe e medium. Nume ous hu-
man p ima y sphe es o med a e 7–10 days in cul u e (Fig-
u e 3A) and exp essed no only NESTIN bu also high messenge
RNA (mRNA) le els o s em cell ac o (KITL) and lep in ecep o
(LEPR;Figu e 3B), which was ecen ly shown in he mouse o
ma k KITL-p oducing BMSCs ha c i ically suppo HSCs
(Ding e al., 2012). Fe al human BM mesensphe es we e capable
o obus ly di e en ia ing in o os eoblas s (Figu es 3C and 3D),
adipocy es (Figu es 3E and 3F), and chond ocy es (Figu es 3G
and 3H), p o iding u he p oo o hei BMSC o igin.
These esul s indica e ha cul u e condi ions simila o hose
p e iously op imized in he mouse (Me
´ndez-Fe e e al., 2010)
can p omo e he selec i e g ow h o human mesensphe es
om e al and adul BM s omal (immunomagne ically en iched)
CD45
cells wi hou he need o addi ional selec ion ma ke s,
al hough, in he adul human BM, mesensphe e-ini ia ing cells
a e highly en iched wi hin he CD45
CD31
CD71
CD105
+
CD146
+
cell popula ion.
Expansion o Undi e en ia ed Human BM
Mesensphe es
Adul human BM mesensphe es we e g own and expanded
unde wo di e en cul u e condi ions, i.e., in cy okine-en iched
medium con aining ei he human se um (HS) o chicken emb yo
ex ac (CEE, a cul u e supplemen ha s imula es s em cell
sel - enewal; S emple and Ande son, 1992). The CEE mesen-
sphe es we e ela i ely undi e en ia ed compa ed wi h he HS
mesensphe es, as shown by inc eased NESTIN exp ession and
educed mRNA le els o genes equi ed o os eoblas ic and adi-
pocy ic di e en ia ion (Figu e 4A). The sphe es could be se ially
expanded unde bo h cul u e condi ions, al hough he CEE mes-
ensphe es gene a ed mo e sphe es o e se ial passages (Fig-
u e 4B; 550 e sus 320 sphe es a passage 3). In addi ion,
whe eas he CEE mesensphe es emained in suspension and
mo phologically iden ical o e >2 mon hs in cul u e (Figu e 4C),
he HS sphe es a ached e en o ul alow-adhe ence dishes
and disappea ed in he long e m as a esul o cell ou g ow h
om he sphe es as an adhe en monolaye (Figu e 4D).
Expanded CEE mesensphe e- o ming cells main ained exp es-
sion o CD105 and CD146 (17% o sphe e- o ming cells; Fig-
u e 4E) and did no exp ess CD45, CD31, o CD71 (da a no
Figu e 2. Human BM Mesensphe es A e De i ed om CD45
CD31
CD71
CD105
+
CD146
+
Cells
(A and B) Immunopheno ype o human BM mesensphe e- o ming cells.
Heal hy human BM CD45
CD31
CD71
cells (A) we e so ed acco ding o
CD105 and CD146 exp ession in o h ee popula ions (B).
(C–E) F equency o (C) clonogenic ib oblas ic colonies and (D and E) sphe es
de i ed om human BM nonhema opoie ic/endo helial/e y h oid CD105
+
CD146
+
, CD105
+
CD146
, and CD105
CD146
cells pla ed (C and D)
a clonal densi y o (E) by single-cell deposi ion. The equency is ela ed o BM
mononuclea cells a e sepa a ion wi h he Rose eSep Ki (S emCell Tech-
nologies) (n = 3 dono s; e o ba s indica e SEM).
(F–I) Rep esen a i e examples o human BM mesensphe es ob ained by
(F and G) single-cell deposi ion o (H and I) bulk cul u e a low densi y.
(J and K) His ology o H&E-s ained pa a in sec ions o human BM mesen-
sphe es.
Scale ba s, 200 mm (F, G, and I), 500 mm (H), and 100 mm (J and K). See also
Figu e S2.
1716 Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s
shown). A e being diges ed and epla ed h ee imes, mo e han
hal o he CEE mesensphe es di e en ia ed in o os eoblas s and
adipocy es, and all o he sphe es ha we e analyzed (n = 20)
di e en ia ed in o oluidine blue
+
chond ocy es (da a no shown),
indica ing p ese a ion o mul ipo ency unde hese cul u e
condi ions (Figu e 4F). To es whe he human mesensphe es
we e capable o in i o main enance and hema opoie ic suppo ,
single p ima y sphe es we e allowed o a ach on o phosphocal-
cic ce amic ossicles and hen implan ed subcu aneously in
immunode icien mice, as p e iously desc ibed (Me
´ndez-Fe e
e al., 2010). Ossicles we e ha es ed a e 2 mon hs and enzy-
ma ically diges ed, and he eco e ed cells we e isola ed by luo-
escence-ac i a ed cell so ing (FACS) acco ding o exp ession
o mCD45, hCD146, and hCD105, and subcul u ed a clonal den-
si y in mesensphe e- o ming medium (Figu es S3A and S3B). The
sphe e- o ming e iciency o hCD105
+
/hCD146
+
cells (21% ±
8%) was 80- old highe han ha o hCD105
hCD146
cells
(0.3% ±0.2%) eco e ed om he ossicles (Figu e S3C). Al o-
ge he , hese esul s indica e p ese ed mul ipo ency and main-
enance o human BMSCs g own as mesensphe es.
We pe o med a quan i a i e p o eomics analysis o he supe -
na an s o CEE and HS sphe es o iden i y he mos abundan
Figu e 3. Human Mesensphe es Can Be
De i ed om Fe al BM
(A) Rep esen a i e sphe es om immunomagne i-
cally en iched e al human BM CD45
cells.
(B) Fe al human BM mesensphe es exp ess
NESTIN, s em cell ac o (KITL), and lep in ecep o
(LEPR) mRNA; qPCR, n = 3.
(C–H) Mul ilineage di e en ia ion o e al human
BM mesensphe es. Sphe es we e enzyma ically
dissocia ed and cul u ed o 3 weeks in (C and D)
os eoblas ic, (E and F) adipocy ic, and (G and H)
chond ocy ic di e en ia ion media.
(C, E, and G) qPCR o genes associa ed wi h (C)
os eoblas ic (os e ix, SP7; os eoglycin, OGN; Run -
ela ed ansc ip ion ac o 2, RUNX2; os eoac i in,
GPNMB; os eocalcin, BGLAP), (E) adipogenic
(pe oxisome p oli e a o -ac i a ed ecep o
gamma, PPARG; adipsin, CFD; adiponec in,
ADIPOQ), and (G) chond ogenic (sex de e mining
egion Y-box 9, SOX9; collagen 11a2, COL11A2;
collagen 2a1, COL2A1) di e en ia ion (n = 3). E o
ba s in (B), (C), (E), and (G) indica e SEM.
(D, F, and H) His ological analyses showing a
di e en ia ed pheno ype o (D) aliza in ed
+
os e-
oblas s, (F) oil ed O
+
adipocy es, and (H) Alcian
blue
+
chond ocy es. Scale ba , 100 mm.
p o eins ha we e di e en ially p oduced
by human BMSCs unde bo h condi ions.
The quan i ied p o eins we e subjec ed o
a sys ems biology compa a i e analysis.
HS mesensphe es displayed highe
exp ession o p o eins in ol ed in me a-
bolic p ocesses, gene exp ession and
ansla ion, mi ochond ial and nucleola
unc ion, and ibonucleop o ein complex
and ac in cy oskele on o ma ion. In
con as , CEE mesensphe es p oduced mo e p o eins in ol ed
in cell adhesion, collagen and basemen memb ane p oduc ion,
ex acellula ma ix-p o ein in e ac ion, axon guidance, esponse
o hypoxia, wounding, and o ganic cyclic compound p oduc ion
(Figu e 4G). CEE mesensphe es also p oduced mo e p o eins
in ol ed in angiogenesis, sec e o y g anule p oduc ion, and
signal ansduc ion, as well as mo e glycosyla ed and plasma in-
eg al p o eins. In e es ingly, HS sphe e-de i ed sec e ome con-
ained mo e p o eins in ol ed in cell-cycle p og ession, whe eas
CEE mesensphe es exp essed mo e p o eins in ol ed in cell-
cycle a es , such as hose in ol ed in TGF-bsignaling (Figu e 4H;
Table S1). These esul s indica e ha nonadhe en human mes-
ensphe es a e quali a i ely dis inc bu may change hei pheno-
ype upon adhe ence o plas ic. When cul u ed in condi ions ha
educe hei a achmen o he dish, hey display educed p oli -
e a ion and me abolic a e, bu inc eased main enance and p o-
duc ion o ac i e molecules as compa ed wi h mo e-di e en i-
a ed BMSCs ha exhibi inc eased adhesion o plas ic.
Human BM Mesensphe es Suppo HSCs
Ou p e ious s udies indica ed ha undi e en ia ed nes in
+
BMSCs ha e essen ial unc ions in he egula ion o mu ine
Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s 1717
HSCs (Me
´ndez-Fe e e al., 2010). The e o e, we s udied
whe he cul u e condi ions ha be e p ese e BMSC sel -
enewal could also enhance hei suppo o HSCs. Cocul u e
o 100 esh CEE mesensphe es wi h 2 310
5
co d-blood-
de i ed CD34
+
cells did no exceed he cell expansion achie ed
only wi h cy okines, bu doubled he numbe o p imi i e hema o-
poie ic p ogeni o s, as measu ed using he long- e m cul u e-
ini ia ing cell assay (LTCIC; Figu e S4A), sugges ing ha human
BM mesensphe es could sus ain co d blood HSC sel - enewal.
Indeed, he equency o co d-blood-de i ed HSCs, de ined as
CD34
+
CD38
CD49
+
cells (No a e al., 2011), was al eady
inc eased 3 days a e cocul u e wi h human BM mesensphe es
(Figu e S4B).
Imma u e Human BM Mesensphe es Expand
Mul ipo en and Long-Te m Eng a ing Co d Blood HSCs
We di ec ly compa ed he abili y o undi e en ia ed (CEE) and
mo e-di e en ia ed (HS) mesensphe es o expand unc ional
HSCs capable o eng a ing in o immunode icien mice. Co d
Figu e 4. Expansion o Sel -Renewing Hu-
man BM Mesensphe es
(A) qPCR analyses o he exp ession o nes in
(NES) o di e en ia ion ma ke s (RUNX2 and CFD)
in human BM mesensphe es cul u ed wi h CEE
( ed) o HS (black), no malized o GAPDH.
(B–F) In i o sel - enewal o human mesen-
sphe es, showing he numbe o CEE ( ed) and HS
(black) BM sphe es o e h ee passages. *p < 0.05;
unpai ed wo- ailed es ; e o ba s indica e SEM.
(C and D) CEE mesensphe es (C) we e p ese ed
o e passages, in con as o HS sphe es (D),
which a ached e en o ul alow-adhe ence
plas ic.
(E) Rep esen a i e FACS his og ams o CEE
mesensphe es expanded o e ou passages,
which emained CD105
+
( ed) and CD146
+
(blue);
g ay, con ol iso ype s aining.
(F) A e h ee passages, 55% o mesensphe es
(11/20) di e en ia ed in o wo o mo e mesen-
chymal lineages.
(G and H) Compa ison o he sec e ome o CEE
and HS mesensphe es by quan i a i e p o eomics.
The cumula i e equency dis ibu ions o s an-
da dized log
2
a ios (Zq) o p o eins g ouped in o
on ological ca ego ies om GO, KEGG, o
REACTOME da abases show he coo dina ed in-
c ease o dec ease o p o eins belonging o hese
ca ego ies. Rela ed on ological ca ego ies ha e
been con en ionally dis ibu ed in o wo g aphs o
acili a e iewing o he pa hways en iched in CEE
(Zq > 0, igh shi o sigmoidal cu es) o HS
(Zq < 0, le shi o cu es) sphe es when
compa ed wi h he p edic ed cumula i e no mali y
plo o he s anda dized a iable a he p o ein le el
o all p o eins (black cu es).
See also Figu e S3 and Table S1.
blood CD34
+
cells we e cul u ed alone
wi h cy okines o in he p esence o
expanded HS o CEE adul BM mesen-
sphe es (Figu e 5A). The HS and CEE
mesensphe es espec i ely yielded a 6- and 40- old expansion
o co d blood CD34
+
cells as compa ed wi h he con ol medium
(Figu es 5B, 5C, and S5A). The capaci y o cul u ed human
CD34
+
cells o con ibu e o hema opoiesis was es ed by ans-
plan a ion in o immunode icien mice. Co d blood CD34
+
cells
expanded in cocul u e wi h HS mesensphe es did no inc ease
he p esence o human CD45/CD71
+
hema opoie ic cells in he
BM o NOD/SCID mice when compa ed wi h CD34
+
cells
cul u ed only wi h cy okines. In con as , co d blood CD34
+
cells
expanded in cocul u e wi h CEE mesensphe es yielded a 38- old
inc ease in BM human hema opoie ic chime ism o NOD/SCID
mice 2 mon hs la e (a e age 6.1% eng a men ; Figu es 5D
and S5B). In addi ion, subcu aneous implan a ion o 2 310
6
co-
cul u ed hema opoie ic cells did no gene a e e a omas in NOD/
SCID mice (da a no shown).
We es ed he sel - enewal and mul ilineage di e en ia ion po-
en ial o expanded HSCs by ansplan a ion in o NOD/SCID/
gamma (NSG) mice. Two mon hs a e ansplan a ion in o
NSG mice, co d blood CD34
+
cells expanded in cocul u e wi h
1718 Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s
CEE mesensphe es yielded a 12.5- old inc ease in human he-
ma opoie ic chime ism o NSG BM when compa ed wi h
CD34
+
cells cul u ed only wi h cy okines. Al hough only 40% o
mice (2/5) ansplan ed wi h CD34
+
cells cul u ed alone showed
human hema opoie ic eng a men abo e he 0.1% h eshold,
62.5% (5/8) showed eng a men o CD34
+
cells ha had been
cocul u ed wi h CEE mesensphe es (Figu es 5E and 5F).
Expanded co d blood CD34
+
cells con ained unc ional HSCs
capable o accomplishing mul ilineage di e en ia ion in o
B-lymphoid (Figu e 5G), myeloid, and e y h oid cells (Figu e 5H).
We es ed he capaci y o cul u ed HSCs o eng a in o second-
a y ecipien mice. BM cells om mice ha showed a simila
hCD45 eng a men a e ansplan a ion o CD34
+
cells cul u ed
alone o in he p esence o CEE sphe es we e used o second-
a y ansplan s. The BM eng a men o hCD45
+
cells 2 mon hs
a e seconda y ansplan a ion demons a es ha expanded
co d blood CD34
+
cells ha yielded a simila eng a men in p i-
ma y ecipien s we e equally able o econs i u e seconda y
ecipien mice (Figu e 5I).
Human BM Mesensphe es Expand Long-Te m Co d
Blood HSCs h ough Soluble Sec e ed Fac o s
Al hough some s udies ha e sugges ed ha cell con ac be-
ween BMSCs and hema opoie ic p ogeni o s is equi ed o
expand he la e (Ha ey and Dzie zak, 2004;Kawada e al.,
1999;Thiemann e al., 1998), o he s udies ha e shown an e ec
ha is la gely independen o di ec cell con ac (B eems e al.,
1997;Ve aillie, 1992). We es ed he capaci y o expanded hu-
man mesensphe es o suppo co d blood CD34
+
cells in di ec
cocul u e o sepa a ed by answell. A simila expansion o co d
Figu e 5. Inc eased HSC Eng a men in
Immunode icien Mice om Co d Blood
CD34
+
Cells Cocul u ed wi h P imi i e
Human BM Mesensphe es
(A) Co d blood CD34
+
cells we e cul u ed in se um-
ee medium con aining cy okines in he absence
(Alone, black) o p esence o human BM mesen-
sphe es p e iously expanded wi h HS (blue) o
CEE ( ed).
(B and C) HS and CEE mesensphe es p e iously
expanded o e 4 weeks espec i ely yielded 6-
and 40- old expansion o co d blood CD34
+
cells
as compa ed wi h CD34
+
cells cul u ed alone (A).
(C) Numbe o co d blood CD34
+
cells measu ed
by FACS a e 16 days o cul u e (n = 3).
(D–I) CEE mesensphe es can expand human HSCs
ha a e capable o mul ilineage econs i u ion and
se ial eng a men in immunode icien mice.
(D) Pe cen age o human CD45
+
/CD71
+
cells in
he BM o NOD/SCID mice 2 mon hs a e
ansplan a ion o 10
4
eshly hawed (F) co d
blood CD34
+
cells o hei p ogeny a e 16 days
o cul u e wi h cy okine-supplemen ed se um-
ee medium alone (A) o in he p esence o
HS (+HS) o CEE (+CEE) sphe es; n = 3;
*p < 0.05; unpai ed wo- ailed es . E o ba s
indica e SEM.
(E) Pe cen age o human CD45
+
cells in he BM o
NSG mice 2 mon hs a e ansplan a ion o he
p ogeny o non ozen 10
4
co d blood CD34
+
cells
cul u ed o 2 weeks in cy okine-supplemen ed
se um- ee medium alone (A) o in he p esence o
CEE (+CEE) sphe es.
(F) Rep esen a i e FACS diag ams o hCD45-
s ained BM cells om NSG mice ansplan ed wi h
CD34
+
cells ha we e cul u ed alone (A) o in he
p esence o CEE (+CEE) sphe es.
(G and H) Mul ilineage di e en ia ion o co d
blood CD34
+
cells in NSG mice. Rep esen a i e
FACS diag ams o BM cells 2 mon hs a e
ansplan a ion o cul u ed co d blood CD34
+
cells in NSG mice. BM cells we e s ained
wi h (G) an i-hCD45 (hema opoie ic) and an i-hCD19 (B-lymphoid) o wi h (H) an i-hCD33 (myeloid) and an i-hCD235a/glycopho in A (e y h oid) an ibodies.
(I) Se ial econs i u ion o expanded co d blood CD34
+
cells in NSG mice. Seconda y ansplan o BM cells ha es ed om mice ha showed a simila hCD45
eng a men 2 mon hs a e ansplan a ion o CD34
+
cells ha we e cul u ed alone (A) o in he p esence o CEE (+CEE) sphe es.
(E and I) Following p e ious s udies, eng a men was conside ed posi i e ( ed do s) when R0.1% hCD45 cells we e eng a ed (g ay do ed line) and nega i e
(black do s) below his h eshold.
See also Figu es S4 and S5, and Table S2.
Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s 1719
blood CD34
+
cells was de ec ed in he di ec and answell co-
cul u es wi h CEE mesensphe es (Figu es 6A–6E), sugges ing
ha soluble ac o s sec e ed by he sphe es migh be su icien
o p omo e HSC expansion. We compa ed he di ec and ans-
well cocul u es in e ms o hei capaci y o expand co d blood
HSCs wi h long- e m eng a men in immunode icien mice.
Co d blood CD34
+
cells ha had been cocul u ed o 3 weeks
wi h human mesensphe es, ei he di ec ly o sepa a ed by ans-
well, we e i. . ansplan ed in o NSG mice. S ikingly, a 4- old
highe human hema opoie ic chime ism was de ec ed 16 weeks
la e in he BM o mice ansplan ed wi h co d blood cells ha
had been cocul u ed using answell as compa ed wi h di ec
cocul u e. No ably, human chime ism was >20% in 50% o
mice om he answell g oup (Figu es 6F and 6G). Long- e m
eng a men om cocul u ed co d blood HSCs was also associ-
a ed wi h mul ilineage econs i u ion (Figu e 6H). These esul s
indica ed ha soluble ac o s sec e ed om human BM mesen-
sphe es a e esponsible o HSC expansion. The e o e, we pe -
o med high- h oughpu mass spec ome y analyses o p o eins
sec e ed by CEE human mesensphe es o be e cha ac e ize
CEE-mesensphe e- o ming cells and candida e HSC-suppo -
ing ac o s. These analyses iden i ied >1,100 p o eins in h ee
independen samples. Po en ially sec e ed p o eins we e
de e mined and quan i a ed as p e iously desc ibed (Bonzon-
Kulichenko e al., 2011). P o eins o he TGF-bpa hway ha ,
simila ly o some o hei a ge genes (e.g., os eopon in), ha e
been epo ed o induce HSC quiescence (La sson and Ka ls-
son, 2005;Nilsson e al., 2005;S ie e al., 2005) we e also de-
ec ed. G ow h ac o s such as hepa ocy e, basic ib oblas ic
and connec i e issue g ow h ac o s, and o he molecules ha
ha e been shown o suppo HSCs, such as insulin-like g ow h
ac o (IGF), angiopoie in-like g ow h ac o (Zhang e al.,
2008), and p os aglandin- ela ed p o eins (Goessling e al.,
2011;Hogga e al., 2009;No h e al., 2007), we e also de ec ed
and may ep esen o he candida e ac o s ha suppo HSCs.
The mos abundan sec e ed p o eins we e hose in ol ed in
basemen memb ane and ex acellula ma ix o ma ion, such
as collagens, laminins, ib onec in, p o eoglycans, nidogens,
Figu e 6. Human BM Mesensphe es Sup-
po Co d Blood HSCs Mainly h ough
Sec e ed Fac o s
(A–C) Co d blood CD34
+
cells we e cul u ed o
16 days in se um- ee medium con aining cy o-
kines (A) in he absence o o (B and C) in cocul u e
wi h human e al mesensphe es, ei he (B) in di ec
con ac o (C) sepa a ed by answell.
(D) G ow h cu e o co d blood CD34-en iched
cells cul u ed alone (black line), in con ac wi h
human mesensphe es (blue line) o sepa a ed om
hem by answell ( ed line).
(E) Numbe o CD34
+
cells measu ed by FACS
(n = 3; e o ba s indica e SEM).
(F–H) Soluble sec e ed ac o s p oduced by
human BM mesensphe es expand co d blood
HSCs capable o long- e m econs i u ion and
mul ilineage di e en ia ion in immunode icien
mice.
(F) Pe cen age o human CD45
+
cells in he BM o
NSG mice 16 weeks a e ansplan a ion o co d
blood CD34
+
cells cul u ed o 3 weeks wi h
human BM mesensphe es in di ec con ac o
sepa a ed by answell. *p < 0.05; unpai ed wo-
ailed es .
(G) Rep esen a i e FACS diag ams o hCD45-
s ained BM cells om NSG mice 4 mon hs a e
ansplan a ion o CD34
+
cells cul u ed in bo h
condi ions.
(H) Rep esen a i e FACS diag am showing long-
e m mul ilineage econs i u ion. BM cells we e
s ained wi h an i-hCD45 (hema opoie ic), an i-
hCD19 (B-lymphoid), and an i-hCD33 (myeloid)
an ibodies.
(I) P o ein-p o ein in e ac ions p edic ed om
he lis o p o eins de ec ed in he sec e ome
o human CEE BM mesensphe es (n = 3 inde-
penden dono s). Only p o eins o which wo
o mo e pep ides we e de ec ed in wo o
mo e samples we e conside ed o he analyses using STRING 9.0. In e ac ions we e iden i ied om expe imen al da a and da abases using a high
con idence le el (0.700). The clus e ing was pe o med using he MCL algo i hm (in la ion pa ame e se a 2).
(J) En ichmen in ele an canonical pa hways de ec ed using Ingenui y Pa hway Analysis.
See also Figu es S4 and S5.
1720 Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s
and ma ix me allop o einases. In eg ins and o he ela ed p o-
eins, such as inculin and alin, we e also abundan . Di e en
calcium anspo and me abolism p o eins, as well as neu al
guidance/ou g ow h and no epineph ine ca abolism p o eins,
we e also de ec ed. The BMSC ma ke s ound in hese samples
included endoglin, CD90, CD276, VCAM-1, CD44, and CD151.
Filamen p o eins associa ed wi h nes in, such as imen in, des-
min, plec in, and ilamin-A, we e abundan in hese samples (Fig-
u es 6I and 6J; Table S2). These esul s show ha human BM
mesensphe es sec e e many di e en p o eins ha a e able o
a ec HSC unc ion, and sugges ha mos likely mul iple ac o s
sec e ed by hem con ibu e o main ain HSCs in cul u e.
In summa y, his s udy demons a es ha BMSCs wi h ea-
u es simila o hose p e iously cha ac e ized in he mouse
(Me
´ndez-Fe e e al., 2010) a e p esen in human e al and adul
BM. When expanded in no el nonadhe en condi ions ha p e-
se e hei imma u e pheno ype, hey can p omo e ex i o
HSC sel - enewal and expansion, which may po en ially be o
he apeu ic use.
DISCUSSION
We desc ibe he e a me hod o isola ing and cul u ing human
BMSCs in a way ha be e p ese es hei sel - enewal and
HSC niche unc ions. O e he pas ew yea s, di e en s a e-
gies ha e been used o ex i o expansion o HSCs, especially
hose de i ed om co d blood. These include ea men s wi h
angiopoie in-like p o eins, IGF-binding p o ein 2 (Zhang e al.,
2008), he No ch ligand Del a1 (Delaney e al., 2010), an agonis s
o he a yl hyd oca bon ecep o (Boi ano e al., 2010), dime hyl-
p os aglandin E2 (Goessling e al., 2011;Hogga e al., 2009;
No h e al., 2007), and inhibi o s o apop o ic p o eases (San-
gee ha e al., 2010). BM eng a men o co d blood HSCs is
also accele a ed a e inhibi ion o CD26/dipep idylpep idase IV
(Campbell e al., 2007;Ch is ophe son e al., 2007) and ex i o
ucosyla ion (Robinson e al., 2012).
BMSCs cul u ed unde s anda d adhe en condi ions ha e
been shown o suppo he ex i o expansion o hema opoie ic
p ogeni o s (B eems e al., 1997;Ha ey and Dzie zak, 2004;
Sha ma e al., 2011;Ve aillie, 1992) bu may be insu icien o p e-
se e p imi i e HSCs (de Lima e al., 2012). Howe e , BMSCs a e
s ill e ospec i ely isola ed based on hei high plas ic adhe ence
and/o cul u ed in minimal medium supplemen ed wi h se um.
We p e iously showed ha sel - enewing mu ine mesen-
sphe es a e able o ans e hema opoie ic ac i i y o an engi-
nee ed bone sca old (Me
´ndez-Fe e e al., 2010). In he cu en
s udy, we cha ac e ized a human BM cell popula ion ha was
able o o m mesensphe es, and assessed i s capaci y o sup-
po co d blood HSCs. Ou esul s demons a e ha nes in
+
BMSCs wi h p ope ies simila o hose p e iously cha ac e ized
in he mouse a e p esen in human e al and adul BM. We did no
need o use speci ic su ace ma ke s o g ow human mesen-
sphe es, because we could simply de i e hem om immuno-
magne ically en iched CD45
cells using a speci ic cul u e
medium. Like colony-ini ia ing cells (CFU-F), mesensphe e-
o ming cells we e highly en iched in he CD45
CD31
CD71
CD105
+
CD146
+
popula ion. A subse o he cells con ained in
he sphe es emained CD105
+
and CD146
+
o e passages.
Al hough only a small numbe o p ima y mesensphe es could
be ob ained om human BM aspi a es, hey could be p opaga ed
o e 2 mon hs du ing mul iple di isions. No ably, sphe es
cul u ed wi h CEE showed inc eased in i o sel - enewal as
compa ed wi h hose cul u ed wi h HS. Mo eo e , he p esence
o HS ins ead o CEE in he medium led o educed nes in exp es-
sion, up egula ion o di e en ia ion ma ke s, and cell a achmen ,
e en o ul alow-adhe en plas ic. In ag eemen wi h his, o he
s udies ha e sugges ed ha nonadhe en human BM mesen-
chymal p ogeni o s display a highe di e en ia ion po en ial
(Baksh e al.,2003) and canno be main ained wi h se um (DiMag-
gio e al., 2012). These changes we e associa ed wi h a ma ked
di e ence in HSC suppo . Al hough cocul u e wi h HS sphe es
inc eased he co d blood CD34
+
cell yield, only CEE sphe es
could inc ease he eng a men in immunode icien mice.
No ably, co d blood HSC expansion was igge ed by soluble
sec e ed ac o s p oduced by CEE mesensphe es. Al hough
some s udies ha e sugges ed ha cell con ac be ween BMSCs
and hema opoie ic p ogeni o s is equi ed o expand he la e
(Ha ey and Dzie zak, 2004;Kawada e al., 1999;Thiemann
e al., 1998), o he s udies ha e sugges ed ha di ec cell-cell
in e ac ion is no equi ed (B eems e al., 1997;Ve aillie,
1992). Ou esul s suppo he las con en ion, since a 4- old
highe long- e m econs i u ion was de ec ed om co d blood
cells cocul u ed wi h human mesensphe es ha we e sepa a ed
by answell. This was also associa ed wi h a highe eco e y o
mesensphe es cocul u ed on he answell compa ed wi h hose
di ec ly cocul u ed using ul alow-adhe en plas ic dishes (da a
no shown). Since he cocul u e medium in all cases was a he-
ma opoie ic-suppo ing medium a he han a mesensphe e me-
dium, hese di e ences sugges ha he me e a achmen o he
sphe es o he plas ic dish du ing he di ec cocul u e changed
hei pheno ype and impai ed hei abili y o sec e e ac o s
capable o expanding HSCs.
Candida e ac o s ha we e abundan ly p oduced by CEE
sphe es and could con ibu e o HSC expansion include p o eins
o he TGF-bpa hway ha ha e been epo ed o p ese e HSC
quiescence (La sson and Ka lsson, 2005;Nilsson e al., 2005;
S ie e al., 2005); hepa ocy e, basic ib oblas ic and connec i e
issue g ow h ac o s, IGF, and angiopoie in-like g ow h ac o
(Zhang e al., 2008); and p os aglandin- ela ed p o eins (Goes-
sling e al., 2011;Hogga e al., 2009;No h e al., 2007). The e-
o e, mul iple ac o s could con ibu e o he enhanced HSC sup-
po displayed by CEE mesensphe es. Elucida ion o c i ical
media o s in he u u e migh acili a e he op imiza ion o chem-
ically de ined media o human HSC expansion. The abili y o
expand co d blood HSCs in answell cocul u e wi h human
mesensphe es migh inc ease hei sa e y in humans and acili-
a e es ing o hei po en ial clinical bene i . In addi ion, cells
ha expanded unde hese condi ions did no gene a e e a-
omas in immunode icien mice.
In summa y, hese esul s demons a e ha p imi i e nes in
+
BMSCs a e p esen in human e al and adul BM. Wi hou he
need o complex isola ion p o ocols, hey can gi e ise o sel -
enewing mesensphe es wi h a mo e p imi i e pheno ype and
enhanced ex i o HSC suppo as compa ed wi h mo e-di e en-
ia ed BMSCs. These s udies pa e he way o new isola ion and
cul u e s a egies o be e p ese e he sel - enewal o human
Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s 1721
BMSCs and hei abili y o suppo HSCs, which will c i ically
de e mine he success o hese app oaches in he clinical a ena.
EXPERIMENTAL PROCEDURES
Mouse S ains
NOD.CB17-P kdc
scid
/J, NOD/L Sz-scidIL2Rg
null
(Jackson Labo a o ies) and
Nes in-G p (Mignone e al., 2004) ansgenic mice we e used in hese s udies.
Expe imen al p ocedu es we e app o ed by he Animal Ca e and Use Commi -
ee o he Spanish Na ional Cen e o Ca dio ascula Resea ch.
Cell Isola ion and Cul u e
BM aspi a es we e ob ained in he Hospi al Clı
´nic (Uni e si y o Ba celona,
Spain) and Ska
˚ne Uni e si y Hospi al (Lund, Sweden) wi h w i en consen
om heal hy indi iduals, dono s o BM hema opoie ic p ogeni o cells o allo-
geneic ansplan a ion o pa ien s ee o hema ological diseases. The samples
ob ained we e p ocessed in wo di e en ways. Fo so ing expe imen s,
human BM mononuclea cells we e en iched om 60 ml human BM aspi a es
using he Rose eSep ki (S emCell Technologies). Cells we e s ained wi h
7-Aminoac inomycin D (7AAD) solu ion (Sigma) and he ollowing mouse
an i-human an ibodies (BD Pha Mingen): CD31-FITC (clone WM59), CD45-
FITC (clone 2D1), CD71-FITC (clone M-A712), CD105-APC (clone 266), and
CD146-PE (clone P1H12). 7AAD
CD45
CD31
CD71
cells we e isola ed
acco ding o CD105 and CD146 exp ession in o CD105
+
CD146
+
, CD105
+
CD146
and CD105
CD146
cells. Fo o he expe imen s, he il e s used
o ap bone spicules and cell agg ega es we e ca e ully and asep ically
washed wi h cold PBS se e al imes. Cells we e eco e ed by cen i uga ion
and e y h ocy es we e lysed wi h 0.8% NH
4
Cl. A e a wash and cen i uga ion,
he pelle was enzyma ically dissocia ed in a solu ion con aining 0.25% ype I
collagenase and 20% e al bo ine se um (FBS) in PBS (S emCell Technolo-
gies) o 30 min a 37C, wi h mechanical agi a ion e e y 10 min.
Human e al BM samples we e ob ained om Leiden Uni e si y Medical
Cen e (The Ne he lands) unde a p o ocol app o ed by he e hics commi ee
o ha ins i u ion. Fe al BM samples we e mechanically dispe sed and he
mononuclea cells we e isola ed by Ficoll g adien , ozen, and kep in liquid
ni ogen. Cells we e hawed, washed wi h PBS once, and s ained wi h an i-
bodies o immunomagne ic en ichmen using an i-CD45 magne ic beads
(Mil enyi Bio ec) acco ding o he manu ac u e ’s ecommenda ions.
Fo sphe e o ma ion, so ed cells we e pla ed a low densi y (<1,000
cells/cm
2
) o by single-cell deposi ion in ul alow-adhe ence 35 mm dishes
(S emCell Technologies). The g ow h medium o bo h CEE and HS sphe es
con ained 0.1 mM b-me cap oe hanol; 1% nonessen ial amino acids (Sigma);
1% N2 and 2% B27 supplemen s (In i ogen); ecombinan human ib oblas
g ow h ac o (FGF)-basic, ecombinan human epide mal g ow h ac o
(EGF), ecombinan human pla ele -de i ed g ow h ac o (PDGF-AB), ecom-
binan human oncos a in M (227 aa OSM, 20 ng/ml) and ecombinan human
IGF-1 (40 ng/ml; Pep o ech) in Dulbecco’s modi ied Eagle’s medium (DMEM)/
F12 (1:1) / human endo helial (1:2) se um- ee medium (In i ogen). CEE
medium was supplemen ed wi h 15% CEE p epa ed as desc ibed p e iously
(Paj le e al., 2010;S emple and Ande son, 1992), and HS medium was
ins ead supplemen ed wi h 10% HS. The cul u es we e kep a 37C wi h
5% CO
2
, 20% O
2
in a wa e -jacke ed incuba o and le un ouched o
1 week o p e en cell agg ega ion. A e wa d, hal -medium changes we e
pe o med wice a week. Fo passage, sphe es we e enzyma ically dissoci-
a ed in 100 ml o a solu ion con aining 0.25% ype I collagenase and 20%
FBS in PBS (S emCell Technologies) o 30 min a 37C, wi h mechanical
dispe sion e e y 10 min. Cells we e washed wi h PBS once and epla ed
wi h mesensphe e medium in ul alow-adhe ence 35 mm dishes (S emCell
Technologies) a 37C in a wa e -jacke ed incuba o wi h 5% CO
2
and 20% O
2
.
Human CD34
+
cells pooled om co d blood samples we e pu chased
(S emCell Technologies) o isola ed as ollows: Co d blood samples we e ob-
ained in acco dance wi h p ocedu es app o ed by he ins i u ional e hics
commi ee. Two o h ee co d blood uni s we e pooled and e y h ocy es
we e emo ed by Ficoll-Paque densi y g adien cen i uga ion. CD34
+
cells
we e isola ed by magne ic cell selec ion using an i-hCD34 magne ic mic obe-
ads (Mil enyi) and ozen immedia ely a e pu i ica ion. Co d blood CD34
+
cells we e cul u ed as p e iously desc ibed (Zhang e al., 2006) in S emSpan
se um- ee medium (S emCell Technologies) supplemen ed wi h 10 mg/ml
hepa in (Sigma), ecombinan human s em cell ac o , ecombinan human
FGF-1 (10 ng/ml), ecombinan human h ombopoie in (20 ng/ml; Pep o ech),
and 1% penicillin-s ep omycin (In i ogen) a 37C in a wa e -jacke ed incu-
ba o wi h 5% CO
2
. Cul u e dishes we e dilu ed 1:1 wi h esh medium wo
o h ee imes a week o main ain cell densi y. Sphe es we e di ec ly added
o he cul u e dishes o placed in he uppe chambe o a 0.2 mm po e polyca -
bona e answell (Co ning) in he same medium. Sphe es we e cocul u ed wi h
CD34
+
cells a a s a ing a io o 1:1,000.
Immunos aining and His ology
Human BM biopsies we e ob ained in he Clı
´nic Hospi al (Ba celona, Spain)
om heal hy dono s wi h hei w i en consen . The biopsy specimens we e
embedded in pa a in, sec ioned wi h a mic o ome, and moun ed on o slides.
A e depa a iniza ion wi h xylol, he sec ions we e ehyd a ed and washed
wi h PBS. Sec ions we e washed and blocked wi h blocking bu e (PBS
wi h 0.1% T i on X-100, 10% donkey se um) o 30 min and incuba ed o e -
nigh a 4C wi h p ima y an ibodies in blocking bu e ( abbi an i-human Nes-
in [Millipo e 1:200], goa an i-human endoglin [R&D 1:20] o con ol abbi o
goa immunoglobulin G [IgG] a he same concen a ions). The samples
we e washed in PBS se e al imes and incuba ed wi h species-speci ic luo-
escen seconda y an ibodies (Alexa 488-conjuga ed donkey an i- abbi IgG
[In i ogen] and Cy3-conjuga ed donkey an i-goa IgG [Jackson Labo a-
o ies]). The samples we e hen washed in PBS wi h 0.1% T i on X-100 and
moun ed (Vec ashield luo escen moun ing medium wi h DAPI; Vec o
Labs). Images we e cap u ed using a con ocal mic oscope (Sp5; Leica) wi h
acquisi ion so wa e (LAS AF; Leica). Fo his ological analysis, human mesen-
sphe es we e ixed in suspension wi h 2% pa a o maldehyde in PBS and
embedded in pa a in o sec ioning. Mic o ome sec ions we e p ocessed
wi h ou ine hema oxylin and eosin (H&E) s aining and images we e acqui ed
using a Zeiss Axio e 200 mic oscope.
In Vi o Di e en ia ion
Os eoblas ic di e en ia ion was induced by cul u ing he cells o 4 weeks wi h
50 mg/ml L-asco bic acid 2-phospha e, 10 mM glyce ophospha e (Sigma), and
15% FBS in a-MEM wi h 1% penicillin-s ep omycin (In i ogen). Adipocy e
di e en ia ion was induced wi h 1 mM dexame hasone, 10 mg/ml insulin
(Sigma), and 10% FBS in a-MEM wi h 1% penicillin-s ep omycin (In i ogen).
Chond ogenic di e en ia ion was induced in cell pelle s wi h 10
7
M dexa-
me hasone, 10
4
M L-asco bic acid 2-phospha e, 1 mM sodium py u a e,
nonessen ial amino acids, 13SITE+3 supplemen (1 mg/ml human insulin,
0.55 mg/ml human ans e in, 0.5 mg/ml sodium seleni e, 0.2 mg/ml e hanol-
amine, 470 mg/ml linoleic acid, 470 mg/ml oleic acid, and 50 mg/ml bo ine
se um albumin; Sigma) and 10 ng/ml TGF-b3 (Pep o ech) in DMEM wi h 1%
penicillin-s ep omycin (In i ogen) o e 2–4 weeks. All cul u es we e main-
ained wi h 5% CO
2
in a wa e -jacke ed incuba o a 37C, and hal -medium
changes we e pe o med wice a week. To assess in i o di e en ia ion in o
mesenchymal lineages, cells we e b ie ly washed wi h PBS and ixed o
10 min wi h 2% pa a o maldehyde in PBS (Sigma). Fo he de ec ion o alkaline
phospha ase ac i i y, cells we e washed wice wi h PBS and incuba ed o
20 min a oom empe a u e wi h 50 mg/ml Naph hol AS-MX phospha e,
0.5% N,N-Dime hyl o mamide, and 0.6 mg/ml Fas Red Viole LB in 0.1 M
T is-HCl, pH 8.9. Aliza in ed s aining o calcium deposi s was pe o med in
ixed cul u es washed wi h dis illed wa e and incuba ed wi h 2% aliza in ed
o 5 min, ollowed by washes in dis illed wa e . Adipocy es we e s ained
wi h oil ed O as ollows: cells we e washed wi h 60% isop opanol and allowed
o d y comple ely. An oil ed wo king solu ion was p epa ed as a 6:4 dilu ion in
dis illed wa e o a 0.35 g/ml oil ed O solu ion in isop opanol (Sigma) and
il e ed 20 min la e . Cells we e incuba ed o 10 min wi h oil ed wo king solu-
ion and insed 4 imes. To s ain mucopolysaccha ides associa ed wi h chon-
d ocy ic di e en ia ion, ixed cul u es and cell pelle s we e incuba ed o
30 min a oom empe a u e wi h 1% Alcian blue 8GX (Sigma) in 3% ace ic
acid, pH 2.5, and insed ou imes. Fo oluidine blue s aining o chond ocy es,
ixed sphe es o hei ehyd a ed pa a in sec ions we e incuba ed in 1% olu-
idine blue (Am esco) in 1% sodium e abo a e decahyd a e (Sigma) o 5 o
2 min, espec i ely, and washed a e wa d.
1722 Cell Repo s 3, 1714–1724, May 30, 2013 ª2013 The Au ho s
